Visit https://ebooknice.

com to download the full version and

explore more ebooks

(Ebook) High-Speed Photonic Devices by Nadir Dagli

(Editor) ISBN 9780367390273, 9780367850326,
9780429144103, 9780750308892, 9781420012071,
0367390272, 036785032X, 0429144105, 0750308893
_____ Click the link below to download _____
https://ebooknice.com/product/high-speed-photonic-
devices-12194334

Explore and download more ebooks at ebooknice.com

Here are some recommended products that might interest you.
You can download now and explore!

(Ebook) Biota Grow 2C gather 2C cook by Loucas, Jason; Viles, James

ISBN 9781459699816, 9781743365571, 9781925268492, 1459699815,
1743365578, 1925268497

https://ebooknice.com/product/biota-grow-2c-gather-2c-cook-6661374

ebooknice.com

(Ebook) Matematik 5000+ Kurs 2c Lärobok by Lena Alfredsson, Hans

Heikne, Sanna Bodemyr ISBN 9789127456600, 9127456609

https://ebooknice.com/product/matematik-5000-kurs-2c-larobok-23848312

ebooknice.com

(Ebook) SAT II Success MATH 1C and 2C 2002 (Peterson's SAT II Success)

by Peterson's ISBN 9780768906677, 0768906679

https://ebooknice.com/product/sat-ii-success-
math-1c-and-2c-2002-peterson-s-sat-ii-success-1722018

ebooknice.com

(Ebook) Master SAT II Math 1c and 2c 4th ed (Arco Master the SAT
Subject Test: Math Levels 1 & 2) by Arco ISBN 9780768923049,
0768923042

https://ebooknice.com/product/master-sat-ii-math-1c-and-2c-4th-ed-
arco-master-the-sat-subject-test-math-levels-1-2-2326094

ebooknice.com
(Ebook) Cambridge IGCSE and O Level History Workbook 2C - Depth Study:
the United States, 1919-41 2nd Edition by Benjamin Harrison ISBN
9781398375147, 9781398375048, 1398375144, 1398375047

https://ebooknice.com/product/cambridge-igcse-and-o-level-history-
workbook-2c-depth-study-the-united-states-1919-41-2nd-edition-53538044

ebooknice.com

(Ebook) High-Speed Photonic Devices (Series in Optics and

Optoelectronics) by Nadir Dagli ISBN 0750308893

https://ebooknice.com/product/high-speed-photonic-devices-series-in-
optics-and-optoelectronics-2209358

ebooknice.com

(Ebook) Vagabond, Vol. 29 (29) by Inoue, Takehiko ISBN 9781421531489,

1421531488

https://ebooknice.com/product/vagabond-vol-29-29-37511002

ebooknice.com

(Ebook) Semiconductor Devices for High-Speed Optoelectronics by

Giovanni Ghione ISBN 9780521763448, 0521763444

https://ebooknice.com/product/semiconductor-devices-for-high-speed-
optoelectronics-1433634

ebooknice.com

(Ebook) Electro-optic Modulation for Photonic Networks: Precise and

high-speed control of lightwaves by Kawanishi, Tetsuya ISBN
9783030867195, 3030867196

https://ebooknice.com/product/electro-optic-modulation-for-photonic-
networks-precise-and-high-speed-control-of-lightwaves-37441480

ebooknice.com
High Speed Photonic Devices 1st Edition Nadir Dagli
(Editor) Digital Instant Download
Author(s): Nadir Dagli (Editor)
ISBN(s): 9781420012071, 0750308893
Edition: 1
File Details: PDF, 5.43 MB
Year: 2006
Language: english
IP369_C000.fm Page i Friday, August 25, 2006 1:43 PM

High-Speed Photonic Devices

IP369_C000.fm Page ii Friday, August 25, 2006 1:43 PM

Series in Optics and Optoelectronics

Series Editors: R G W Brown, University of Nottingham, UK
E R Pike, Kings College, London, UK

Other titles in the series

Lasers in the Preservation of Cultural Heritage

C Fotakis, D Anglos, V Zafiropulos, S Georgiou, V Tornari
Modeling Fluctuations in Scattered Waves
E Jakeman, K D Ridley
Fast Light, Slow Light and Left-Handed Light
P W Milonni
Diode Lasers
D Sands
Diffractional Optics of Millimetre Waves
I V Minin
Handbook of Electroluminescent Materials
D R Vij
Handbook of Moire Measurement
C A Walker
Next Generation Photovoltaics
A Martí
Stimulated Brillouin Scattering
M J Damzen
Laser Induced Damage of Optical Materials
Roger M Wood
Optical Applications of Liquid Crystals
L Vicari
Optical Fibre Devices
J P Goure
Applications of Silicon-Gremanium Heterostructure Devices
C K Maiti
Optical Transfer Function of Imaging Systems
Tom L Williams
Super-radiance
M G Benedict
IP369_C000.fm Page iii Friday, August 25, 2006 1:43 PM

Series in Optics and Optoelectronics

High-Speed
Photonic Devices

Nadir Dagli
University of California
Santa Barbara, USA

New York London

Taylor & Francis is an imprint of the

Taylor & Francis Group, an informa business
IP369_C000.fm Page iv Friday, August 25, 2006 1:43 PM

CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
© 2007 by Taylor & Francis Group, LLC
CRC Press is an imprint of Taylor & Francis Group, an Informa business

No claim to original U.S. Government works

Printed in the United States of America on acid-free paper
10 9 8 7 6 5 4 3 2 1

International Standard Book Number-10: 0-7503-0889-3 (Hardcover)

International Standard Book Number-13: 978-0-7503-0889-2 (Hardcover)

This book contains information obtained from authentic and highly regarded sources. Reprinted
material is quoted with permission, and sources are indicated. A wide variety of references are
listed. Reasonable efforts have been made to publish reliable data and information, but the author
and the publisher cannot assume responsibility for the validity of all materials or for the conse-
quences of their use.

No part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any
electronic, mechanical, or other means, now known or hereafter invented, including photocopying,
microfilming, and recording, or in any information storage or retrieval system, without written
permission from the publishers.

For permission to photocopy or use material electronically from this work, please access www.
copyright.com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC)
222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that
provides licenses and registration for a variety of users. For organizations that have been granted a
photocopy license by the CCC, a separate system of payment has been arranged.

Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and
are used only for identification and explanation without intent to infringe.

Library of Congress Cataloging-in-Publication Data

Dagli, N. (Nadir)
High-speed photonic devices / author : N. Dagli.
p. cm.
Includes bibliographical references and index.
ISBN-13: 978-0-7503-0889-2 (alk. paper)
ISBN-10: 0-7503-0889-3 (alk. paper)
1. Optoelectronic devices. 2. Very high speed integrated circuits. I. Title.

TK8304.D33 2006
621.381’045--dc22 2006011669

Visit the Taylor & Francis Web site at

http://www.taylorandfrancis.com
and the CRC Press Web site at
http://www.crcpress.com
IP369_C000.fm Page v Friday, August 25, 2006 1:43 PM

Editor
Nadir Dagli received his Ph.D. degree in electrical engineering from Massachusetts
Institute of Technology in 1987.
After graduation he joined the electrical and computer engineering department
at University of California at Santa Barbara, where he is currently a professor. He
has done pioneering work on finite difference beam propagation method; ultra fast
substrate removed compound semiconductor electro-optic modulators, slow wave
electrodes for efficient optical modulators, electron waveguides and couplers, total
internal reflection mirrors for compact photonic integrated circuits and compact ring
resonators integrated with semiconductor optical amplifiers. His current interests are
design, fabrication and modeling of guided-wave components for optical inte-
grated circuits, ultrafast electro-optic modulators, WDM components, and photonic
nanostructures.
Dagli was awarded NATO science and IBM predoctoral fellowships during his
graduate studies. He received the 1990 UCSB Alumni Distinguished Teaching Award
and 1990 UC Regents Junior Faculty Fellowship. He served as a technical program
committee member, program and conference chair, and advisory board member of
many international conferences including the Integrated Photonics Research Con-
ference, SPIE’s Photonics West, International Topical Meeting on Microwave Pho-
tonics, CLEO, CLEO/Pacific Rim, IEEE Lasers and Electro Optics Society Annual
Meeting. He was the associate editor of IEEE Photonics Technology Letters from
1997 to 2000 and the editor-in-chief of IEEE Photonics Technology Letters from
2000 to 2005. He also served as an elected member of the IEEE-LEOS board of
governors from 2003 to 2005. He is a fellow of IEEE.
IP369_C000.fm Page vi Friday, August 25, 2006 1:43 PM
IP369_C000.fm Page vii Friday, August 25, 2006 1:43 PM

Contributors
Nadir Dagli Taiichi Otsuji
Electrical and Computer Engineering Center for Microelectronic Systems,
Department Faculty of Computer Science and
University of California at Santa Systems Engineering
Barbara Kyushu Institute of Technology
Santa Barbara, CA Fukuoka, Japan
P. Fay
Department of Electrical Masatoshi Saruwatari
Engineering National Defense Academy
University of Notre Dame Yokosuka-shi, Kanagawa Japan
Notre Dame, IN
Harold R. Fetterman William H. Steier
Department of Electrical Department of Electrical
Engineering Engineering
University of California, University of Southern California
Los Angeles Los Angeles, CA
Los Angeles, CA
Osamu Mitomi Koichi Wakita
Ubiquitous Network Group Department of Electronic Engineering
NGK Insulators, LTD Chubu University
Nagoya, Japan Kasugai, Aichi, Japan
IP369_C000.fm Page viii Friday, August 25, 2006 1:43 PM
IP369_C000.fm Page ix Friday, August 25, 2006 1:43 PM

Table of contents
Chapter 1 Introduction ...........................................................................................1
Nadir Dagli

Chapter 2 Electroabsorption (EA) Modulators and EA Modulators

Monolithically Integrated with Distributed-Feedback Lasers ..................................5
Koichi Wakita

Chapter 3 High-Speed LiNbO3 Optical Modulators............................................31

O. Mitomi

Chapter 4 III-V Compound Semiconductor Electro-Optic Modulators ..............63

Nadir Dagli

Chapter 5 High-Speed Polymer Optical Modulators and Their

Applications ...........................................................................................................109
Harold R. Fetterman and William H. Steier

Chapter 6 High-Speed Photodetectors and Photoreceivers................................155

P. Fay

Chapter 7 IC Technologies for Future Lightwave Communication

Systems ..................................................................................................................185
Taiichi Otsuji

Chapter 8 High-Speed All-Optical Technologies for Photonics........................217

Masatoshi Saruwatari

Index ......................................................................................................................249
IP369_C000.fm Page x Friday, August 25, 2006 1:43 PM
IP369_C001.fm Page 1 Tuesday, August 1, 2006 3:21 PM

1 Introduction

Nadir Dagli

Ever since the Internet was opened to public use in 1993, the amount of information
transmitted over communication networks has increased drastically. Most of the
traffic in communication networks today is due to data transmission rather than
voice transmission, and the amount of data transmitted continues to increase
rapidly. For such high-volume applications, fiber is the most natural transmission
medium. Standard single-mode fiber offers more than 25 terahertz bandwidth. The
availability of such high bandwidth enables very high-bit-rate transmission. Fiber
optic networks operating at 40 Gbit/second are being installed all around the globe.
The technology that provides such high-bit-rate transmission is crucial and is being
developed in many different research laboratories around the world. This technol-
ogy requires high-speed transmitters, high-speed receivers, high-speed electronics,
and high-speed all-optical techniques. It is obvious that these topics require very
diverse backgrounds, and it is best to rely on the expertise of individuals who are
leaders in their fields. This book presents this technology by combining several
chapters written by the experts working on different aspects of high-speed photonic
devices.
High-speed transmission starts at the source end. In principle, a high-speed
transmitter can be obtained by directly modulating a semiconductor laser diode.
Although semiconductor lasers with small signal bandwidths on the order of 40
GHz have been demonstrated, direct modulation is typically not used for trans-
missions over 2.5 Gbit/second. This is mainly due to chirping of the laser output
under direct modulation. As the current supplied to the laser diode is modulated,
the number of carriers in the device changes. This changes the index of the material
and the laser output frequency. This undesired frequency modulation increases the
bandwidth of the transmitted signal and severely limits the transmission distances.
Therefore, for high-speed transmission systems, external modulators are used, and
the laser output is kept unchanged. High-speed external modulation can be achieved
using different technologies. These technologies are the topics of Chapter 2 through
Chapter 6.
Chapter 2 describes electroabsorption (EA) modulators and EA modulators
integrated with distributed feedback (DFB) lasers. The EA modulator is a very
compact semiconductor device. The absorption through this device depends on the
voltage applied to the device. This absorption modulation can be done very effi-
ciently; even a few volts can be sufficient. EA modulators typically use specially

1
IP369_C001.fm Page 2 Tuesday, August 1, 2006 3:21 PM

2 High-Speed Photonic Devices

designed quantum-well absorption regions. Therefore, proper material design and

growth is crucial. They work efficiently over a limited range of wavelengths and
demonstrate sensitivity to temperature changes and high optical input powers. They
also tend to have high optical-insertion loss. EA modulators are compact (in most
cases less than 200 µm long) and can be integrated with DFB lasers. This results in
a very compact source that can be modulated very efficiently. Such devices at bit
rates of 10 Gbit/second are commercially available from several different vendors.
EA modulators also have chirp, which varies during modulation. However, their
chirp can be kept small and negative, which actually may help in data transmission
to a certain degree.
The third chapter describes LiNbO3 modulators. LiNbO3 is a ferroelectric
material that possesses an electro-optic coefficient. The index of refraction of the
material changes in response to an applied electric field without changing the
optical loss. Therefore, it is possible to make interferometric or Mach-Zehnder–type
modulators in this material. Such modulators can have zero chirp and very low
insertion loss. Furthermore, LiNbO3 has a very large electro-optic coefficient.
However, even in this case an efficient modulator requires a traveling-wave elec-
trode. In such designs the modulator electrode is designed as a transmission line,
and the modulating voltage and the modulated optical wave travel in the same
direction and interact along the length of the electrode. If the optical and electrical
velocities are matched, this interaction is most efficient and the small phase
changes can be integrated along the length of the electrode. Typical electrode
lengths for LiNbO3 modulators are approximately 3 to 5 centimeters. This approach
helps to reduce the drive voltage to a few volts but results in large devices. One
fundamental difficulty of this material is its dielectric constant dispersion. The
dielectric constant at microwave and millimeter wave frequencies is very high,
but it decreases significantly at optical frequencies. As a result, the electrical signal
travels slower than the optical signal, and velocity matching requires increasing
the electrical signal velocity. Therefore, electrode capacitance should be reduced,
which in turn requires reducing the electric field for a given voltage. The large
electro-optic coefficient of LiNbO3 partially offsets these difficulties. Such devices
are commercially available from several vendors.
The fourth chapter covers III-V compound semiconductor modulators. These
materials also possess an electro-optic coefficient unlike the most common semi-
conductor, silicon, which does not. Their electro-optic coefficient is smaller than
that of LiNbO3, but they have high indices of refraction and very low index
dispersion between microwave and optical frequencies. Since electro-optic index
change is proportional to the cubic power of the index of refraction, the high index of
refraction makes up for the low electro-optic coefficient to a certain extent. Fur-
thermore, the low-index dispersion makes the electrical signal travel faster than
the optical signal. As a result, traveling wave designs with slow wave electrodes
become possible, and large capacitive loading can be used. The high index of
refraction also makes tight optical confinement possible. When these two effects
are combined, it becomes possible to have tightly confined optical modes over-
lapping very well with very high modulating fields. Therefore even bulk III-V
compound semiconductor modulators with better voltage-length products than
IP369_C001.fm Page 3 Tuesday, August 1, 2006 3:21 PM

Introduction 3

LiNbO3 are possible. Furthermore, the electro-optic coefficient of these materials

can be improved using quadratic effects that are related to electroabsorption. These
devices also have the potential of integration with other III-V compound semi-
conductor devices such as lasers and detectors. However they have larger insertion
loss mainly due to small-mode size-related coupling loss. Mode transformers can
solve this problem, but at increased cost.
Chapter 5 discusses polymer modulators. Polymers offer the potential of both a
high electro-optic coefficient and a low index of refraction. Their high electro-optic
coefficient has the potential to reduce the drive voltage. Their low index of refraction
should help to increase the modulation bandwidth since velocity matching becomes
easier and microwave electrode loss tends to be lower on lower-index material.
Synthesizing high electro-optic coefficient polymers has been hampered due to
reliability issues for quite some time. However, recent advances have solved this
problem to a certain extent, and good-quality polymers with reasonably high electro-
optic coefficients are available. The design of modulators and other guided wave
components is described in detail in this chapter.
The critical components at the end of a transmission system are the photodetectors
and photoreceivers, which are described in Chapter 6. These components convert optical
information into electrical information while minimizing noise and distortion. There are
many different types of photodetectors used for different types of applications. A pho-
todetector is followed by an electronic circuitry for amplification and signal processing.
The receiver design requires the selection of appropriate technology, devices, and archi-
tecture. Recently, different modulation formats have been introduced to improve the
spectral efficiency of optical transmission. Special receivers are needed for such trans-
mission systems.
Chapter 7 covers integrated circuit technologies for fiber optic systems. High-
speed electronic integrated circuits are essential for the realization of fiber optic
systems. Both receiver and transmitter ends require high-speed electronics not only
for modulator drivers and photoreceivers, but also for clock recovery and decision
making. Currently, electronic integrated circuits operating at 40 Gbits/second are
available in different technologies including Si-, Si-Ge-, GaAs- and InP-based mate-
rials. Their fundamental limitations, advantages, and architectures are the topic of
this chapter. This chapter also discusses the limiting factors for electronic integrated
circuits for 100-Gbit/s and faster operation.
The last chapter is on all-optical high-speed technologies. As demand for infor-
mation increases, electronic limitations could be the bottleneck for high-speed fiber
optic networks. This limitation could be overcome using all-optical techniques not
only in transmission but also in processing information at the network nodes. Optical
signal processing has much wider bandwidth than electronic processing and has the
potential of providing extremely high-speed transmission systems. In addition to
speed, transparency for modulation format is another very important property of
networks using all-optical techniques. Ultrashort and ultrahigh-speed optical pulse
generation, modulation, multiplexing, and demultiplexing are technologies for high-
speed transmission. These technologies coupled with other all-optical technologies
such as optical timing extraction, optical signal monitoring, all-optical repeating,
and all-optical regenerating could enable ultrafast networks.
IP369_C001.fm Page 4 Tuesday, August 1, 2006 3:21 PM

4 High-Speed Photonic Devices

After reading the book, it will become obvious that there is strong potential for
significant further improvement in all the topics considered. Therefore, all high-
speed photonic components represent active and exciting research areas. Although
it is hard to predict which particular technology will prevail in the future, it is
expected that these research efforts will result in low-cost, high-performance fiber
optic networks covering the entire globe. Such improved and easily accessible
communications capability will improve the quality of life for everyone.
Finally, the editor takes this opportunity to thank all the contributors to this book
for taking their time to write very informative chapters about their work.
IP369_C002.fm Page 5 Tuesday, August 8, 2006 12:28 PM

2 Electroabsorption (EA)
Modulators and EA
Modulators
Monolithically Integrated
with Distributed-
Feedback Lasers
Koichi Wakita

CONTENTS

2.1 Introduction .......................................................................................................6

2.1.1 Franz-Keldysh Effect ............................................................................6
2.1.2 Quantum-Confined Stark Effect ...........................................................7
2.1.3 Strong and Weak Quantum Confinement .............................................8
2.1.4 Optimization of MQW Structure........................................................10
2.2 EA Modulator Design Principles ...................................................................12
2.2.1 EA Modulator Characteristics ............................................................13
2.2.1.1 On/Off Ratio........................................................................13
2.2.1.2 Required Voltage .................................................................14
2.2.1.3 3-dB Bandwidth ..................................................................14
2.2.1.4 Insertion Loss ......................................................................15
2.2.1.5 Chirping...............................................................................16
2.2.1.6 Figure of Merit ....................................................................17
2.3 Monolithic Integration of EA Modulators and DFB Lasers..........................19
2.3.1 Monolithic Integration of Franz-Keldysh Modulators
and DFB lasers....................................................................................20
2.3.2 Monolithic Integration with MQW
Modulators ..........................................................................................20
2.3.3 Advances for Monolithic Integration..................................................21
2.4 Recent New Applications of EA Modulators.................................................22
2.4.1 UltraHigh-Speed Operation ................................................................22
2.4.2 High Incidental Optical Power Operation ..........................................24

5
IP369_C002.fm Page 6 Tuesday, August 8, 2006 12:28 PM

6 High-Speed Photonic Devices

2.4.3 Chirping Reduction for EA Modulators.............................................26

2.4.4 Optical Gate Integrating a UTC-PD and TW-EA Modulator ............27
References................................................................................................................28

ABSTRACT
This chapter describes the recent research and development of high-speed, low-driving-
voltage EA modulators and their integrated distributed-feedback lasers for high-bit-
rate and long-haul optical fiber transmission systems. Low insertion loss and low
chirping as well as polarization insensitivity are also discussed. Increased saturation
intensity, high-input optical power allowance, and wide-wavelength operation for
wavelength-division multiplexing are described. Particular attention has been paid to
multiple quantum well (MQW) modulators operating at long wavelengths, taking into
account the losses and dispersion in silica fibers occurring at around 1.3 and 1.55 µm.
The future prospects for optical modulators and their monolithically integrated light
sources for ultrahigh-bit-rate OTDM are also discussed.

2.1 INTRODUCTION
The research and development of optical fiber communication systems, satellite
communication systems, and radio transmission systems based on large-scale inte-
grated circuit technologies has opened the door to a new age in the application of
Internet and multimedia communications. In the early 21st century, telecommuni-
cations and computers will be combined into information network systems to satisfy
the demands of the enormous increases in information and communication required
by society and industry. Semiconductor modulators mainly used as electroabsorption
(EA) modulators are one of the important devices playing a role as key components
of this new age. EA modulators integrated with DFB lasers have been developed
and applied to commercial uses for metropolitan high-bit-rate and long-haul optical
transmission systems. Moreover, in optical-time-division-multiplexed (OTDM) sys-
tems based on high-speed optical switching (over 100 Gbit/sec), EA modulators are
indispensable, and much interest has been focused on them.
In this chapter, the progress of the semiconductor optical modulators throughout
eight recent years [1] will be reviewed from the viewpoints of high-speed and low-
driving-voltage operation. New directions of modulator research such as low-chirp
operation, high-input optical power allowance, and wide wavelength operation for
wavelength-division multiplexing (WDM) as well as low insertion loss will also be
discussed.

2.1.1 FRANZ-KELDYSH EFFECT

The research and development of EA modulators has a long history. In 1958, Franz
and Keldysh [2,3] independently reported groundbreaking theoretical studies of the
effect of an electric field on the absorption edge of a semiconductor. They predicted
that, in the presence of an electric field, an absorption would occur at photon energies
lower than the forbidden energy gap (Franz-Keldysh effect), as shown in Figure 2.1.
IP369_C002.fm Page 7 Tuesday, August 8, 2006 12:28 PM

Electroabsorption (EA) Modulators 7

Conduction band
Electron

hw

Hole Ec
Eg

Valence band Ev

FIGURE 2.1 Energy band diagram of a semiconductor under high electric field.

However, the bulk structure used at that time had a large operating voltage and large
insertion loss. The limitations of directly modulated laser diodes have caused us to
take a new look at this mechanism. In high-bit-rate (over a few Gbit/s), long-haul
optical transmission systems, the frequency chirping associated with high-speed
direct modulation of semiconductor laser diodes is a serious problem that limits
transmission length and modulation speed. Advances in double heterostructure
growth by liquid-phase-epitaxy and waveguide theory brought about renewed inter-
est in applying the effect for light modulation, and there were many attempts at
devising the effect for light modulators [4–8]. Several volts operation for an on/off
ratio of more than 10 dB has been achieved. Optical fiber transmission using EA
modulators that were monolithically integrated with DFB laser diodes was also
demonstrated [9–11]. However, driving voltages are large compared with those of
MQW modulators.

2.1.2 QUANTUM-CONFINED STARK EFFECT

The multiple quantum well (MQW) structure exhibits strong excitonic effects that
modify the fundamental absorption edge of materials. The exciton effect results from
the Coulomb interaction between the electron-hole pair in the crystal, and it manifests
itself as an increased steepness in the absorption coefficient. These effects are also
present in covalent semiconductors, but they can be observed only at very low
temperatures due to the low excitonic binding energy (typically 3 meV for Ge and
GaAs). In MQW structures, the binding energy can be as large as 10 meV and the
excitonic structure can be observed at room temperature. When an electric field is
applied perpendicularly to such structures, the absorption coefficient near the band
edge decreases significantly, as shown in Figure 2.2. This is similar to the Franz-
Keldysh effect; however, because it is associated with quantum well (QW) structures,
it is called the quantum-confined Stark effect (QCSE) [12]. The difference between
the two is that, due to the two-dimensional confinement, excitonic peaks and their
energy-shifts are observed for MQW structures, while no such effects are observed
for the Franz-Keldysh effect. The energy shift is approximately proportional to the
IP369_C002.fm Page 8 Tuesday, August 8, 2006 12:28 PM

8 High-Speed Photonic Devices

100

+0.5
0V
80

Photocurrent (a.u.)
−1
−2
−3
60
λ = 1550 nm −4

40

1200 1300 1400 1500 1600

Wavelength (nm)

FIGURE 2.2 Absorbed photocurrent spectra for InGaAs/InAlAs MQWs as a function of

applied biases.

fourth power of the QW thickness, according to the variational calculation [13],

whereas the increase in size can result in a significant change in the electron-hole
overlap, reducing the absorption strength of the exciton line. Therefore, there is an
optimum QW thickness at which a highly efficient EA effect will arise; this is
discussed in the next section. The QCSE has received a lot of attention because it
is very large even at room temperature, allowing high-speed, low-driving voltage
modulators and switches.

2.1.3 STRONG AND WEAK QUANTUM CONFINEMENT

From the viewpoint of practical applications to modulators, QWs operating in the
long-wavelength region are important; in this region, optical fibers are transparent and
their dispersion is low. The InGaAsP/InGaAsP quantum well typically used for oper-
ation in that region has less electron confinement than InGaAs/InP or InGaAs/InAIAs
MQWs. This is because, in an attempt to overcome optical saturation at higher optical
intensities due to hole pile-up, the QWs are designed so that the valence band offset
is small. In such situations, the discrete energy level for the electron becomes quasi-
bound states with a finite energy width due to the weakened confinement.
Figure 2.3 shows the calculated shapes of the absorption spectra in a QW
structure under an electric field strength of 0 and 150 kV/cm, respectively, together
with the wave functions [14]. In the presence of an electric field, the wave functions
have an oscillatory tail that extends out of the well, as shown in the inset of Figure 2.3(a).
As a result, the field-induced broadening is remarkable, i.e., the weakened confine-
ment makes the discrete level quasi-bound with a finite width.
As mentioned above, the behavior of the excitons is very different in QWs with
different barrier components even if the field is applied perpendicularly to the layers.
Experimental results and calculations are shown for a 6-nm-thick InGaAsP QW with
InGaAsP barriers whose photoluminescent (PL) wavelength is 1.3 µm (Figure 2.4)
IP369_C002.fm Page 9 Tuesday, August 8, 2006 12:28 PM

Electroabsorption (EA) Modulators 9

3 C.B.
(a) 0 kV/cm
2.5 50

Absorption coefficient (× 104 cm–1)

2 100 ∆EC
150
1.5 200
1 V.B.
0.5
0
(b) 0 kV/cm
2.5 50 C.B.
100
2 150
200 ∆EC
1.5
1
V.B.
0.5
0
1.45 1.5 1.55 1.6
Wavelength (µm)

FIGURE 2.3 Variation in the calculated absorption spectra under 5 different electric field
strengths with an increment of 50 kV/cm. Upper lines in (a) show the spectra of QW structure
with tilt barriers, i.e., weak confinement, and lower lines in (b) show those with flat barrier,
i.e., strong confinement.

and a l0-nm-thick InGaAsP QW whose PL wavelength is 1.2 µm (Figure 2.5). In

the calculations, conduction-band offsets of 0.5∆Eg for the 6-nm QW structure and
0.55∆Eg for the l0-nm QW structure were assumed, where ∆Eg is the band gap
energy difference. Hence, the estimated barrier heights from the ground level of the
conduction band are 40 and 130 meV, respectively. As can be seen from the figures,

2 10
Absorption coefficient (× 104 cm−1)

Lw = 6 nm
36.6 kV/cm
F = 0 kV/cm
Photocurrent (Arb. units)

30
53.1
69.2
1 5
102.0

60
90
120

0 0
1.4 1.5 1.6 1.7 1.4 1.5 1.6 1.7
Wavelength (µm) Wavelength (µm)
(a) (b)

FIGURE 2.4 Comparison of the calculated absorption spectra (a) and the measured photo-
current spectra (b) for a 6-nm InGaAsP/InGaAsP single QW structure.
IP369_C002.fm Page 10 Tuesday, August 8, 2006 12:28 PM

10 High-Speed Photonic Devices

10
Absorption coefficient (x104 cm-1)
1
0 kV/cm
18.4 kV/cm

Photocurrent (arb. units)

30 34.8
51.1
60 83.9
0.5 5 116.7
90 133.1
120 149.5
150

(a) (b)
0 0
1.4 1.5 1.6 1.7 1.4 1.5 1.6 1.7
Wavelength (µm) Wavelength (µm)

FIGURE 2.5 Comparison of the calculated absorption spectra (a) and the measured photo-
current spectra (b) for a 10-nm InGaAsP/InGaAsP single QW structure.

the barrier height difference has an important effect on the field-induced broadening,
and carrier tunneling dominantly contributes to the spectral broadening. Other broad-
ening mechanisms such as inhomogeneous field distribution and thickness and
composition fluctuations play a minor role.
In contrast to InGaAsP/InGaAsP MQWs, InGaAs/InAlAs MQWs have a large
∆Ec and show strong electron confinement. Therefore, when an electric field is applied
perpendicular to the layers, the exciton resonance and shifts can be observed clearly.
This quantum confinement difference affects device characteristics. In particular, when
tensile strain is introduced to obtain polarization-insensitive modulation, both MQWs
can operate as polarization-insensitive modulators, but the power saturation level is
very different between them. This is because tensile strain introduction increases
valence band offset between wells and barriers, resulting in hole pile-up under high
incidental optical power. The details of this will be described in the next section.

2.1.4 OPTIMIZATION OF MQW STRUCTURE

In InGaAs MQW modulators on InP substrates, the barrier and waveguide may be made
from quaternary alloys of either InGaAsP or InGaAlAs. If a lattice-matched alloy with
a particular bandgap is selected, the other properties are fixed at values that depend on
the choice of InGaAsP or InGaAlAs. Although the bandgap discontinuity ∆Eg at the
heterojunctions may be the same in both cases, the discontinuities in the conduction and
valence bands ∆Ec and ∆Ev are different. For InGaAs/InGaAsP, ∆Ec:∆Ev is thought to
be about 40:60 [15], whereas for InGaAs/InGaAlAs, this ratio is about 72:28 [16].
Because the effective masses of electrons and heavy holes differ by an order of magni-
tude, the band discontinuities have a significant effect on the well thickness appropriate
for a desired output wavelength. The well thickness must be greater for InGaAlAs
barriers. In the quantum-confined Stark effect (QCSE), the exciton peaks shift in pro-
portion to the fourth power of the well thickness when the applied field is weak [17].
IP369_C002.fm Page 11 Tuesday, August 8, 2006 12:28 PM

Electroabsorption (EA) Modulators 11

Moreover, ∆Ev has an effect on modulator characteristics such as frequency response

and absorption saturation under high incidental optical power because of hole pile-up at
the heterointerface. Especially for InGaAs/InP MQW structures with ∆Ev of 0.38 eV,
the frequency response is degraded because of hole pile-up [18]. InGaAsP/InGaAsP
MQW structures with a ∆Ev of 0.2, eV is used instead to reduce this degradation.
On the other hand, for the same quaternary alloys we can tailor the well thickness
and well bandgap energy even though the energy band edge of the MQW is constant.
Because larger absorption variations can therefore be obtained with similar applied
electric fields, it is necessary to optimize the well and barrier structure for modulators.
The first proposal to improve the EA effect in MQW structures by replacing the
ternary InGaAs well with a quaternary InGaAsP well was reported in 1987 [19].
The optimizations of the InGaAsP/InP [20,21] and InGaAlAs/InAlAs [20] systems
have also been reported.
Quaternary QW materials are believed to be more advantageous than the
conventional ternary QW materials for enhancing excitonic EA effects. Since the
bandgap of quaternary materials is larger than that of ternary materials, the exciton
transition energy is also greater. The QW thickness must therefore be increased
for the quaternary QWs to keep the exciton transition energy the same as that for
the ternary QWs. This is necessary for operating the device at a certain fixed
wavelength, such as the 1.55-µm wavelength. The energy shift of quantum levels
(Stark shift) induced by application of an external electric field increases with
well thickness. The quaternary QW is therefore thought to have field effects more
pronounced than those of the conventional ternary QWs when both are operated
at the same wavelength. The increase in well thickness, however, implies a decrease
in the oscillator strength of the excitonic transition. This factor may impede the
above enhancement.
We define the figure of merit as Γ∆a/∆F, where Γ is the optical confinement
factor, and we calculate its value under the condition that it includes the field-
induced broadening and the transitions between the conduction ground subband and
the topmost three valence subband (1e-1hh, 1e-1lh, 1e-2hh) [21]. The former effect,
which has long been neglected, is critical for practical use in weak-confinement
MQWs such as InGaAsP/InGaAsP, where under a relatively weak electric field the
conduction band offset is too small to confine electrons.
As a result, we obtain the optimum conditions for MQW structures, such as the
QW thickness and configuration. Figure 2.6 shows the figure of merit versus QW
thickness and composition. The operating wavelength is 1.55 µm and the barrier
thickness is 7 nm. There are two peaks: one is for a thin well and small energy gap
(Eg = 0.7 eV), and the other is for a thick well and large energy gap (Eg = 0.8 eV).
As shown in Figure 2.6, the former corresponds to the first heavy-hole excitonic
transition and the latter to the light-hole exciton. The former has relatively large
transmission loss because of smaller detuning energy, whereas the latter has a larger
figure of merit and lower transmission loss. That is, the thicker QW (about 11–12 nm)
is better. This is because of the larger overlap integral of the 1e-1lh transition due
to the light-hole character and to the enhanced joint density of state due to the
valence band-mixing in the 1lh subband. Thicker wells also increase the optical
confinement and lower the operating voltages.
IP369_C002.fm Page 12 Tuesday, August 8, 2006 12:28 PM

12 High-Speed Photonic Devices

500 F = 150 kV/cm

400

300

200

100
16
131415
0 12
9 1011 )
0.70 0.75 7 8 (nm
0.80 0.85 6 Lz
0.90 0.95 5
Eg (eV
)

FIGURE 2.6 Figure of merit Γ∆a for excitonic electroabsorption effects as a function of
well thickness Lz and bandgap Eg of QW materials for InGaAsP/InGaAsP QWs. The applied
electric fields are 150 kV/cm.

This result gives us insight into the efficient EA effect in QWs. The introduction
of tensile strain into wells, which has been used to reduce the anisotropic modulation
of polarization dependence, will also affect the reduction of drive voltage according
to this calculated model because the hh and lh transitions cross under a certain tensile
strain, and a thicker well is used to keep the detuning energy suitable. In fact,
polarization insensitive MQW modulators are operated at a lower drive voltage and
high speed [22].

2.2 EA MODULATOR DESIGN PRINCIPLES

In this section, we introduce the operation principle of an EA modulator as a guide.
First, the operating wavelength must be determined. The propagation loss and dis-
persion of silica-based single-mode fiber depend on the wavelength. We have two
kinds of optical fibers whose dispersion-minimum wavelength is 1.3 mm or 1.55 mm.
The former is called normal fiber, and the latter is called dispersion-shifted fiber.
Generally, the operating wavelength and the materials are not independent. For
the wavelengths of low-loss region, InGaAsP or InGaAs are determined. Then we
determine the waveguide thickness and its fractional content of mixture of well and
barrier. Waveguide characteristics are strongly dependent on the waveguide core and
cladding materials, because the optical confinement depends on the refractive index
profile. The suitable detuning energy between the modulated light wavelength and
the absorption band edge of EA modulators is about 40–50 meV.
For MQW structures, the thickness of an InGaAsP QW is about 7.5 nm when
the well and barrier are lattice-matched to an InP substrate. If we use an InP barrier,
the InGaAs well thickness is 5.0 nm when this modulator operates at 1.55 mm.
Next, we determine the total thickness of the waveguide core or the MQW layer.
IP369_C002.fm Page 13 Tuesday, August 8, 2006 12:28 PM

Electroabsorption (EA) Modulators 13

p-contact

Polyimide

p-InGaAs
SiO2
p-InAlAs
n-InAlAs MQW

n-InP n-contact

FIGURE 2.7 Schematic view of MQW EA modulator.

The barrier thickness must be thin to ensure large optical confinement and small
driving voltage as far as the electron confinement is achieved. The thickness of the
waveguide layer (summation of well and barrier thickness multiplied by their number)
is determined from a practical point of view. That is, it depends on the operating
voltage, 3-dB bandwidth, insertion loss, and so on. For typical devices, the waveguide
thickness is between 0.1 and 0.2 mm, so that the operating voltage is a few volts
and the 3-dB bandwidth is around 10 GHz.
A schematic diagram of an InGaAs/InAlAs MQW EA modulator is shown in
Figure 2.7.

2.2.1 EA MODULATOR CHARACTERISTICS

Electroabsorption modulators have five important parameters: on/off ratio, driving
voltage, 3-dB bandwidth, insertion loss, and chirping. We discuss these factors as
follows.

2.2.1.1 On/Off Ratio

The most important parameter is the on/off ratio for the intensity modulator. At least
15 or 20 dB is usually needed for applying the practical systems. The on/off ratio
is defined as the ratio of incident light intensity Pin to transmitted light intensity Pout
and is given as

[on/off] = −10 log10 (Pout/Pin) [dB].

In EA materials, the transmitted light intensity is given by using the absorption

coefficient with the incident light intensity as

Pout/Pin = exp (−Γ ∆aL),

where Γ, ∆a, and L are the optical confinement factor, absorption coefficient change,
and sample length, respectively. Therefore, the on/off ratio is given as

[on/off] = 0.434 Γ∆aL [dB].

IP369_C002.fm Page 14 Tuesday, August 8, 2006 12:28 PM

14 High-Speed Photonic Devices

2.2.1.2 Required Voltage

As shown in Figure 2.2, the absorption coefficient change ∆a strongly depends on the
applied voltage and the wavelength used. Based on the QCSE, the absorption coeffi-
cient peak shifts in proportion to the square of the applied electric field. In practical
use, the smaller the applied voltage, the better for driving electronics circuits. Up until
now, 2-V peak-to-peak operation has been necessary for high-speed (more than 10 Gb/s)
operation. The applied voltage is determined by the total thickness of the i-layer, and
the barrier thickness is also important. For lumped modulators, the total thickness
limits the device capacitance that results in 3-dB bandwidth.

2.2.1.3 3-dB Bandwidth

The bandwidth is usually determined by the device capacitance when the device is
operated in the reverse bias condition, except when it is operated by carrier injection,
which is operated at very slow speed (corresponding to a few ns). When the speed
is limited by the device capacitance, the 3-dB bandwidth is given as

[f3dB]1 = 1/pRC,

where R and C are the load resistance and capacitance, respectively. The device
capacitance is proportional to sample length L and width W, and it is inversely
proportional to i-region thickness d when the stray capacitance is neglected and the
electric field is applied perpendicularly to the i-region. The last assumption is
discussed in the last section, where a new structure of lateral p-i-n using a parallel
field is proposed. As the capacitance decreases, the frequency response increases.
Usually, stray capacitance induced from the bonding pad cannot be neglected and
some countermeasures, such as polyimide with small dielectric constant, are used
to reduce the stray capacitance.
The above discussion is based on the assumption that the sample is short enough
and the transit time of light through the sample is very short. When we use long
samples, the bandwidth is limited by the transit time t of the light through the sample
and the bandwidth is given as

[f3dB]2 = 1.39/pt = 1.39c0 /p ngL,

where ng is the refractive index of the waveguide and c0 is the speed of light in a
vacuum. As an example, we calculate this value for InGaAsP with a refractive index
of 3.24 as

[f3dB]2 L = 4.13 GHz cm.

That is, the 3-dB bandwidth for the lumped InGaAsP device is limited to above
[f3dB]2 even if [f3dB]1 is increased.
To improve this limitation, a traveling-wave-type structure is proposed and
demonstrated. In this structure, a traveling line is arranged along the optical waveguide
IP369_C002.fm Page 15 Tuesday, August 8, 2006 12:28 PM

Electroabsorption (EA) Modulators 15

to match the transmission velocity of microwave and optical waves. In this case, the
bandwidth is [23]

[f3dB] = 1.39c0 /p [ng − nm]L,

where nm is the refractive index of the transmission line for the modulated microwave.
Based on this equation, [f3dB] seems to increase monotonously if we increase L when
the velocity matching condition is fulfilled. Practically, the losses for optical and
microwave transmission are not neglected; however, the maximum value in the
reported data is 350 GHz, which was obtained by using superconducting materials
for the ohmic contact [24].

2.2.1.4 Insertion Loss

Insertion loss consists of transmission loss, reflection loss, and coupling loss. The
transmission loss consists of residual absorption loss of the material, free carrier
loss, and scattering loss. The coupling loss is due to the mode-spot size mismatch
between the incident light and the guided light. The reflection loss is at most 3 dB
and is eliminated by antireflection coating.
The transmission loss atran consists of the inherent absorption ag due to the
energy difference between the incident light and the absorption edge of the
waveguide material, free carrier absorption afc due to the free carrier in the waveguide
and the cladding layer, and light scattering as, and is given by

atran = Γag + (1 − Γ)afc + as,

where Γ is the optical confinement factor (Figure 2.8). This value depends on the
refractive index profile. The free carrier absorption in the p-doped cladding layer is
much larger than that in the n-doped cladding layer, by as much one order of
magnitude for InP. The doping level is required to be small to provide low trans-
mission loss, while high doping is required so that the applied voltage is biased only
to the i-region, resulting in small bias voltage. Therefore, the p-doping is chosen to
be 5 × 1017 to 1 × 1018 cm−3.

Refractive index Relative intensity

InGaAs cap

InAlAs clod

MQW
InAlAs clod

InP sub.

FIGURE 2.8 Electric field distribution and optical confinement factor Γ for an InGaAs/InAlAs
MQW modulator three-layer waveguide.
IP369_C002.fm Page 16 Tuesday, August 8, 2006 12:28 PM

16 High-Speed Photonic Devices

In a semiconductor waveguide, the mode spot size is generally very small,

especially in the direction perpendicular to the layer, due to the large refractive
index difference between the guide layer and the cladding layer. This strong optical
confinement produces small driving voltage, causing effective overlap between
the optical field and the electric field which results in small spot-size. Therefore,
the large insertion loss in semiconductor optical devices is mainly due to the
coupling loss; and some counter-plans, such as monolithic integration with laser
diodes or mode-spot size transformers, have been tried in an attempt to minimize
this loss.

2.2.1.5 Chirping

Any change of absorption coefficient in a material structure will be accompanied

by a phase shift of light, since the real and imaginary parts of a dielectric constant
constitute a Kramers-Kronig transform pair. This causes frequency chirping in inten-
sity modulation and intensity fluctuation in phase modulation. When chirping is
induced, the transmitted optical pulses are broadened through the fiber due to the
dispersion. Though the chirping for external modulators is thought to be much
smaller than that for direct-modulated laser diodes, it is a limiting factor for the
capacity of long-haul high-bit-rate optical-fiber communication systems due to fiber
dispersion, as shown in Figure 2.9.
The magnitude of chirping is defined as the ratio of the change of refractive
index ∆n to the change of the extinction coefficient ∆k. It has been proposed that
waveguide Mach-Zehnder amplitude modulators may be operated in a perfectly
chirpless mode. Devices designed to operate chirp-free have been reported, whereas
some chirping exists for EA modulators; however, its magnitude is at most 1.5 based
on previously reported data (Table 2.1).
However, the lowest fiber-dispersion penalty is generally not obtained for a
frequency chirp parameter equal to zero. It can be advantageous to choose a nonzero
value for the chirp parameter, depending on the fiber dispersion coefficient and
distance, to provide some amount of pulse compression.
Transmission length, L (km)

Bit-rate-length product

102 103
BL (Gb/s km)

101 Penalty = +1dB 102

B = 10 Gb/s
D = 16 ps/km nm
−4 −2 0 2 4
Chirp parameter αc

FIGURE 2.9 Calculated transmission length versus chirping parameter. Modulation speed,
power penalty, and fiber dispersion are assumed to be 10 Gbit/s, 1 dB, and 16 ps/km nm,
respectively.
IP369_C002.fm Page 17 Tuesday, August 8, 2006 12:28 PM

Electroabsorption (EA) Modulators 17

TABLE 2.1
Review of  Parameter for EA Modulators
Detuning
 Energy,
Author Firm Material Parameter meV Ref.

Y. Noda KDD InGaAs/InP F-K 1–2 51.5 JLT, LT-4, 1445, ’87
T.H. Wood AT&T GaAs/AlGaAs QCSE 1 12 (TM) APL, 50, 798, ’87
K. Wakita NTT InGaAs/InAlAs QCSE 0.6–0.7 22–27 JJAP, 26, 1169, ’87
H. Soda Fujitsu InGaAsP/InP F-K 1 68 EL, 24, 1194, ’88
T. Saito NEC InGaAsP/InP F-K 0.6 85.7 OEC ’90, 13A2-4, ’90
K. Wakita NTT InGaAs/InAlAs QCSE 0.7 27 PTL, 3, 138, ’91
M.S.Whalen AT&T InGaAs/ InP QCSE 0.6–0.8 24.7–28.3 PTL, 3, 451, ’91
M. Suzuki KDD InGaAsP/InP F-K 0.2–0.4a 56–71 PTL, 3, 451, ’91
J. Langanay Alcatel InGaAsP/InGaAsP QCSE −0.2–0.2 24.7–27.4 APL, 62, 2067, ’93
F. Devaux CNET InGaAsP/InGaAsP QCSE 1.2 — PTL, 4, 720, ’92
F. Koyama TIT InGaAs/InGaAsP QCSE 0.4–1 5.4–21.2 PTL, 5, 1389, ’93
I. Kotaka NTT InGaAsP/InGaAsP QCSE 0.8 30.5 PTL, 5, 61, ’93
F. Devaux CNET InGaAsP/InGaAsP QCSE −2.0–3.0 — PTL, 3, 1288, ’93
T. Kataoka NTT InGaAsP/InGaAsP QCSE 0.2–1.4 30.5 EL, 30, 872, ’94
T.H. Wood AT&T InGaAsP /InP QCSE 0.5 28.3 JLT, LT-12, 1152, ’94
T. Ido Hitachi InGaAs/InAlAs QCSE 0.6 32.2 PTL, 6, 1207, ’94
J. Shimizu Hitachi InGaAlAs/InGaAlAs QCSE −2.5–1.0 26.7 EL, 38, 821, ’02
Y. Miyazaki Mitsubishi InGaAsP/InGaAsP QCSE −0.7–0.7 — J.QE., 38, 1075, ’02
a Underestimated; revised values 0.585–0.795(JLT, LT-12, ’94).

The chirp parameter for the EA modulator is shown in Figure 2.10, where the
transverse axis indicates the transmission loss with pre-bias. Note that the chirp
parameter decreases as the operating wavelength approaches the absorption edge
energy, while the absorption at zero bias (transmission loss) increases. That is, the
chirp parameter and the transmission loss contradict each other, and some compro-
mise is necessary. In general, the transmission loss is too large to be used under the
condition of negative chirp parameter. Recent advanced technologies produce highly
efficient optical amplifiers and monolithically integrated light sources with EA
modulators and DFB lasers, which enable us to use EA modulators with large
propagation loss. The recent improvements of the chirping in the external modulators
will be discussed in the last section.

2.2.1.6 Figure of Merit

As described in the previous sections, there are trade-offs among the above five
parameters. For example, on/off ratio is proportional to the product of absorption
coefficient change and sample length L, whereas 3-dB bandwidth f3dB is inversely
proportional to L and proportional to i-region thickness d, and the operating voltage
is proportional to d.
IP369_C002.fm Page 18 Tuesday, August 8, 2006 12:28 PM

18 High-Speed Photonic Devices

Alpha parameter
0

−1
1.535 um
−2 1.54 um
1.55 um
−3 1.557 um

−4
−10 −15 −20 −25 −30 −35
Insertion loss with pre-bias (dB)

FIGURE 2.10 Chirp parameter versus insertion loss as a function of incident light wavelengths.

Moreover, high-speed and low-driving-voltage operation is a typical trade-off rela-

tionship, because device speed is usually limited by capacitance (for lumped modula-
tors) that is associated with the guide layer thickness, and the thickness limits the
operating voltage. The on/off ratio is given by the product of the absorption coefficient
change and the sample length. Therefore, maximization of the absorption coefficient
change associated with applied voltage is necessary. For waveguide structures, the
optical confinement factor must be taken into account. For MQW structures, thick QW
and thin barriers are effective for this purpose as far as the quantum size effect can work.
Tensile-strained quantum wells can reduce the operating voltage drastically. This intro-
duction can also achieve both polarization insensitivity and low chirp [22].
Figure 2.11 shows the summary of the required RF power vs. capacitance-limited
3-dB bandwidth for recently developed modulators. A drive impedance of 50 Ω is

1000
Al-MQW
F–K
Required power (mW)

LN
100
GaAs–TW
P-MQW
10

1
1 10 100
Bandwidth (GHz)

FIGURE 2.11 Required power versus capacitance-limited bandwidth for recently developed
modulators. Solid squares, F-K, Franz-Keldysh; open triangles, LiNbO3 modulator; circle
with bar, solid diamond circle within circle, open circle modulator; solid circle, InGaAs/InAIAs
MQW. Required drive power Pac = Vac2 /50, Vac = Vpp/2 2 , Vpp = Vp or V10dB.
IP369_C002.fm Page 19 Tuesday, August 8, 2006 12:28 PM

Electroabsorption (EA) Modulators 19

assumed, and the operating voltage is the total voltage swing for either a p-phase shift
for a phase modulator or 10 dB on/off ratio for an EA modulator. F-K stands for Franz-
Keldysh modulators, TW indicates traveling waveguide structure modulators [25–28],
respectively, and the triangles indicate LiNbO3 modulators [29,31]. Reduction of device
capacitance results in high-speed operation, whereas drive voltage increases. The ratio
of device voltage to bandwidth figure of merit for InGaAs/InAIAs MQW modulators
is lower than for any other existing optical modulators.
Recently, a well-established approach to obtain very wide-band modulators, the
so-called traveling-wave design, has been reported [26,33,34]. In this design, the
electrode is designed as a transmission line so electrode capacitance is distributed
and does not limit the modulator speed due to constant limitation of RC time.
Although this method has been applied for a long time to LiNbO3 and semiconductor
phase modulators because of their relatively long waveguide length, it has been
applied to shorter EA modulators as well. Details will be discussed in the last section
of this chapter.

2.3 MONOLITHIC INTEGRATION OF EA

MODULATORS AND DFB LASERS
We saw in Section 2.2 that semiconductor modulators have a large insertion loss,
mainly in the form of coupling loss because of their small spot size. One approach
to reducing the coupling loss is to integrate the modulators with light sources. In
this section, we consider the monolithic integration of modulators with light
sources.
Monolithic integration of opto-electronic devices has been studied for many
years. Since the first concept of OEICs (opto-electronic integrated circuits) was
proposed by Miller in 1969 [35], a great deal of research on developing OEICs has
been done in many laboratories. Though the OEICs concept incorporates advanced
technical research and development, and many devices have been fabricated, prac-
tical applications have been realized for only a few devices. One exception, however,
is monolithic integration of modulators and laser diodes. This integration has many
advantages: it offers low insertion loss, high output power, low chirping, and, by
reducing the number of parts, increased reliability. This should reduce the overall
system cost. Usually the operating light wavelength for a laser is different from that
for a modulator, and the former is forward-biased whereas the latter is reverse-biased.
Therefore, the two devices need to be isolated both electrically and optically.
Many attempts to integrate modulators and laser diodes have been reported. How-
ever, it was not until the development of DFB lasers that success in integrating modu-
lators and lasers was demonstrated [36, 37], because conventional Fabry-Perot lasers
needed optical cavities and were never isolated optically from modulators. Since then,
integrated light sources using the EA effect have been reported [38–43] with the
emphasis on their ease of fabrication. In particular, a novel yet simple method for
monolithic integration of a DFB laser and an EA modulator operating in the 1.5-mm
window has been proposed and demonstrated [38]; in this method, the two devices are
made of a single active layer.
IP369_C002.fm Page 20 Tuesday, August 8, 2006 12:28 PM

20 High-Speed Photonic Devices

2.3.1 MONOLITHIC INTEGRATION OF FRANZ-KELDYSH

MODULATORS AND DFB LASERS
Since the first demonstration of an integrated light source was reported in 1987, many
laboratories in Japan have worked to develop this technology. System experiments
using integrated light sources [42–46] have been reported, and laboratories in other
countries have begun to investigate their use in multigigabit long-haul transmission
systems [47,48]. Figure 2.12 shows another integrated light source [40]. The power
emitted from the modulator facet is 17 mW and the 3-dB bandwidth is over 10.3 GHz.
Limited wavelength broadening (chirping) within 0.01-nm under 10-Gbit/s NRZ
modulation was also demonstrated. As an accelerated life test, under 70°C ambient
temperature and high-output power operation, the integrated light source showed stable
operation over 3000 hours without any degradation. The device voltage of EA mod-
ulators utilizing the Franz-Keldysh effect was still too high for high-speed IC drivers
and it was necessary to incur increased insertion loss to reduce drive voltage because
of the small detuning energy. At present, this integrated light source is suitable for
systems with relatively low modulation speed of about 2.4 Gbit/s.

2.3.2 MONOLITHIC INTEGRATION WITH MQW MODULATORS

Optical modulators made with MQW structures can produce excellent performance
because of their large EA effect (QCSE). DFB lasers with lower threshold current
densities, improved high frequency response, and narrower linewidth have been
developed by introducing MQW structures in place of the bulk in the active layer
[49]. Here we describe a low drive-voltage and high-speed monolithic light source
with an MQW structure for both the DFB laser and the external modulator. At first,
the MOVPE/MBE hybrid growth technique was used for the integration [50]; today,
the MOVPE technique is more popular.

DFB laser

Modulator

SI-InP
Active layer
Butt-joint coupling
Absorption layer
AR coat

FIGURE 2.12 Schematic drawing of a DFB laser/electroabsorption modulator integrated

light source.
IP369_C002.fm Page 21 Tuesday, August 8, 2006 12:28 PM

Electroabsorption (EA) Modulators 21

2.3.3 ADVANCES FOR MONOLITHIC INTEGRATION

The previous hybrid method of monolithically integrating an MQW modulator and
a DFB laser is too complicated to permit easy fabrication. However, easy and simple
methods using a vertical mode coupling structure [51,52] and/or a simultaneous
selective area growth (SAG) method [42,43,53–57] have been reported.
Figure 2.13 shows the schematic diagram of the vertical mode coupling integrated
light source [52]. In this case, the two MQW core layers (the upper for a laser and
the lower for a modulator) are stacked on top of each other in one growth step, separated
by a thin InP spacer layer. After the grating on the waveguide layer is fabricated, the
upper core to the etching stop layer is etched off selectively by wet chemical etching
and the cladding layer is regrown. The crystal quality and optical coupling efficiency
are improved by the regrowth on the flat layer and two-core structure, compared with
those of the butt-joint-coupling method. The main advantage remains the independent
optimization of the two elements.
Recently, simultaneous selective area growth has been reported [53–57]. With this
method, both the laser section and the modulator section can be grown simultaneously
by using MOVPE, based on the difference in growth speed between the dielectric mask
widths. Optical coupling efficiency is almost 100% and the fabrication process is easier
than butt-joint processing, which has been studied for a long time. One potential draw-
back, however, may lie in the difficulty of independent optimization of each component.
The reported device characteristics have been fair and transmission experiments were
demonstrated at 2.5 Gbit/s over 517 km [54] and 600 km [55] of standard fiber.

FIGURE 2.13 Schematic drawing of a strained InGaAsP/InGaAsP MQW DFB laser/

electroabsorption modulator integrated light source and its cross-section; (a) and (b) indicates
laser side and modulator side, respectively.
IP369_C002.fm Page 22 Tuesday, August 8, 2006 12:28 PM

22 High-Speed Photonic Devices

Moreover, butt-joint coupling between an MQW laser active layer and an MQW
modulator waveguide has been reported and 10-Gb/s, 120-km transmission has been
achieved in one experiment [56]. The butt-joint coupling has previously been con-
sidered to be too difficult for applying MQW structures due to the excessive growth
at the interface. That is, thick quantum wells will produce a small energy band gap
causing increasing transmission loss. However, by using MOVPE techniques, high-
coupling efficiency and high-emitting power from the facet of the modulator have
been obtained and over 40-Gb/s modulation has been demonstrated [59–61].

2.4 RECENT NEW APPLICATIONS OF EA MODULATORS

In this section, we discuss recent research and development for EA modulators:
ultrahigh-speed modulation, high allowability of incidental optical power, and low
chirping modulation.

2.4.1 ULTRAHIGH-SPEED OPERATION

Low-driving-voltage operation is the key point bringing high-speed (more than 40 GHz)
modulators into practical use because it eliminates the need for high-speed electrical
amplifiers. There is generally a trade-off between modulator speed and driving
voltage for lumped modulators. One way to achieve high-speed operation without
increasing voltage is to use short samples with small capacitance as shown in
Figure 2.14 [58,62,63], where an MQW modulator is integrated with nondoped
waveguides. This structure enables us to fabricate samples less than 100 mm long,
but the extinction ratio decreases. Table 2.2 shows the high-speed modulators with
modulation bandwidth of over 40 GHz reported so far [34,58–66]. As mentioned in
the previous section, there is an optimum condition at which EA modulators operate
both at low driving and high speed, and our conclusion is that a well thickness of
12 nm is the best when we introduce tensile strain in the wells [22].
The introduction of tensile strain enhances the EA effect, resulting in reduced
drive voltage. This is due to heavy- and light-hole degeneracy, where the absorption
of heavy-hole excitons transition coincides with that of the light-hole, as well as to

Electrode Polyimide

Passive region

Modulation region (ᐍ)

Passive region
n-InP

p-InAlAs
MQW n-InAlAs
i - InP
i - InGaAsP

FIGURE 2.14 Structure of an MQW modulator with integrated waveguides.

IP369_C002.fm Page 23 Tuesday, August 8, 2006 12:28 PM

Electroabsorption (EA) Modulators 23

TABLE 2.2
Reported High-Speed Semiconductor Modulators
Ext. Ratio,
Year/ 3dB Drive
Month (GHz) Voltage L(m) Material Affiliation Ref.

1990.5 40 20 dB, 7V 100 InGaAs/InAlAs MQW NTT CLEO

1993.10 38 10 dB, 8V 1 GaAs/AlGaAs MQW UCSB DRC
1994.3 35 13 dB, 6V 3 mm InGaAs/InP bulk HHI OFC
1994.8 35 19 dB, 4.2V 120 InGaAsP/InGaAsP MQW CNET EL, 30, 1347
28 10 dB, 2.5V 115
1995.2 40 10 dB, 2.6V 50 InGaAs/InAlAs MQW Hitachi PTL, 7, 170
30 20 dB, 2.6V 100
1995.3 40 Vp = 25V 2 cm GaAs/AlGaAs bulk UCSB OFC
(70) 10V
1995.7 50 20 dB, 3.4V 63 InGaAs/InAlAs MQW Hitachi IOOC
42 19 dB, 4.4V 75 (tensile) CNET IOOC
1995.9 42 20 dB, 2V 107 InGaAs/InAlAs MQW NTT ECOC
1997.8 50 15 dB, 9V 200 InGaAs MQW NTT EL, 33, 1580
2000.7 50 20 dB, 3V 100 InGaAsP MQW Oki OECC
2001.3 70 20 dB, 3V 225 InGaAsP MQW NTT EL, 37, 1
2001.3 40 10 dB, 3V 75 InGaAsP MQW Mitsubishi OFC

the electric field effect difference in the absorption edge shift and oscillator strength
decrease between them [67]. This enhances the oscillator strength at the zero-bias
condition and absorption coefficient change. Based on the data, a strain of −0.45%
gives the best figure of merit. Polarization insensitivity was achieved and extinction
ratio was about 20 dB at the swing voltage of 1.0 V. The frequency response of this
modulator was measured. The ratio of 3-dB bandwidth to drive voltage required for
a 20-dB on/off ratio is over 40 GHz/V. Clear eye opening of 40 Gbit/s was obtained
for a 0.9 Vpp pseudo-random modulation signal, and 1.5V DC bias was observed.
This modulator was used in a 320-km long dispersion-shifted fiber transmission
experiment with four-channel wavelength division multiplexing, in the first experi-
ment of its kind [68]. In this experiment, no power penalty was observed. This
indicates that the strained MQW modulator gets an advantage over power-hungry
LiNbO3 modulators over 40 Gbit/s without a high-speed electrical amplifier.
Another way to achieve high-speed operation without increasing voltage is to use
traveling-wave-electrode configuration, where the device speed is free from capaci-
tances and over 40 Gbit/s large-signal modulation has been reported [30,60]. Figure 2.15
shows a schematic design of a TW-EA modulator integrated with a DFB laser. How-
ever, their inevitably small characteristic impedance (about 25 Ω) is much smaller than
that of the standard 50-Ω rf connections. A novel traveling-wave modulator electrically
matched for InP-HBT drivers was demonstrated, and a 40-Gbit/s, 2-km SMF trans-
mission with a 0.3-dB penalty at a 1.3-mm wavelength was reported [69].
A new electroabsorption modulator consisting of a lateral p-i-n configuration
whose i-region is composed of multiple quantum wells (MQWs) has been proposed
IP369_C002.fm Page 24 Tuesday, August 8, 2006 12:28 PM

24 High-Speed Photonic Devices

DFB-LD

Electrical
Input

TW-EA

Electrical
Input

Optical Output

FIGURE 2.15 Schematic diagram of TW-EA-DFB.

[70] as shown in Figure 2.16; the device characteristics, polarization insensitivity,

and low insertion loss are simulated for high-speed, low-driving-voltage operation.
The 3-dB bandwidth and the driving voltage required to achieve a 20-dB extinction
ratio are estimated to be over 250 GHz and less than 2V, respectively, assuming that
the speed is limited by the device capacitance. This structure enables us to apply a
parallel field for MQW layers. The exciton resonance can easily broaden and dis-
appear with a smaller field intensity than that of a perpendicular field (QCSE) [12].
On the other hand, the i-region width between an n- and p-doped region can be
designed to optimize the field intensity and coupling efficiency between an optical
fiber and the modulator. This is another characteristic for this structure, because the
previous structure (see Figure 2.7) has limited thickness of MQWs for low driving
voltages that results in large device capacitance and limited optical confinement and
coupling efficiency between an optical fiber and the modulator.

2.4.2 HIGH INCIDENTAL OPTICAL POWER OPERATION

Another topic of current research is high allowability of incidental optical power.
At present, the incidental optical power increases with the need for dense wavelength

P-InP N-InP

S. I. InP

FIGURE 2.16 Schematic diagram of a lateral p-i-n MQW modulator.

IP369_C002.fm Page 25 Tuesday, August 8, 2006 12:28 PM

Electroabsorption (EA) Modulators 25

p-contact layer

AuZnNi
p-InP cladding layer

Fe-InP Fe-InP

Al MQW layer

n-InP sub.

AuGeNi

(a)

(b)
FIGURE 2.17 A cross-section of a fabricated modulator buried in semi-insulating InP (a),
and its near field pattern (b).

division multiplexing (DWDM) systems; however, there are few reports on this topic.
It was confirmed that the upper limit of allowability is determined by the product
of absorbed photocurrent and applied voltage. The value of an EA modulator with
a semi-insulating buried heterostructure (SIBH) (see Figure 2.17) is superior to that
of the high-mesa structure that almost all reported discrete EA modulators are made
from [71]. Figure 2.18 shows the relationship between breakdown voltage and
absorbed photocurrent. Compared with high-mesa structures, the SIBH can with-
stand high-input power and is not observed to deteriorate even at the greatest
experimentally obtained power level (20 dBm). This is considered to be due to the
difference of thermal conductivity/capacity between the two structures, and SIBH
modulators are expected to have high reliability.
It has been a serious problem for a long time that modulation characteristics of
the usual Fe-doped SIBH EA modulators are inferior to those of high-mesa modu-
lators because of the interdiffusion between the transition metal (Fe) and the zinc
(p-type dopant in the cladding layer). Ruthenium-(Ru) doped SIBH modulators have
been developed [72] and their modulation characteristics such as extinction ratio,
IP369_C002.fm Page 26 Tuesday, August 8, 2006 12:28 PM

26 High-Speed Photonic Devices

20

High mesa SIBH

Breakdown voltage, V
0.09 um
15

0.09
10 0.2
0.3

5
4 6 8 10 12 14
Photocurrent, mA

FIGURE 2.18 Breakdown voltage versus absorbed photocurrent for high-mesa structure and
SIBH modulators with MQW guide thickness as a parameter. Incidental light is TE-polarized
with a wavelength of 1.55 µm.

frequency response, and dark current are superior to those of Fe-doped SIBH mod-
ulators and similar to those of high-mesa structures.
In EA modulators, photocarriers are generated inevitably, and piling up of the
carriers is induced by high-input power degrades rf response. To reduce the density
of photocarriers, a small valence-band offset and thin barrier have been designed as
discussed in Section 2.1.3 and Section 2.1.4. A strain-compensated structure with a
compressive InGaAsP QW and a tensile InGaAsP barrier to provide small valence
band discontinuities has been proposed to reduce the hole escape time from QWs
[73] at the expense of polarization insensitivity. This choice is permitted when
applying this technology to the application of monolithically integrated light sources.
Moreover, the introduction of InGaAlAs barrier [74] or tensile-strained asymmetric
QW [75] has been proposed and demonstrated for this purpose. These trials have
been made possible thanks to a better understanding of the chirping behavior of
these structures, as discussed in the following section.

2.4.3 CHIRPING REDUCTION FOR EA MODULATORS

The primary factor for developing external modulators is considered to be a strong
demand for devices that operate with low or negative chirp. At present, low-chirp
operation has been achieved at the expense of increased insertion loss and reduced
extinction ratio [76]. Although polarization-insensitive, strain-compensated InGaAlAs/
InGaAlAs MQW modulators have been developed, negative-chirp operation without
increased transmission loss has not yet been accomplished. Based on the applied bias
dependence of chirp parameter a for lattice-matched QWs with 20-nm thickness, the
a is almost zero even at zero bias [22], although the quantum size effect is small. This
indicates that poor electron and/or hole confinement will give negative a parameters.
In fact, negative a parameters have been reported for bulk modulators based on
IP369_C002.fm Page 27 Tuesday, August 8, 2006 12:28 PM

Electroabsorption (EA) Modulators 27

Franz-Keldysh effect (no QCSE) even zero bias [77]; however, the detuning energy
was small (about 28 meV) and insertion loss was large (about 15 dB). Various trials
for reducing the escape time of photocarriers and the degradation of frequency response
have been done as discussed in the previous section; these result in reducing the a
parameters. However, at present the a parameter is still larger than desired under
conditions of nonzero bias. More investigation is necessary.

2.4.4 OPTICAL GATE INTEGRATING A UTC-PD AND TW-EA

MODULATOR
Ultrashort optical pulses synchronized to an electrical clock are required in ultrahigh-
bit-rate soliton transmission and optical signal processing systems. When we apply
a large signal to the modulator, the transmitted light is narrower than half the width
of the driving sinusoidal wave, and short optical pulses can thus be obtained under
a high-frequency operation [78]. The width depends on the repetition frequency and
driving voltage to the extent that the modulator can respond to the driving frequency.
Low-driving-voltage EA modulators enable us to directly drive optical signals by a
high-output uni-traveling carrier photodiode (UTC-PD) [79] without electrical
amplifiers. This optical gate demonstrated a 40-Gbit/s 4:1 demultiplexing operation
[80]. Recently, monolithic integration of a UTC-PD and an EA modulator has been
reported and demonstrates a 160 Gbit/s 16:1 demultiplexing operation [81] and a
320-Gbit/s 32:1 demultiplexing operation [82]. Figure 2.19 shows the layout of
PD–EA modulator, and good eye opening was observed for the eye diagrams of the
input 160- and 320-Gbit/s data, respectively, and demultiplexed 10-Gbit/s data. An
optical gate integrating a UTC-PD and TW-EA modulator will be one of the key
devices for signal processing in high-bit-rate TDM channels.

TW-EAM Terminal resistor

RT

UTC-PD

C C

PD bias EA bias GND

PD UTC- EAM
bias PD TW-EAM bias
200 µm
C C

GND
RT

FIGURE 2.19 Equivalent circuit and layout of PD-EA modulator.

IP369_C002.fm Page 28 Tuesday, August 8, 2006 12:28 PM

28 High-Speed Photonic Devices

REFERENCES
1. K. Wakita, Semiconductor Optical Modulators, Kluwer Academic Publishers, 1998.
Contains detailed references of semiconductor modulators up to 1995.
2. W. Franz, Z. Naturforsch., 13a, 484, 1958.
3. L. V. Keldysh, Zh. Eksp. Teor. Fiz., 34, 1138, 1958 [Sov. Phys.—JETP, 7, 788, 1958.
4. N. Bottka and L. D. Hucheson, J. Appl. Phys., 46, 2645, 1975.
5. J. C. Campbell, J. C. DeWinter, M. A. Pollack, and R. E. Nahory, Appl. Phys. Lett.,
32, 471, 1978.
6. N. K. Dutta, and N. A. Olessen, Electron. Lett., 20, 634, 1984.
7. Y. Noda, M. Suzuki, Y. Kushiro, and S. Akiba, J. Lightwave Technol., LT-4, 1445,
1986.
8. H. Soda, K. Nakai, H. Ishikawa, and H. Imai, Electron. Lett., 21, 1232, 1987.
9. T. Okiyama, I. Yokota, H. Nishimoto, K. Hironishi, T. Horimatsu, T. Touge, and H.
Soda, ECOC’ 89, MoA1-3, 1989.
10. N. Henmi, S. Fujita, T. Saito, M. Yamaguchi, M. Shikada, and J. Namiki, in Digest
of Conference on Optical Fiber Communication, OFC ’90, 1990, Paper No. PD8-1.
11. G. Mak, C. Rolland, K. E. Fox, and C. Baauw, Photon. Technol. Lett., 2, 10, 1990.
12. D. A. B. Miller, D. S. Chemla, T. C. Damen, A. C. Gossard, W. Wiegmann, T. H.
Wood, and C. A. Burrus, Phys. Rev., B32, 1043, 1985.
13. G. Bastard, Phys. Rev., B30, 3547, 1985.
14. T. Yamanaka, K. Wakita, and K. Yokoyama, Appl. Phys. Lett., 65, 1540, 1995.
15. B. Broberg, and S. Lindgren, J. Appl. Phys., 55, 3376, 1984.
16. S. R. Forrest, P. H. Schmidt, P. B. Wilson, and M. L. Kaplan, Appl. Phys. Lett., 45, 1199,
1984.
17. G. Bastard, E. E. Mendez, L. L. Chang, and L. Esaki, Phys. Rev., B28, 3241, 1983.
18. U. Koren, B. I. Miller, T. L. Koch, G. Eisenstein, R. S. Tucker, I. Bar-Joseph, and
D. S. Chemla, Appl. Phys. Lett., 51, 1132, 1987.
19. H. Temkin, D. Gershoni, and M. Panish, Appl. Phys. Lett., 50, 1985.
20. S. Nojima and K. Wakita, Appl. Phys. Lett., 53, 1958, 1988.
21. T. Yamanaka, K. Wakita, and K. Yokoyama, InP and Related Materials, IPRM95,
ThP46, May 9–13, 1995, Sapporo, Hokkaido, Japan, p. 544.
22. K. Wakita, SPIE 97, 3038, 39, 1997.
23. A. Yariv, Quantum Electronics, 2nd ed., John Wiley & Sons, 1975, p. 347.
24. J. Nees, S. Williamson, and G. Mourou, Appl. Phys. Lett., 54, 196–1964, 1989.
25. R. G. Walker, IEEE J. Quantum Electron., 27, 654, 1991. Information can also be
found in the GEC-Marconi catalog.
26. N. Dagli, IEEE Trans. Microwave Theory Tech., 47, 1151, 1999. Contains references
of wide-band modulators up to the time of publication.
27. R. Spickermann, N. Dagli, and M. G. Peters, Electron. Lett., 31, 915, 1995.
28. R. Spickermann, S. R. Sakamoto, M. G. Peters, and N. Dagli, Electron. Lett., 32,
1095, 1996.
29. O. Mitomi, K. Noguchi, and H. Miyazawa, IEEE Trans. Microwave Theory Tech.,
43, 2203, 1995.
30. K. Noguchi, O. Mitomi, and H. Miyazawa, IEEE J. Lightwave Technol., 16, 615, 1998.
31. R. C. Alferness, IEEE Trans. Microwave Theory Tech., MTT-30, 1121, 1982.
32. H. H. Liao, X. B. Mei, K. K. Loi, C. W. Tu, P. M. Asbeck, and W. S. C. Chang,
SPIE. Optoelectron. Integrated Circuits, Vol. 3006, 291, 1997.
33. K. Kawano, M. Kohtoku, M. Ueki, T. Ito, S. Kondoh, Y. Noguchi, and Y. Hasumi,
Electron. Lett., 33, 1580, 1997.
IP369_C002.fm Page 29 Tuesday, August 8, 2006 12:28 PM

Electroabsorption (EA) Modulators 29

34. S. Z. Zhang, Y. J. Chiu, P. Abram, and J. E. Bowers, IEEE Photon. Technol. Lett.,
11, 191, 1999.
35. S. E. Miller, Bell Syst. Tech. J., 48, 2059, 1969.
36. Y. Kawamura, K. Wakita, Y. Itaya, Y. Yoshikuni, and H. Asahi, Electron. Lett., 22,
242, 1986.
37. Y. Kawamura, K. Wakita, Y. Yoshikuni, Y. Itaya, and H. Asahi, IEEE J. Quantum
Electron. Lett., QE-23, 915, 1987.
38. A. Ramdane, F. Devaux, N. Souli, D. Delprat, and A. Ougazzaden, IEEE J. Quantum
Electron., 2, 326, 1996. Contains references for monolithic integration of MQW lasers
and modulators up to the time of publication.
39. M. Suzuki, Y. Noda, H. Tanaka, S. Akiba, Y. Kushiro, and H. Isshiki, IEEE Lightwave
Technol., LT-5, 1277, 1987.
40. H. Soda, K. Nakai, and H. Ishikawa: ECOC’88, Technical Digest, 1988, Brighton, p. 227.
41. N. Henmi, S. Fujita, T. Saito, M. Yamaguchi, M. Shikata, and J. Namiki, OFC ’90,
PD8-1, 1990.
42. T. Kato, T. Sasaki, N. Kida, K. Komatsu, and I. Mito, in 17th European Conference
on Optical Communication, ECOC’91, Paper No. WeB7-1.
43. M. Aoki, M. Takahashi, M. Suzuki, H. Sano, K. Uomi, T. Kawano, and T. Takai,
Photon. Technol. Lett., 4, 580, 1992.
44. T. Saito, N. Hemmi, S. Fujita, OEC’90, Tech. Dig., 13A2–4, 1990.
45. T. Okiyama, I. Yokota, H. Nishimoto, K. Hironishi, T. Horimatsu, T. Touge, and H.
Soda, ECOC’ 89, MoA1–3, 1989.
46. M. Suzuki, H. Tanaka, H. Taga, S. Yamamoto, and Y. Matsushima, J. Lightwave
Technol., 10, 90, 1992.
47. P. Ojala, C. Pettersson, B. Stoltz, A. -C. Morner, and M. Janson, ECOC’ 93,
WeC7.4.
48. M. Gibbon, G. H. B. Thompson, D. Boyle, J. P. Stagg, B. Patel, D.J. Moule, S.
Wheeler, E. J. Thrush, K. Warbrich, and P. Gurton, ECOC ’93, WeC7.5.
49. Y. Arakawa and A. Yariv, IEEE J. Quantum Electron., QE-21, 1666, 1985.
50. K. Wakita, I. Kotaka, H. Asai, M. Okamoto, and Y. Kondo, Photon. Technol. Lett.,
4, 16, 1992.
51. U. Koren, B. Glance, B. I. Miller, M. G. Young, M. Chien, T. H. Wood, L. M. Ostar,
T. L. Koch, R. M. Jopson, J. D. Evankow, G. Raybon, C. A. Burrus, P. D. Magill,
and K. C.Reichmann, in Tech. Dig. Opt. Fiber Communication Conf. (OFC ’92), San
Jose, CA, 1992, p.124.
52. K. Sato, I. Kotaka, K. Wakita, Y. Kondo, and M. Yamamoto, Electron. Lett., 29, 1087, 1993.
53. E. Zielinski, D. Bauns, H. Haisch, M. Klenk, E. Satzke, and M. Schilling, in Proc.
InP and Related Compounds, 1994, ME-1, p. 68.
54. J. E. Johnson, T. Tanbun-Ek, Y. K. Chen, D A. Fishman, P. A. Morton, S. N. G. Chu,
A. Tate, A. M. Sergent, P. F. Sciortino, and K. W. Wecht, IEEE Semiconductor Conf.,
1994, M. 47.
55. K. C. Reichmann, P. D. Magill, U. Koren, B. I. Miller, M. Young, M. Newkirk, and
M. D. Chien, Photon. Technol. Lett., 5, 1093, 1993.
56. K. Morito, R. Sahara, K. Sato, Y. Kotaki, and H. Soda, Electron. Lett., 31, 975, 1995.
57. J. Shimizu, M. Shirai, T. Tsuchiya, A. Taike, and S. Tsuji, Tech. Dig. CLEO/Pacific
Rim 2001, Chiba, Japan, 2001, Vol. 2, p. 54.
58. T. Ido, S. Tanaka, M. Suzuki, M. Koizumi, H. Sano, and H. Inoue, J. Lightwave
Technol., 14, 2026, 1996.
59. H. Takeuchi, K. Tsuzuki, K. Sato, M. Yamamoto, Y. Itaya, A. Sano, M. Yoneyama,
and T. Otsuji, Photon. Technol. Lett., 9, 572, 1997.
IP369_C002.fm Page 30 Tuesday, August 8, 2006 12:28 PM

30 High-Speed Photonic Devices

60. Y. Akage, K. Kawano, S. Oku, R. Iga, H. Okamoto, Y. Miyamoto, and H. Takeuchi,

Proc. ECOC 2000, Munich, PD004, 2000; Y. Akage, K. Kawano, S. Oku, R. Iga, H.
Okamoto, Y. Miyamoto, and H. Takeuchi, Electron. Lett., 37, 1, 2001.
61. H. Kawanishi, Y. Yamauchi, N. Mineo, Y. Shibuya, H. Murai, K. Yamada, and H.
Wada, Photon. Technol. Lett., 13, 954, 2001.
62. N. Mineo, K. Yamada, S. Sakai, and T. Ushikubo, OECC 2000, postdeadline paper
2-7.
63. K. Takagi, Y. Miyazaki, H. Tada, T. Aoyagi, T. Nishimura, and E. Omura, in Proc.
2001 Int’l Conf. InP and Related Materials, WA3-2, 2001, p. 432.
64. F. Devaux, S. Chelles, J. C. Harmand, N. Bouadma, F. Huet, M. Carre, and M.
Foucher, in Proc. IOOC95, 1995, p. 56, paper FB3-2.
65. R. Weinmann, D. Baums, U. Cebulla, H. Haisch, D. Kaiser, E. Kuhn, E. Lack, K.
Satzke, J. Weber, P. Wiedemann, and E. Zielinski, Photon. Technol. Lett., 8, 891,
1996.
66. K. Wakita, K. Yoshino, I. Kotaka, S. Kondo, and Y. Noguchi, in Proc. ECOC96, Th.
B.3.2, Brussels, Belgium, p. 1011.
67. B. N. Gomatam and N. G. Anderson, IEEE J. Quantum Electron. Lett., 28, 1496, 1992.
68. S. Kuwano, N. Takachio, K. Iwashita, T. Otsuji, Y. Imai, and T. Enoki, K. Yoshino,
and K. Wakita, in Proc. OFC96, 1996, paper PD25.
69. N. Shirai, H. Arimoto, K. Watanabe, A. Taike, K. Shinoda, J. Shimizu, H. Sato, T.
Ido, T. Tsuchiya, M. Aoki, and S. Tsuji, ECOC 2002.
70. K. Wakita, Y. Toyoda, and T. Kojima, Jpn. J. Appl. Phys., Vol. 41, Part 1, No. 2B,
2002, p. 1175.
71. K. Wakita, I. Kotaka, S. Matsumoto, R. Iga, S. Kondo, and Y. Noguchi, Jpn. J. Appl.
Phys., Vol. 37, Part 1, No. 3B, 1998, p. 1432.
72. S. Kondo, and Y. Noguchi, K. Tsuzuki, M. Yuda, S. Oku, Jpn. J. Appl. Phys., Vol.
41, Part 1, No. 2B, 2002, p. 1171.
73. R. Sahara, K. Morito, K. Sato, Y. Kotaki, H. Soda, and N. Okazaki, IEEE Photon.
Technol. Lett., 7, 1004, 1995.
74. J. Shimizu, M. Aoki, T. Tsuchiya, M. Shirai, A. Taike, T. Ohtoshi, and S. Tsuji,
Electron. Lett., 38, 821, 2002.
75. Y. Miyazaki, H. Tada, S. Tokizaki, K. Takagi, Y. Hanamaki, T. Aoyagi, and Y. Mitsui,
ECOC 2002, paper 10.5.6.
76. J. F. Fells, M. A. Gibbon, I. H. White, G. H. B. Thompson, R. V. Peny, C. J. Armistead,
E. M. Kimber, J. D. Moule, and E. J. Thrush, Electron. Lett., 30, 1168, 1994.
77. K. Yamada, N. Nakamura, Y. Matsui, T. Kunii, and Y. Ogawa, IEEE Photon. Technol.
Lett., 7, 1157, 1995.
78. K. Kawano, O. Mitomi, K. Wakita, I. Kotaka, and M. Naganuma, IEEE J. Quantum
Electron., QE-28, No. 1, pp. 224–230, 1992.
79. T. Ishibashi, N. Shimizu, H. Ito, T. Nagatsuma, and T. Furuta, Tech. Dig., UEO’97,
1997, p. 166.
80. M. Yoneyama, Y. Miyamoto, K. Hagimoto, N. Shimizu, T. Ishibashi, and K. Wakita,
Electron. Lett., 34, 1, 1998.
81. S. Kodama, T. Yoshimatsu, and H. Ito, Electron. Lett., 38, 1575, 2002.
82. S. Kodama, T. Yoshimatsu, and H. Ito, Optical Fiber Communication Conference and
Exhibit (OFC 2003), paper ThX5, Atlanta, Georgia, USA, March 23–28, 2003.
IP369_C003.fm Page 31 Tuesday, August 8, 2006 10:56 AM

3 High-Speed LiNbO3
Optical Modulators
O. Mitomi

CONTENTS

3.1 Introduction .....................................................................................................31

3.2 Device Design .................................................................................................32
3.2.1 Operation Principles ............................................................................32
3.2.2 Design of Ti-Diffused Optical Waveguides ........................................33
3.2.3 Electrode Configuration.......................................................................35
3.2.3.1 Lumped-Type Devices..........................................................36
3.2.3.2 Traveling-Wave Type ............................................................37
3.3 Device Fabrication ..........................................................................................41
3.3.1 Optical Waveguide...............................................................................41
3.3.2 Buffer Layer and Electrode.................................................................43
3.3.3 Fiber Coupling and Packaging........................................................... 44
3.4 Broadband Optical Modulators...................................................................... 45
3.4.1 Methods for Velocity Matching ......................................................... 45
3.4.2 Device Structures and Characteristics................................................ 46
3.4.2.1 Planar z-cut LiNbO3 Substrate ............................................ 46
3.4.2.2 Ridged z-cut LiNbO3 Substrate........................................... 49
3.4.2.3 Thinned x-cut LiNbO3 Substrate......................................... 53
3.5 Functional Modulators ................................................................................... 54
3.5.1 Push-Pull Type.................................................................................... 54
3.5.2 Multi-Input Type................................................................................. 54
3.5.3 Band-Pass Type .................................................................................. 54
3.6 Considerations for Designing Ultrahigh-Speed Modulators......................... 56
References................................................................................................................58

3.1 INTRODUCTION
The ferroelectric material LiNbO3 (LN) has been extensively applied to optical
devices. It has excellent electro-optic and optical properties, that is, it has a large
electro-optic effect and is capable of a high-speed response. It is also transparent
for infrared light and it is easy to fabricate into low-loss channel waveguides by
diffusing titanium. Consequently, various high-performance optical waveguided

31
IP369_C003.fm Page 32 Tuesday, August 8, 2006 10:56 AM

32 High-Speed Photonic Devices

LiNbO3 devices have been developed for the terminal functions of external intensity
modulators, phase modulators, and multi/demultiplexers, as well as switch arrays
for optical fiber network systems. In particular, these LiNbO3 devices are very useful
for optical wavelength-division-multiplexing (WDM) systems because of the possi-
bility of operation in the range of wide-wavelength infrared light with a single device
due to its transparence.
LiNbO3 external modulators have been developed for extensive use in high-
speed and long-distance optical fiber transmission systems. This is because they can
offer the advantages of modulation exceeding 10 Gbits/s combined with a low driving
voltage, and they can eliminate the dynamic laser wavelength chirping which limits
the span-rate system product due to their fiber dispersion characteristics. LiNbO3
external modulators can also offer pure phase modulation in coherent systems and
can realize various optical signal processors. As the bit rate of optical network
systems becomes higher, it becomes more difficult to drive a modulator with a high
voltage due to the restrictions of electrical instruments, in particular, electrical
driving amplifiers. Therefore, reduction of the driving voltage of an LN modulator
with a broadband characteristic is an extremely important issue for realizing future
high-speed optical transmission systems.
This chapter describes linear electro-optic Ti-diffused LiNbO3 devices, particu-
larly traveling-wave high-speed modulators, with respect to the device design and
the fabrication procedures for optical network systems operating within the 1.3–1.6 µm
wavelength region.

3.2 DEVICE DESIGN

3.2.1 OPERATION PRINCIPLES
The LiNbO3 crystal demonstrates the linear electro-optic effect (Pockels effect) pro-
viding a change in refractive index proportional to the applied electric field E [1], that is,

∆n = −n3 ⋅ r ⋅ E/2, (3.1)

where n and r are the relevant refractive index and electro-optic coefficients, respec-
tively, both based on crystal orientation. The largest electro-optic coefficient in
LiNbO3 is r33, which is introduced when the polarized fields of the light and applied
voltage are in the c-axis (z-axis) of the LiNbO3 crystal. Using numerical values of
r33 ~ 30 × 10−12 m/V and n33 (= ne ) ~2.15 at a wavelength l of 1.5 µm, an electric
field of 10 V/µm is found to produce a refractive index change of about 1.5 ×10−3.
This index change induces an optical phase modulation of

∆ f = ko ⋅ L ⋅ ∆ n, (3.2)

where ko = 2p/l and L is the electrode length of the device. The optical phase
modulation can be used to accomplish optical intensity modulation in several ways,
namely, (1) interferometrically (Mach-Zehnder modulator, balanced bridge switch),
and (2) phase match control (directional coupler). Figure 3.1 shows these Ti-diffused
IP369_C003.fm Page 33 Tuesday, August 8, 2006 10:56 AM

High-Speed LiNbO3 Optical Modulators 33

y
LiNbO3
V x Phase-
z(c-axis) modulated
Electrode light
Light
in
Ti-diffused
L optical
waveguide
(a)
Intensity
modulated
Light light
in

(b)

Light
in Intensity
modulated
light

(c)

FIGURE 3.1 Scheme of typical waveguided z-cut LiNbO3 devices. (a) phase modulator, (b)
intensity modulator using Mach-Zehnder interferometer, and (c) intensity modulator using
directional coupler.

single-mode channel waveguided device structures using z-cut LiNbO3 crystals,

where a low-refractive-index buffer layer of SiO2 film is formed between the elec-
trode and the LiNbO3 substrate to prevent light absorption by the metal.
For external modulators, the absence of optical wavelength chirping in modu-
lated signals is essential in high-speed and long-distance optical transmission sys-
tems. A Mach-Zehnder interferometric modulator has an ideal zero chirp [2] with
push-pull operation, i.e., an increase in the refractive index of one waveguide causes
a corresponding decrease in that of the other. Consequently, a Mach-Zehnder inter-
ferometer is normally utilized for high-speed intensity modulators, whereas a direc-
tional coupler is used for integrated switching devices.

3.2.2 DESIGN OF TI-DIFFUSED OPTICAL WAVEGUIDES

The LiNbO3 optical waveguides are formed through thermal diffusion (at about 1000°C
for over 10 hr) of titanium as will be described in a later section. The refractive
index increase ∆ nTi is related to [3]

∆ nTi = A ⋅ CTia, (3.3)

where A is a constant dependent on optical wavelengths, CTi is the titanium concentration

in LiNbO3, and a is a constant (a = 0.8 for extraordinary and 0.5 for ordinary indices).
IP369_C003.fm Page 34 Tuesday, August 8, 2006 10:56 AM

34 High-Speed Photonic Devices

WTi Ti

D x
LiNbO3
Optical waveguide

n2(x)
n02

nx2

n2s

xt1 0 xt2 x
xi xi + 1

FIGURE 3.2 Approximate index profile of optical waveguide for numerical analysis.

From the diffusion equation, the Ti concentration distributions in LiNbO3 can be

expressed using Gaussian functions for the depth direction and error functions for
the lateral direction of the LiNbO3 substrate.
For analyzing Ti-diffused graded-index LiNbO3 optical waveguides, as shown in
Figure 3.2, many calculation methods have been extensively used, for example, the WKB
method, the linear segment method (LSM) [4], and the modified step segment method
(MSSM) [5]. On the other hand, computer numerical analyses by the finite element
method (FEM) and the finite differential method (FDM) are useful for various waveguide
structures. Figure 3.3 shows the calculated results for (a) the propagation constants of
optical waveguides and (b) insertion losses of butt couplings between optical waveguides
and single-mode fibers, where dl and dd , respectively, are the lateral and depth diffusion
lengths. The Ti-pattern width and thickness should be selected under the conditions of
a single-mode optical waveguide and low-loss butt coupling with a fiber.
Low insertion losses in a LiNbO3 device are of fundamental importance for optical
network systems. Factors contributing to loss include waveguide propagation (both
absorption and scattering), reflection, waveguide bending, and Y-branching as well as
fiber-waveguide coupling. Propagation loss can be reduced using optimal fabrication
conditions, as will be described in a later section. Reflection loss is eliminated using an
antireflection coating. Bending loss has been reduced by determining optimal bend
geometries using S-shaped or raised-cosine-shaped transitions for Ti-diffused
waveguides. Conventional Y–branches have a relatively large loss of more than 0.5 dB.
Low-loss branch waveguides have been developed, which use decreasing refractive
indices near the branching region and raised-cosine-shaped waveguides to reduce radi-
ation losses [5,6,7,8]. The Y-branches achieve a low loss of less than 0.2 dB.
Discovering Diverse Content Through
Random Scribd Documents
again now with him. And after all, what is there to envy, anyway?
What sort of a future if I do succeed? The life of a flunkey—a menial
in gold lace, playing for money—and to whom? I’ve been studying
my fellow-creatures this winter—musical people—my audience-to-be.
Copenhagen’s not the world I know; but human beings are much the
same everywhere, I take it, though their looks and manners may
differ somewhat in detail. Grahl has been taking me about. He hates
‘society,’ I know, but he took it all up again for my sake—that’s the
sort of man he is. It all helps, he says. Oh, and you should have
heard their talk, their hard-and-fast opinions, and the views of the
professional critics. Sometimes I feel I simply can’t go on living.
Simply can’t stand it. What wretched caricatures we all are—myself
included. No I’ve finished with this sort of life. There’s not a thing in
the world I care for now, except to go back home. If only I could be
sure that was a genuine feeling, and not another delusion. Don’t
look down on me, old man—Heaven knows, I’ve no great thoughts
about myself just now. You know me well enough to see that I’m not
drunk. But I feel—oh, just worthless. All these years—and living like
this—it’s too contemptible. I feel as if I hadn’t an atom of will-power
left. Just sick and tired of everything ... and longing, aching for
something.... Good of you to listen so patiently. Have a drink.”
Blad was silent for some time, and when at last he spoke it was in
a low voice.
“There’s something I should like to say to you,” he said quietly.
“And I’m half afraid to begin. I’ve been thinking a lot, and some of it
I mustn’t say at all. But I will say this: When we have been together
anywhere—out in the country, or on the sea, or in the town—
anywhere, I always had a feeling that we lived as it were on
different levels, you and I. To me, you were always the born leader;
I felt if you took it into your head to order me about, I should have
to obey. Things seemed somehow to belong to you. Then at other
times, I could feel as if you were a distinguished visitor—one can’t
help these stray thoughts, you know—as if Nature herself put on her
best and did all she could to please you—while I was just an
ordinary person, not worth making a fuss about. I belonged to her,
as one of her children, and could stray about unnoticed among the
trees like any other creature in the forest; it never came into my
head to look on her in that gay lordly way of yours. And sometimes
it seemed you were the better off; sometimes that it was better to
be as I was. It’s all only fancies, of course, but still it does prove one
thing: that we are utterly different. I am quite content to live an
ordinary uneventful life; as long as I can ramble about in Nature’s
garden and cultivate the modest growths of my art, it is enough for
me. I don’t care for anything that calls for greater energy than I
generally give, whether it be the way of pleasure, or pain, or work.
I’ve no ambition worth mentioning. I can sit in my garden, and enjoy
the scent of the flowers, or go out in a boat, and watch the sunlight
on the water; walk in the woods in spring and see the delicate green
of the beech leaves against the sky—I am happy enough with such
things. There are heaps of little trifling things of that sort that please
me every day. But it’s all different with you. It may sound theatrical,
perhaps, but it’s as if you had mountains—glaciers and volcanoes—in
your soul. And I shouldn’t care to change with you—it’s all too big
for me. But then again, if you were like me, I shouldn’t care about
you. You must live and act in a different way; I see that. You could
stand suffering better than I; I’m sure of that. But I’m not quite sure
that you have the power of being really happy. Anyhow—well, you
know I’m your friend, and always will be.”
“I know that, Blad.”
Ormarr got up, switched on the light, looked through a bundle of
newspapers and found the one he was looking for. Nervously he
turned the pages till he came to the shipping intelligence.
“There is a boat leaving the day after tomorrow.”
He dropped the paper, walked up and down the room several
times, shaking his head defiantly, as if at his own thoughts, then
threw himself down in a chair. A moment later he glanced at his
watch, and rose reluctantly.
“It’s time I went round now—to Grahl. The final rehearsal....”
 In the big room where, ten years before, a curious figure of a boy
in ill-fitting clothes had called on him for the first time, Abel Grahl sat
at the piano accompanying the later stage of that same youth—now
a slender, pale-faced young man. They were playing a nocturne—the
only one of Ormarr’s own compositions on the morrow’s program.
The theme was that same one of the sunset with which Ormarr had
introduced himself to his master, only the technique was different.
Ormarr looked out through the window as he played, seeing
nothing in particular. As long as he held his violin, his soul lived only
in the magic world of melody that flowed from the strings.
Grahl’s accompaniment was strangely absent and mechanical. His
figure was bowed at the shoulders, and the black coat he wore
accentuated his thinness. He had aged much of late, and looked
haggard and worn. Now and again he turned his head towards his
pupil with a searching glance.
When they had been through the whole of the programme, Grahl
remained seated at the instrument, striking one chord repeatedly,
his eyes fixed on nothing. The corners of his mouth dropped in a
bitter smile. Then, turning to Ormarr, he said in a queer, strained
voice:
“Play that Andante once more, will you? Not that you need it—it
couldn’t be better. Just play it for me.”
And Ormarr played.
When he had finished, Grahl spoke, without looking up, as to
himself:
“That was one of the things I played at my first concert. I did not
play it as well as you—no, not half so well. I doubt if Beethoven
himself ever played it better!”
For a while he sat with bowed head. Then raising himself
suddenly, he ran his fingers over the keyboard, and the gay tones of
the “Valse d’Espagne” danced like demons out upon the silence that
had followed Beethoven’s Andante.
Ormarr, who had been standing deep in thought, looked round
with a start; Grahl rose from the music-stool with a harsh laugh.
 “A fancy of mine,” he said shortly, “to let Waldteufel loose on the
heels of Beethoven.”
He went across to the table, lit a cigar, and slipped into an easy-
chair.
Ormarr followed his movements intently. There was a strange
expression in his eyes, and the lines on his forehead and face
seemed deeper than usual.
Grahl paid no heed to him; he was smoking, and evidently
occupied with his own reflections. When Ormarr moved, he looked
up, and pointed to a chair.
“Sit down, Ormarr; not time to go home yet. Take a cigar.”
“Thanks.”
Ormarr took a cigar and lit it, covertly watching the expression of
the old man’s face.
“Sit there, Ormarr, where I can see you; that’s it. I was thinking,
there’s not much left of the peasant lad who came up here that
morning ten years ago. The eyes are the same, yes; and a look
about the face—I’ve noticed it the last few days.... Anyhow, it was as
well I didn’t send you away that day after all.”
Ormarr felt his cheeks flush, and bent forward in his chair.
“My dear Grahl, I feel myself a man now in most things, but
there’s one thing that has stuck to me since I was a child. I never
could thank any one in words. And I don’t know how to thank you in
any other way.... I’m sure no father ever did more for his son than
you have done for me. I hardly know how any one could do more for
a fellow-creature than you have.”
“Oh.... And what is this, if you please, if not thanking me in
words?”
“You know yourself how much I owe you—you know I don’t
exaggerate things as a rule....”
“There, Ormarr, that’s enough. You must have seen what it meant
to me all along—the joy and delight of teaching you. No more pupils
now for Abel Grahl. You are my last—and my greatest. If I could find
one greater still...? I don’t think I shall live to be roused from my
bed a second time at six in the morning by a lad with his fiddle in a
calfskin bag and the promise of fame in his eyes.”
Ormarr laughed at the thought. A moment later he was serious
once more. And Grahl went on:
“You’ll go travelling about the world, giving concerts here, there,
and everywhere. I wish I were strong enough to go with you.”
Ormarr laughed again, but without heartiness.
“Grahl, my dear master, why not? Come with me! Nowadays, with
trains de luxe and floating palaces, it will be pleasant as could be.
And at least I should have some one to play for.”
“I ... to travel ... after all? It’s late in the day ... and not exactly
the way I had once thought....”
Ormarr sprang to his feet, but sat down again.
“Grahl, you are my friend—the best I have, I think. I must tell you
something now—something that has happened to me. Listen: I do
not care about the concert tomorrow—it means nothing. Fame is
nothing to me now. To tell the truth, I shudder at the thought of
going about playing for people I do not know, and should not care to
know. Strangers—foreigners! It makes me a piece of common
property; one of the artistic wonders of the world. And then to see
my name, my portrait, on huge posters everywhere ... read
interviews with myself, criticisms of my art—Grahl, the thought of it
sickens me. I won’t—I can’t—oh, if only I could get out of it now,
before....”
“Why, boy ... Ormarr, my dear lad, what is this? what has come
over you? Surely you do not—you could not think of throwing
everything away now—burning your ships? Ten years of hard work—
yours and mine.... If there were any risk, I could understand
perhaps your being afraid ... but as it is ... you have only to show
yourself—one first appearance, and the thing is done. No, Ormarr,
you could not draw back now. It would be madness—nothing else.”
“That may be. But none the less, that is how I feel. I have lost all
desire to show myself, to appear in public. I do not care for any
‘conquest.’ I could do it, I know. But that means that in reality I have
already conquered. It is satisfaction enough to me; I need not show
myself on a platform to utter strangers who have paid so much for
the right to hear me play this or that. Every item on the programme
as a right—and extras in return for their applause. No—if you cared,
I should not mind playing to you every day, for hours together—to
you alone. Or to any others that I cared about. Come back with me
to Iceland. I will look after you, be a son to you, take care of you, in
every way. But spare me this; release me from the burden of that
concert and all that should come after it.”
“Ormarr—you must be out of your senses.”
“Whether or no, I am what I am. And I can’t be otherwise. I am
furious with myself too; blind fool that I have been—oh, you don’t
know what I feel at this moment.”
Ormarr noticed that Grahl was feeling for his watch.
“Don’t,” he put in hastily. “I don’t want to see any one tonight. I
can’t stand it. I don’t know what may happen....”
Abel Grahl rose from his seat. When he spoke, his voice was calm
and earnest.
“Ormarr, remember I stand to you in a father’s stead. You cannot
get away from this. Where is my son, who had grown to be a man of
the world? We had grown out of stage fright, nerves and all that
nonsense, surely? Tomorrow is our concert. We must not forget it,
we must be there in time. But beyond that, we need not give the
matter a thought. There—that’s the way to look at it. Don’t forget.”
Ormarr paled slightly.
“Very well—have it your own way.”
A car was heard hooting outside, and they went out.

Ormarr stood on the platform of the Concert Hall, playing the

Andante from Beethoven’s Sonata. This was the third item on the
programme. The first had been a show piece, from Tchaikowsky,
which had given him an opportunity of displaying his extraordinary
skill and masterly technique. After the second, his own nocturne, it
seemed as if the applause would never end. The audience was
delirious. The atmosphere of the nocturne, with its melancholy
depths and wild heights of joy, its bewildering beauty and strange
transitions, moved the dense crowd as if by magic.
The appearance of the young artist had fascinated his listeners
from the outset. Despite the air of superiority and composure, there
was nothing of arrogance in his bearing. At the first entry of this
young man, with the pale, lean face and the half-closed eyes that
yet seemed to see everything, and see through every one, the
audience felt the magnetism of an extraordinary personality.
Success was certain, inevitable. From the very first, the audience
had surrendered unconditionally.
As he stood there playing, Ormarr appeared quite calm and
collected. Not the slightest tremor of the body, no trace of
expression on his smooth face, betrayed the struggle raging within.
But Ormarr himself knew that it was merely a question of time; up
to a certain point he might control himself—after that, the deluge.
Two men there were, however, among those in the hall, who
suspected something of the strain it cost him to keep his rebellious
temperament in check: they knew that his apparent calm was but a
mask. The two were Blad and Abel Grahl, sitting together in the
front row.
The serene progress of the Andante was undisturbed by any
sound from those in front. Ormarr felt as if his listeners were turned
to stone, and his playing was caressing them like a gentle breeze.
Then suddenly there came over him an irresistible desire to jerk
them back to life—to startle them, set them fluttering and cackling
like a pack of frightened fowls. To tear at their sense, to render their
innermost souls, to fling at them, like a fiery volcanic eruption,
something unexpected and terrible—something unheard of.
In a fraction of a second it had come. A bursting of all bonds that
chained his ungovernable mind: reason, duty, ambition, the fear of
consequences. It was as if in a moment he flung from him the
prejudices and traditions in which men are wont to dress, and stood
there before them in primeval nakedness.
He saw Grahl trying to rise: trying to prevent something he knew
was coming....
And half unconsciously, as if it had been the most natural thing in
the world, he plunged blasphemously from Beethoven’s Andante into
Waldteufel’s “Valse d’Espagne.”
Ormarr was cool and calm as ever, but pale as a ghost. The music
raced away madly into the waltz, laughing and crying in complete
abandon.
A feeling of something uncanny seized the audience for a second;
as if icy waters had overwhelmed them in flood, depriving them of
movement, suffocating all cries for help.
Grahl rose to his feet, and opened his mouth as if to cry aloud.
Then he fell back in his chair, without a sound.
Suddenly Ormarr stopped playing; his arms fell to his sides, and
he stood on the platform laughing—a tremulous, uneasy laugh. Then
he turned and fled.
A storm of shouts and noise rose up from the audience. The
silence of enraptured listeners had given place to the confusion of a
disturbed ant-hill. Some questioned, others raged, a few broke down
entirely.
“Scandalous!” “Mad!” sounded through the din. Several minutes
passed before any thought of leaving. Then suddenly the word
“dead” began to circulate. And gradually the crowd grew quiet, and
dispersed, moved to forgiveness by the thought that the madman
had ceased to live. Only a few were aware that it was not the player
who was dead.

Ormarr reached home and let himself in—not until then did he
notice that he had walked all the way without hat or overcoat, still
carrying his violin.
 After all, what did it matter? His mind was in a state of utter
indifference to everything; completely numbed.
His shoes were muddy, his dress coat wet through; he raised his
hand to his forehead and wiped the rain from his face.
His throat was parched; he felt nervous and ill. He fumbled about
for whisky and a syphon, drained one glass at a draught and poured
out another. Then, drenched and dirty as he was, he threw himself
down on the divan, without a thought of changing his wet things.
The blood throbbed in his temples; there was not a clear thought
in his mind. When he shut his eyes, he felt as if a wheel were
tearing round at a furious rate inside his head.
The door bell rang—it was Blad.
“Grahl is dead!”
Blad threw down Ormarr’s hat and coat, which he had been
carrying; he himself was out of breath, and overpowered with
emotion.
“Grahl—dead?” Ormarr sat bowed forward, his hands clasped, his
eyes staring vacantly before him. Blad stood watching him for a
moment. Then he burst out:
“You—you must be mad!”
“I suppose so—yes.”
“And—you don’t care in the least?”
Ormarr made no reply.
“Think of the scandal of it all!”
Still Ormarr said nothing.
“And then—Grahl! That ought never to have happened.”
“I suppose not.”
“Do you mean to say it is all nothing to you—that you have ruined
your own career for ever, and killed Grahl—your friend—your
teacher? After that—oh, but you must be insane, there’s no other
word for it.”
“Very well, then.”
 “Were you drunk?”
“Drunk? No, I wasn’t drunk. But do let’s talk of something else. It’s
no good discussing this any more. It’s done, and can’t be undone. I
am going back home—to Iceland. There’s a boat leaving tomorrow.
Take off your coat, won’t you—you’re going to stay now? Mix
yourself a drink, man, do.”
“No, thank you.” Blad spoke coldly, flinging out his words, and
pacing the floor excitedly.
“Have I hurt you too? I can’t think how I could have done that.
Surely you can’t feel hurt at my being what I am, and doing what I
can’t help doing? I asked you to stay just now, because I thought
you were my friend. If you are no longer my friend, then you had
better go.”
“Really, I almost fancy you would like to turn me out now because
I decline to drink with you to Grahl’s happy decease. By Heaven, you
do not deserve that I should stay.”
“Oh, you damned fool—who’s talking about what I deserve!”
Blad stopped suddenly, as if paralysed by the word. Then in a
voice heavy with emotion, he said:
“Ormarr—that was the first ugly word I have ever heard you use.
And it was said to me—to me!”
“To you—yes. But you made me angry, you know. Up to then, I
was only miserable—and so hopelessly tired. And here you are
reproaching me for things I could not help. And really, you know,
when you are so utterly foolish as to measure me by your standards,
I can’t call you anything else. I don’t repent what I did tonight. How
can a man repent things that happen—things over which he had no
control whatever? But I do repent—or at least, I am sorry—for what
happened before—for what brought it all about. Grahl was my friend
and benefactor—and yet I cannot feel any grief at his death. I simply
can’t think at all at the present moment; haven’t a single atom of
emotion in me. I’m just a wilderness. Oh, if you knew what I am
suffering now—death would be welcome; a relief. There’s just one
thing that grows and grows in me now—the need to go back, to go
home.”
“And your father—what will he say, do you think?”
“My father? I don’t know. I wonder what he will say. It will be a
big disappointment to him, this. How could I ever have done it? I
don’t understand myself now—it all seems so ridiculous; to lose
control of oneself like that.”
Blad started.
“Then—then you didn’t do it on purpose?”
“Good heavens, no! Did you—could you think that of me? I
suppose you fancied it was a new sort of advertising trick—well, why
not?”
“Ormarr—forgive me. But you were so cool about it all—I never
thought....”
“All right, never mind. We won’t worry about it any more. I’m
dead tired. Stay here tonight, won’t you? I’m not going to bed; no
good trying to sleep. Stay and see me off; the boat goes at nine.
Thanks, that’s good of you. Get some sleep, if you can, yourself.
There’s a lot of things I’ll want you to do for me while I’m away.
Send me—no ... no, I won’t have any of these things here. You can
take them over—keep what you care about and sell the rest. I want
to forget these years—as far as I can. Though I’ve learned much in
the time—and paid dearly for it. Now I am going home—going home
to Iceland, and then ... what next, I wonder?”
 CHAPTER V

I t was a bright wintry day when Ormarr, watching from the

captain’s bridge, saw his native land rise snow-clad from the blue-
green sea against a high, clear sky. The captain noticed that the fur-
clad man who had been up on the bridge since early that morning to
get the first glimpse of land, seemed strangely moved at the sight of
it. Well, it was none of his business....
Never before had Ormarr seen Iceland rising thus out of the sea;
he had but a dim notion of the grandeur of the sight. Unconsciously,
he had always thought of Iceland in the green of spring or summer,
and had looked forward to seeing it so on his return. Being winter, of
course, there would be snow. But he had never thought to see it all
so white and clean and brilliant as now.
A vague joy filled him as he looked; he felt that his soul was come
of the race of those great mountains, as of a line of kings.
Iceland—his country! Like a cathedral, a consecrated pile of
granite, pure and holy in the seas of the far north. And the snow—
how he loved it! And the rocks, the hills and valleys ... the brooks
and streams, sleeping their winter sleep now, under the ice. And fire
too, the marvellous, merciless fire, smouldering quietly in its lava
bed, yet strong enough to melt the ice of a hundred years in less
than a minute and hurl it in huge floods of boiling water and redhot
rocks and lava down the mountain-side, through the valleys, out into
the sea. What did it care for men, or their goods or their lives! All
had to die. And better to die by fire or ice than on a bed of sickness.
Far better to die young in some mighty upheaval than to drag
palsied bones through a dreary wilderness of old age.
Ormarr smoothed his brow.
 Why think of dying now? He was still young, and fit for action. Yet
if Mother Iceland should think fit to crush him to his death in her
embrace, well, he was ready. Well for him, perhaps, to find death on
her icebound, fiery heart, if the road of life proved too wearisome.
Strange thoughts—was he mad, after all? He was thinking now as
he had done so often when a child. But his dreams had changed.
Then, Iceland had been the starting-point of his imaginings; it had
been as a weight at his heel, keeping him in bondage, holding him
back from all that he thought made life worth living. Now it was
changed—now all his dreams turned towards it, centred round it—
Iceland now was his home. Home? No, he had no home anywhere
on earth. Yet he felt drawn towards it none the less; longing for his
country....
But what was this—Iceland—hovering above him, looking down at
him—would she no longer receive him? Was he her child no more?
Had the world worn away the marks by which his mother had known
him?
Foolishness—his brain was running wild. And yet—how was it with
him, after all? Was it not true that he was unworthy of love—a
failure, self-condemned?
Iceland, towering in shining armour, in glittering floes and spotless
mantle of snow. And one coming to her from the outer world, with
the dirt of alien countries on his feet, and the pain and weariness of
the world in his heart. Her sacred places were no longer open to him
now; closed, locked; the keys hidden far away, not there. Perhaps in
the place whence he had come, perhaps far distant, on some other
continent. Or hidden, maybe, on the other side of life.
Iceland! As he watched the land rise from the cold blue waves, he
felt that he, who once had been her child, was no longer worthy to
be so. He had sinned in coming back at all. And he vowed in his
heart to set out once more in quest of the key that might unlock its
holy places to him once more. Whatever happened, he must go
away again. And if he could not find what he sought, then there
could be no return. Only let him first breathe the air here for a little
while, tread the soil that had been his father’s—men who had never
shamed their native land.
Again he smoothed his forehead—the movement had become a
habit with him whenever he wished to check or change a train of
thought. And he laughed harshly.
“Well, Ormarr Ørlygsson, my friend and brother,” he thought to
himself, “this time you are certainly mad ... mad beyond cure ...
caught in the act—hysteria pure and simple.”
He sighed deeply—there was an ache at his heart.
“What is it?” he thought. “If I go on like this ... if I let my thoughts
and fancies play at will like this, I shall end as a lunatic: lose all
control over myself, and be shut up somewhere—a pleasant
prospect! Or at best, be allowed to go about at home in a living
death: a beast with instincts and no soul, on the place I was born to
rule. And father—to see his son an object of pity or contempt.... No:
I must get away now, before something happens. Better perhaps not
to land at all, but go on round the coast, and back with the steamer
to Copenhagen.
“Well, we shall see. Most likely it would be the wisest thing to do.
On the other hand, it would be cruel to father....
“Wait and see. Let me at least feel the soil of my own country
under my feet: touch the snow, drink its water, and breathe its air—
satisfy myself that it is not a vision merely, no fairy tale, but a
reality.”
At the first port Ormarr went ashore. He felt happy as a child, and
laughed and joked with the crew. And when the boat neared the
pier, he waved his hand to the crowd there, though he did not know
a soul among them. They shrank back a little at the gay familiarity
on the part of a stranger—but Ormarr did not care.
He set out on foot to explore the neighbourhood, a poor enough
place it was. It was only with an effort that he restrained himself
from walking up to the windows of the little houses and looking in,
or knocking at the doors, just to breathe the atmosphere of a home
in his own country.
 On an open space some boys were racing about playing
snowballs. This was too much for Ormarr; before he knew it, he was
in the thick of the fight, and in a moment he had all the lads on top
of him. With shouts and laughter they pelted him from all sides, and
ended by fairly burying him in the loose snow.
The boys stood around laughing heartily when at last, gasping for
breath, he emerged; this was a first-rate playmate that had suddenly
appeared from nowhere. Eager queries were hurled at him.
A tall, freckled peasant lad came up and asked his name, others
equally inquisitive put their questions without giving him time to
reply to the first. Was he from the steamer just come in? Where had
he come from? From Copenhagen? What had he been doing there?
Was he going on with the steamer again? If so, he would have to
hurry; the second whistle had already gone.
And the whole crowd followed him down to the harbour, two of
the smaller boys taking each a hand. When he gave them some
small coin, they decided that he must be the new Governor at the
very least, and felt some tremors at the disrespectful manner in
which they had treated such a personage.
As the boat rowed off to the steamer, they stood on the pier
waving their caps, and stayed there, waving and shouting as the
vessel moved off.
Ormarr felt unspeakably grateful for this welcome from his country
—a welcome of smiles, and snow, and youth; the glowing warmth
that was in its element amid the biting cold. He felt himself akin to
these lads, with their hands and faces warm and wet from
perspiration and melting snow; who rolled about in the snowdrifts
despite their clothing, braved the cold and the roughness of the
elements, enjoying themselves in the depth of an arctic winter as
well as in any tropical summer heat. They had no idea of modern
precautions against climate.
There they stood, waving to him, acknowledging him as one of
their own, never dreaming that he had been about to drift away into
an artificial life that nursed the frailties of the body regardless of
health, until the body became a thing to loathe, unless the soul itself
were cynically hardened.
This was the moment for action, the time to pull oneself together
and decide; here was the way to follow—follow it!
But first of all, to find the right way.
Ormarr felt now that he could go back to his father. Could tell him
all, confess that he had chosen a wrong path, a way whereby his
body might have passed unscathed, but his soul never—it was never
meant that the two should be divided. He must rest and think for a
while and find a new road.
Once more Ormarr had climbed to the bridge, and remained there
till the steamer touched at the next port. It would be a couple of
days before he could reach home.
The day wore away, and night came down, but it was still quite
light. The moon was high, right over the land, its white glow
hovering over the landscape and giving it an air of unreality, like a
spell that held all things in the bonds of sleep. The ship itself,
chained to a silver beam, was the captive of this enchanted country,
for all that it kept on its course; sooner or later, it seemed, the time
would come when it must crash on a rocky coast.
Ormarr turned from the moon, forgetting the base designs which
he had just attributed to its dull red bridge of rays. He looked at the
stars—and suddenly he remembered the summer nights at home,
when he had lain out among the hay in the fields, unable to draw his
eyes from the twinkling golden points of light.
The northern lights flickered and faded, and showed up anew; like
fiery clouds, appearing suddenly on one horizon, to vanish in a
flaming trail behind another. Ormarr loved them—their restlessness,
their capricious, fantastic shapes, the play of mood through every
imaginable shade of colour—it was a silent musical display of
heavenly fire.

Next day, the captain and Ormarr were alone on the bridge. Each
was occupied with his own thoughts, and both were gazing towards
the shore.
The captain broke the silence.
“See there, Hr. Ørlygsson—that ring of mist there round the peak.
Now, mist, I should say, is white as a rule, but looking at it there,
against the snow, it looks just grey.”
Ormarr made some brief reply; he was studying the face of the
little Danish captain.
The latter spoke again:
“I don’t know if you know this part of the country at all. When we
round that point just ahead, you will see one of the strangest fjords
all round the coast, though that’s saying a good deal. Rocks sticking
up out of the sea, sharp as needles some of them, and some all
tumbled about in groups; some look like houses, and there are a few
that make gateways, as it were, real arches, that you can take a ship
through if you like.”
“Then we shall be in very soon, I suppose—and up to time for
once.”
The little Dane drew himself up stiffly, glanced coldly at Ormarr,
and said:
“Begging your pardon, sir, my ship is always up to time.”
“Why, then, it is I who must ask your pardon, Captain Jantzen.”
“Always excepting pack ice and being hung up by a gale,” added
the captain in a milder tone. “Otherwise, I admit you’re right about
being up to time generally—my ship’s an exception, that’s all. I put it
plainly to the owners: either give me a time-table that I can keep to,
or find another skipper. It’s a point of honour with me, as you might
say. As a matter of fact, there was another Iceland boat once came
into port on the day fixed—only it was just a month late.”
The captain laughed at his own jest, and Ormarr joined in. Then
Captain Jantzen went on:
“Really, you know, it is a shame that there should be such a
wretched service of steamers in these waters. There are several
companies, I know, but they simply agree that there’s no sense in
competition, so they keep up freights, and run their ships as they
please. You may often have to wait weeks for a boat, and then find
the sailing’s cancelled for some reason or other. Yes, there’s a
chance for a man with energy and capital, that’s certain.”
Ormarr started at the other’s words; it was as if a mist faded from
before his eyes; here before him was a chance to redeem himself.
He turned to the captain and looked at him searchingly; a good
man, by the look of him, and with determination in his face.
Suddenly he noticed that the man lacked one finger on his left hand
—strange, Abel Grahl too had lost a finger. The coincidence seemed
to form a bond between himself and the captain. Fate, perhaps—
why not?
He shook his head, smiling at himself for the superstition.
Nevertheless, he asked the captain:
“Ever taken a turn with Fate, Captain Jantzen?”
The captain smiled, a mirthless smile that might have been a
setting of his teeth.
“I should think so,” he said, with an air of definite certainty, as if
answering question about a harbour he knew blindfolded. “And if
you haven’t, I’ll give you a bit of advice: take it by the horns straight
away; don’t wait on the defensive, attack at once. There’s this about
it: when luck favours a man, and he’s sound enough not to get
spoiled by it at once, sure enough, Fate will try to get a foot on his
neck.”
He stretched out his left hand towards Ormarr, showing the index
finger missing, and went on:
“It cost me that. I was a deck hand on a fishing-boat at the time,
though I knew the sea, and had many a rough turn with it, and
saved more than one from drowning. And that’s a thing the sea
won’t forgive. One day I was alone on the foredeck, getting the
anchor ready, when there was a hitch in the cable. And then a thing
happened that I’ve never known before or since—my feet slipped
sheer away from under me, as if some one had pulled them. I came
down headlong, and the anchor tore away to the bottom of the sea,
taking me with it. My finger was caught between two links of the
cable—there was no getting it free. I thought to myself, ‘Not this
time, anyway,’ and managed to get at my knife, and hacked it off. It
didn’t seem to hurt much while I was in the water—but when I came
up—the men—believe me or not, as you will, but they started back
when they saw my face. I hurried down below—I had a sort of
feeling what it was. And I tell you, sir, there was the mark of death
in my face when I looked; the mark Fate puts on a man before
handing him over. And it was twenty-four hours before it passed off.”
Captain Jantzen laughed.
“Since then, Fate’s left me alone. Maybe she never found out how
I’d cheated her. And if she has forgotten, why, maybe I shall live to
be an old man after all.” And as if repenting his levity, the little
captain became serious once more.
“All the same, it’s not right to joke about that sort of thing.”
Ormarr had listened with interest to the captain’s story. When he
had finished, he was silent for a moment, then asked:
“How long have you been captain of ‘Bjørnen,’ Captain Jantzen?”
“Why, it’ll be twelve years this spring.” And in a tone of some
resignation he went on:
“It’s not likely I’ll have her for another dozen years. Though I’d
like to. She’s a fine boat, and somehow we sort of belong to one
another. But the owner’s getting on now, and his health’s not what it
might be. And no sons. I fancy the other shareholders are not quite
pleased with things as it is.”
Ormarr walked up to the captain, and looking straight at him,
asked abruptly:
“What about buying them out?”
Jantzen started, and looked inquiringly at Ormarr.
“I mean it.”
“Well—yes, I dare say. It’s a limited company. The biggest
shareholder is the owner—and if any one were to buy up all the
other shares on the quiet, well, there’s no saying....”
 Ormarr and the captain seemed suddenly to have become
remarkably intimate with each other—so, at least, it seemed to the
others on board.
They remained for a long time in the captain’s cabin, bending over
a map of Iceland, discussing routes, tariffs, and traffic in a half-
whisper. They talked of nothing but how many vessels and what size
would be needed if one company were to take over the whole of the
goods and passenger traffic between Iceland-Denmark, Iceland-
Norway, and Iceland-Great Britain.
It was late when Ormarr shook hands with the captain and went
to his bunk, with the parting words:
“Then the first thing you have to do is to buy up all the shares on
the market. After that, get the old man to sell his holding—but to me
and no one else!”

The following morning, Ørlygur à Borg was standing on the

borders of his land, deep in thought. He had dreamed a strange
dream the night before, and was trying hard to remember the
details. One thing only stood out plainly in his memory. He had been
standing on this very spot, a little hill just outside Borg, one day
towards the end of summer. And there he had fought—with what, he
could not say. But it was against something stronger than himself,
something which would overpower him unless Ormarr, his son, came
to his aid. Then suddenly he had seen a viking ship rounding the
point, steering straight up the fjord. The sight of the vessel gave him
new strength; he knew that Ormarr was coming to help him, and the
ship was sailing faster than any he had ever seen.... Here the dream
had ended abruptly.
Ørlygur stood on the hill, trying hard to recall more of the vision.
As if to aid his memory, he looked out in the direction of the fjord....
A steamer was rounding the point.
Ørlygur à Borg lost no time; he ran to the stables, and saddled his
horse. He was about to saddle another in addition, but checked
himself—possibly it was only an important message. Anyhow,
instead of mounting, he had a sleigh brought out, and drove off
towards the snow-covered valley at full speed, reaching the trading
station just before “Bjørnen” came in.
Ormarr was not a little surprised to find his father among the
crowd of people gathered on the shore. Most of those present had
recognized Ormarr where he stood on the bridge, and there was a
general surprise at his appearance. No one had expected him. Only
his father seemed to regard his homecoming as natural, and showed
no sign of astonishment.
Ormarr was in high spirits and full of pleasant anticipation; he
shook hands right and left. Ørlygur found it hard to conceal his
emotion at the meeting.
Ormarr introduced Captain Jantzen to his father, but the latter
spoke only a few words to the captain; he seemed intent on getting
home without delay, where he could have his son to himself.
Before taking his seat in the sleigh, Ormarr took the captain aside:
“Remember,” he said, “you must get everything ready beforehand.
First of all, a detailed scheme and tariff rates, for our calculations. I
shall be here all winter. After that, I am going to England and
France, to get the money. I shall get it, never fear. Anyhow, I shall
see you next summer in Copenhagen. And then we can set to work
in earnest. Be ready for a struggle when the time comes—it will take
some doing, but we can do it. Au revoir.”
On the way out to Borg, the horse was allowed to choose its own
pace; father and son were too engrossed in their talk to trouble
about anything else.
Ørlygur could not quite understand his son’s attitude towards
music and fame—possibly because Ormarr himself was loth to lay
bare all the trouble of his mind. Moreover, he felt a different man
already, far healthier in mind and body, after the last few days, as if
separated by a wide gulf from the Ormarr who had left Copenhagen
after the scandal at the Concert Hall, a broken man, to seek rest and
idleness in his own country.
 Ørlygur could not altogether grasp his son’s changed attitude
towards the question of his musical career, which had cost ten years
of his life and several thousand pounds. But he thoroughly
understood and approved of his new plan for a better and cheaper
and more reliable service of steamers between Iceland and abroad.
Ormarr pointed out the advantage of having an independent
national steamship service, and Ørlygur at once perceived the
possibilities of the scheme for furthering the development of Iceland
commerce and industry. The idea of excluding other countries from
participating here appealed to him, and gained his entire support for
the scheme. The very thought thrilled the old chieftain’s heart. Ay,
they deserved no better, those slack-minded, selfish traders—they
would only be reaping the results of their own shortcomings. They
should no longer be allowed to monopolize trade, send up prices,
make unreasonable profits, and do what they liked generally. There
would be an end of their ill-found, ramshackle vessels, coming and
going at their own convenience without the slightest regard for the
public or their own advertised times. It was war—and he rejoiced at
it. No question but that the people of Borg must win in the end.
As they were nearing home, Ormarr said:
“I am going to stay here this winter, father, before I set out again
—Heaven knows how long it may be before I come back after that. I
should like to live to enjoy one more spring here in Iceland. But after
that, I must go abroad; work, work. It will take best part of the
summer, I reckon, to raise the money—it will need a lot of money.”
Ørlygur gazed thoughtfully at the landscape, and answered:
“Well, well—I suppose you are right.”
For a while no sound was heard but the beat of the horse’s hoofs
and the creaking of the sleigh. Then Ørlygur said in a half-whisper:
“But—we have some money here, you know, ourselves.”
Ormarr looked at his father keenly, and after a moment’s thought
he said:
“Look here, father, I will tell you what I have thought of doing
about the money part of the business. I want to get the money
without offering shares. It will be difficult, I dare say. But I must be
independent here; I cannot bear to be bound by considerations of
credit, or other men’s interests, and that sort of thing. It would spoil
the whole thing. The business must be my property; I will not have
a thing that can be ruined by others after I have built it up. But if I
should be unable to get the capital in the way I want it—why, then, I
may come to you. Provided, of course, I can be sure of running no
risk in the investment. I owe you too much already.—My inheritance,
you say? I have not come into the property yet. But suppose we put
it that way; that I owe so much to the estate. Anyhow, I owe it; it is
money that must be paid, if things do not go altogether against us.
For the present, I must fall back on you. But I shall not want much—
nothing like what I have been drawing up to now. And I am proud
that you are willing to help me, when I know I must have
disappointed you by what I have done up to now.”
“I trust you, Ormarr,” his father said. “I do not quite understand,
but I feel sure you were obliged to act as you did. The rest does not
concern me. I know that you are honest and sincere, and I know
that your aim now is not a selfish one.”
For a time no more was said; both men seemed anxious to let it
appear that their minds were occupied with anything rather than
with each other. But for all his apparent calmness, Ormarr was
overwhelmed with gratitude to his father; to the fate that had given
him such a father; given him Borg for his inheritance, and suffered
him to be born a son of this little nation. Ørlygur, on his part,
concealed beneath an expression of indifference a feeling of pride
and love for his son.
As the sleigh drove up in front of the house, all the servants came
out to welcome Ormarr, with a heartiness that showed plainly
enough for all their quiet manner. A tall girl of about thirteen, with
lovely flaxen hair flowing loose about her shoulders, appeared; this
was Gudrun, a daughter of Pall à Seyru, now adopted by Ørlygur.
Ketill was nowhere to be seen; Ormarr asked where his brother was.
Ørlygur smiled.
 “Have you forgotten already? I wrote you in my last letter that I
had sent him to the school at Rejkjavik. He wants to enter the
Church, I understand. And I have been thinking that it would not be
a bad idea later on, if he took over the living here. If, then, you
decide to live abroad, as seems likely, and give up the estate here,
then he could manage that as well. For the present, I have my
health and strength, and hope to look after it myself for many years.
We shall see.”

Of Ormarr’s stay at Borg that winter there is little to be said. Every

Sunday the people of the parish came up to hear him play the violin.
He was delighted to play to them, and touched at their grateful,
almost devotional, reception of his playing.
Spring came. The snow melted, and the rivers sent floods of
muddy water and blue ice towards the sea. A great unrest came
over Ormarr, and he left earlier than he had planned. So, after all, he
missed the soft purity of the Iceland spring, the beautiful white
nights with the glow of light on the fields and ridges pearled with
dew. He missed the sight of the butterflies fluttering in gaudy flocks,
and the birds among the little hillocks where their nests lay hid.
He had already felt the grip of spring at his heart when he saw the
wild swans and other fowl heading for the still frozen heights farther
inland, driving their wedges through the air, and crying aloud in joy
of life. And that same viking spirit which had driven his fathers
before him came on him now and drove him abroad in haste.
As he left Iceland for the second time, his father stood on the pier
with moist eyes. Ørlygur remained there, watching till nothing was
to be seen of the vessel but a few grey wisps of smoke. Then he
tore himself away, mounted his horse, and rode home, deep in
thought.
If his blessing carried any weight, then surely matters would go
well with his son.
He slept but ill that night; he was sorry he had not prevailed upon
Ormarr to accept the money from him. It would have saved much
trouble, and, at any rate, a certain amount of time.
If only Ormarr had come to him, rather than procure the funds he
needed from others, and upon doubtful terms....
 CHAPTER VI

T he cold, pure light of an autumn morning found the electric

lamps still burning in a villa by the Sound. It was the residence
of Ormarr Ørlygsson, company director, a man well known in the
business world, and bearer of sundry decorations.
The light shone through the rose-coloured curtains of the French
windows opening on to a verandah facing the sea. The room was
large; the arrangement marked its owner as a bachelor. It served as
office, sitting-room, and study. The wall opposite the window was
occupied entirely by shelves filled with books: works of reference
and lighter literature. The other walls, each with a heavily curtained
door, were hung with paintings, all representing Icelandic
landscapes. In one corner was a heavy piece of bronze statuary,
likewise Icelandic, “The Outlaw.” The floor was covered with an
Oriental carpet.
Ormarr sat at the big writing-table, his head buried in his hands.
Lights burned in a crystal globe above his head, and in a reading-
lamp at his elbow. The glow from the green shade of the latter,
blending with the light of day, created a weird effect.
Ormarr had been sitting at his desk the whole night, going
through piles of accounts and business papers.
For some time he sat thus, motionless. When at last he looked up,
it was plain that thirteen years of work as a business man had left
their mark on him. His face was thinner; his dark, rough hair was
longer than was customary among men on the bourse, and the fact
gave a touch of independence to his otherwise faultless appearance.
His expression was changed; the large, dark eyes were restless—a
dreamy, far-away look alternating rapidly with a glance of keen
alertness. When alone, his look varied continually with his varying
moods, but in the presence of others he kept rigid control over his
features; the severest scrutiny could detect nothing of the workings
of his mind. Two deep furrows slanted down on either side the
mouth, completing the impression of resolute firmness combined
with melancholy resignation and bitterness.
As he looked round the room, his eyes betrayed the trouble in his
mind. He glanced deliberately at each of the things around him,
works of art and furnishings, as if in farewell, dwelling now and then
on some single item as if trying to fix it in his mind.
Gradually he began to realize that his first impression of the
previous day was correct—he was a stranger in his own place. And
he shuddered at the thought. Unconsciously he picked up the cable
he had received the day before, smoothed it out before him, and
read it over with bitter, scornful eyes.
“What a fool I have been!” he muttered. “I might have known....”
And he laughed—a choking, unnatural laugh, and rose slowly to his
feet. Languidly he drew back the curtain, opened the window, and
stepped out on to the verandah.
Leaning on the railing, he looked out over the shore, with the
troubled sea and the Swedish coast beyond. The view had calmed
him often, but there was no rest in it now; he looked at it all
impatiently, no longer able to find any comfort in visions.
All was changed now.
His clothes irked him; his hands were soiled with dust from the
papers he had been busied with; a general sense of bodily
discomfort pervaded him. And as if to escape from his emotional
self, he left the room hurriedly; a bath and a change of clothes
would be something at least....

The housekeeper received her master’s orders to serve lunch on

the verandah with some surprise. It was a way of hers to appear
mildly surprised at things and today there certainly seemed some
reason for astonishment: for thirteen years her master had never
been at home to a meal at that hour of the day—why was he not at
the office as usual? Ormarr’s manner, however, forebade all
questioning, and she did not venture to ask if anything were wrong.
Ormarr went to the telephone, and rung up the office, speaking
coolly enough.
“That you Busck? Good morning. Captain Jantzen there? Morning,
Captain.... No, nothing wrong, but something has happened. Yes ...
listen! You must hand over ‘Bjornen’ to the first mate this voyage....
What? Lose half an hour? Can’t be helped; I want you here. Come
out here at once, please, but first get the chief clerk to tell you what
I want done about the shares, and do as he says. Then out here to
me as quick as you can. I’ll tell you all about it when you arrive.
Right—good-bye.”
A few minutes later the telephone bell rang. Ormarr took up the
receiver with a gesture of annoyance, but on recognizing the
speaker’s voice, his manner changed.
“Yes—yes. Morning, Ketill. Ill? No, not a bit. Are you both there?
Well, come out and have lunch with me instead. Don’t know what
we’ve got in the house, but come anyway. Eh? No, not a bit. I have
been rather busy—up all night.... No, never can sleep in the daytime.
Right, then. Au revoir.”

Ketill, now getting on for thirty, was already in orders, and was to
be presented to the living of Hof in Hofsfjordur in the autumn, Sera
Daniel being about to retire on account of age.
The original plan had been that Ketill should have spent a few
days only in Copenhagen when going abroad in the spring, on his
way to Switzerland and Italy, returning via England. But Ketill, who
had preferred staying at an hotel rather than at his brother’s, had
soon found friends, largely owing to his brother’s introductions. One
of the acquaintances thus made was that of a banker, Vivild, whose
daughter Alma had quickly captured Ketill’s heart.
His tour of Europe, then, came to consist of but a few short trips,
with Copenhagen as his headquarters. Ormarr had been surprised at
this, but his brother gave him no enlightenment as to the attraction
which drew him constantly back to the capital. Until one fine day
Ketill announced his engagement and forthcoming marriage.
Ormarr had always looked on Alma as a tender plant, that could
never be transplanted and live; the news surprised him. But he
made no comment. Without realizing it himself, he had been deeply
in love with dainty, sweet-natured Alma, but for no other reason
apparently than a sense of his own unworthiness, had said no word
of it to her. And here was his brother, holding the blossom himself,
and tantalizingly inviting him to admire its sweetness.
The part of brother-in-law was by no means a pleasant prospect
to Ormarr, but he reconciled himself to the thought.
Ketill—Sera Ketill, as we should now call him—was young and
good-looking, with a pleasant and genial bearing. At times Ormarr
could not help feeling that there was something a trifle insincere in
his brother’s geniality. Still, Ketill was a nice enough fellow to all
outward seeming, albeit a trifle stouter of build than need be.
There was never any exchange of confidence between the two
brothers; they knew, indeed, but little of each other. Ormarr was
conscious of an involuntary dislike of Ketill; he tried in vain to
subdue the feeling; it remained unaltered. Ketill, on the other hand,
appeared not to notice any lack of brotherly love and sympathy.
Neither of the two men realized that Ketill’s nature not only did not
invite, but rendered impossible any real confidence.
The first to notice this, albeit but vaguely to begin with, was Alma.
The discovery troubled her a little, but she let it pass.
From all appearances, the union was a promising one, and the
wedding was looked forward to by both parties with equal
anticipation. The ceremony was to take place on the day before
Ketill’s entering upon his new dignity, and the bride was to
accompany him to their new home.
Alma and Ketill arrived at Ormarr’s house half an hour after Ketill
had rung up. Alma promptly went out to assist the housekeeper with
the lunch.