Advances in Semiconductor

Technologies: Selected Topics Beyond

Conventional CMOS An Chen
Visit to download the full and correct content document:
https://ebookmass.com/product/advances-in-semiconductor-technologies-selected-to
pics-beyond-conventional-cmos-an-chen/
Advances in Semiconductor Technologies
 IEEE Press
445 Hoes Lane
Piscataway, NJ 08854

IEEE Press Editorial Board

Sarah Spurgeon, Editor in Chief

Jón Atli Benediktsson Andreas Molisch Diomidis Spinellis

Anjan Bose Saeid Nahavandi Ahmet Murat Tekalp
Adam Drobot Jeffrey Reed
Peter (Yong) Lia Thomas Robertazzi
Advances in Semiconductor Technologies

Selected Topics Beyond Conventional CMOS

Edited by
An Chen
IBM Research – Almaden, CA, USA
Copyright © 2023 by The Institute of Electrical and Electronics Engineers, Inc. All rights reserved.

Published by John Wiley & Sons, Inc., Hoboken, New Jersey.

Published simultaneously in Canada.

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, scanning, or otherwise, except as permitted under Section
107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or
authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222
Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com.
Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons,
Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at http://www.wiley.com/
go/permission.

Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing
this book, they make no representations or warranties with respect to the accuracy or completeness of the contents
of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose.
No warranty may be created or extended by sales representatives or written sales materials. The advice and
strategies contained herein may not be suitable for your situation. You should consult with a professional where
appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages,
including but not limited to special, incidental, consequential, or other damages. Further, readers should be aware
that websites listed in this work may have changed or disappeared between when this work was written and when it
is read. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial damages,
including but not limited to special, incidental, consequential, or other damages.

For general information on our other products and services or for technical support, please contact our Customer
Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or fax (317)
572-4002.

Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be
available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com.

Library of Congress Cataloging-in-Publication Data:

Names: Chen, An (Electronics engineer), editor.

Title: Advances in semiconductor technologies : selected topics beyond
conventional CMOS / An Chen.
Description: Hoboken, New Jersey : Wiley-IEEE Press, [2023] | Includes
bibliographical references and index.
Identifiers: LCCN 2022018684 (print) | LCCN 2022018685 (ebook) | ISBN
9781119869580 (cloth) | ISBN 9781119869597 (adobe pdf) | ISBN
9781119869603 (epub)
Subjects: LCSH: Semiconductors.
Classification: LCC TK7871.85 .A3565 2023 (print) | LCC TK7871.85 (ebook)
| DDC 621.3815/2 – dc23/eng/20220627
LC record available at https://lccn.loc.gov/2022018684
LC ebook record available at https://lccn.loc.gov/2022018685

Cover image: © Blue Andy/Shutterstock

Cover design: Wiley

Set in 9.5/12.5pt STIXTwoText by Straive, Chennai, India

 v

Contents

Preface xi
List of Contributors xv

1 Heterogeneous Integration at Scale 1

Subramanian S. Iyer and Boris Vaisband
1.1 Introduction 1
1.2 Technology Aspects of Heterogeneous Integration 4
1.2.1 Interconnect Pitch 4
1.2.2 Substrate Material 5
1.2.3 Inter-Die Spacing 6
1.2.4 Die Size Considerations 6
1.2.5 Dielet to Substrate Pitch Considerations 8
1.2.6 Backward Compatibility 8
1.3 Design and Architecture of Heterogeneous Integration Platforms 9
1.3.1 Power Delivery and Thermal Management 9
1.3.2 Floorplanning 11
1.3.3 Communication 12
1.4 Reliability of Heterogeneous Integration Systems 14
1.5 Application Space of Heterogeneous Integration 17
1.6 Future of Heterogeneous Integration 18
1.7 Summary 20
References 21

2 Hyperdimensional Computing: An Algebra for Computing with Vectors 25

Pentti Kanerva
2.1 Introduction 25
2.2 Overview: Three Examples 26
2.2.1 Binding and Releasing with Multiplication 26
2.2.2 Superposing with Addition 26
2.2.3 Sequences with Permutation 28
2.3 Operations on Vectors 28
2.4 Data Structures 30
2.5 Vector Sums Encode Probabilities 32
2.6 Decoding a Product 33
2.7 High-Dimensional Vectors at Large 34
vi Contents

2.8 Memory for High-Dimensional Vectors 35

2.9 Outline of Systems for Autonomous Learning 36
2.10 Energy-Efficiency 37
2.11 Discussion and Future Directions 37
References 40

3 CAD for Analog/Mixed-Signal Integrated Circuits 43

Ahmet F. Budak, David Z. Pan, Hao Chen, Keren Zhu, Mingjie Liu, Mohamed B. Alawieh,
Shuhan Zhang, Wei Shi, and Xiyuan Tang
3.1 Introduction 43
3.2 Front-End CAD 45
3.2.1 Circuit Architecture and Topology Design Space Exploration 45
3.2.2 Device Sizing 46
3.2.2.1 AMS Circuit Sizing: Problem Formulation 46
3.2.2.2 Methods for AMS Circuit Sizing 47
3.3 Layout Automation 48
3.3.1 Procedural Layout Generation 49
3.3.2 Optimization-Based Layout Synthesis 50
3.4 Post-Layout Extraction and Verification 52
3.5 Conclusion 53
Acknowledgments 53
References 53

4 Magnetoelectric Transistor Devices and Circuits with Steering Logic 61

Andrew Marshall and Peter A. Dowben
4.1 Introduction 61
4.2 Simple Logic Functions with the MEFET “Steering Logic” 62
4.3 Logic Functions – Majority Gate 64
4.4 The Full Adder and the Dual XOR (Sum) Gates 67
4.5 Latch and Memory 70
4.6 The JK Master–Slave Flip-Flop 72
4.7 Conclusion 75
Acknowledgments 75
References 75

5 Nonvolatile Memory Based Architectures Using Magnetoelectric FETs 79

Shaahin Angizi, Deliang Fan, Andrew Marshall, and Peter A. Dowben
5.1 Introduction 79
5.2 Magnetoelectric Field Effect Transistor (MEFET) 79
5.3 1T-1M Memory Design Based on the MEFET 81
5.3.1 Read Operation 81
5.3.2 Write Operation 82
5.4 2T-1M Memory Design Based on the MEFET 84
5.4.1 Read Operation 85
5.4.2 Write Operation 86
5.5 MEFET Steering Memory 87
5.6 Evaluation 90
5.6.1 Comparative Read Time 90
5.6.2 Comparative Write Time 90
5.6.3 Comparison of Cell Areas 90
 Contents vii

5.7 Conclusion 91
Acknowledgments 91
References 91

6 Organic Electronics 93
Hagen Klauk
6.1 Introduction 93
6.2 Organic Light-Emitting Diodes 94
6.3 Organic Solar Cells 96
6.4 Organic Thin-Film Transistors 97
6.5 Outlook 101
References 102

7 Active-Matrix Electroluminescent Displays 109

Xiaojun Guo, Li’ang Deng, and Arokia Nathan
7.1 Introduction 109
7.2 Light-Emitting Diodes for Displays 110
7.2.1 Thermally Evaporated OLEDs 110
7.2.2 Realization of Full Color Displays 111
7.2.3 Printed Displays 113
7.2.4 Micro-LED 114
7.3 TFT Backplanes 115
7.4 Driving Schemes and Pixel Circuits 116
7.4.1 Analog Driving 117
7.4.2 Compensation for Voltage Programming 118
7.4.2.1 Internal Compensation 118
7.4.2.2 External Compensation 119
7.4.3 Digital Driving 120
7.4.4 Hybrid Driving 123
7.5 Conclusion 124
References 124

8 Organic and Macromolecular Memory – Nanocomposite Bistable Memory

Devices 133
Shashi Paul
8.1 Introduction 133
8.1.1 What Is an Electronic Memory Device? 133
8.2 Organic Memory and Its Evolution 137
8.2.1 Molecular Memory 137
8.2.2 Polymer Memory Devices 139
8.3 Summary 146
Acknowledgment 147
References 147
Further Reading/Resources 151
Related Articles (See Also) 151

9 Next Generation of High-Performance Printed Flexible Electronics 153

Abhishek S. Dahiya, Yogeenth Kumaresan, and Ravinder Dahiya
9.1 Introduction 153
9.2 Printing Technologies 155
viii Contents

9.3 High-Performance Printed Devices and Circuits Using Nano-to-Chip Scale

Structures 158
9.3.1 Nanoscale Structures 158
9.3.2 Microscale Structures 162
9.3.3 Chip-Scale (or Macroscale) Structures 165
9.4 Challenges and Future Directions 168
9.4.1 Integration of Nano-to-Chip Scale Structures 168
9.4.2 Technological Challenges 169
9.4.3 Robustness 169
9.4.4 Disposability 170
9.4.5 Modeling for Flexible Electronics 170
9.4.6 Power Consumption 170
9.5 Summary 171
References 171

10 Hybrid Systems-in-Foil 183

Mourad Elsobky
10.1 Introduction 183
10.1.1 System-Level Concept 183
10.2 Emerging Applications 185
10.2.1 Smart Labels for Logistics Tracking 185
10.2.2 Electronic Skin 187
10.2.2.1 E-Skin with Embedded AI 187
10.2.3 Biomedical 189
10.3 Integration Technologies 191
10.3.1 Substrate and Interconnect Materials 191
10.3.2 Flex-PCB 191
10.3.3 Wafer-Based Processing 193
10.3.3.1 ChipFilm Patch 193
10.3.4 Challenges 194
10.3.4.1 Surface Topography 194
10.3.4.2 Thermal Stress 196
10.3.4.3 Assembly and Positioning Errors 196
10.4 State-of-the-Art Components 196
10.4.1 Active Electronics 197
10.4.1.1 Microcontrollers 197
10.4.1.2 Sensor Frontends 197
10.4.1.3 Sensor Addressing and Multiplexing 200
10.4.2 Passive Components and Sensors 200
10.4.2.1 Flexible Antennas 200
10.4.2.2 HySiF-Compatible Sensors 201
10.5 HySiF Testing 202
10.6 Conclusion and Future Directions 204
References 204

11 Optical Detectors 211

Lis Nanver and Tihomir Knežević
11.1 Introduction 211
 Contents ix

11.2 Si Photodiodes Designed in CMOS 213

11.3 Ultraviolet Photodetectors 217
11.4 Infrared Optical Detectors 219
11.4.1 Detectors for Photonic Integrated Circuits 220
11.4.2 Infrared Photoconductive Detectors 223
11.4.3 Thermal Infrared Detectors 224
11.5 Emerging Devices 225
11.6 Concluding Remarks 226
References 227

12 Environmental Sensing 231

Tarek Zaki
12.1 Motivation 231
12.1.1 Air Pollution 231
12.1.2 Hazardous Pollutants 232
12.1.3 Air Quality Index 233
12.1.4 Air Monitoring Network 234
12.1.5 Hand-Held Devices 236
12.2 Particulate Matter (PM) Sensing 238
12.2.1 Particulate Matter (PM) 238
12.2.2 Sensing Mechanisms 239
12.2.3 Optical Particle Counter (OPC) 241
12.2.4 Particle Size Distributions 242
12.2.5 Miniaturized Optical PM Sensing 243
12.3 Volatile Organic Compounds (VOCs) Sensing 244
12.3.1 Volatile Organic Compounds (VOCs) 244
12.3.2 Sensing Mechanisms 246
12.3.3 MOX-Based Sensors 247
References 249

13 Insulated Gate Bipolar Transistors (IGBTs) 255

Thomas Laska
13.1 Introduction 255
13.2 State-of-the-Art IGBT Technology 257
13.2.1 Structural Basics with Respect to Blocking, ON State and Switching 257
13.2.2 Cell and Vertical Design 258
13.2.3 Wafer Technology 259
13.2.4 Reverse-Blocking and Reverse-Conducting IGBTs 259
13.2.5 Increasing Maximum Junction Temperature 260
13.2.6 Assembly and Interconnect Technology 260
13.2.7 Power Density Increase 260
13.3 Future Prospect of IGBT 261
13.3.1 Application Requirement Aspects 261
13.3.2 Next Generation Cell Design Including Gate Driving Schemes 262
13.3.3 Next Generation Vertical Structure Concepts 265
13.3.4 Next Level of Thermal Management and Interconnect Technique Innovation 266
13.4 Outlook 268
Acknowledgment 268
References 268
x Contents

14 III–V and Wide Bandgap 273

Mohammed Alomari
14.1 Introduction 273
14.2 Diamond Power Devices 276
14.3 SiC Power Devices 277
14.4 GaN Power Devices 279
14.5 Wide Bandgaps for High-Temperature Applications 285
14.6 Conclusion 286
References 287

15 SiC MOSFETs 295

Peter Friedrichs
15.1 Introduction to Silicon Carbide for Power Semiconductors 295
15.2 SiC Schottky Barrier Diodes 295
15.3 SiC Transistors 298
15.4 SiC Power MOSFETs 299
15.4.1 Possible Cell Concepts 299
15.4.2 SiC MOS Channel Challenges 301
15.4.3 Typical MOSFET Device Characteristics – Static Behavior, Switching Performance, and
Body Diode Aspects 302
15.4.4 Gate-Oxide Reliability Aspects 308
15.4.5 Short-Circuit Aspects and Avalanche Ruggedness of SiC MOSFETs 312
15.5 SiC MOSFETs in Power Applications – Selected Aspects and Prospects 316
References 317

16 Multiphase VRM and Power Stage Evolution 321

Danny Clavette
16.1 Evolution of the First Multiphase Controllers 321
16.2 Transition from VRMs to “Down” Solutions 324
16.3 Intel Xeon Generations Challenges Moore’s Law 326
16.4 Increased System Digitization Enables Digital Control 327
16.5 DrMOS 1.0: Driver + MOSFETs 328
16.6 DrMOS 4.0 and International Rectifier’s Power Stage Alternative 330
16.7 International Rectifier’s “Smart” Power Stage 334
16.8 DrMOS 5 × 5 mm and 4 × 4 mm De-standardization 335
16.9 5 × 6 mm Smart Power Stage: Industry Driven Standardization 336
16.10 Latest SPS Activities 337
16.11 Trending Back to VRMs 338
16.12 Summary 339
References 340
Abbreviations 341

Index 343
 xi

Preface

Since the invention of the solid-state transistors, the semiconductor technologies have advanced
at an exponential pace and become the foundation for numerous industries, e.g. computing,
communication, consumer electronics, autonomous systems, and defense. Guided by Moore’s
law, the scaling of transistors has provided new generations of chips every one to two years, with
ever-increasing density and better performance. Today, silicon transistors are approaching some
fundamental limits of dimensional scaling. The semiconductor industry has also transformed
through several phases and foundational technologies. The emergence of Internet of Things (IoT),
big data, artificial intelligence (AI), and quantum computing has created new opportunities for
advanced semiconductor technologies. The complementary metal-oxide-semiconductor (CMOS)
technology dominates the semiconductor industry today, but there are numerous technologies and
active research beyond conventional CMOS. Although semiconductors are often associated with
high-performance computing chips such as central processing unit (CPU) and graphics processing
unit (GPU), there is a wide range of applications beyond computing for semiconductor products,
e.g. sensors, displays, and power electronics. Silicon (Si) is the most important semiconductor, but
the semiconductor research also covers a variety of materials, e.g. germanium (Ge), III–V com-
pounds, organic materials, carbon nanotube, 2D materials, magnetic materials, and topological
materials.
This book is a collection of articles reviewing advanced semiconductor technologies beyond con-
ventional Si CMOS for various applications. These articles written by the experts in the fields can
be read independent of each other. The variety of topics reflects the breadth of the semiconductor
R&D and applications today, but these articles only cover a very small fraction of semiconductor
technologies.
With the transistor scaling approaching the fundamental limits, heterogeneous integration is a
promising direction to sustain the improvement of performance and functionalities without relying
on reducing transistor sizes. Chapter 1, “Heterogeneous Integration at Scale,” provides a compre-
hensive review of technologies, design/architecture considerations, reliability issues, applications,
and future directions of large-scale heterogeneous integration.
While technology innovation has been a primary driver for the semiconductor industry, the
future of semiconductor systems will increasingly resort to novel computing paradigms. Chapter 2,
“Hyperdimensional Computing: An Algebra for Computing with Vectors,” presents an example of
entirely new ways of computing inspired by the information processing in the brain. Instead of
traditional model of computing with numbers, hyperdimensional (HD) computing encodes infor-
mation in a holographic representation with wide vectors and unique operations. HD computing
is extremely robust against noise, matches well with 3D circuits, and is uniquely suitable to process
a variety of sensory signals without interference with each other.
xii Preface

The majority of semiconductor chips are digital circuits; however, analog and mixed-signal
circuits are crucially important. The physical world is analog; therefore, analog circuits are always
needed to connect digital chips with real world, e.g. sensory data, power management, and
communication. Although digital circuit design is highly automated, analog circuit design still
relies on manual effort. Chapter 3, “CAD for Analog/Mixed-Signal Integrated Circuits,” reviews
the progress toward automated computer-aided design (CAD) of analog and mixed-signal circuits.
Modern computers are built based on the von Neumann architecture with separate logic/
computing units and memory/storage units. Emerging memory devices not only provide new
technologies to improve memory systems but also enable novel computing architectures, e.g.
in-memory computing. One of the most promising emerging memories is based on magnetic
materials and properties. Chapters 4 and 5 focus on a so-called magnetoelectric field effect
transistor (MEFET) based on the programming of the polarization in a 2D semiconductor channel
with large spin-orbit coupling, via the proximity effect of a magnetoelectric gate. Chapter 4,
“Magnetoelectric Transistor Devices and Circuits with Steering Logic,” presents various logic gate
designs based on a one-source two-drain MEFET configured with a steering function. Chapter 5,
“Nonvolatile Memory Based Architectures Using Magnetoelectric FETs,” describes MEFET
memory designs with the performance and size suitable to fulfill the application space between
static random-access-memory (SRAM) and dynamic random-access-memory (DRAM).
Novel materials beyond Si, Ge, and III–V compounds may enable new semiconductor products
and applications. Among them, organic semiconductors are promising materials for low-cost, flex-
ible, and bio-compatible electronics. Chapter 6, “Organic Electronics,” discusses the opportunities
of organic semiconductors for large-area flexible electronics, including organic light-emitting diode
(OLED), organic displays, organic solar cells, and thin-film transistors. Chapter 7, “Active-Matrix
Electroluminescent Displays,” delves into the details of flat panel electroluminescent displays
based on light-emitting diodes (LEDs) that have been utilized in a wide range of applications
including smart phones, tablets, laptops, and TVs. Various underlying LED technologies, associ-
ated circuits, and design considerations are reviewed. Another interesting application of organic
materials is memory. Chapter 8, “Organic and Macromolecular Memory – Nanocomposite Bistable
Memory Devices,” discusses the mechanisms, characteristics, and current status of organic mem-
ories. One of the advantages of organic materials is their low-cost processing and the potential
to stack up multiple layers. Chapter 9, “Next Generation of High-Performance Printed Flexible
Electronics,” summarizes different printing technologies for flexible electronics, showcases
the state-of-the-art printed flexible electronic circuits, and discusses the challenges and future
directions of large-scale cost-effective printed electronics. The vision of integrating electronic
components onto polymer foils leads to the flexible electronics version of systems-on-chip (SoC),
known as systems-in-foil (SiF). A wide range of applications can benefit from SiF, e.g. smart labels,
intelligent electronic skin, and implanted devices. Chapter 10 “Hybrid Systems-in-Foil” reviews
the opportunities of SiF and challenges in materials, integration, and testing.
The electronic systems need an interface with the physical world. Semiconductor chips rely on
sensors to “see,” “hear,” and “smell.” Optical sensing is utilized in a wide range of applications, e.g.
camera, fiber optics and communication, light source and laser, data storage, medical monitoring
and diagnostics, and manufacturing. Chapter 11, “Optical Detectors,” reviews the photodiodes
based on Si, III–V, and emerging materials as the essential components for highly sensitive
detectors for a broad spectrum of wavelengths. Chapter 12, “Environmental Sensing,” covers com-
prehensively different air pollution sources, air quality metrics, and various sensing approaches
for particulate matters and volatile organic compounds. The advancement of semiconductor
 Preface xiii

technologies contributes to the miniaturization of the sensing equipment and the improvement of
their performance.
Unlike computer chips operating with very low voltage and current, power electronics handle
very high voltage (e.g. thousands of volt or higher) and current required to operate machinery,
vehicles, appliances, etc. Special device designs and unique material properties are required to
sustain such high voltage and current in semiconductor chips. Chapter 13, “Insulated Gate Bipolar
Transistors (IGBTs),” reviews an important high-power device known as Si insulated gate bipolar
transistors (IGBTs). IGBT not only dominates power electronics today but also continues to be
innovated for further gains in power density and efficiency. At the same time, significant progress
has been made on wide bandgap semiconductors. Chapter 14, “III–V and Wide Bandgap,” reviews
promising materials (e.g. diamond, GaN) and their applications in high-frequency power conver-
sion and high-temperature electronics. While wide bandgap power modules may be combined with
Si-based control circuits in near-term solutions, considerable effort is made to advance integrated
circuits based on wide bandgap semiconductors. Chapter 15, “SiC MOSFETs,” reviews SiC-based
power semiconductor devices including diodes and transistors. SiC is well positioned to fulfill the
requirements of power electronics, e.g. energy efficiency, scaling, system integration, and reliability.
The unique ability of SiC to form a native SiO2 as the gate dielectric makes it particularly attractive
for power metal-oxide-semiconductor field-effect-transistors (MOSFETs). At the end, Chapter 16,
“Multiphase VRM and Power Stage Evolution,” provides a detailed overview of the evolution of
CPU power delivery technologies and explains the reasons driving the technology shifts.
This book could be considered as a small-scale reference of advanced semiconductor
technologies, which may potentially be expanded into a large-scale reference with more
comprehensive coverage. It is our wish that this collection of chapters will provide useful tutorials
on selected topics of advanced semiconductor technologies.

An Chen
IBM Research – Almaden, CA, USA
September 2022
 xv

List of Contributors

Mohamed B. Alawieh Danny Clavette

Department of Electrical and Computer Infineon Technologies
Engineering Americas Corporation
The University of Texas at Austin El Segundo, CA
Austin, TX USA
USA
Abhishek S. Dahiya
Mohammed Alomari University of Glasgow
Institut für Mikroelektronik Stuttgart James Watt School of Engineering, Bendable
Stuttgart Electronics and Sensing Technologies (BEST)
Germany Group
Glasgow
Shaahin Angizi UK
Department of Electrical and Computer
Engineering Ravinder Dahiya
New Jersey Institute of Technology University of Glasgow
Newark, NJ James Watt School of Engineering, Bendable
USA Electronics and Sensing Technologies (BEST)
Group
Ahmet F. Budak Glasgow
Department of Electrical and Computer UK
Engineering
The University of Texas at Austin Li’ang Deng
Austin, TX Department of Electronic Engineering
USA Shanghai Jiao Tong University
China
Hao Chen
Department of Electrical and Computer Peter A. Dowben
Engineering Department of Physics and Astronomy
The University of Texas at Austin Jorgensen Hall, University of Nebraska
Austin, TX Lincoln, NE
USA USA
xvi List of Contributors

Mourad Elsobky Thomas Laska

Institut für Mikroelektronik Stuttgart, Sensor Infineon Technologies AG
Systems Germany
Stuttgart
Germany Mingjie Liu
Department of Electrical and Computer
Peter Friedrichs Engineering
IFAG IPC T, Infineon Technologies AG The University of Texas at Austin
Neubiberg Austin, TX
Germany USA

Xiaojun Guo Andrew Marshall

Department of Electronic Engineering Department of Electrical Engineering, The
Shanghai Jiao Tong University Erik Johnson School of Engineering and
China Computer Science
University of Texas at Dallas
Subramanian S. Iyer Richardson, TX
Electrical Engineering Department USA
University of California
Los Angeles, CA Lis Nanver
USA Faculty of Electrical Engineering
Mathematics & Computer Science
Pentti Kanerva MESA+ Institute of Technology
University of California at Berkeley University of Twente
Redwood Center for Theoretical Neuroscience Enschede
Berkeley, CA The Netherlands
USA
Arokia Nathan
Hagen Klauk Darwin College
Max Planck Institute for Solid State Research University of Cambridge
Stuttgart Cambridge
Germany UK

Tihomir Knežević David Z. Pan

Faculty of Electrical Engineering Department of Electrical and Computer
Mathematics & Computer Science Engineering
MESA+ Institute of Technology The University of Texas at Austin
University of Twente Austin, TX
Enschede USA
The Netherlands
Shashi Paul
Yogeenth Kumaresan Emerging Technologies Research Centre
University of Glasgow De Montfort University
James Watt School of Engineering, Bendable Leicester
Electronics and Sensing Technologies (BEST) UK
Group
Glasgow, UK
 List of Contributors xvii

Wei Shi Shuhan Zhang

Department of Electrical and Computer Department of Electrical and Computer
Engineering Engineering
The University of Texas at Austin The University of Texas at Austin
Austin, TX Austin, TX
USA USA

Xiyuan Tang Keren Zhu

Department of Electrical and Computer Department of Electrical and Computer
Engineering Engineering
The University of Texas at Austin The University of Texas at Austin
Austin, TX Austin, TX
USA USA

Boris Vaisband Deliang Fan

McGill University Department of Electrical, Computer and
Electrical and Computer Engineering Energy Engineering
Department Arizona State University
Montreal, QC Tempe, AZ
Canada USA

Tarek Zaki
Munich, Germany
 1

Heterogeneous Integration at Scale

Subramanian S. Iyer 1 and Boris Vaisband 2
1 Electrical Engineering Department, University of California, Los Angeles, CA, USA
2
McGill University, Electrical and Computer Engineering Department, Montreal, QC, Canada

1.1 Introduction
Microelectronics has made tremendous progress over the last several decades adhering to what is
popularly called Moore’s Law. One measure of Moore’s law is the scaling factor of minimum fea-
tures on a silicon integrated circuit (IC). This trend is shown in the dark gray (left-hand y-axis) curve
in Figure 1.1 [1] exhibiting over a 1,000-fold decrease in minimum feature size, corresponding to a
million-fold transistor density improvement. This improvement corresponds to reduction in power
per function as well as reduction of cost and price per function. Nonetheless, until recently, pack-
aging did not scale as seen in the light gray (right-hand y-axis) curve in Figure 1.1. For example,
in 1967, when flip-chip bonding was first introduced, the bump pitch was 400 μm. Even today, the
pitch of the bump (die to laminate) has scaled to about 130 μm, while ball grid array (BGA) pitch
and trace pitch on laminates and printed circuit boards (PCBs) have not fared better. However, in
the last few years, we have seen an acceleration of these metrics as shown in the inset in Figure 1.1.
Note that the silicon (Si) scale is in nanometers, while the packaging scale is in micrometers. There
are two key factors that have mediated this acceleration: (i) the adoption of silicon-like process-
ing materials and methods to achieve scaling, including silicon interposers, and, importantly, (ii)
fan-out wafer-level packaging (FOWLP).
This trend has manifested itself in two ways: (i) The extensive use of interposers, which is an
additional level in the packaging hierarchy as shown in Figure 1.2. At a basic level, interposers pro-
vide a first-level platform for the integration of several (eight) heterogeneous dielets on a thinned
silicon substrate that is then further packaged on a laminate and attached to a PCB. This allows
the dielets on the interposer to communicate intimately within the interposer, though communi-
cation outside the interposer is more conventional. (ii) Three-dimensional (3D) integration, where
dies are stacked one on top of the other, typically face to back with through silicon vias (TSVs)
or alternatively face to face through surface connections. These face-to-face connections can be at
high bandwidth and low latency. Both of these techniques have transformed packaging, especially
when it comes to the memory subsystem. A roadmap using the memory subsystem as a paradigm
for advanced packaging is depicted in Figure 1.3. Another area in which interposers and 3D inte-
gration can play a big role is the integration of analog and mixed signal functions. Moore’s law
scaling does a good job in scaling digital logic, but is at best marginal when it comes to analog and
mixed signal functions. This is shown in Figure 1.4, where the analog/mixed signal components
can occupy an increasing percentage of real estate at finer geometries. In these cases, retaining the
Advances in Semiconductor Technologies: Selected Topics Beyond Conventional CMOS, First Edition. Edited by An Chen.
© 2023 The Institute of Electrical and Electronics Engineers, Inc. Published 2023 by John Wiley & Sons, Inc.
2 1 Heterogeneous Integration at Scale

25 160
(nm) (μm)
Advanced 1400
10,000 20
2
350 packaging 120

15 100
300

Package scaling (μm)

80
1,000 10

μ
60
Si scaling (nm)

250
40
5
200 20
100 Chiplet/dielet revolution
0 0
2011 201
2012 2013 2014 2015 2016 2017 2018 2019 2020 2021
20
202
150

Pack
10 100

5500

1 0
1960 1970 1980 1990 2000 2010 0
2020 2030
Year

Figure 1.1 Scaling trends for CMOS features as well as package features. Package scaling has lagged
significantly as compared to Si scaling. Adoption of silicon-like technology for packaging has somewhat
accelerated scaling.

Heat sink (Cu, AI)

Heat sink (Cu, AI)

Underfill Die 3D IC

μ-bumps
MCM Die Die
Interposer
C4 bumps
TSV
Laminate Laminate
BGA

Printed circuit board

Figure 1.2 Current packaging hierarchy. Left: die-laminate-PCB. Right: die-interposer-laminate-PCB.

analog/mixed signal functions in an older node makes sense as long as one can provide compatible
voltage domains and ensure low latency as well as low analog signal distortion. These are not very
difficult to do on interposers.
Why is it important to scale packaging? Packaging dimensions determine the size of the sys-
tem especially since the scale is 10–100 times larger than chip dimensions. Power too is a major
consideration. Communication power between chips accounts for 30–40% of total system power.
So for size, weight, and power (SWaP), as well as cost, scaling the package has advantages. The
key parameters that affect SWaP are dominated by packaging metrics, and scaling the package has
greater impact on SWaP than additional Si scaling. For flexible hybrid electronics (FHE), form fac-
tor and power play a critical role. Most FHE devices are mobile and dependent on battery power.
As such, FHE packages will benefit immensely from scaling.
 1.1 Introduction 3

Interposers Massively integrated

silicon-like board
Level of integration

Integration

“Prehistoric” Future
Stacked die

Stacked memory
Now Wafer stacking

Time

Figure 1.3 Packaging evolution – the memory paradigm.

100 100%

90 90%

80 80%

70 70%

60 60%
Die area (%)

Analog (%)

50 50%

40 40%

30 30%

20 20%

10 10%

0 0%
0 50 100 150 200
Node (nm)

Figure 1.4 Percent of area dedicated to analog circuits is increasing with scaling (squares). Thus, practical
die area (normalized to 180 nm technology) is increasing with scaling (rhombuses) as compared to ideal die
area (triangles).
4 1 Heterogeneous Integration at Scale

Another aspect of advanced packaging that has the focus of attention in recent years is hetero-
geneous integration. This term requires some clarification. Most packaging constructs do in fact
achieve heterogeneous integration via the integration of diverse packaged chips on an extended
substrate such as a PCB. Heterogeneous integration, therefore, in general and in itself, is not new.
However, in the context of advanced packaging, heterogeneous integration refers to the integration
of bare dies on a first-level packaging substrate. This could be an organic, ceramic, or silicon inter-
poser. The key features that distinguish heterogeneous integration from classical or conventional
packaging are the pitch of the connections between the bare die and the substrate, the number
of connections between the interconnected bare dies, the size of the dies, and hence a significant
simplification in the communication protocols of interdie signaling. It is generally accepted that
for bump pitches <50 μm, interdie spacing of <2 mm, and trace pitches (wiring between the dies)
of <5 μm, the integration is considered in the regime of advanced packaging.
Finally, “chiplets” and “dielets” are another feature of advanced packaging. A complex system or
large chip design is fragmented into smaller entities called chiplets and then instantiated in Si as
dielets. These dielets are then intimately reintegrated at fine pitch (bump and trace) as well as short
interdie spacing, as previously described, to synthesize a subsystem or a module. This construct can
be further assembled on a PCB or, in the case of wafer scale systems, and can represent the entire
system [2, 3].

1.2 Technology Aspects of Heterogeneous Integration

Technology innovation is the main driver of the various heterogeneous integration platforms. In
the past several decades, packaging technology (e.g. vertical interconnect pitch) has been scaling
at a significantly slower rate as compared to IC technology (e.g. device dimensions), as shown in
Figure 1.1. Specifically, on-chip dimensions have scaled approximately 200 times more than pack-
age features. This disparity in scaling of parts of a single system led the package hierarchy to become
the bottleneck of modern integrated systems.
Nonetheless, in the last five years, the electronic packaging community has picked up the pace
proposing various novel integration technologies to reduce the dimensions of the packaging hier-
archy. Specifically, novel heterogeneous integration platforms have been proposed, significantly
driving down the features of the packaging hierarchy. Realization of the heterogeneous integration
concept is predicated on several important technology considerations. A review of the vertical inter-
connect pitch (between dielets and substrate), substrate material, interdielet spacing, and dielet
termination is provided in this section.

1.2.1 Interconnect Pitch

Typical package-level interconnect pitch is several hundreds of micrometers. These are
solder-based BGA or land grid array (LGA) connections between the package and the PCB.
Connections between dies and package laminate, i.e. solder-based C4/pillar, exhibit a smaller
pitch (∼50 to 100 μm). Nonetheless, comparing to the last metal levels on the die (pitch of
2–10 μm), the package-level interconnects are approximately between 10 and 500 times larger.
In fact, the main purpose of the packaging hierarchy is to fan-out the interconnect pitch from
the small die-level pitch, to the large PCB-level pitch, and then vice versa when connecting to
the neighboring packaged die. Solder has the advantage of deforming at low temperature and
pressure, to accommodate warpage of the PCB, laminate, and die. A typical Cu pillar capped with
 1.2 Technology Aspects of Heterogeneous Integration 5

(a) (b) (c)

Figure 1.5 (a) A 50 mm diameter Cu pillar capped with solder. (b) After mass reflow with compression
showing the solder extrusions. (c) Micrograph showing extrusions that cause adjacent pillars to become
electrically shorted. Source: Photo courtesy: Eric Perfecto.

solder is shown in Figure 1.5a, a cross section after mass reflow is shown in Figure 1.5b, and a
micrograph of two shorted pillars is shown in Figure 1.5c. Shorting of neighboring solder balls is
the main challenge to the continuous scaling of solder-based interconnects below about 50 μm.
Another possible integration approach is to attach dielets directly to a substrate without solder
(or other intermetallics), using direct metal-to-metal thermal compression bonding (TCB). In this
integration approach, two metals are bonded together by applying pressure and temperature for
a certain amount of time. After this process, a very strong low-contact resistance bond is formed.
To ensure high-quality connections using TCB, the surfaces of the metals must be pristine and
atomically smooth. Additional conditions must be met when bonding Cu for example, such as
surface preparation to avoid oxidation of the Cu prior to bonding (by using plasma or formic acid).
Alignment is another challenge in small pitch connections for heterogeneous integration.
Solder-based interconnects are easier to align due to the larger dimensions and since they exhibit
a form of self-alignment property due to surface tension at solder melting temperature. Unlike
solder, direct metal-to-metal integration requires a high level of alignment accuracy which is
difficult to achieve by optical means since the die and substrate are typically not transparent. A
second-order optical alignment is therefore required, where the die and the substrate are aligned to
a virtual reference. For example, in the silicon interconnect fabric (Si-IF) technology, an alignment
accuracy of ±1 μm has been achieved [4]. Although this alignment accuracy is good, it is still about
an order of magnitude worse than die-level conventional optical lithography.

1.2.2 Substrate Material

Different substrate materials are currently used in heterogeneous integration. The most commonly
used materials are compared in Table 1.1.
Hybrid substrate materials are also used in industry. For example, the embedded multidie inter-
connect bridge (EMIB) [5] approach supports the integration of small Si bridges within an organic
substrate. The EMIB enables different interconnect pitches, coarse pitch on the organic FR4 and
fine pitch on the Si bridge. This technology, however, exhibits increased complexity and therefore
higher cost.
6 1 Heterogeneous Integration at Scale

Table 1.1 Key structural and thermal properties of Si and other relevant materials.

Young’s Tensile Thermal

modulus strength CTE conductivity
Material (GPa) (MPa) (ppm) (W/m K) Warpage Cost Notes

Organic 0.1–20 2,000–3,000 14–70 0.3–1 High Low Large horizontal and
(FR4) vertical interconnect pitch
Glass 50–90 33–3,500 4–9 1–2 Low High Low electrical losses.
Metallization is difficult
Silicon 130–185 5,000–9,000 3–5 148 Low Low High electrical losses
Steel 190–200 400–500 11–13 16–25
Copper 128 200–350 17 400

The Si-IF [2] is a silicon wafer-scale platform that supports integration of small dies at fine vertical
pitch (2–10 μm). Si is a highly mature substrate that benefits from decades of technology opti-
mization. Furthermore, passive Si with micrometer size interconnects is a relatively inexpensive
construct.

1.2.3 Inter-Die Spacing

In conventional packaging, dies are packaged and placed on PCBs. Interposers, an additional hier-
archical layer, is often utilized in modern systems as a stepping stone to close integration of hetero-
geneous components. However, interposers are expensive and limited to only several components.
In any of the abovementioned integration approaches, the inter-die spacing is large. In interposers
the spacing is a few hundreds of micrometers, whereas the spacing between two packaged dies on
a PCB reaches tens of millimeters. The large inter-die spacing significantly increases the latency
and power of communication. In standard von Neumann architectures, where processor-memory
communication is a key bottleneck, the excessive performance degradation of communication is
especially limiting.
In the Si-IF technology, dies are integrated at high proximity (∼50 μm). This is enabled since
the dies are not packaged. The only limitations on interdielet spacing in the Si-IF platform are the
roughness of the edge of the die (due to wafer dicing), and edge seals and crack stops that are typi-
cally present at the edge of the die, effectively increasing the interdie spacing. Novel technologies,
such as plasma dicing [6], will enable further reduction of the interdie spacing on the Si-IF.

1.2.4 Die Size Considerations

Modern systems-on-chips (SoCs) are typically very large in area (excluding mobile SoCs). For
example, AMD’s Rome server SoC includes up to 1,000 mm2 of cumulative silicon area [7]. The
reason for such a large area is the attempt to integrate as many components as possible at the
IC level to leverage the small interconnect dimensions and intercomponent spacing. Large SoCs
integrated on interposers, are, however, prone to yield degradation and are typically expensive
to fabricate. The concept of heterogeneous integration aims to solve this problem by enabling
platforms that support IC-level interconnect dimensions at the package level and eliminate the
need for complex and expensive integration hierarchy, (e.g. the use of interposers).
 1.2 Technology Aspects of Heterogeneous Integration 7

Mechanical constraints (high) T sting complexity

Te
Testing Electrical/logical constraints (high)

CMOS wire‐like (low) I/O complexity/power SerDes‐like (high)

Reduced defect probability (high) Yield Increased defect probability (low)

Tiny components (difficult) Handling Large components (easy)

Basic functions (high) IP reuse Complex system (low)

100 nm 2 1–100 mm2 1,000 mm 2

Gate‐size die Optimal die size SoC‐size die

Figure 1.6 Parametric space to determine optimal die size for heterogeneous integration. The key plotted
parameters are: intellectual property (IP) reuse, die handling, yield, I/O complexity/power, and testing
complexity. Those parameters are plotted as a function of the die size. An optimal parametric space in the
center of the figure drives the optimal die size to be 1–100 mm2 .

Since integration at fine pitch and small inter-component spacing is enabled at the package level
by various heterogeneous integration platforms, the components (or dies) need not be very large.
Several parameters are key to determine the optimal die size for integration, as shown in Figure 1.6.
The optimal die size is represented by the light gray block in the center of Figure 1.6, driven by the
optimal parametric space. Following is a discussion of the parameters that determine the die size
for heterogeneous integration.

● IP reuse: Enables reduced nonrecurring engineering (NRE) cost and faster time to market. Abil-
ity to reuse IP significantly increases with smaller die size. The smaller the die, the simpler and
potentially more fundamental the function, the higher probability to reuse the die. For example,
at the leftmost side of the bar in Figure 1.6, a simple logic block (e.g. multiplexer), has a high
probability to be reused many times in a other projects or systems. Alternatively, a very large
system (e.g. an entire SoC), will not likely to be reused in other projects or systems. Therefore,
according to the IP reuse parameter, the smaller the die, the better.
● Handling: This parameter prefers large dies since handling very small dielets (smaller than
1 mm2 ) is difficult. Special tools, alignment techniques, and handling procedures are required
to handle such small dies. On the other hand, handling large dies is easy and established tooling
can be used.
● Yield: A key parameter and potentially the main driver of any fabrication facility. Statistically,
large dies are prone to high probability of defects leading to low yield and therefore high cost.
Alternatively, small dies typically exhibit a very high yield in an established process, driving down
the cost of the dies and the entire system. Small dies are therefore preferred to optimize the yield
parameter.
● I/O complexity/power: Rent’s rule drives this parameter, i.e. the number of I/Os is related to
the complexity of the component (further described in Section 1.2.5). Small components (or
dies) require fewer I/Os and therefore less I/O-related power. Integration of small dies sup-
ports local highly parallel communication, as compared to large dies that required high-speed
8 1 Heterogeneous Integration at Scale

serializer/deserializer (SerDes) communication, which is both power and area-hungry. Smaller

are therefore preferred to satisfy this parameter.
● Testing complexity: This is a unique parameter that is prohibitive for both very small and very
large dies. Small dies, although exhibit low complexity and require simpler testing approach, are
difficult to probe and expensive. Alternatively, large dies are highly complex requiring sophisti-
cated testing approaches and significantly limit testing at speed.

1.2.5 Dielet to Substrate Pitch Considerations

Rent’s rule [8] determines a relation between the number of I/Os of a chip and the complexity of
that chip. Specifically, the number of signal terminals T is related to the number of internal chip
components, g (e.g. gates, blocks)
T = t ⋅ gp (1.1)
where t and p are constants that represent the technology and circuit complexity. Typical values
of p are 0 < p < 1 and approach unity for low-complexity dies. In microprocessors, for example
the values of t and p are, respectively, 0.8 and 0.45 [9]. Current system integration technologies do
not support the number of I/O terminals that are required according to Rent’s rule. Power-hungry
SerDes circuitry is used to “bypass” Rent’s rule. In [2], the following expression is derived for the
pitch P of the I/Os between the die and substrate
( )p
4 AA
T
P= 1
(1.2)
t ⋅ Ap− 2
where A is the area of the die, and AT is the area per transistor (for the specific technology).
From (1.2) and assuming SerDes circuitry is to be eliminated, an I/O pitch of ∼3 to 7 μm will be
required for SoCs (in technologies of 45 nm and smaller). Note that this pitch is similar to the fat
wire pitches of ICs. Advanced heterogeneous integration platforms will, therefore, have to support
a similar range of vertical interconnect pitch between the dies and substrates. The Si-IF platform,
for example borrows heavily from standard Si fabrication techniques and supports a fine integration
pitch of 2–10 μm. Heterogeneous integration platforms that utilize solder-based vertical intercon-
nects will not be able to reduce the pitch to the required range, as previously discussed. The notion
of the fine pitch and small interdie spacing, positions the Si-IF as a natural platform to realize an
SoC-like system-on-wafer (SoW).
In addition to the power savings due to the elimination of SerDes, the fine I/O pitch on the Si-IF
supports a significant increase in the number of I/Os. The data rate of each I/O can, therefore,
be lower as compared to the data rate per I/O in current packaging technologies. The aggregate
bandwidth of the IBM POWER9 chip, for example is 1,206 GB/s [10] with 2,359 C4 pads dedicated
to differential signaling [11], resulting in a data rate per pad of 8.2 Gb/s. Whereas, if integrated
using a fine pitch of 3.5 μm [2], the number of pads dedicated to signaling could be increased by a
factor of approximately 35. The increase in the number of pads for signaling will result in reduced
data rate requirements per pad of 0.23 Gb/s to support the same aggregate bandwidth. Similar to
the signal I/Os, the number of I/Os dedicated to power delivery will significantly increase in the
Si-IF platform.

1.2.6 Backward Compatibility

Heterogeneous integration strongly promotes the concept of IP reuse leveraging components-off-
the-shelf (COTS). These components may be fabricated in various technologies with a wide range of
 1.3 Design and Architecture of Heterogeneous Integration Platforms 9

integration pad pitches and materials. Heterogeneous integration platforms must support integra-
tion of these disparate technologies within the same platform. For example, the Si-IF technology
supports integration of dies of different height, area, pad pitches, and attachment material, i.e.
solder-based and direct metal). This feature supports utilization of hardened legacy IP and elimi-
nates the need to redesign circuits in newer technologies unnecessarily, significantly reducing NRE.

1.3 Design and Architecture of Heterogeneous Integration

Platforms
Several heterogeneous integration platforms are at somewhat mature stages of fabrication technol-
ogy [5, 12]–[15]. Nonetheless, many system-level design and architectural challenges must still be
addressed. In this section, we will discuss power delivery and thermal management, floorplanning,
and communication for heterogeneous integration systems. These design aspects pose some of the
key challenges on the path to functional and efficient heterogeneous systems.

1.3.1 Power Delivery and Thermal Management

While power delivery design has recently shifted toward distributed power delivery schemes with
point-of-load (POL) voltage regulation, in modern heterogeneous ICs power is primarily managed
in centralized manner, at the expense of limited real-time control and significant power dissipation.
In large-scale heterogeneous integration platforms, where many dielets can be potentially inte-
grated, the physical horizontal distances become very large, rendering centralized power delivery
and management approaches impractical.
To enable efficient power supply within these large heterogeneous systems, a fundamental
change in power delivery and management approach is necessary. (i) Global high-voltage power
should be distributed across the wafer and locally converted and regulated within a wide range of
dynamically scaled supply voltages and high nonlinear load currents; (ii) Power should be man-
aged locally, with fine spatial, temporal, and voltage granularity; (iii) Local power management
decisions should be communicated across the system, optimizing system-wide power efficiency
[16]. The distributed power delivery and management approach is, however, area-hungry, requir-
ing space for local power regulators and limiting scaling of interdielet spacing and therefore
communication. Three potential power delivery approaches are, therefore, to be considered:
(1) Vertically integrated miniature voltage regulators: In this approach, novel high-efficiency
and high-current density miniature voltage regulators [17] are integrated on top of functional
dielets in a 3D manner. These regulators will receive power from peripheral connectors at high
voltage and convert and regulate locally. The Si-IF platform supports such integration as part of
the network on interconnect fabric (NoIF) methodology [18], where utility dies (UDs) are floor-
planned alongside with the functional dies on the wafer. The NoIF, with UDs as nodes within
the network, is expected to provide all of the services required by the heterogeneous system
integrated on the Si-IF, including power delivery. The voltage regulator dies can be integrated
on top of the UDs with power management circuitry located at the UDs. This approach supports
high quality of delivered power and high efficiency (i.e. low resistive losses), since relatively
low current is distributed in the long cross-wafer interconnect. For the same reason, heat dis-
sipation is also favorable in this approach. Locally, heat can be managed using heat sinks and
active cooling techniques. The Si-IF platform is effectively a Si wafer with significantly better
thermal conductivity as compared to classical organic packaging materials. The Si-IF serves,
Another random document with
no related content on Scribd:
 schools. … The only important amendment to this Act was passed
in 1875, and provided that the legislative grant, instead of
being divided between the Protestant and Catholic schools as
heretofore, should in future be distributed in proportion to
the number of children of school age in the Catholic and
Protestant districts. Already immigration had begun to upset
the balance of numbers and power, and as the years went on it
became evident that the Catholics were destined to be in a
permanent minority in Manitoba. This trend of immigration,
which in 1875 made legislation necessary, has continued ever
since; and to-day the Catholics of the province number only
20,000 out of a total population of 204,000. No further change
was made in the educational system of Manitoba until the
memorable year of 1890. In that year the provincial
legislature boldly broke all moorings with the past, and,
abolishing the separate denominational schools, introduced a
system of free compulsory and unsectarian schools, for the
support of which the whole community was to be taxed. … To
test the legality of the change, what is known as Barrett's
case was begun in Winnipeg. It was carried to the Supreme
Court of Canada, and the Canadian judges by a unanimous
decision declared that the Act of 1890 was ultra vires and
void.
{60}
The city of Winnipeg appealed to the Privy Council, and that
tribunal in July 1892 reversed the decision of the Canadian
Court and affirmed that the Act was valid and binding. … The
second subsection of the 22nd section of the Manitoba Act
already quoted says: 'An appeal shall lie to the
Governor-General in Council from any Act or decision of the
legislature of the province, or of any provincial authority,
affecting any right or privilege of the Protestant or Roman
Catholic minority of the Queen's subjects in relation to
education.' But if the legislation of 1890 was intra vires,
and expressly declared to be so on the ground that it had not
prejudicially affected the position which the minority held at
the time of the Union, how could there be an appeal from it? …
The Governor-General, however, consented to refer the question
as to his jurisdiction to the courts of justice. What is known
as Brophy's case was begun, and in due course was carried to
the Supreme Court of Canada. The decision of that tribunal,
though not unanimous, was in accord with public expectation.
The majority of the judges felt that the previous judgment of
the Privy Council had settled the matter beforehand. The Act
of 1890 had been declared intra vires on the ground that it
had not interfered with the rights which the minority
possessed before the Union, and therefore there could be no
appeal from it. …

"Still the undaunted Archbishop of St. Boniface went on, and

for a last time appealed to that Judicial Committee of the
Privy Council which two years and a half before had so spoiled
and disappointed the Catholic hopes. In January 1894 the final
decision in Brophy's case was read by the Lord Chancellor. For
a second time the Lords of the Council upset the ruling of the
Supreme Court of Canada, and treated their reasoning as
irrelevant. It will be remembered that both the appellant
prelates and the Canadian judges had assumed that the clause
in the Manitoba Act, which conferred the right of appeal to
the Governor-General, was limited to one contingency, and
could be invoked only if the minority were robbed at any time
of the poor and elementary rights which they had enjoyed
before the Act of Union. But was the clause necessarily so
limited? Could it not be used to justify an appeal from
legislation which affected rights acquired after the Union? …
In the words of the judgment: 'The question arose: Did the
sub-section extend to the rights and privileges acquired by
legislation subsequent to the Union? It extended in terms to
"any" right or privilege of the minority affected by any Act
passed by the legislature, and would therefore seem to embrace
all the rights and privileges existing at the time when such
Act was passed. Their lordships saw no justification for
putting a limitation on language thus unlimited. There was
nothing in the surrounding circumstances or in the apparent
intention of the legislature to warrant any such limitation.'
… In other words, the dispute was referred to a new tribunal,
and one which was free to consider and give effect to the true
equities of the case. The Governor-General and his responsible
advisers, after considering all the facts, found in favour of
the Catholic minority, and at once issued a remedial Order to
the Government of Manitoba, which went far beyond anything
suggested in the judgment in Brophy's case. The province was
called upon to repeal the legislation of 1890, so far as it
interfered with the right of the Catholic minority to build
and maintain their own schools, to share proportionately in
any public grant for the purposes of education, and with the
right of such Catholics as contributed to Catholic schools to
be held exempt from all payments towards the support of any
other schools. In a word, the Governor-General and Sir
Mackenzie Bowell's Administration, exercising, as it were,
appellate jurisdiction, decided that the minority were
entitled to all they claimed. The Government of Manitoba,
however, had hardened their hearts against the minority in the
province, and refused to obey the remedial Order. …

"The refusal of the provincial Government 'to accept the

responsibility of carrying into effect the terms of the
remedial Order' for the first time brought the Parliament of
Canada into the field, and empowered them to pass coercive
legislation. A remedial Bill was accordingly, after an
inexplicable delay, brought into the Federal Parliament to
enforce the remedial Order. … The Cabinet recognised that the
Federal Parliament had no power to spend the money of the
province, and so all they could do was to exempt the minority
from the obligation to contribute to the support of schools
other than their own. The Bill bristled with legal and
constitutional difficulties; it concerned the coercion of a
province; it contained no less than 116 clauses; it was
introduced on the 2nd of March 1896, when all Canada knew that
the life of the Federal Parliament must necessarily expire on
the 24th of April. Some fifteen clauses had been considered
 when the Government admitted, what all men saw, the
impossibility of the task, and abandoned the Bill. … While the
fate of the remedial Bill was still undecided, Sir Donald
Smith and two others were commissioned by the Federal
Government to go to Winnipeg and see if by direct negotiations
some sort of tolerable terms could be arranged. … Sir Donald
Smith proposed that the principle of the separate school
should be admitted wherever there were a reasonable number of
Catholic children—thus, wherever in towns and villages there
are twenty-five Catholic children of school age, and in
cities where there are fifty such children, they should have
'a school-house or school-room for their own use,' with a
Catholic teacher. … In the event the negotiations failed; the
baffled Commissioners returned to Ottawa, and on the 24th of
April 1896 Parliament was dissolved. The Government went to
the country upon the policy of the abandoned Bill. On the
other hand, many of the followers of Mr. Laurier in the
province of Quebec pledged themselves to see justice done to
the Catholics of Manitoba, and let it be understood that they
objected to the remedial Bill only because it was not likely
to prove effective in the face of the combined hostility of
the legislature and the municipalities of the province. …
Catholic Quebec gave Mr. Laurier his majority at Ottawa. …

{61}

"When the Liberal party for the first time for eighteen years
found itself in power at Ottawa, Mr. Laurier at once opened
negotiations with Manitoba. The result was a settlement which,
although it might work well in particular districts, could not be
accepted as satisfactory by the Catholic authorities. It arranged
that where in towns and cities the average attendance of
Catholic children was forty or upwards, and in villages and
rural districts the average attendance of such children was
twenty-five or upwards, one Catholic teacher should be
employed. There were various other provisions, but that was
the central concession. … Leo the Thirteenth, recognising the
 difficulties which beset Mr. Laurier's path, mindful, perhaps,
also that it is not always easy immediately to resume friendly
conference with those who have just done their best to defeat
you, has sent to Canada an Apostolic Commissioner."

J. G. Snead Cox,
Mr. Laurier and Manitoba
(Nineteenth Century, April, 1897).

CANADA: A. D. 1895.
Northern territories formed into provisional districts.

"The unorganized and unnamed portion of the Dominion this year

was set apart into provisional districts. The territory east of
Hudson's Bay, having the province of Quebec on the south and
the Atlantic on the east, was to be hereafter known as Ungava.
The territory embraced in the islands of the Arctic Sea was to
be known as Franklin, the Mackenzie River region as Mackenzie,
and the Pacific coast territory lying north of British
Columbia and west of Mackenzie as Yukon. The extent of Ungava
and Franklin was undefined. Mackenzie would cover 538,600
square miles, and Yukon 225,000 square miles, in addition to
143,500 square miles added to Athabasca and 470,000 to
Keewatin. The total area of the Dominion was estimated at
3,456,383 square miles."

The Annual Register, 1895,

page 391.

CANADA: A. D. 1895.
Negotiations with Newfoundland.

Negotiations for the entrance of Newfoundland into the

federation of the Dominion of Canada proved ineffectual and
were abandoned in May. The island province refused the terms
proposed.
CANADA: A. D. 1896 (June-July).
Liberal triumph in Parliamentary elections.
Formation of Ministry by Sir Wilfred Laurier.

General elections held in Canada on the 23d of June, 1896,

gave the Liberal Party 113 seats out of 213 in the Dominion
House of Commons; the Conservatives securing 88, and the
Patrons of Industry and other Independents 12. Much to the
general surprise, the scale was turned in favor of the
Liberals by the vote of the province of Quebec,
notwithstanding the Manitoba school question, on which
clerical influence in the Roman church was ranged against that
party. The effect of the election was to call the Liberal
leader, Sir Wilfred Laurier, of Quebec, to the head of the
government, the Conservative Ministry, under Sir Charles
Tupper, retiring on the 8th of July.

CANADA: A. D. 1896-1897.
Policy of the Liberal Government.
Revision of the tariff, with discriminating duties
in favor of Great Britain, and provisions for reciprocity.

"The position of the Canadian Liberals, when they came into

power after the General Election of 1896, was not unlike that
of the English Liberals after the General Election of 1892.
Both Liberal parties had lists of reforms to which they were
committed. The English measures were in the Newcastle
Programme. Those of the Canadian Liberals were embodied in the
Ottawa Programme, which was formulated at a convention held at
the Dominion Capital in 1893. … A large part of the Ottawa
Programme was set out in the speech which the Governor-General
read in the Senate when the session of 1897 commenced. There was
then promised a measure for the revision of the tariff; a bill
providing for the extension of the Intercolonial railway from
Levis to Montreal; a bill repealing the Dominion Franchise Act
and abolishing the costly system of registration which goes
with it; and a measure providing for the plebiscite on the
Prohibition question. Neither of these last two measures was
carried through Parliament. Both had to be postponed to
another session; and the session of 1897 was devoted, so far
as legislation went, chiefly to the tariff, and to bills, none
of which were promised in the Speech from the Throne, in
retaliation for the United States Contract Labor Laws, and the
new United States tariff. …

"The new tariff was a departure from the tariffs of the

Conservative regime in only one important direction.
Protective duties heretofore had been levied on imports from
England, in the same way as on imports from the United States
or any other country. The 'National Policy' had allowed of no
preferences for England; and during the long period of
Conservative rule, when the Conservatives were supported by
the Canadian manufacturers in much the same way as the
Republican party in the United States is supported by the
manufacturing interests, the Canadian manufacturers had been
as insistent for adequate protection against English-made
goods, as against manufactured articles from the United States
or Germany. The Conservative party had continuously claimed a
monopoly of loyalty to England; but in its tariffs had never
dared to make any concession in favour of English goods. In
the new tariff, preferences for England were established; and
with these openings in favour of imports from Great Britain,
there came a specific warning from the Minister of Finance
that Canadian manufacturers must not regard themselves as
possessing a vested interest in the continuance of the
protective system. …

"When the Minister of Finance laid the tariff before the House
of Commons, he declared that the 'National Policy,' as it had
been tried for eighteen years, was a failure; and … claimed
that lowering the tariff wall against England was a step in
the direction of a tariff 'based not upon the protective
system but upon the requirements of the public service.'
During the first fifteen months of the new tariff, the
 concession to England consists of a reduction by one-eighth of
the duties chargeable under the general list. At the end of
that time, that is on the last of July, 1898, the reduction
will be one-fourth. The reductions do not apply to wines, malt
liquors, spirits and tobacco, the taxes on which are
essentially for revenue. While England was admitted at once to
the advantages of the reduced tariff, this tariff is not to be
applicable to England alone. In July, it was extended to the
products of New South Wales, the free-trade colony of the
British Australasian group; and any country can come within
its provisions whose government can satisfy the Comptroller of
Customs at Ottawa, that it is offering favourable treatment to
Canadian exports, and is affording them as easy an entrance
through its customs houses as the Canadians give by means of
the reciprocal tariff. It is also possible, under a later
amendment to the Tariff Act, for the Governor in Council to
extend the benefits of the reciprocal tariff to any country
entitled thereto by virtue of a treaty with Great Britain.
{62}
Numerous alterations were made in the general list of import
duties. Some of these involved higher rates; others lowered
the duties. But if the changes in the fiscal system had been
confined to these variations, the new tariff would not have
been noteworthy, and it would have fulfilled few of the
pledges made by the Liberals when they were in Opposition. It
owes its chief importance to the establishment of an inner
tariff in the interests of countries which deal favourably
with Canada."

E. Porritt,
The New Administration in Canada
(Yale Review, August, 1897).

CANADA: A. D. 1897 (June-July).

Conference of colonial premiers with
the British Colonial Secretary.
 See (in this volume)
ENGLAND: A. D. 1897 (JUNE-JULY).

CANADA: A. D. 1897 (October).

Self-government for the Northwestern Territories.

By an Act passed in October, a system of self-government,

going far towards the full powers of a provincial government,
but having some limitations, was provided for the Northwest
Territories.

CANADA: A. D. 1898 (January).

Encyclical Letter of the Pope on the Manitoba School Question.

On the report made by his delegate, Monsignor Merry del Val,

Pope Leo XIII. addressed an encyclical letter to the Roman
Church in Canada, concerning the duty of Catholics in the
matter of the Manitoba schools (see above: A. D. 1890-1896),
which was made public at Quebec on the 9th of January, 1898.
The letter has great general importance, as defining with
precision the attitude of the Church towards all secular
school systems. With a few unessential passages it is given in
what follows:

"It was with extreme solicitude," wrote the Pope, "that we

turned our mind to the unhappy events which in these later
years have marked the history of Catholic education in
Manitoba. … And since many expected that we should make a
pronouncement on the question, and asked that we should trace
a line of conduct and a way to be followed, we did not wish to
decide anything on this subject before our Apostolic delegate
had been on the spot, charged to proceed to a serious
examination of the situation, and to give an account to us of
the state of affairs. He has faithfully and diligently
fulfilled the command which we had given him. The question
agitated is one of great and exceptional importance. We speak
of the decision taken seven years ago by the parliament of
Manitoba on the subject of education. The act of Confederation
had secured to Catholic children the right of education in public
schools in keeping with their conscientious convictions. The
parliament of Manitoba abolished this right by contrary law.
By this latter law a grave injury was inflicted, for it was
not lawful for our children to seek the benefits of education
in schools in which the Catholic religion is ignored or
actively combated, in schools where its doctrine is despised
and its fundamental principles repudiated. If the Church has
anywhere permitted this, it was only with great reluctance and
in self-defense, and after having taken many precautions,
which, however, have too often been found unequal to parrying
the danger. In like manner one must at all cost avoid, as most
pernicious, those schools wherein every form of belief is
indifferently admitted and placed on an equal footing—as if in
what regards God and Divine things, it was of no importance
whether one believed rightly or wrongly, whether one followed
truth or falsehood. You well know, venerable brothers, that
all schools of this kind have been condemned by the Church,
because there can be nothing more pernicious nor more fitted
to injure the integrity of faith and to turn away the tender
minds of youth from the truth. … For the Catholic there is but
one true religion, the Catholic religion; hence in all that
concerns doctrine, or morality, or religion, he cannot accept
or recognize anything which is not drawn from the very sources
of Catholic teaching. Justice and reason demand, then, that
our children have in their schools not only scientific
instruction but also moral teachings in harmony, as we have
already said, with the principles of their religion, teachings
without which all education will be not only fruitless but
absolutely pernicious. Hence the necessity of having Catholic
teachers, reading books, and textbooks approved of by the
bishops, and liberty to organize the schools, that the
teaching therein shall be in full accord with Catholic faith
as well as with all the duties that flow therefrom. For the
rest, to decide in what institutions their children shall be
instructed, who shall be their teachers of morality, is a
right inherent to parental authority. When, then, Catholics
demand, and it is their duty to demand, and to strive to
obtain, that the teaching of the masters shall be in
conformity with the religion of their children, they are only
making use of their right; and there can be nothing more
unjust than to force on them the alternative of allowing their
children to grow up in ignorance, or to expose them to
manifest danger in what concerns the supreme interests of
their souls. It is not right to call in doubt or to abandon in
any way these principles of judging and acting which are
founded on truth and justice, and which are the safe-guards
both of public and private interests. Therefore, when the new
law in Manitoba struck a blow at Catholic education, it was
your duty, venerable brothers, to freely protest against the
injury and disaster inflicted; and the way in which you all
fulfilled that duty is a proof of your common vigilance, and
of a spirit truly worthy of bishops; and, although each one of
you will find on this point a sufficient approbation in the
testimony of his own conscience, learn, nevertheless, that you
have also our conscience and our approbation, for the things
which you sought and still seek to protect and defend are most
sacred. The difficulties created by the law of which we speak by
their very nature showed that an alleviation was to be sought
for in a united effort. For so worthy was the Catholic cause
that all good and upright citizens, without distinction of
party, should have banded themselves together in a close union
to uphold it. Unfortunately for the success of this cause, the
contrary took place. What is more deplorable still, is that
Catholic Canadians themselves failed to unite as they should
in defending those interests which are of such importance to
all—the importance and gravity of which should have stilled
the voice of party politics, which are of much less
importance. We are not unaware that something has been done to
amend that law. The men who are at the head of the federal
government and of the Province of Manitoba have already taken
certain measures with a view to decreasing the difficulties of
which the Catholics of Manitoba complain, and against which
 they rightly continue to protest.
{63}
We have no reason to doubt that these measures were taken from
love of justice and from a laudable motive. We cannot, however,
dissimulate the truth; the law which they have passed to
repair the injury is defective, unsuitable, insufficient. The
Catholics ask—and no one can deny that they justly ask—for
much more. Moreover, in the remedial measures that have been
proposed there is this defect, that in changes of local
circumstances they may easily become valueless. In a word, the
rights of Catholics and the education of their children have
not been sufficiently provided for in Manitoba. Everything in
this question demands, and is conformable to justice, that
they should be thoroughly provided for, that is, by placing in
security and surrounding with due safe-guards those
unchangeable and sacred principles of which we have spoken
above. This should be the aim, this the end to be zealously
and prudently sought for. Nothing can be more injurious to the
attainment of this end than discord; unity of spirit and
harmony of action are most necessary. Nevertheless since, as
frequently happens in things of this nature, there is not only
one fixed and determined but various ways of arriving at the
end which is proposed and which should be obtained, it follows
that there may be various opinions equally good and
advantageous. Wherefore let each and all be mindful of the
rules of moderation, and gentleness, and mutual charity; let
no one fail in the respect that is due to another; but let all
resolve in fraternal unanimity, and not without your advice,
to do that which the circumstances require and which appears
best to be done. As regards especially the Catholics of
Manitoba, we have every confidence that with God's help they
will succeed in obtaining full satisfaction. This hope is
founded, in the first place, in the righteousness of the
cause, next in the sense of justice and prudence of the men at
the head of the government, and finally in the good-will of all
upright men in Canada. In the meantime, until they are able to
obtain their full rights, let them not refuse partial
 satisfaction. If, therefore, anything is granted by law to
custom, or the good-will of men, which will render the evil
more tolerable and the dangers more remote, it is expedient
and useful to make use of such concessions, and to derive
therefrom as much benefit and advantage as possible. Where,
however, no remedy can be found for the evil, we must exhort
and beseech that it be provided against by the liberality and
munificence of their contributions, for no one can do anything
more salutary for himself or more conducive to the prosperity
of his country, than to contribute, according to his means, to
the maintenance of these schools. There is another point which
appeals to your common solicitude, namely, that by your
authority, and with the assistance of those who direct
educational institutions, an accurate and suitable curriculum
of studies be established, and that it be especially provided
that no one shall be permitted to teach who is not amply
endowed with all the necessary qualities, natural and
acquired, for it is only right that Catholic schools should be
able to compete in bearing, culture, and scholarship with the
best in the country. As concerns intellectual culture and the
progress of civilization, one can only recognize as
praiseworthy and noble the desire of the provinces of Canada
to develop public instruction, and to raise its standard more
and more, in order that it may daily become higher and more
perfect. Now there is no kind of knowledge, no perfection of
learning, which cannot be fully harmonized with Catholic
doctrine."

CANADA: A. D. 1898 (September).

Popular vote on the question of Prohibition.

Pursuant to a law passed by the Dominion Parliament the

previous June, a vote of the people in all the Provinces of
the Dominion was taken, on the 29th of September, 1898, upon
the following question: "Are you in favor of the passing of an
act prohibiting the importation, manufacture or sale of
spirits, wine, ale, beer, cider, and all other alcoholic
 liquors for use as beverages?" The submitting of this question
to a direct vote of the people was a proceeding not quite
analogous to the Swiss Referendum, since it decided the fate
of no pending law; nor did it imitate the popular Initiative
of Swiss legislation, since the result carried no mandate to
the government. It was more in the nature of a French
Plébiscite, and many called it by that name; but no Plebiscite
in France ever drew so real an expression of popular opinion
on a question so fully discussed. The result of the voting was
a majority for prohibition in every Province except Quebec,
Ontario pronouncing for it by more than 39,000, Nova Scotia by
more than 29,000, New Brunswick by more than 17,000, Manitoba
by more than 9,000, Prince Edward's Island by more than 8,000,
and the Northwest Territories by more than 3,000, while
British Columbia gave a small majority of less than 600 on the
same side. Quebec, on the other hand, shouted a loud "No" to
the question, by 93,000 majority. The net majority in favor of
Prohibition was 107,000. The total of votes polled on the
question was 540,000. This was less than 44 per cent of the
total registration of voters; hence the vote for Prohibition
represented only about 23 per cent of the electorate, which
the government considered to offer too small a support for the
measure asked for.

CANADA: A. D. 1898-1899.
The Joint High Commission for settlement of all unsettled
questions between Canada and the United States.

As the outcome of negotiations opened at Washington in the

previous autumn by the Canadian Premier, relative to the
seal-killing controversy, an agreement between Great Britain,
Canada and the United States was concluded on the 30th of May,
1898, for the creation of a Joint High Commission to negotiate
a treaty, if possible, by which all existing subjects of
controversy between the United States and Canada should be
settled with finality. Appointments to the Commission by the
three governments were made soon afterwards, Great Britain
 being represented by the Lord High Chancellor, Baron
Herschell; Canada by Sir Wilfred Laurier, Premier, Sir Richard
Cartwright, Minister of Trade and Commerce, and Sir Louis
Henry Davies, Minister of Marine and Fisheries; the United
States by Honorable John W. Foster, ex-Secretary of State,
Senator Charles W. Fairbanks, Senator George Gray,
Representative Nelson Dingley, and the Honorable John A.
Kasson, Reciprocity Commissioner. Senator Gray having been
subsequently appointed on the Commission to negotiate peace
with Spain, his place on the Anglo-American Commission was
taken by Senator Faulkner.
{64}
The Joint Commission sat first in Quebec and later in
Washington. Among the questions referred to it were those
relating to the establishment of the boundary between Alaska
and British Columbia; the issues over Bering Sea and the catch
of fur seals; the unmarked boundary between Canada and the
United States near Passamaqnoddy Bay in Maine and at points
between Wisconsin and Minnesota and Canada; the northeast
fisheries question, involving the rights of fishing in the
North Atlantic off Newfoundland and other points; the
regulation of the fishing rights on the Great Lakes;
alien-labor immigration across the Canadian-American border;
commercial reciprocity between the two countries; the
regulation of the bonding system by which goods are carried in
bond across the frontier and also the regulation of traffic by
international railways and canals of the two countries;
reciprocal mining privileges in the Klondyke, British North
America and other points; wrecking and salvage on the ocean
and Great Lakes coasting waters; the modification of the
treaty arrangement under which only one war vessel can be
maintained on the Great Lakes, with a view to allowing
warships to be built on the lakes and then floated out to the
ocean. The sessions of the Joint Commission were continued at
intervals until February, 1899, when it adjourned to meet at
Quebec in the following August, unless further adjournment
should be agreed upon by the several chairmen. Such further
 adjournment was made, and the labors of the Joint Commission
were indefinitely suspended, for reasons which the President
of the United States explained in his Message to Congress,
December, 1899, as follows: "Much progress had been made by
the Commission toward the adjustment of many of these
questions, when it became apparent that an irreconcilable
difference of views was entertained respecting the
delimitation of the Alaskan boundary. In the failure of an
agreement as to the meaning of articles 3 and 4 of the treaty
of 1825 between Russia and Great Britain, which defined the
boundary between Alaska and Canada, the American Commissioners
proposed that the subject of the boundary be laid aside and
that the remaining questions of difference be proceeded with,
some of which were so far advanced as to assure the
probability of a settlement. This being declined by the
British Commissioners, an adjournment was taken until the
boundary should be adjusted by the two Governments. The
subject has been receiving the careful attention which its
importance demands, with the result that a modus vivendi for
provisional demarcations in the region about the head of Lynn
Canal has been agreed upon [see (in this volume) ALASKA
BOUNDARY QUESTION] and it is hoped that the negotiations now
in progress between the two Governments will end in an
agreement for the establishment and delimitation of a
permanent boundary."

CANADA: A. D. 1899 (October).

Modus Vivendi, fixing provisional boundary line of Alaska.

See (in this volume)

ALASKA BOUNDARY QUESTION.

CANADA: A. D. 1899-1900.
Troops to reinforce the British army in South Africa.

A proposal from the Canadian government to assist that of the

Empire in its South African War was gratefully accepted in the
 early stages of the war, and a regiment of infantry called the
Royal Canadian, numbering a little more than 1,000 men, sailed
from Quebec, October 30. In the following January a second
contingent of more than 1,000 men was sent to the field. This
latter comprised squadrons of mounted rilles and rough-riders,
and three batteries of field artillery. In the same month the
Canadian government accepted an offer from Lord Strathcona to
raise, equip and transport at his own expense a body of 500
mounted men from the Northwest.

CANADA: A. D. 1900 (November).

General election.

The general election of members of the Dominion House of

Commons was held November 7, resulting as follows:

Provinces. Liberal. Conservative.

Independent. Total.

Nova Scotia. 15 5
0 20
New-Brunswick. 9 5
0 14
Prince Edward Island. 3 2
0 5
Quebec. 57 8
0 65
Ontario. 33 54
5 92
Manitoba. 2 3
2 7
Northwest Territories. 2 0
2 4
British Columbia. 3 2
1 6

Totals. 124 79
10 213

As in the election of 1896, the Liberal Ministry of Sir

Wilfred Laurier found its strong support in the province of
Quebec. Its party suffered unexpected losses in Ontario. The
slight meaning of the election was summed up by Professor
Goldwin Smith as follows: "The net result of the elections
seems to be a Government resting on French Quebec and an
Opposition resting on British Ontario. The minor provinces
have been carried, as usual, by local interests rather than on
general questions. Apart from the distinction of race between
the two great provinces and the antagonism, before dormant but
somewhat awakened by the war, there was no question of importance
at issue between the parties. Both concurred in sending
contingents to South Africa. The Liberals, though they went in
at first on the platform of free trade—at least, of a tariff
for revenue only—have practically embraced protection under
the name of stability of the tariff, and are believed to have
received from the protected manufacturers contributions to
their large election fund. The other special principles, such
as the reduction of expenditure and discontinuance of the
bonus to railways, proclaimed by Liberals before the last
election, have been dropped. So has reform of the Senate. It
is not likely that the Liberal victory will be followed by any
change either in legislation or government, or by any special
reform. Mr. Bourassa and Monet, of the French-Canadian members
who protested against the contingent, have been re-elected.
Great as may be the extent and warmth of British feeling, the
statement that Canadians were unanimously in favour of
participation in the war must not be taken without
qualification. For myself, I felt that so little principle was
at stake that I voted for two Conservatives on their personal
merits."
{65}

CANAL, The new Bruges.

See (in this volume)

BRUGES: A. D. 1900.

CANAL, The Chicago Drainage.

See (in this volume)

CHICAGO: A. D. 1900.

CANAL, City of Mexico Drainage.

See (in this volume)

MEXICO: A. D. 1898.

CANAL, The Elbe and Trave.

See (in this volume)

GERMANY: A. D. 1900 (JUNE).

CANAL, Interoceanic, The Project of the: A. D. 1581-1892.

The early inception of the project.
Movements towards its realization.

"The thought of uniting the two great oceans by means of a

canal across the American isthmus sprang up, as is known, from
the moment the conviction was reached that the passage which,
from the days of Columbus, was thought to exist towards the
Southern Sea, was not a reality. … Nevertheless the first
survey of the land was not carried out until the year 1581,
when, in obedience to superior instructions, Captain Antonio
Pereira, Governor of Costa Rica, organized an expedition and
explored the route by way of the San Juan river, the lake, and
the rivers emptying into Gulf Nicoya, Costa Rica. Thirty-nine
years later Diego de Mercado submitted to King Philip III his
famous report of January 23, 1620, suggesting the route by the
river and lake, and thence through Costa Rican territory along
the Quebrada or Barranca Honda to Salinas Bay, then called
Puerto del Papagayo. Either because the magnitude of the
undertaking was at that time superior to the necessities of
trade, or, as was said, because Spain considered the canal
antagonistic to her interests, the era of independence arrived
without the execution of the project ever having been entered
upon. After independence the Congress of Central America, in
which Costa Rica and Nicaragua were represented as States of
the Federation which succeeded the Colonial Government,
enacted on June 16, 1825, a decree providing for the
construction of the canal, and in that same year Don Antonio
José Cañas, Diplomatic Representative of Central America in
Washington, addressed the Secretary of State, Mr. Henry Clay,
informing him of this resolution and stating that: 'A company
formed of American citizens of respectability was ready to
undertake the work as soon as a treaty with the United States
insuring the coöperation of the latter was signed; that he was
ready to enter into negotiations for the treaty, and that
nothing would be more pleasant for Central America than to see
the generous people of the United States joining her in the
opening of the canal, sharing the glory of the enterprise, and
enjoying the great advantages to be derived from it.' The
Government of Central America could not carry the undertaking
into effect, notwithstanding that among the means employed to
reach the desired result there figures the arrangement
concluded with the King of Holland in October, 1830. But,
though the hopes centered in the undertaking were frustrated,
to the honor of Central America the declarations of that
Congress, which constitute, like the concession for the canal
itself, one of the loftiest public documents ever issued by
any nation of the earth, have become a matter of record. The
Central American Federation dissolved, this important matter
attached to Nicaragua and Costa Rica directly, and the
boundary line between the two republics having been determined