Related Titles

Ohtsu, M. (ed.)

Frequency Control of Semiconductor Lasers

1996

Print ISBN: 978-0-471-01341-9

Physics of Optoelectronic Devices

1995

Print ISBN: 978-0-471-10939-6

eMobi-lite ISBN: 978-0-470-30140-1

Coldren, L.A., Corzine, S.

Diode Lasers and Photonic Integrated Circuits

1995

Print ISBN: 978-0-471-11875-6

eMobi-lite ISBN: 978-0-470-30154-8

Fukuda, M.

Optical Semiconductor Devices

1999

Print ISBN: 978-0-471-14959-0

eMobi-lite ISBN: 978-0-470-29851-0

Saleh, B.E., Teich, M.C.

Fundamentals of Photonics, Online Version

2nd Edition

2001

Print ISBN: 978-0-471-21374-1

Choi, H.K. (ed.)

Long-Wavelength Infrared Semiconductor Lasers

2004

Print ISBN: 978-0-471-39200-2

eMobi-lite ISBN: 978-0-470-31334-3

Adobe PDF ISBN: 978-0-471-64980-9

ISBN: 978-0-471-64981-6

Luryi, S., Xu, J., Zaslavsky, A. (eds.)

Future Trends in Microelectronics

The Nano, the Giga, and the Ultra

2004

Print ISBN: 978-0-471-48405-9

May, G.S., Spanos, C.J.

Fundamentals of Semiconductor Manufacturing and Process Control

2006

Print ISBN: 978-0-471-79028-0

Korvink, J.G., Greiner, A.

Semiconductors for Micro- and Nanotechnology

An Introduction for Engineers

2002

Print ISBN: 978-3-527-30257-4

ISBN: 978-3-527-60022-9

Adobe PDF ISBN: 978-3-527-61625-1

Epstein, R., Sheik-Bahae, M. (eds.)

Optical Refrigeration

Science and Applications of Laser Cooling of Solids

2009

Print ISBN: 978-3-527-40876-4

ISBN: 978-3-527-62804-9

Adobe PDF ISBN: 978-3-527-62805-6

Okhotnikov, O.G. (ed.)

Semiconductor Disk Lasers

Physics and Technology

2010

Print ISBN: 978-3-527-40933-4

ISBN: 978-3-527-63039-4

Adobe PDF ISBN: 978-3-527-63040-0

Khanh, T., Bodrogi, P., Vinh, Q., Winkler, H. (eds.)

LED Lighting

Technology and Perception

2016

Print ISBN: 978-3-527-41212-9

ISBN: 978-3-527-67014-7

eMobi ISBN: 978-3-527-67015-4

ePub ISBN: 978-3-527-67016-1

Adobe PDF ISBN: 978-3-527-67017-8

Preface and Introduction

The title of this book is “Compact Semiconductor Lasers” and, in accordance with this title, it aims to provide a nearly comprehensive review of efforts in various research laboratories to create semiconductor lasers that are progressively smaller. In our view, efforts to make semiconductor lasers ever more compact are mainly driven by the need to integrate semiconductor lasers in photonic integrated circuits (PICs) or optoelectronic integrated circuits (OEICs). Two major technical advantages can be derived from being small – taking up less space on a chip, and consuming less power per laser – both of which are necessary for higher density and higher device-count integration.

The primary industrial relevance of the compact semiconductor laser possibly comes from its potential for application in advanced PICs aimed at communication applications. PICs, such as those integrating high quality laser sources, modulators, and on-chip monitoring devices, have already been applied commercially for long distance optical communications. While not yet a major constraint on their application, compactness, and power-efficiency are nevertheless high on the wish list for such components. Yet one needs to be aware that compactness may be an attribute that conflicts with other requirements. For instance, the spectral purity, frequency precision, and tunability that are essential for high precision sources such as distributed-feedback (DFB) and distributed Bragg reflector (DBR) waveguide lasers imply the need for device lengths that are specified as several hundred micrometers, if not several millimeters.

Other emerging important information transmission applications may impose very different requirements on laser sources. Optical interconnects have at their base the familiar silicon very-large-scale integration (VLSI) circuit, or more colloquially, the chip. The driving concept of the optical interconnect is that light waves rather than electronic wires should provide the primary means for the transport of information between silicon chips or, more speculatively, from one part of a silicon chip to another. Used properly, optical interconnects could be a vitally important way of bypassing and/or overcoming bottlenecks in data processing and transmission. Therefore, optical interconnects are one form of integrated circuit where photonics might play a role. The compact semiconductor laser could, for instance, simply be organized, in large or small arrays, around the edges of the silicon VLSI chip. Useful performance could mean that each compact laser was both dissipating and delivering sub-microwatt-scale power levels, together with signal bandwidths of 10 Gb s⁻¹ and beyond in each channel. It is obvious that existing communication lasers cannot be suitable, because of their size and power consumption. Here compact semiconductor lasers in the 1300–1600 nm band are the best candidates, since their development is more mature and, being in the transparency window of silicon, they are more complementary metal-oxide semiconductor (CMOS)-compatible than their counterparts at other, shorter, wavelengths.

With an increasing amount of information carried by optical waves, there has also been strong interest in the realization of what we may call “digital” PICs/OEICs. One can envisage the use of compact semiconductor lasers to realize photonic digital logical functionalities that are direct counterparts of several of the electronic logical functionalities. These are routinely built into the silicon VLSI chip, utilizing the nonlinear optical response of compact semiconductor lasers such as optical bistability for binary logic functions including Boolean logic, buffering, and storage. Such nonlinearity stems from the fundamental photon–charge-carrier interaction dynamics of the semiconductor and depends strongly on the structure of the laser cavity. The compact semiconductor laser based PIC/OEICs that might emerge will probably be quite different from the PICs and OEICs that have been most well developed so far for the transmission of information.

Alternative forms of functional PIC/OEIC, such as those being developed for sensing applications, may require different wavelengths and hence different semiconductor materials and structures, yet will still benefit from reduced laser size. For example, quantum cascade lasers (QCLs) that operate in the far infrared can benefit from compactness in applications where low power consumption is required including, for instance, vehicle embarked systems. Finally, it is possible to envisage a number of other situations where compact near-infrared semiconductor lasers could play a role. Examples could include single and paired photon sources that are relevant to various quantum processing applications and novel optomechanical device structures.

The fundamental mechanism that supports both reduced size and low power consumption is the high optical gain provided by direct bandgap semiconductor active materials and the quantum structures (quantum wells, quantum dots, etc.) based on these materials, making it possible for lasing to take place even with a tiny volume of active material. In this perspective, the effort toward smaller size and higher energy efficiency is largely an effort to realize ever-tighter simultaneous confinement of both the photon and the electron in this active volume. The reader may infer that all the structures covered in this book are aimed at strong confinement, which optically benefits from the high refractive index of the semiconductor materials used and at its most extreme is exemplified by the use of metal clad structures to overcome the optical diffraction limit. Truly subwavelength semiconductor lasers are indeed attainable with the incorporation of metal, provided that the resulting absorption losses are sufficiently low as applies, for instance, in the long wavelength region of the spectrum. Whatever may be the structure, much effort must also be dedicated to the electronic confinement, since the reduced active volume typically means a much-increased surface-area to active-volume ratio that leads to higher levels of nonradiative carrier recombination. Making semiconductor lasers more and more compact is, by itself, a worthwhile technological challenge at the level of targeted applications.

A broader commentary on semiconductor lasers and their applications at this point may help our reader to place the contents of this book in the wider context of a field that has been established for 50 years and yet is still evolving rapidly. The compact semiconductor lasers described in the different chapters of this book could enter progressively into the armory of contemporary optoelectronics and photonics technologies, where “conventional” semiconductor lasers are already commonplace items. The latter are routinely manufactured, en masse, with production volumes that are measured in hundreds of millions per year. In fact, almost all of the lasers in the world are semiconductor lasers. Furthermore, semiconductor laser diodes are the primary light source, that is, the primary converters between electrical power and optical power, for a large fraction of all other lasers, for example, where the characteristic gain medium of the laser is a doped fiber, a single-crystal slab, or a rod of glass. In terms of the conversion efficiency between electrical power (current) and optical power, no other laser comes close to the semiconductor laser.

It is salutary to contrast production volume numbers for individual lasers including, for example, the vertical cavity surface-emitting lasers (VCSELs) that are often organized in arrays with financial numbers. Approximately half of the financial volume for the production of “lasers,” around the world, is made up of semiconductor diode lasers. In contrast, the remaining half is almost totally made up of the “big” lasers that are used for applications such as numerically controlled machining and welding. In the case of such large lasers, the actual laser is, of course, built into a physically large machine that is a complete system of electronics, precision mechanisms, and optics.

New applications for semiconductor lasers continue to appear and they may, for instance, even displace the conventional light-emitting diode (LED) in parts of the display technology market. But there is also the possibility that some large-scale applications will disappear or greatly diminish in importance, at least in their most typical form for example, the use of blue diode lasers in compact disk (CD) equipment. The optical spectrum covered by semiconductor lasers is vast, although wavelength tunability remains an important issue. The spectral range of semiconductor lasers that are already being exploited commercially – or will probably be exploited commercially in the near-future – extends from the ultra-violet (UV), through the visible and a long way on into the infra-red (IR). With the inclusion of the quantum-cascade (QC) laser, the IR spectral coverage possible with semiconductor lasers ranges all the way from the shortest IR wavelengths, through the mid-IR to the far-IR, which last spectral region itself extends down in frequency to the terahertz (THz) region.

The list of applications where semiconductor lasers are used is, as already mentioned, extensive and we shall do no more than mention some of the more obvious and important applications. One important area, with financial values (revenue, turnover, and capitalization) measured in tens of billions of dollars, is fiber-optical telecommunications. In this domain, DFB and DBR lasers are the source of choice for large information bandwidth transmission over long distances and seem likely to remain so. On the other hand, the rapidly growing short-reach optical fiber communications market, for example, in access networks (“fiber-to-the-home” or “last mile”) and in the very large-scale data centers that support the Internet “cloud” services, demands large numbers of low-cost semiconductor lasers that are mainly conventional Fabry–Perot (FP) lasers and VCSELs. Newer types of integrated laser, such as those that can emit multiple wavelengths at a low cost, could prove to be vital in the future upgrade of such systems.

Line-of-sight, building-to-building communication using modulated semiconductor lasers, over kilometer scale distances, is of potential importance, although the atmospheric transmission is often impeded by adverse weather conditions. Useful communication in space over thousands of kilometers has already been demonstrated, which is more challenging because of other factors. Although the scale of the market and activity in the domain of free-space optical telecommunications has, so far, been much smaller than for fiber-optics based telecommunications, semiconductor lasers will surely form an important ingredient in this field.

The CD, as used in CD players for audio reproduction (e.g., for music) and in DVD (digital video/versatile disk) players, has been much the largest market in volume for the semiconductor laser, with annual demand amounting to several tens of millions of suitably packaged individual lasers. The semiconductor laser in a CD player is used to read the fine pattern of holes that has been created in the thin sheet of metal deposited on the surface of the plastic substrate. This quite specialized application of the semiconductor laser, despite its historical success, may yet diminish considerably in importance – or even disappear – because of trends in information transmission, storage, and delivery associated with the Internet. More generally, data storage by optical methods that include holographic techniques must compete with electronic and magnetic memory methods that have their own areas of strength, as well as weakness. Our view is that the intrinsic merits of the semiconductor laser are sufficiently strong that it will surely continue to play an important role in major markets, but it is not easy to predict what these will be in, say, 20 years time. For the CD player and DVD player, the optical wavelength used to read the pattern in the disk should in principle be as short as possible, since the resolution-determined hole-packing density (i.e., the information density) increases at least as strongly as the inverse of the wavelength squared. This requirement has led to the use of gallium nitride-based diode lasers that emit at wavelengths in the blue-violet region of the visible spectrum. Laser diodes with even shorter emission wavelengths, some way into the UV, are a credible future possibility.

At the other end of the spectrum from the UV and visible regions, penetrative imaging with electromagnetic waves at long enough wavelengths to be safe for moderate exposure of human beings, that is, THz imaging, since THz radiation is nonionizing, is clearly of emerging importance. But there is a need for sources of THz frequency electromagnetic waves that have much greater efficiency than has so far been possible, as well as other desirable properties such as rapid switching and modulation, compactness, and room temperature operation. This requirement indicates the desirability of a suitably scaled semiconductor-based source for the electromagnetic energy required, but so far this efficient semiconductor source for THz electromagnetic waves has largely been conspicuous by its absence.

For Lidar and optical radar applications in the mid- and near IR, semiconductor lasers, because of their characteristically high efficiency, are of potentially great importance in situations where the output light beam is required to propagate over large distances in “free-space,” in gaseous atmospheres (in particular, the earth's atmosphere), and fluid environments such as the sea. Probing the atmosphere, for example, for gaseous or particle pollutants, using optical radar techniques and sources of coherent short light pulses provides a challenging domain of potential applications for the semiconductor laser. Issues of beam-quality and directional control and the special optical systems that are likely to be required become important. For some of these applications, QCLs that offer wide spectral coverage through much of the IR spectrum have been researched quite intensively. QCLs are still heterostructure semiconductor lasers but have a dramatically different basic epitaxial layer structure, as well as distinctly different physical principles of operation because they are based on unipolar transitions, instead of electron–hole recombination. Polarization of the emission is naturally TM (transverse magnetic), that is, with the predominant magnetic field component of the emitted light parallel to the defining device plane while it is mostly TE (transverse electric) in conventional diode lasers. The range of possible wavelengths obtainable with the QCL extends from the mid-IR, through the far-IR and down to THz frequencies, although the coherent emitted radiation that justifies the epithet “laser” becomes progressively more difficult to generate, as the wavelength increases. The range of emission wavelengths possible with the QCLs overlaps, at near-IR wavelengths, with what is obtainable from conventional diode heterostructure lasers that have a substantial fraction of antimony in the composition of the light emitting III–V semiconductor region.

We return to the important, indeed basic, question: “What is to be gained from the pursuit of compactness in the semiconductor laser?” – a pursuit that is at the center of the research that is analyzed in this compact book. A partial answer to the question of “why compactness?” or “what is compactness for?” comes from considerations that also apply for the classic and central device of modern electronics, the transistor. Transistors are, in their standard format, three-terminal devices that are capable of both switching and amplification and, with feedback, also of oscillation. With appropriate organization, semiconductor lasers are capable of providing the same three functions, and it should also be born in mind that semiconductor lasers are devices that involve both electronic and photonic functionality. For example, when used as an amplifier in the “semiconductor optical amplifier (SOA),” the semiconductor laser takes an optical input and produces a higher power optical output that replicates the input but the amount of amplification of the light depends on electrical control of the gain of the optical amplifier.

In modern integrated electronics, the transistor is firstly a high-speed electronic switch. In a transistor, a controlling data stream goes into the transistor and is processed onto or into another data stream. The maximum speed (i.e., the rate) at which the switching operations involved can be carried out is restricted by the device size, from “large” down to very small sizes. As is well known, the silicon integrated circuit that is at the heart of modern electronics typically has several million interconnected transistors on it, as well as other components such as resistors and capacitors.

Semiconductor lasers can be organized to behave like their electronic counterpart just mentioned, the transistor. In a standard configuration, a heterostructure diode laser has its light output level modulated, in the simplest case, in amplitude by varying the level of the injection current that drives the laser into oscillation. Reducing the dimensions of the laser reduces the lasing threshold and the power consumption needed to modulate the optical output level, just as the reduction of the transistor dimensions implies a smaller energy dissipation per bit in an electronic circuit. For both devices, local heating, crosstalk effects, and low on-state/off-state contrast must be palliated to maintain high performance as the device size becomes progressively smaller, in order to produce a higher integration density. The use of semiconductor lasers to realize logical functionality and their possible integration into multi-stage logic circuits naturally push the semiconductor laser toward compactness. Considerations of speed of operation together with propagation delay, switching energy, and power consumption also provide the pressures that dictate compactness.

The chapters of this book, in order are

Chapter 1: Nano-scale metallo-dielectric coherent light sources.
Chapter 2: Optically pumped semiconductor photonic crystal lasers.
Chapter 3: Electrically pumped photonic crystal lasers: laser diodes and quantum cascade lasers.
Chapter 4: Photonic crystal VCSELs.
Chapter 5: III–V compact lasers integrated onto silicon (SOI).
Chapter 6: Semiconductor microring lasers.
Chapter 7: Nonlinearity in semiconductor microring lasers.

Although we do not claim that the organization of the present contribution to the literature on semiconductor lasers is totally systematic – and even less do we claim that the book is encyclopedic – we believe that what the book contains is organized appropriately, with a logical evolution of topics, and that it provides a good sample of the subtopics that constitute the whole field. Examples of subtopics that might have been included are the pillar geometry VCSEL-like laser and the promising vertical nanowire.

The book begins, in Chapter 1, with the topic of the metal enclosed nanoscale semiconductor laser. Detailed and rigorous electromagnetic analysis, together with creative design, leads to coherent light sources that can be substantially smaller than a free-space wavelength in all three space dimensions, while exploiting essentially the same III–V semiconductor based heterostructure gain medium as that in the conventional “macrolaser.” The combination of the intrinsically high gain available with III–V semiconductor structures and careful minimization of the potentially overwhelming propagation losses that occur for metals at optical frequencies produces the possibility of efficient coherent light sources that could viably be packed in million-scale numbers on a single, modest-area, wafer section.

Chapter 2 shares with the other chapters the basic aspect that the “natural” gain medium for coherent semiconductor light sources is invariably an epitaxial III–V semiconductor heterostructure. As in Chapter 1, the work described is primarily reliant on optical pumping processes for the gain medium. But the device structure is the radically different 2D photonic crystal (PhC) patterned membrane that provides strong confinement in all three space dimensions. The use of optical pumping has allowed a wide-ranging basic (i.e., fundamental) exploration of the characteristics and behavior of very compact lasers that exploit PhC principles for the generation of the feedback required for an optical oscillator.

Chapter 3 shares with Chapter 2 the fact that PhC structures are at the heart of the laser devices that have been investigated and that are described in detail in this chapter. The vital difference between the content of Chapter 3 and that of Chapter 2 is that the optical gain medium is pumped by means of electric charge-carrier injection. The research described in Chapter 3 is important because it directly addresses the technological issues that must be “solved” if the compact PhC-structured semiconductor laser is to be useful in a wide variety of situations – situations where the intrinsic efficiency of electrical pumping and the direct control that it provides are of paramount importance. Both PhC laser diodes and PhC QCLs are considered in this chapter, thereby covering a wide range of emission wavelengths from near infrared to THz waves. It is incidentally shown that the use of microwave- or electronics-inspired resonator structures provides the opportunity to design THz semiconductor lasers that can be much smaller than the emitted wavelength.

Chapter 4 continues with the PhC structuring theme, but with the clear alternative configuration of “vertical” emission that is mediated by the PhC structure. The title of the chapter immediately identifies the compact lasers involved as being a form of VCSEL but one in which the resonant cavity and laser performance stem from the use of a PhC structure on the lateral surface of the laser. The compact PhC VCSEL structure provides control of the single-mode power and polarization, while simultaneously optimizing the output power and coupling through the top surface. Using detailed modeling of the 3D structure, it is also shown that PhC-VCSELs with true photonic bandgap characteristics are feasible.

Chapter 5 departs radically from the previous chapters because it is centrally concerned with a hybrid situation in which the gain medium remains a III–V semiconductor epitaxial heterostructure, but the supporting substrate is a silicon-on-insulator (SOI) wafer section, to which the thinned-down III–V semiconductor laser structure is bonded and the III–V semiconductor substrate on which it has been grown is (almost) completely removed. This chapter suggests the strong plausibility of the transition from compact semiconductor lasers in the research lab to lasers in the real world of optical interconnect applications. The characteristic compact laser geometry of the work in Chapter 5 is the microdisk, which supports the so-called whispering gallery mode (WGM). The crucial – and so far not fully solved – challenge for this hybrid configuration is the need to drive the laser by electrical current injection.

Chapters 6 and 7 are concerned with ring-geometry semiconductor lasers and therefore share, in the simplest limiting case, the same circular geometry as that of the microdisk described in Chapter 5. Furthermore, there is considerable similarity between the mode structure of a disk and a ring with similar diameter and the ring can meaningfully be considered as a limiting case of a disk with the inner wall serving to limit the number of transverse (in this case meaning the radial direction) modes. By demonstrating intrinsically credible electrical pumping in such microring lasers and gaining an in-depth understanding of the photon–charge-carrier interaction dynamics in such structures, Chapters 6 and 7 also demonstrate – both from the experimental and the theoretical points of view – that compact semiconductor lasers could enter the real world of laser applications.

All three chapters – Chapters 5–7 – are substantially concerned with the bi-stable operating characteristics that make such lasers of interest as optical switches with a latching capability. If the electrical pumping and heat-dissipation issues associated with the hybrid configuration of disk lasers mounted on planar silicon waveguides can be addressed satisfactorily, it might eventually be that the compact hybrid semiconductor laser will become the preferred light source and optical switching device in high-density photonic integration based on silicon. Hybrid III–V/SOI integration might therefore win-out over III–V semiconductor monolithic integration.

As already mentioned, this book is by no means encyclopedic. We have not explicitly addressed the micropillar cavity geometry that has received considerable levels of attention for more than a decade. Only part of the research carried out on micropillar cavities was actually concerned with lasers as opposed, for example, to cavity quantum dynamics investigations. Issues such as the need to suppress nonradiative carrier recombination at dry-etch process-exposed surfaces that intersect with the quantum-well active region are arguably of particular importance in this case, as well as in the cases of the microring and microdisk lasers. The situation might well be different for vertical nanowires grown by the vapor–liquid–solid (VLS) method, because this approach generates structures with a much lower number of surface defects. Recent results [1] obtained using optical pumping on these laser structures, – including, for instance, spontaneous emission factors, β, close to unity over very wide bandwidths – have considerable promise for future developments of compact semiconductor lasers. Lasers based on vertical nanowires and micropillars could well be an appropriate additional topic in a future edition of this book.

Reference

1. Claudon, J., Bleuse, J., Malik, N.S., Bazin, M., Jaffrennou, P., Gregersen, N., Sauvan, C., Lalanne, P., and Gerard, J.-M. (2010) A highly efficient single-photon source based on a quantum dot in a photonic nanowire. Nat. Photonics, 4, 174–177.