Photonic

PHOTONIC INTEGRATION AND
PHOTONICS-ELECTRONICS
CONVERGENCE ON SILICON
PLATFORM
EDITED BY : Koji Yamada, Jifeng Liu, Toshihiko Baba, Laurent Vivien
and Dan-Xia Xu
PUBLISHED IN : Frontiers in Materials and Frontiers in Physics
Frontiers Copyright Statement
About Frontiers
© Copyright 2007-2015 Frontiers
Media SA. All rights reserved. Frontiers is more than just an open-access publisher of scholarly articles: it is a pioneering
All content included on this site,
such as text, graphics, logos, button
approach to the world of academia, radically improving the way scholarly research
icons, images, video/audio clips, is managed. The grand vision of Frontiers is a world where all people have an equal
downloads, data compilations and
software, is the property of or is opportunity to seek, share and generate knowledge. Frontiers provides immediate and
licensed to Frontiers Media SA permanent online open access to all its publications, but this alone is not enough to
(“Frontiers”) or its licensees and/or
subcontractors. The copyright in the realize our grand goals.
text of individual articles is the property
of their respective authors, subject to
a license granted to Frontiers. Frontiers Journal Series
The compilation of articles constituting The Frontiers Journal Series is a multi-tier and interdisciplinary set of open-access, online
this e-book, wherever published,
as well as the compilation of all other journals, promising a paradigm shift from the current review, selection and dissemination
content on this site, is the exclusive
property of Frontiers. For the
processes in academic publishing. All Frontiers journals are driven by researchers for
conditions for downloading and researchers; therefore, they constitute a service to the scholarly community. At the same
copying of e-books from Frontiers’
website, please see the Terms for time, the Frontiers Journal Series operates on a revolutionary invention, the tiered publishing
Website Use. If purchasing Frontiers system, initially addressing specific communities of scholars, and gradually climbing up to
e-books from other websites
or sources, the conditions of the broader public understanding, thus serving the interests of the lay society, too.
website concerned apply.
Images and graphics not forming part
of user-contributed materials may
Dedication to quality
not be downloaded or copied
Each Frontiers article is a landmark of the highest quality, thanks to genuinely collaborative
without permission.
Individual articles may be downloaded interactions between authors and review editors, who include some of the world’s best
and reproduced in accordance academicians. Research must be certified by peers before entering a stream of knowledge
with the principles of the CC-BY
licence subject to any copyright or that may eventually reach the public - and shape society; therefore, Frontiers only applies
other notices. They may not be
re-sold as an e-book.
the most rigorous and unbiased reviews.
As author or other contributor you Frontiers revolutionizes research publishing by freely delivering the most outstanding
grant a CC-BY licence to others to research, evaluated with no bias from both the academic and social point of view.
reproduce your articles, including any
graphics and third-party materials By applying the most advanced information technologies, Frontiers is catapulting scholarly
supplied by you, in accordance with
the Conditions for Website Use and
publishing into a new generation.
subject to any copyright notices which
you include in connection with your What are Frontiers Research Topics?
articles and materials.
All copyright, and all rights therein, Frontiers Research Topics are very popular trademarks of the Frontiers Journals Series:
are protected by national and
international copyright laws. they are collections of at least ten articles, all centered on a particular subject. With their
The above represents a summary unique mix of varied contributions from Original Research to Review Articles, Frontiers
only. For the full conditions see the
Conditions for Authors and the
Research Topics unify the most influential researchers, the latest key findings and historical
Conditions for Website Use. advances in a hot research area! Find out more on how to host your own Frontiers
ISSN 1664-8714
Research Topic or contribute to one as an author by contacting the Frontiers Editorial
ISBN 978-2-88919-693-7
DOI 10.3389/978-2-88919-693-7 Office: researchtopics@frontiersin.org
Frontiers in Materials and Frontiers in Physics 1 October 2015 | Photonic integration and photonics-electronics convergence
PHOTONIC INTEGRATION AND
PHOTONICS-ELECTRONICS
CONVERGENCE ON SILICON PLATFORM
Topic Editors:
Koji Yamada, National Institute of Advanced Industrial Science and Technology, Japan
Jifeng Liu, Thayer School of Engineering, USA
Toshihiko Baba, Yokohama National University, Japan
Laurent Vivien, Institute of Fundamental Electronics, France
Dan-Xia Xu, National Research Council, Canada
Silicon photonics technology, which has

the DNA of silicon electronics technology,
promises to provide a compact photonic
integration platform with high integration
density, mass-producibility, and excellent
cost performance. This technology has
been used to develop and to integrate
various photonic functions on silicon
substrate. Moreover, photonics-electronics
convergence based on silicon substrate is
now being pursued. Thanks to these features,
silicon photonics will have the potential
to be a superior technology used in the
Multifunctional integration on a silicon photonic
platform. Photograph by Dr. Patrick Lo Guo-Qiang.
construction of energy-efficient cost-effective
apparatuses for various applications, such as
Taken from: Luo X, Cao Y, Song J, Hu X, Cheng communications, information processing,
Y, Li C, Liu C, Liow T-Y, Yu M, Wang H, Wang QJ and sensing.
and Lo PG-Q (2015) High-throughput multiple
dies-to-wafer bonding technology and III/V-on-
Considering the material characteristics of
Si hybrid lasers for heterogeneous integration of
silicon and difficulties in microfabrication
optoelectronic integrated circuits. Front. Mater. 2:28.
doi: 10.3389/fmats.2015.00028 technology, however, silicon by itself is not
necessarily an ideal material. For example,
silicon is not suitable for light emitting devices because it is an indirect transition material. The
resolution and dynamic range of silicon-based interference devices, such as wavelength filters,
are significantly limited by fabrication errors in microfabrication processes.
For further performance improvement, therefore, various assisting materials, such as indium-
phosphide, silicon-nitride, germanium-tin, are now being imported into silicon photonics by
using various heterogeneous integration technologies, such as low-temperature film deposition
and wafer/die bonding. These assisting materials and heterogeneous integration technologies
would also expand the application field of silicon photonics technology. Fortunately, silicon
photonics technology has superior flexibility and robustness for heterogeneous integration.
Moreover, along with photonic functions, silicon photonics technology has an ability of
integration of electronic functions. In other words, we are on the verge of obtaining an ultimate
technology that can integrate all photonic and electronic functions on a single Si chip.
This e-Book aims at covering recent developments of the silicon photonic platform and novel
functionalities with heterogeneous material integrations on this platform.
Citation: Yamada, K., Liu, J., Baba, T., Vivien, L., Xu, D.-X., eds. (2015). Photonic integration and
photonics-electronics convergence on silicon platform. Lausanne: Frontiers Media. doi: 10.3389/978-2-
88919-693-7
Cover image: Large-scale photonic integration on silicon wafer.

Photograph by Koji Yamada
Table of Contents
05 Editorial: Photonic integration and photonics–electronics convergence

on silicon platform
Koji Yamada
07 Silicon photonic integration in telecommunications
Christopher R. Doerr
23 Small sensitivity to temperature variations of Si-photonic Mach–Zehnder
interferometer using Si and SiN waveguides
Tatsurou Hiraki, Hiroshi Fukuda, Koji Yamada and Tsuyoshi Yamamoto
28 Ultrahigh temperature-sensitive silicon MZI with titania cladding
Jong-Moo Lee
32 Silicon-nitride-based integrated optofluidic biochemical sensors using a
coupled-resonator optical waveguide
Jiawei Wang, Zhanshi Yao and Andrew W. Poon
45 High-throughput multiple dies-to-wafer bonding technology and
III/V-on-Si hybrid lasers for heterogeneous integration of optoelectronic
integrated circuits
Xianshu Luo, Yulian Cao, Junfeng Song, Xiaonan Hu, Yuanbing Cheng,
Chengming Li, Chongyang Liu, Tsung-Yang Liow, Mingbin Yu, Hong Wang,
Qi Jie Wang and Patrick Guo-Qiang Lo
66 Group IV light sources to enable the convergence of photonics and electronics
Shinichi Saito, Frederic Yannick Gardes, Abdelrahman Zaher Al-Attili, Kazuki Tani,
Katsuya Oda, Yuji Suwa, Tatemi Ido, Yasuhiko Ishikawa, Satoshi Kako,
Satoshi Iwamoto and Yasuhiko Arakawa
81 Group IV direct band gap photonics: methods, challenges, and opportunities
Richard Geiger, Thomas Zabel and Hans Sigg
99 Direct growth of Ge1-XSnx films on Si using a cold-wall ultra-high vacuum
chemical-vapor-deposition system
Aboozar Mosleh, Murtadha A. Alher, Larry C. Cousar, Wei Du, Seyed Amir Ghetmiri,
Thach Pham, Joshua M. Grant, Greg Sun, Richard A. Soref, Baohua Li,
Hameed A. Naseem and Shui-Qing Yu
106 Room-temperature near-infrared electroluminescence from boron-diffused
silicon pn-junction diodes
Si Li, Yuhan Gao, Ruixin Fan, Dongsheng Li and Deren Yang
Editorial
published: 14 October 2015
doi: 10.3389/fmats.2015.00065
Editorial: Photonic integration and

photonics–electronics convergence
on silicon platform
Koji Yamada*
National Institute of Advanced Industrial Science and Technology, Tsukuba, Japan
Keywords: silicon photonics, photonic integration, additional waveguide system, III–V semiconductors,
germanium-based emitter, wafer bonding, telecommunications applications, bio-chemical applications
Silicon-based photonics technology, which is based on the same paradigm of silicon (Si) electronics
technology, promises to provide us with a compact photonic integration platform with high integra-
tion density, mass manufacturing, and excellent cost performance. This technology has been used
to develop various photonic devices based on silicon, such as waveguides, filters, and modulators.
In addition, germanium (Ge) photodetectors have been built on a silicon-based photonic platform.
These photonic devices have already been monolithically integrated on silicon chips. Moreover, pho-
tonics–electronics convergence based on silicon photonics is now being pursued. These emerging
compact photonics–electronics convergent modules have the potential to be used in the fabrication
of energy-efficient cost-effective systems for various applications, such as communications, informa-
tion processing, and sensing.
The last decade first saw the development of Si-based photonic technologies for communication
applications, and commercial products are now available for short-range data communications.
For medium-/long-range telecommunication applications, in which stringent technical standards
are applied to guarantee long-distance data transmission, intensive R&D is now providing us with
technologies for high-performance Si-based photonic modules with complex device integrations
(Doerr, 2015). In such high-performance applications, various assisting technologies should be
implemented on the silicon photonic platform. For example, the resolution and dynamic range of
Edited and reviewed by: silicon-based interference devices, such as wavelength filters, are considerably limited by fabrication
Lorenzo Pavesi, errors in microfabrication processes. To overcome such limitations, additional waveguide systems,
University of Trento, Italy based on silicon nitride and silicon-rich silica, have been implemented (Yamada et al., 2014; Doerr,
*Correspondence: 2015).
Koji Yamada Additional waveguide systems can also provide novel functionalities for further performance
yamada.koji@aist.go.jp improvements. For example, the thermo-optic response of photonic devices can be controlled by
combining silicon nitride and silicon waveguides, which could guarantee temperature-insensitive
Specialty section: operation of data transmission systems (Hiraki et al., 2015). Thermo-optic responses can also be
This article was submitted to widely controlled by using titania as a cladding material in a Si waveguide (Lee, 2015). Moreover,
Optics and Photonics, an additional waveguide system can expand the application field of the Si photonic platform. For
a section of the
instance, silicon nitride waveguides, which are transparent to visible light, can be used to construct
journal Frontiers in Materials
compact bio-sensing systems on a small Si chip (Wang et al., 2015).
Received: 24 September 2015 Light-source integration, which is the most important open issue for the Si photonic platform,
Accepted: 29 September 2015
requires the help of other materials. For this purpose, III–V semiconductor materials have been
Published: 14 October 2015
bonded on silicon by using various hybrid integration techniques, such as direct die-to-wafer
Citation:
bonding (Fang et al., 2006; Luo et al., 2015). For monolithic integration, Ge-based light sources are
Yamada K (2015) Editorial: Photonic
integration and photonics–electronics
now being studied intensively. The most important technology for Ge-based light sources is band-
convergence on silicon platform. gap engineering, which aims to achieve a direct transition in Ge, which is originally an indirect-
Front. Mater. 2:65. transition material. The recent status of Ge-based light sources on Si is reviewed in this special issue
doi: 10.3389/fmats.2015.00065 (Saito et al., 2014; Geiger et al., 2015). Mechanical stress and heavily doped n-type carriers
Frontiers in Materials | www.frontiersin.org 5 October 2015 | Volume 2 | Article 65

Yamada Photonic integration on silicon platform
would significantly contribute to making Ge a direct-transi- Fortunately, it has superior flexibility and robustness for integrat-
tion material. GeSn alloy is also a very attractive material for ing these materials. Along with photonic functions, the Si-based
light sources on a Si platform because Sn, which can easily photonic platform can integrate electronic functions monolithi-
be dissolved in Ge, offers an important degree of freedom in cally. In other words, we are on the verge of obtaining an ultimate
band-gap engineering (Mosleh et al., 2015). Another approach technology that can integrate all photonic and electronic
now being investigated for monolithic integration of light functions on a single Si chip.
sources is Si-based electro-luminescence at room tempera-
ture, although its physical origin has not fully understood (Li ACKNOWLEDGMENTS
et al., 2015).
The Si-based photonic platform requires various assisting I would like to acknowledge all the authors, reviewers, edi-
materials for accomplishing practical photonic functions. tors, and publishers, who have supported this Research Topic.
REFERENCES vacuum chemical-vapor-deposition system. Front. Mater. 2:30. doi:10.3389/

fmats.2015.00030
Doerr, C. R. (2015). Silicon photonic integration in telecommunications. Front. Saito, S., Gardes, F. Y., Al-Attili, A. Z., Tani, K., Oda, K., Suwa, Y., et al. (2014).
Phys. 3:37. doi:10.3389/fphy.2015.00037 Group IV light sources to enable the convergence of photonics and electronics.
Fang, A. W., Park, H., Cohen, O., Jones, R., Paniccia, M. J., and Bowers, J. E. (2006). Front. Mater. 1:15. doi:10.3389/fmats.2014.00015
Electrically pumpled hybrid AlGaInAs-silicon evanescent laser. Opt. Express Wang, J., Yao, Z., and Poon, A. W. (2015). Silicon-nitride-based integrated optoflu-
14, 9203–9216. doi:10.1364/OE.14.009203 idic biochemical sensors using a coupled-resonator optical waveguide. Front.
Geiger, R., Zabel, T., and Sigg, H. (2015). Group IV direct band gap photonics: Mater. 2:34. doi:10.3389/fmats.2015.00034
methods, challenges, and opportunities. Front. Mater. 2:52. doi:10.3389/ Yamada, K., Tsuchizawa, T., Nishi, H., Kou, R., Hiraki, T., Takeda, K.,
fmats.2015.00052 et al. (2014). High-performance silicon photonics technology for tele-
Hiraki, T., Fukuda, H., Yamada, K., and Yamamoto, T. (2015). Small sensitivity to communications applications. Sci. Technol. Adv. Mater. 15, 024603.
temperature variations of Si-photonic Mach-Zhender interferometer using Si doi:10.1088/1468-6996/15/2/024603
and SiN waveguides. Front. Mater. 2:26. doi:10.3389/fmats.2015.00026
Lee, J.-M. (2015). Ultrahigh temperature-sensitive silicon MZI with titania clad-
Conflict of Interest Statement: The author declares that the research was con-
ding. Front. Mater. 2:36. doi:10.3389/fmats.2015.00036
ducted in the absence of any commercial or financial relationships that could be
Li, S., Gao, Y., Fan, R., Li, D., and Yang, D. (2015). Room-temperature near-infrared
construed as a potential conflict of interest.
electroluminescence from boron-diffused silicon pn-junction diodes. Front.
Mater. 2:8. doi:10.3389/fmats.2015.00008
Luo, X., Cao, Y., Song, J., Hu, X., Cheng, Y., Li, C., et al. (2015). High-throughput Copyright © 2015 Yamada. This is an open-access article distributed under the terms
multiple dies-to-wafer bonding technology and III/V-on-Si hybrid lasers for of the Creative Commons Attribution License (CC BY). The use, distribution or
heterogeneous integration of optoelectronic integrated circuits. Front. Mater. reproduction in other forums is permitted, provided the original author(s) or licensor
2:28. doi:10.3389/fmats.2015.00028 are credited and that the original publication in this journal is cited, in accordance
Mosleh, A., Alher, M. A., Cousar, L. C., Du, W., Ghetmiri, S. A., Pham, T., et al. with accepted academic practice. No use, distribution or reproduction is permitted
(2015). Direct growth of Ge1-xSnx films on Si using a cold-wall ultra-high which does not comply with these terms.
Frontiers in Materials | www.frontiersin.org 6 October 2015 | Volume 2 | Article 65

REVIEW
published: 05 August 2015
doi: 10.3389/fphy.2015.00037
Silicon photonic integration in

telecommunications
Christopher R. Doerr *
Acacia Communications, Hazlet, NJ, USA
Silicon photonics is the guiding of light in a planar arrangement of silicon-based materials

to perform various functions. We focus here on the use of silicon photonics to create
transmitters and receivers for fiber-optic telecommunications. As the need to squeeze
more transmission into a given bandwidth, a given footprint, at a given cost increases,
silicon photonics makes more and more economic sense.
Keywords: integrated optics, silicon photonics, optical fiber, optical communications, coherent, gratings,
waveguides
1. Introduction
Edited by:
Qiaoliang Bao, Until circa 2002, fiber-optic communication for metropolitan distances (80—600 km) and long-
Soochow University, China haul distances (600–15,000 km) employed mostly simple on-off keying (OOK) transmission. On-off
Reviewed by:
keying is simply turning on and off the light to transmit “1” s and “0” s. Higher performance, i.e., a
Lukas Novotny, lower bit-error rate (BER) for the same received optical power and/or for the same optical signal-to-
ETH Zurich, Switzerland noise ratio (OSNR), can be obtained by using phase-modulated formats, such as binary phase-shift
Satoshi Iwamoto, keying (BPSK) or quadrature phase-shift keying (QPSK). They maximize the distance between
The University of Tokyo, Japan constellation points for the same average signal power. In these “advanced” modulation formats
Xiangping Li, [1], the term “symbol” is used to represent each data portion in time, because each symbol can
Swinburne University
carry multiple bits of information. Early BSPK and QPSK were detected by differential detection,
of Technology, Australia
i.e., by interfering one symbol with the previous symbol in an interferometer in the receiver.
*Correspondence:
However, bandwidth needs have been constantly growing exponentially. It is expensive to install
Christopher R. Doerr,
Acacia Communications, 1301 Route
new optical fibers, ∼ $30 k per mile [2], so carriers and data-center operators needed to send
36, Hazlet, NJ 07730 USA more bits per second in the same fiber in the same optical bandwidth. One key way is to use both
chris.doerr@acacia-inc.com optical polarizations, because this doubles the available bandwidth. Although signal orthogonality
is maintained, their polarizations are essentially randomly changed during propagation through
Specialty section: fiber. To unscramble them requires significant signal processing. Optical coherent detection allows
This article was submitted to this to be done by digital electronics.
Optics and Photonics, Optical coherent detection was a hot topic in the 1980s, because it is a form of optical
a section of the journal amplification. However, the invention of the erbium-doped fiber amplifier (EDFA) eliminated
Frontiers in Physics
that advantage and coherent interest died away. Another advantage of coherent detection is the
Received: 11 February 2015 ability to receive the full optical field, both the real and imaginary parts of both polarizations. With
Paper pending published: improvements in complementary metal-oxide-semiconductor (CMOS) electronics, digital signal
23 March 2015
processing (DSP) became available circa 2002 to handle coherent detection even up to 100-Gb/s,
Accepted: 13 May 2015
Published: 05 August 2015
causing a revival of coherent detection. In the past, coherent detection was simply single quadrature
and single polarization. Now it is dual quadrature and dual polarization.
Citation:
100-Gb/s coherent systems have proven to be extremely compelling. They allow an upgrade of
Doerr CR (2015) Silicon photonic
integration in telecommunications. a 10-Gb/s channel to a 100-Gb/s channel with actually improved reach. Industry analyses show
Front. Phys. 3:37. the number of metro and long-haul 100-Gb/s coherent transceivers sold per year to be on a steep
doi: 10.3389/fphy.2015.00037 upwards ramp as 10-Gb/s OOK transceivers are replaced by 100-Gb/s coherent transceivers.
Frontiers in Physics | www.frontiersin.org 7 August 2015 | Volume 3 | Article 37

Doerr Silicon photonic integration in telecommunications
FIGURE 1 | 100-Gb/s coherent transceiver form-factor evolution. It went from a full line card, to a multi-source agreement (MSA) module, to a 100-Gb/s
form-factor pluggable (CFP) module. The D in D-CFP means that it is a digital module and the DSP is included inside, as opposed to a module that contains only the
optics.
However, the price of 100-Gb/s coherent transceivers is TABLE 1 | Node size and first year of commercial introduction for CMOS
expected to drop significantly. This is because users want to electronics.
pay the same price per connection even though the bit rate Node size Year
keeps increasing. However, not only must the price drop, but the
footprint and power consumption as well. As seen in Figure 1, 10 µm 1971
100-Gb/s modules have gone from full line cards to 5 × 7 6 µm 1974
in2 screwed-on modules to 3.2 × 5.7 in2 pluggable modules 3 µm 1977
today. Today’s 100-Gb/s pluggable form factor is called a CFP. 1.5 µm 1982
Tomorrow’s will be a CFP2, which is half the size, and eventually a 1 µm 1985
CFP4, which is a quarter the size. Power consumptions have gone 800 nm 1989
from more than 100 W on a line card, to 70 W for the screwed- 600 nm 1994
on module, to 28 W for the CFP. The next step, the CFP2, allows 350 nm 1995
only 12W. 250 nm 1997
There are two main components in a coherent transceiver— 180 nm 1999
the DSP chip and the optics. Today’s coherent CFP contains both. 130 nm 2001
There is a, possibly temporary, trend to take the DSP out of the 90 nm 2004
module and put it on the line card. Such modules are called 65 nm 2006
“analog” modules, rather than digital. With today’s technology, it 45 nm 2008
is not possible to have both the optics and DSP be under 12 W, the 32 nm 2010
maximum power in a CFP2. However, in 1–2 years, technology 22 nm 2012
will likely be ready for a “digital” CFP2. 14 nm 2014
To meet these requirements of lower price, lower power and 10 nm 2016
smaller footprint, one must make advancements in technology. 7 nm 2018
For the DSP, one can take advantage of the steady reduction in 5 nm 2020
transistor size in industry, which reduces power and footprint.
Node size and introduction year are shown in Table 1 [3]. Today’s
years. The main advantages of photonic integration are a small
coherent DSPs use 20–28 nm. Tomorrow’s will use 14 nm.
footprint, due to strongly confining waveguides and lens-free
For the optics, one must use photonic integration, the focus of
connections between parts; low power, due to an obviation of
this article. Most of today’s coherent transceivers are built using
50- RF lines; higher bandwidth RF connections; and low price,
separate LiNbO3 /planar lightwave circuit (PLC) modulators and
due to fewer touch points, no mechanical adjustments, less test
InP/PLC receivers, as shown in Figure 2. More and more, smaller
equipment, and less material. The main disadvantages of PICs
InP modulators and InP receivers are being used. In today’s
are typically a higher insertion loss and the inability to optimize
coherent CFP, there is a single silicon photonic (SiPh) integrated
components independently.
circuit (PIC) containing both the transmitter and receiver [4].
Not shown is a separate tunable laser.
Finally, a dominant cost for the DSP and optics is the 2. PIC Material Systems
packaging; one can further reduce cost, power, and footprint by
co-packaging the DSP and optics. Such transceivers are expected Figure 4 shows the most popular PIC material systems. From left
in 2–3 years. to right there are silica-on-silicon PICs, also called PLCs; silicon-
Figure 3 shows many of the elements that may be integrated on-insulator PICs, also called silicon photonics; lithium niobite
in a PIC. The blue are passive, the red are active (have an intended (LiNbO3 ); and III–V PICs, such as InP and GaAs. This article
dynamic interaction between light and matter), and the green focuses on silicon photonics. In silicon photonics, the light is
are electronic components. PICs have been around more than 20 mostly guided in silicon, which has an indirect bandgap of 1.12

eV (1.1 µm). The silicon is a pure crystal grown in a boule and

then sliced into wafers, today typically 300 mm in diameter, as
shown in Figure 5. The surfaces are oxidized to form SiO2 layers.
One wafer is bombarded with hydrogen atoms to a specified
depth. Then the two wafers are placed together in a vacuum,
and the oxide layers bond to each other. The assembly is cracked
at the hydrogen implantation line. Then the silicon layer where
the crack was is polished, and one is left with a thin layer of
crystalline silicon on a layer of oxide on a full silicon “handle”
wafer. The waveguides are formed from this thin crystalline layer.
While these silicon-on-insulator (SOI) wafers are what makes
low-loss silicon photonic waveguides possible, they are actually
used mostly for low-power CMOS circuits, because of the low
leakage currents they offer.
There is a wide family of possible silicon-based optical
waveguides, shown in Figure 6. They range from micro-scale
Ge-doped SiO2 waveguides to nano-scale Si wire waveguides. By
FIGURE 2 | 100-Gb/s coherent optics evolution, going from LiNbO3 adding Ge, one can make photodetectors and electro-absorption
modulators and planar lightwave circuit (PLC)-based receivers to modulators. Potentially even optical amplifiers. By doping the
InP-based modulators and receivers to silicon photonic modulators silicon one can make optical modulators. From left to right at the
and receivers.
bottom are silicon wire waveguides, silicon nitride waveguides,
FIGURE 3 | Some of the possible PIC integration elements.
FIGURE 4 | Popular PIC material systems. The second from the left is a silicon wire waveguide and is the focus of this article.

silicon oxynitride waveguides, thick silicon rib waveguides, thin Another key element is a spot-size converter, which converts
silicon nitride waveguides, and doped silica waveguides. From left the ∼ 0.5 × 1 µm2 mode of a Si wire waveguide to the ∼ 10 ×
to right at the top are a depletion modulator, a Ge photodetector, 10 µm2 mode of an optical fiber. A typical method is to use
and a Ge optical amplifier. an inverse taper, in which the waveguide is narrowed down to
a small tip, causing the optical mode to expand very large [8].
3. Si Photonic Passive Elements The mode can be captured by a suspended glass waveguide, such
as in Figure 8 [9]. Coupling losses less than 1.5 dB are readily
There are several key silicon photonic passive elements. One achievable with such spot-size converters.
is the surface-emitting grating coupler, as shown in Figure 7A Another key passive element is a polarization splitter. Some
[5, 6]. It consists of a strong grating in the waveguide with a pitch polarization splitter examples are shown in Figure 9. The first
approximately equal to the wavelength in the waveguide. This is a Mach-Zender interferometer with a different birefringence
causes light to emit or be received vertical to the surface, which is in each arm [10]. The second is a simple directional coupler
well-suited for wafer level measurements and/or coupling to an [11]. The shape birefringence is so high in typical silicon wire
optical fiber. The grating coupler is somewhat unique to silicon waveguides, that the transverse-magnetic (TM) polarization can
photonics because it requires a high vertical index contrast. For couple fully while the transverse-electric (TE) polarization has
example, if one tried to do a grating coupler in traditional InP barely begun to couple. The third is a grating coupler in which the
waveguides, the light would simply leak away into the substrate fiber is placed at an angle such that TE couples in one direction
rather than be emitted vertically, because the average index of the
grating waveguide would be below that of the substrate. To make
it work in InP, one must undercut the material under the grating,
suspending it, as shown in Figure 7B [7].
FIGURE 7 | Surface-emitting 1-D grating couplers in silicon (A) and

InP (B). In (A), the gray and light blue represent silicon and silicon dioxide,
FIGURE 5 | How a silicon-on-insulator (SOI) wafer is made. Each wafer respectively. In (B), the red and orange represent InGaAsP and InP,
is made from two silicon wafers. The wafers are oxidized, bonded, and one is respectively. (C,D) SEM pictures of the InP suspended cantilever grating
cut and polished to a thin layer. coupler.
FIGURE 6 | Cross sections of the family of Si-based optical waveguides. Also shown are typical propagation losses and refractive indices.

and TM the other [12]. The fourth is a 2D grating coupler number of free electrons and holes, either by doping, electrical
[13]. The fiber mode with its electric field perpendicular to the means, or optical means, as shown in Equations (1, 2), obtained
waveguide propagation direction will couple to that waveguide. by fitting to data in Soref and Bennett at 1550-nm wavelength
The fiber can be either tilted and couple to two waveguides [14]. The holes have a larger ratio of real to imaginary index
or be normal to the surface and couple to four waveguides. change, i.e., more phase change for a given loss change, and
The 2D grating coupler has the added advantage of acting as a thus are usually favored for making the phase modulators in
polarization rotator, in that all the light on the chip has the same Mach-Zehnder and ring modulators.
polarization yet was two orthogonal polarizations in the fiber.
1nr = −8.8 × 10−22 Ne − 8.5 × 10−18 Nh0.8 (1)
−22 −23
ni = 1.0 × 10 Ne + 7.4 × 10 Nh (2)
4. Si Photonic Active Elements
Various Si modulator types are shown in Figure 10A. In the
As mentioned above, a photonic active element has an intentional carrier injection modulator, the light is in intrinsic silicon
dynamic interaction between light and matter. A typical photonic inside a very wide p-i-n junction, and electrons and holes are
active element is an optical modulator. All the Si optical injected. Such a modulator is slow, however, typically 500-MHz
modulators today are based on the plasma free carrier effect. The bandwidth, because it takes a long time for the free electrons
complex refractive index of the silicon changes by changing the and holes to recombine after injection. Thus, such structures
are usually used as variable optical attenuators (VOAs) rather
than modulators [15, 16]. In the carrier depletion modulator,
the light is partly in a narrow p-n junction, and the depletion
width of the p-n junction is varied by an applied electric field.
Such a modulator can operate at over 50 Gb/s [17], but has
a high background insertion loss. A typical Vπ L is 2 V-cm.
The metal-oxide-semiconductor (MOS) (really semiconductor-
oxide-semiconductor) modulator contains a thin oxide layer in
the p-n junction [18]. It allows for some carrier accumulation
as well as carrier depletion, allowing for a smaller Vπ L of ∼0.2
V-cm, but with the drawbacks of higher optical loss and higher
capacitance per unit length. There are also SiGe electroabsorption
modulators [19] that rely on band-edge movement in SiGe. There
are also graphene modulators that rely on switching the graphene
between an absorbing metal and a transparent insulator [20].
FIGURE 8 | Spot-size converter for silicon wire waveguides. The silicon Various Si-based photodetectors are shown in Figure 10B.
is inverse tapered inside a suspended glass waveguide. The silicon substrate
has been etched away under the suspended glass waveguide.
The absorption material is Ge. Ge absorbs light with wavelengths
up to about 1.6 µm. Shown on the left is a p-i-n configuration
FIGURE 9 | Various polarization splitters.

FIGURE 10 | (A) Cross sections of various silicon-based optical modulator designs and (B) photodetector designs.
FIGURE 11 | Configurations for integrating optical gain into silicon photonics. Fabrication insertion point becoming later in the process as one moves from left
to right.
[21], the most successful commercially today. It consists of p- monolithic integration, including using epitaxially grown Ge as
doped silicon on which Ge is grown. Ge and Si have a 4% lattice an optical gain material [24], Er-doped glass waveguides, such
mismatch, so to minimize dislocations, a thin layer of SiGe is as Al2 O3 , (which require optical pumping) [25], and epitaxially
grown first. The top of the Ge is n doped. Shown in the middle is grown GaAs quantum dots [26]. The next column is wafer-
a metal-semiconductor-metal (MSM) photodiode [22] and at the to-wafer assembly, including oxide bonding [27] and organic
right avalanche photodiodes (APDs) [23]. The APD avalanche bonding [28] of III–V gain regions. The next column is die-to-
region is in Si, which has a lower noise than avalanche regions wafer assembly, including inserting III–V die into cavities in the
in III–V materials. Si wafer and then patterning the waveguides [29]. The advantages
There is still no clear-winning solution for integrating optical of all the left three columns is that the full device can be tested
gain with silicon photonics. Some of the various option are shown on the wafer level, before it is diced out. The far right column
in Figure 11, organized by assembly level. On the far left is is die-to-die assembly, including butt coupling of a Si die and a

III–V die and coupling with a lens and a grating coupler [30]. Silicon photonics is usually considered only for low-cost,
Commercial deployment is tending to move from the right to left short-reach, high-volume (>1M/year) products. This is because
of this figure. it is assumed that a large number of wafer starts is required
An element that is partway between an active and passive to pay for mask and development costs and that silicon
element is an optical isolator. Optical isolators are required to photonics has a significant performance penalty for metro and
stop back reflections from causing noise and oscillations in lasers long-haul products. However, the real situation is actually the
and optical amplifiers. An isolator requires a non-reciprocal opposite. This is because in low-cost, short-reach, high-volume
element [31]. In silicon photonics, two main types of isolators applications, there is tremendous competition from vertical
have been reported: magneto-optic and modulation-based. cavity surface-emitting lasers (VCSELs) and directly modulated
In magneto-optic isolators, garnets are placed on the side or lasers (DMLs), and silicon photonics’ weakness of not having
top of the waveguide [32, 33]. In a modulation-based isolator, an easy way to integrate lasers is a significant disadvantage.
the optical field is modulated with either a traveling wave or On the other hand, in metro and long-haul applications, it is
a time delay between multiple modulators [34]. Figure 12 better to keep the laser separate anyway as it is preferable to
shows a modulation-based isolator design based on a parallel integrate the silicon photonics and DSP together, which is a hot
arrangement of phase modulators in series [35]. Each modulator environment. Also, coherent detection can make up for many
is driven by a sine wave. In the forward direction, the second
modulator in each arm undoes the modulation of the first
modulator; but in the backward direction, the two modulators
add constructively. Thus, there is no effect at all on the signal in
the forward direction but in the backward direction it is strongly
phase modulated. If the phase modulation amplitude is just right,
then a continuous-wave signal passing backwards is completely
attenuated at its original frequency. This gives narrow-band
isolation. By having multiple such narrow-band isolators in
parallel, driven by the same frequency but appropriate different
RF drive phases in each arm, one can achieve broadband
isolation. A two-arm version was demonstrated in silicon
photonics, achieving ∼3 dB of isolation. The modulation
was done by carrier injection in the silicon waveguide. The
isolation can be improved by reducing the residual amplitude
modulation in the phase modulators, by increasing the speed of
the modulators, and/or by increasing the number of arms in the
interferometer.
FIGURE 12 | (A) configuration of an optical isolator that uses a tandem

5. PIC Material System Comparison arrangement of phase shifters. There are N arms in the interferometer. The
more the arms, the higher the broadband isolation. (B,C) 2-arm version built in
Table 2 shows a comparison between InP and Si. InP is a much silicon photonics.
more expensive material than Si because of the rarity of In. Si
circuits tend to have a higher yield than InP circuits because there TABLE 2 | Pros and cons of InP and Si for photonic integrated circuits.
is much less epitaxy involved in Si circuits. In Si circuits, usually
the only epitaxy is Ge, used in the photodetectors, whereas in InP Si
InP all of the waveguides, even the passive ones, must be grown
Expensive material Cheap material
by epitaxy. Epitaxy tends to have a higher defect density than
• In is scarce • 27% mass Earth’s crust is Si
crystal growth from a boule. InP waveguides have high index
Medium yield High yield
contrast only laterally, whereas Si waveguides have high index
contrast laterally and vertically. This allows much smaller bend • W.g. material from epitaxy • W.g. material from original boule
radii and other more compact structures in Si. InGaAsP has a Small footprint Extremely small footprint
direct bandgap, whereas Si and Ge do not. Thus, the InP material • High index contrast in 1D • High index contrast in 2D
system has a much more efficient laser. The native oxide of the Native laser No native laser
InP system is much less robust than the native oxide of Si, which Poor native oxide Excellent native oxide
is SiO2 . Silicon is a stronger material than InP, allowing for much Low dark current Medium dark current
larger wafers, 75 mm compared to 300 mm (going to 450 mm Small wafers (75 mm typ.) Large wafers (300 mm typ.)
soon). InP modulators usually depend on the quantum-confined • 75 mm typical • 300 mm typical
Stark effect, which is temperature sensitive because of the band • Brittle material • Strong material
edge movement with temperature. Silicon modulators have very Modulator temperature sensitive Modulator temperature insensitive
minimal temperature dependence. • Band edge moves with temperature • Carrier density not v. temp. dep.

FIGURE 13 | Simulation results from 3D sparse FDTD. (A) is the top figures show the top views of the quasi-TE and quasi-TM signals, and the
view of the structure being simulated, which is a directional coupler. (B) lower two figures show the corresponding cross-section views. (C) Shows a
Shows a screen shot from a simulation using a quasi-TE launch. The top two screen shot from a simulation using a quasi-TM launch.
FIGURE 14 | Silicon photonics 8-PSM transceiver. Courtesy of Luxtera.
FIGURE 15 | Silicon photonics 8-WDM receiver. The upper figure shows a photograph of the chip, the lower left figure shows the measured responsivities to the 8
detectors vs. wavelength, and the lower right figure shows the measured bit-error rate at 1.25 Gb/s for one of the channels using a polarization scrambler.

of silicon photonics’ imperfections, such as the dark current state-of-the-art CMOS, and thus the masks and runs are relatively
is much smaller than the local oscillator photocurrent. Also, inexpensive.
the argument that one needs a large number of wafer starts
to pay for mask and development costs is fallacious, because
silicon photonics is done in a very large node size compared to 6. PIC Design
PICs are usually laid out in using mathematical scripts. This
is because usually in PICs, path lengths matter, when in
interferometers or because of skew. The PIC is made by
patterning multiple layers, typically 10 to 30, on a wafer. These
layers consist of many polygon shapes, typically in a GDSII
format. Before sending the files to the photomask shop, there
is a strong desire to be able to simulate the PIC to verify the
design. There are multiple levels of simulation. The lowest level
is 3D electromagnetic (EM) simulation, in which simulation
is done at the sub-wavelength level. Interaction with atoms
in the materials is done on the macroscopic scale. Typical
FIGURE 16 | Measured PAM-2, -4, and -8 optical eye diagrams at 28
Gbaud using a silicon photonics modulator. methods are the 3D finite-difference time domain (3D FDTD)
[36] and eigenmode expansion (EME) methods [37]. These
methods are the most accurate but simulation times for an
entire PIC are prohibitive. The next level is 2.5D EM simulation,
such as the finite-difference beam propagation method (FD-
BPM). These methods are significantly faster, with a tradeoff
of accuracy. Also, BPMs can handle only paraxial propagation,
e.g., they cannot be used to simulate a resonator. The next
level is 2D EM simulation, such as 2D FDTD and 2D BPM.
Again, these are faster, but limited. These cannot simulate
e.g., a polarization rotator. The next level up is transmission
and/or scattering matrix simulation. Each main component is
FIGURE 17 | 80-Gb/s dual-polarization transmitter in InP. It consists of reduced to an element with inputs and outputs, and connecting
two electro-absorption InGaAsP modulators, a polarization splitter, and a
polarization combiner. The incoming laser has its polarization oriented at 45◦ .
waveguides are reduced to phase shift and attenuation elements.
These simulations are extremely fast. A transmission matrix is
FIGURE 18 | DP-DQPSK receiver in silicon photonics. Uses optical polarizations by a 2D grating coupler, a series of couplers and phase shifters
polarization tracking. The upper figure shows a schematic, and the lower demultiplex the two signals, which had been mixed during fiber transmission,
figure a device photograph. The incoming signal is separated into two and two Mach-Zehnder delay interferometers demodulate the signals.

FIGURE 19 | 7-core-fiber silicon photonics receiver. The upper right circuit shown in the bottom figure. The incoming fiber is tilted at the proper
figure shows a photograph of the fiber cross section, showing the seven angle to the 1D grating couplers such that TE polarization couples to the left
cores. The top shows the schematic for each channel in the silicon photonic and TM polarization couples to the right.
FIGURE 21 | First reported vector modulator. It was in GaAs.
FIGURE 20 | PIC for coupling to the multimodes of a ring-core fiber by

using a circular grating coupler connected to a star coupler. The upper
figure shows a schematic of the silicon photonic circuit photograph shown at without sufficient high-order mode rejection, or two waveguides
the bottom. The circuit contains a circular grating coupler connected to an that pass too close to each other and have undesired coupling are
array of waveguides of equal length.
unlikely to be caught.
A technique called sparse FDTD allows one to do 3D and
2D FDTD simulation directly on the entire PIC design to verify
multiplied by the incoming signals to find the outgoing signals. the design [38]. While it is unlikely any EM simulation tool
A scattering matrix (whose elements are called s-parameters) is can simulate a very large PIC, sparse FDTD can simulate quite
multiplied by the incoming and outgoing signals on one side of large portions. In conventional 3D FDTD, one starts with all six
the element to find the incoming and outgoing signals on the components of the EM fields in a specified quantized volume.
other side of the element. Basically, scattering matrices include Time is advanced a step, and the new field components are
reflections within the element. Scattering matrices are typically calculated in the volume, and so on. So many calculations every
twice as large in each dimension as transmission matrices. step takes a very long time. In sparse 3D FDTD, rather than do
However, relying on EM simulation of some elements and calculations for every point in the volume every step, a list of
scattering/transmission matrices to simulate the entire PIC does field components is maintained, theoretically in an arbitrarily
not guarantee that the design is error-free before tape out. For large volume, is maintained and calculations are done on these.
example, a miscalculated path length, a multimode waveguide At each time step points neighboring the field components are

added, and field components with power below a certain level are for a wide diversity of solutions. These various solutions do not
discarded. For certain structures this calculation can be orders of interoperate, but the users do not care so much, as long as prices
magnitude faster than conventional 3D FDTD. However, sparse are low.
FDTD performs poorly with dispersive structures, because then Today, most of the short reach links are based on vertical-
the optical field spreads out too much, making the list too long. cavity surface-emitting lasers (VCSELs) over multimode fiber,
Example screen shots from 3D FDTD simulation of a PBS like i.e., do not involve PICs at all. VCSELs are very inexpensive and
that shown in Figure 9B are shown in Figure 13 [39]. easy to couple to multimode fiber. It is nearly impossible for PICs
to compete against VCSELs on price. However, the bandwidth-
7. Short Reach PICs distance product for a VCSEL over multimode fiber is ∼ 2 GHz-
km. At 25 Gb/s, this limits distances to ∼100 m. Also, multimode
Short reach communications typically means less than 2 km, but fiber (MMF) costs more than standard single-mode fiber (SSMF),
can sometimes include up to 40 km. Short reach is usually for because many more km of SSMF have been produced than MMF.
intra-data center, connecting racks, or client-side optics. There is Thus, when new data centers are built, it can be advantageous
an emerging need for very short reach communications in which to outfit them with SSMF. Single-mode VCSELs are difficult to
boards are connected optically within a rack. Such optics are no make today, so this is a good opportunity for PICs. However,
longer considered “transceivers” and for the sake of focus are left VCSEL technology is constantly improving, providing a constant
out of this article. challenge to PICs in short-reach applications.
Because of the fast growth and turn-over in data centers there A successful PIC short-reach commercial solution today is
is usually insufficient time for standards to develop. This allows based on parallel single-mode fibers (PSM). Figure 14 shows an
FIGURE 22 | Early reported silicon photonic vector modulators.
FIGURE 23 | 2-bit optical DAC in silicon photonics.

8-fiber PSM solution (4 fibers out and 4 fibers in) based on The advantage over PSM is requiring only two fibers instead
silicon photonics from Luxtera [40]. The chip contains a 1.4- of eight, and the disadvantage is requiring four lasers instead
µm laser in a small hermetic assembly on top of the PIC. This of one. WDM makes more sense as the cost of transceivers
wavelength was chosen as optimum for the grating couplers that drops compared to the cost of fiber and installing it, especially
couple the laser light into the PIC. This laser is split four ways ribbon fibers. Figure 15 shows an 8-channel CWDM receiver in
to four 10-Gb/s on-off-keying (OOK) distributed-driven Mach- silicon photonics [41]. It uses a silicon nitride spot-size converter
Zehnder-interferometer modulators (MZMs). The CMOS drive and arrayed waveguide grating (AWG), which is polarization
electronics are monolithically integrated with the photonics. independent via variation of waveguide widths, silicon output
Distributed driven means that the modulator is broken into N multimode waveguides, and Ge photodetectors.
sections in series, each with a separate driver timed appropriately. Yet another solution is to use multi-level modulation, called
This saves power consumption over a traveling-wave modulator, pulse amplitude modulation (PAM). Figure 16 shows PAM4 and
because a traveling-wave modulator has a termination resistor PAM8 eye diagrams at 28 Gb/s generated by a silicon photonics
into which power must be dumped. MZM.
Another successful PIC short-reach solution is based on One can also use polarization-division multiplexing (PDM),
wavelength-division multiplexing (WDM). Typically four also called dual-polarization (DP) transmission. In this case,
wavelengths, each modulated with OOK at 25 Gb/s, are different signals are in each polarization. Figure 17 shows a
multiplexed in the transmitter and demultiplexed at the receiver. dual-polarization 80-Gb/s modulator in InP [42]. Such a design
FIGURE 24 | Early (A) InP and (B) silicon integrated coherent receivers.
FIGURE 25 | Single-chip silicon photonic coherent transceiver.

could be readily made in silicon photonics. In the fiber, the coherent transmission. This is because long fiber routes are
two signals will stay predominantly orthogonally polarized, expensive to install/obtain, and thus the user wants to push
but the polarization will vary unpredictably with time. At the as much information over each fiber as possible. Coherent
receiver, if one does not use coherent detection, one needs to receivers make it possible to receive WDM, PDM, and high-order
optically demultiplex the two polarizations. Figure 18 shows constellations with high-performance, because the complete
a device in silicon photonics that can optically demultiplex optical field is received and acted on by a DSP. In intradyne
polarization [43, 44]. It does this by receiving two orthogonal coherent communications, the transmitted signal comes from
polarizations from the fiber, these polarizations not necessarily a dual-polarization vector modulator, and the received signal
that of the signals, and then interferes the two with a is interfered with a continuous-wave (CW) laser signal whose
controllable phase and coupling ratio to demultiplex them. To frequency is close to the carrier of the signal (within ∼2–3 GHz),
do this in an endless fashion, i.e., without ever needing phase- but does not need to be exact.
shifter resets back to zero, one needs multiple interferometer The first reported vector modulator was a GaAs PIC,
stages. shown in Figure 21 [47]. It consists of two MZMs in a larger
In the far future one may find the data center interconnections interferometer. Figure 22 show some early vector modulators
so crowded that one must reduce the number of fiber strands in silicon photonics. The modulator in Figure 22A contains
and instead put multiple cores and/or modes in a single fiber. two vector modulators, one for each polarization, along with
Figure 19 shows a PIC for receiving from a 7-core fiber, using the polarization splitting optics [48]. The single-polarization
polarization diversity [45]. It includes optical filters for WDM. modulator in Figure 22B uses a thin-oxide layer in the p-n
Figure 20 shows a PIC for receiving from a multi-mode ring- junction to obtain a low Vπ L product and is driven directly
core fiber [46]. A multi-mode ring core fiber is advantageous by CMOS inverters [49]. By using multiple segments in the
because the modes can be accessed without waveguide crossings modulator, one can create an optical digital-to-analog converter
and conveniently demultiplexed by a star coupler. (DAC). The segment lengths are in a geometric sequence.
Figure 23 shows a demonstration that achieved 16-QAM
8. Metro and Long-reach PICs modulation at 13 Gbaud using a silicon photonic optical DAC
[50].
Unlike short-reach links, which we saw have many choices of The first reported coherent receivers were in InP, as
transmission type, metro and long-reach links demand intradyne shown in Figure 24A [51–55]. Figure 24B shows an early
FIGURE 26 | (A) Measurement setup, (B) measured 30-Gbaud eye diagram for DP-QPSK, (C) real-time-processed constellations at 120 Gb/s, and (D–F) bit-error
rate vs. OSNR for various cases for the silicon photonics single-chip coherent transceiver in an optical loop back configuration.

Figure 26D shown for comparison is the performance of discrete

optics. The performance of the silicon PIC is nearly the same
as the discrete optics. Figure 26E shows the performance at
various temperatures, showing that the silicon photonics can
indeed work without temperature control. Figure 26F shows the
performance up to 3000 km without significant penalty. This
shows that the chirp of the silicon photonics modulator is low.
Figure 27 shows this PIC in a 100-Gb/s CFP module. As one
can see, the module is tightly packed and would be very difficult
to make with discrete optics.
FIGURE 27 | 100-Gb/s coherent CFP module using silicon photonics.
This single-chip coherent transceiver contains all the optics
needed for a coherent transmitter except the tunable laser. As
mentioned earlier, it is probably better to keep the laser separate
anyways because this chip can be co-packaged with the DSP,
dual-polarization, dual-quadrature receiver in silicon photonics
which runs very hot.
[56]. It uses a 2-D grating coupler as a fiber coupler, polarization
splitter, and polarization rotator.
Figure 25 shows a recent silicon photonic PIC that contains 9. Conclusion
the full vector modulator and full coherent receiver on a single
chip [4]. This is lower cost and smaller footprint than separate The touted advantage of silicon photonics is the die are lower
transmitter and receiver chips. There are three fibers connected cost than any other solution. While this may be true, it is
to the module: laser input, which is split between transmitter and of limited help in short-reach applications, where the lack
receiver; transmitter output; and receiver input. The fibers are of an integrated laser puts silicon photonics at a significant
connected in a 3-fiber array, reducing cost and assembly time. It disadvantage compared to the incumbents, such as VCSELs and
is co-packaged in a hermetic gold box with four drivers and four DMLs. Instead, the less-touted advantages of silicon photonics:
transimpedance amplifiers. It does not require any temperature high yield, low modulator temperature sensitivity, high chip
control, allowing the total power consumption to be less than strength, and ability to do polarization handling; make it ideal
5 W, −5 to 80◦ C. A silicon photonics modulator does have for metro and long-haul applications. Spending its adolescence
some imperfections compared to Pockels-effect modulators, like in metro and long-haul, silicon photonics will have time to
GaAs and LiNbO3 . It has residual amplitude modulation, diode develop mature laser integration methods, more routine foundry
nonlinearity, capacitance change with voltage, and bandwidth services, and sophisticated packaging solutions so it can later
limitations. A simulation including these effects shows that the take on the short-reach incumbents. By that time, coherent
imperfection performance penalty is only 0.1 dB compared to an transmission may be cost- and power-effective enough to work
ideal modulator. in very short links, bringing its advantages of high sensitivity,
Each PIC was tested in a socket in an optical loop-back high spectral efficiency, high-order modulation, and wavelength
configuration using a 100-Gb/s DSP for real-time measurements, selection.
as shown in Figure 26A. Optical loop-back insures that any
potential crosstalk between the transmitter and receiver would Acknowledgments
show up as degradation. Figure 26B shows a measured 30-
Gbaud DP-QPSK eye diagram. There are five levels in such The author is indebted to Long Chen, Diedrik Vermeulen,
a signal. Figure 26C shows measured real-time-processed 120- Torben Nielsen, Scott Stulz, Saeid Azemati, Greg McBrien, Benny
Gb/s DP-QPSK constellations. Measured BER vs. OSNR curves Mikkelsen, Christian Rasmussen, Mehrdad Givehchi, Seo Yeon
at multiple wavelengths across the C-band are shown in Park, Jonas Geyer, Xiao-Ming Xu, and many others.
References 6. Mekis A, Gloeckner S, Masini G, Narasimha A, Pinguet T. A grating-coupler-

enabled CMOS photonics plaform. J Sel Topics Quant Electron (2011) 17:597–
1. Winzer P. Advanced optical modulation formats. Proc IEEE (2006) 94:952–85. 608.
doi: 10.1109/jproc.2006.873438 7. Chen L, Zhang L, Doerr CR, Dupuis N, Weimann NG, Kopf RF. Efficient
2. Wellbrock G. Where do we go from here? In: OIDA Workshop on State of the membrane grating couplers on InP. IEEE Photon Tech Lett. (2010) 22:890–2.
Art Integrated Photonics. Washington, DC (2014). doi: 10.1109/LPT.2010.2047391
3. Wikipedia. Semiconductor Device Fabrication (2008). 8. Shoji T, Tsuchizawa T, Watanabe T, Yamada K, Morita H. Low
4. Doerr CR, Chen L, Vermeulen D, Nielsen T, Azemati S, Stulz S, et al. loss mode size converter from 0.3 um square Si wire waveguides
Single-chip silicon photonics 100-Gb/s coherent transceiver. In: Optical Fiber to singlemode fibres. Electron Lett. (2002) 38:1669. doi: 10.1049/el:
Communication Conference: Postdeadline Papers. San Jose, CA (2014). 20021185
5. Taillaert D, Van Laere F, Ayre M, Bogaerts W, Van Thourhout D, Bienstman P, 9. Chen L, Doerr CR, Chen YK, Liow TY. Low-loss and broadband
et al. Grating couplers for coupling between optical fibers and nanophotonic cantilever couplers between standard cleaved fibers and high-index-contrast
waveguides. Jpn J Appl Phys. (2006) 45:6071–7. doi: 10.1143/JJAP. SiN or Si waveguides. IEEE Photon Tech Lett. (2010) 22:1744–6. doi:
45.6071 10.1109/LPT.2010.2085040

10. Yamazaki H, Yamada T, Goh T, Kaneko A, Sano A. Integrated 100-Gb/s 31. Jalas D, Petrov A, Eich M, Freude W, Fan S, Yu Z, et al. What is and what is
PDM-QPSK modulator using a hybrid assembly technique with silica- not an optical isolator. Nat Photonics (2013) 7:579–82.
based PLCs and LiNbO3 phase modulators. In: European Conference and 32. Levy M, Osgood RMJr, Hegde H, Cadieu FJ, Wolfe R, Fratello VJ. Integrated
Exposition on Optical Communications Brussels: IEEE (2008). 1–4. doi: optical isolators with sputter-deposited thin-film magnets. IEEE Photonics
10.1109/ECOC.2008.4729119 Technol Lett. (1996) 8:903–5. doi: 10.1109/68.502265
11. Fukuda H, Yamada K, Tsuchizawa T, Watanabe T, Shinojima H, Itabashi 33. Shoji Y, Mizumoto T, Yokoi H, Hseih IW, Osgood R. Magneto-optical isolator
SI. Silicon photonic circuit with polarization diversity. Opt Express (2008) with silicon waveguides fabricated by direct bonding. Appl Phys Lett. (2008)
16:4872–80. doi: 10.1364/OE.16.004872 92:071117. doi: 10.1063/1.2884855
12. Tang Y, Dai D, He S. Proposal for a grating waveguide serving as 34. Bhandare S, Ibrahim SK, Sandel D, Zhang H, Wust F, Noe R. Novel
both a polarization splitter and an efficient coupler for silicon-on-insulator nonmagnetic 30-dB traveling-wave single-sideband optical isolator integrated
nanophotonic circuits. IEEE Photonics Technol Lett. (2009) 21:242–4. doi: in III/V material. IEEE J Sel Top Quantum Electron. (2005) 11:417–21. doi:
10.1109/LPT.2008.2010528 10.1109/JSTQE.2005.845620
13. Taillert D, Chong H, Borel PI, Frandsen LH, Rue RMDL, Baets R. A 35. Doerr CR, Chen L, Vermeulen D. Silicon photonics broadband modulation-
compact two-dimensional grating coupler used as a polarization splitter. based isolator. Opt Express (2014) 22:4493. doi: 10.1364/OE.22.004493
IEEE Photonics Technol Lett. (2003) 15:1249–51. doi: 10.1109/LPT.2003. 36. Yee KS. Numerical solution of initial boundary value problems involving
816671 Maxwell’s equations in isotropic media. Antennas Propagation IEEE
14. Soref R, Bennett B. Electrooptical effects in silicon. IEEE J Quantum Electron. Trans Livermore, CA: IEEE (1966). 302–7. doi: 10.1109/TAP.1966.
(1987) 23:123–9. doi: 10.1109/JQE.1987.1073206 1138693
15. Zheng D, Smith B. Improved efficiency Si-photonic attenuator. Opt Express 37. Bienstman P, Baets R. Optical modelling of photonic crystals and VCSELs
(2008) 16:16754–65. doi: 10.1364/OE.16.016754 using eigenmode expansion and perfectly matched layers. Opt Quantum
16. Dong P, Chen L. Monolithically integrated VOA-MUX with power monitors Electron. (2001) 33:327–41. doi: 10.1023/A:1010882531238
based on submicron silicon photonics platform. Opt Fiber Commun Los 38. Doerr CR. Sparse finite difference time domain method. IEEE Photonics
Angeles, CA: IEEE (2012) 9–11. doi: 10.1364/ofc.2012.otu2i.5 Technol Lett. (2013) 25:2259–62. doi: 10.1109/LPT.2013.2285181
17. Thomson DJ, Gardes FY, Fedeli JM, Zlatanovic S, Hu Y, Ping B, et al. 50-Gb/s 39. Doerr CR. 3D sparse finite-difference time-domain simulation of silicon
silicon optical modulator. IEEE Photonics Technol Lett. (2012) 24:234–6. doi: photonic integrated circuits. In: Optical Fiber Communication Conference. Los
10.1109/LPT.2011.2177081 Angeles, CA (2015).
18. Liu A, Jones R, Liao L, Samara-Rubio D, Rubin D. A high-speed silicon optical 40. Narasimha A, Analui B, Liang Y, Member S, Sleboda TJ, Abdalla S, et al. A
modulator based on a metal âĂŞ oxide âĂŞ semiconductor capacitor. Nature fully integrated 4 x 10-Gb / s DWDM optoelectronic transceiver implemented
(2004) 427:615–8. doi: 10.1038/nature02310 in a standard 1.3 um CMOS SOI technology. IEEE J Solid State Circ. (2007)
19. Liu J, Beals M, Pomerene A, Bernardis S, Sun R, Cheng J, et al. Waveguide- 42:2736–44. doi: 10.1109/JSSC.2007.908713
integrated, ultralow-energy GeSi electro-absorption modulators. Nat 41. Doerr CR, Chen L, Buhl LL, Chen YK. Eight-channel SiO2 /Si3 N4 /Si/Ge
Photonics (2008) 2:433–7. doi: 10.1038/nphoton.2008.99 CWDM receiver. IEEE Photonics Technol Lett. (2011) 23:1201–03. doi:
20. Liu M, Yin X, Ulin-Avila E, Geng B, Zentgraf T, Ju L, et al. A 10.1109/LPT.2011.2158091
graphene-based broadband optical modulator. Nature (2011) 474:64–7. doi: 42. Doerr CR, Zhang L. Monolithic 80-Gb / s dual-polarization On-off-keying
10.1038/nature10067 modulator in InP. J Lightwave Technol. (2008) 2:19–21.
21. Ahn D, Hong Cy, Liu J, Giziewicz W, Beals M, Kimerling LC, et al. High 43. Doerr CR, Chen L. Monolithic PDM-DQPSK receiver in silicon. In: Optical
performance, waveguide integrated Ge photodetectors. Opt Express (2007) Communication (ECOC), 2010 36th European Conference and Exhibition on
15:3916–21. doi: 10.1364/OE.15.003916 (IEEE) (2010). p. 1–3. Available online at: http://ieeexplore.ieee.org/xpls/abs_
22. Assefa S, Xia F, Bedell S, Zhang Y. CMOS-integrated 40GHz germanium all.jsp?arnumber=5621418
waveguide photodetector for on-chip optical interconnects. In: Optical Fiber 44. Doerr CR, Fontaine NK, Buhl LL. PDM DQPSK silicon receiver with
Communication Conference (2009). p. 10–2. Available online at: http://www. integrated monitor and minimum number of controls. IEEE Photonics
opticsinfobase.org/abstract.cfm?id=177916 Technol Lett. (2012) 24:697–9. doi: 10.1109/LPT.2012.2187048
23. Kang Y, Liu H, Morse M, Paniccia M. Monolithic germanium / silicon 45. Doerr CR, Taunay TF. Silicon photonics core-, wavelength-, and polarization-
avalanche photodiodes with 340 GHz gain-bandwidth product. Nat Photonics diversity receiver. IEEE Photonics Technol Lett. (2011) 23:597–9. doi:
(2008) 3:59–63. doi: 10.1038/nphoton.2008.247 10.1109/LPT.2011.2118748
24. Michel J, Camacho-aguilera RE, Cai Y, Patel N, Bessette JT, Dutt BR, et al. 46. Doerr CR, Fontaine N, Hirano M, Sasaki T, Buhl L, Winzer P. Silicon photonic
An electrically pumped Ge-on-Si laser. In: Optical Fiber Communication integrated circuit for coupling to a ring-core multimode fiber for space-
Conference. Los Angeles, CA: IEEE (2012). p. 5–7. division multiplexing. In: European Conference and Exposition on Optical
25. Purnawirman, Sun J, Adam TN, Leake G, Coolbaugh D, Bradley JDB, et al. Communications (2011). Available online at: http://www.opticsinfobase.org/
C-and L-band erbium-doped waveguide lasers with wafer-scale silicon nitride abstract.cfm?URI=ECOC-2011-Th.13.A.3
cavities. Opt Lett. (2013) 38:2–5. doi: 10.1364/OL.38.001760 47. Griffin R, Johnstone R, Walker R, Wadsworth S, Carter A, Wale M. Integrated
26. Liu AY, Zhang C, Norman J, Snyder A, Lubyshev D, Fastenau JM, et al. High DQPSK transmitter for dispersion-tolerant and dispersion-managed DWDM
performance continuous wave 1.3 µ m quantum dot lasers on silicon. Appl transmission. OFC 2003 Optical Fiber Communications Conference (2003). p.
Phys Lett. (2014) 104:3–7. doi: 10.1063/1.4863223 7–8. doi: 10.1109/OFC.2003.1248559
27. Fang A, Park H, Cohen O, Jones R, Paniccia M. Electrically pumped hybrid 48. Dong P, Xie C, Chen L, Buhl LL, Chen YK. 112-Gb/s monolithic
AlGaInAs-silicon evanescent laser. Opt Express (2006) 14:9203–10. doi: PDM-QPSK modulator in silicon. Opt Express (2012) 20:B624–9. doi:
10.1364/OE.14.009203 10.1364/OE.20.00B624
28. Roelkens G, Brouckaert J, Van Thourhout D, Baets R, NoÌĹtzel R, Smit 49. Milivojevic B, Raabe C, Shastri A, Webster M, Metz P, Sunder S, et al. 112Gb / s
M. Adhesive bonding of InPâĹŢInGaAsP dies to processed silicon-on- DP-QPSK transmission over 2427km SSMF using small-size silicon photonic
insulator wafers using DVS-bis-benzocyclobutene. J. Electrochem Soc. (2006) IQ modulator and low-power CMOS driver. In: Optical Fiber Communications
153:G1015. doi: 10.1149/1.2352045 Conference. San Jose, CA (2013). p. 5–7.
29. Marchena E, Creazzo T, Krasulick SB, Yu PKL, Van Orden D, Spann JY, et al. 50. Shastri A, Webster M, Jeans G, Metz P, Sunder S, Chattin B, et al. Experimental
Integrated tunable CMOS laser. In: Optical Fiber Communication Conference demonstration of ultra-low-power single polarization 56 Gb / s QAM-16
(2013). Available online at: http://www.opticsinfobase.org/abstract.cfm?URI= generation without DAC using CMOS photonics. In: European Conference
oe-21-23-28048 and Exposition on Optical Communications. Cannes (2014). p. 16–8.
30. Snyder B, Corbett B, O’Brien P. Hybrid integration of the wavelength-tunable 51. Koch T, Koren U, Gnall R, Choa F, Hernandez-Gil F, Burrus C, et al.
laser with a silicon photonic integrated circuit. J Lightwave Technol. (2013) GaInAs/GaInAsP multiple-quantum-well integrated heterodyne receiver.
31:3934–42. doi: 10.1109/JLT.2013.2276740 Electron Lett. (1989) 25:1621. doi: 10.1049/el:19891086

52. Takeuchi H, Kasaya K, Kondo Y, Yasaka H, Oe K, Imamura Y. Monolithic 56. Doerr CR, Winzer PJ, Chen YK, Chandrasekhar S, Rasras MS, Chen L, et al.
integrated coherent receiver on InP substrate. IEEE Photonics Technol Lett. Monolithic polarization and phase diversity coherent receiver in silicon. J
(1989) 1:398. doi: 10.1109/68.43392 Lightwave Technol. (2010) 28:520–5. doi: 10.1109/JLT.2009.2028656
53. Deri RJ, Pennings E, Scherer A, Gozdz A, Caneau C, Andreadakis N,
et al. Ultracompact monolithic integration of balanced, polarization diversity Conflict of Interest Statement: The author declares that the research was
photodetectors for coherent lightwave receivers. IEEE Photonics Technol Lett. conducted in the absence of any commercial or financial relationships that could
(1992) 4:1238. doi: 10.1109/68.166954 be construed as a potential conflict of interest.
54. Bach H, Matiss A, Leonhardt CC, Kunkel R, Schmidt D, Schell M, et al.
Monolithic 90◦ hybrid with balanced PIN photodiodes for 100 Gbit / s PM- Copyright © 2015 Doerr. This is an open-access article distributed under the terms
QPSK receiver applications. Optical Fiber Communication Conference (2009). of the Creative Commons Attribution License (CC BY). The use, distribution or
p. 4–6. reproduction in other forums is permitted, provided the original author(s) or licensor
55. Doerr CR, Zhang L, Winzer PJ. Monolithic InP multiwavelength coherent are credited and that the original publication in this journal is cited, in accordance
receiver using a chirped arrayed waveguide grating. J Lightwave Technol. with accepted academic practice. No use, distribution or reproduction is permitted
(2011) 29:536–41. doi: 10.1109/JLT.2010.2097240 which does not comply with these terms.

ORIGINAL RESEARCH ARTICLE
MATERIALS
published: 30 March 2015
doi: 10.3389/fmats.2015.00026
Small sensitivity to temperature variations of Si-photonic

Mach–Zehnder interferometer using Si and SiN
waveguides
Tatsurou Hiraki 1,2 *, Hiroshi Fukuda 3 , Koji Yamada 1,2 and Tsuyoshi Yamamoto 1
1
NTT Device Technology Laboratories, NTT Corporation, Kanagawa, Japan
2
NTT Nanophotonics Center, NTT Corporation, Kanagawa, Japan
3
NTT Device Innovation Center, NTT Corporation, Kanagawa, Japan
Edited by: We demonstrated a small sensitivity to temperature variations of delay-line Mach–Zehnder

Toshihiko Baba, Yokohama National
interferometer (DL MZI) on a Si photonics platform. The key technique is to balance a
University, Japan
thermo-optic effect in the two arms by using waveguide made of different materials. With
Reviewed by:
Junichi Fujikata, Photonics Electronics silicon and silicon nitride waveguides, the fabricated DL MZI with a free-spectrum range of
Technology Research Association, ~40 GHz showed a wavelength shift of -2.8 pm/K with temperature variations, which is 24
Japan times smaller than that of the conventional Si-waveguide DL MZI. We also demonstrated
Yosuke Terada, Yokohama National
the decoding of the 40-Gbit/s differential phase-shift keying signals to on-off keying signals
University, Japan
with various temperatures. The tolerable temperature variation for the acceptable power
*Correspondence:
Tatsurou Hiraki , Device Technology penalty was significantly improved due to the small wavelength shifts.
Laboratories, NTT corporation, 3-1, Keywords: silicon photonics, thermo-optic effect, Mach-Zehnder interferometer, waveguide, silicon nitride
Morinosato Wakamiya, Atsugi-shi,
Kanagawa 243-0198, Japan
e-mail: hiraki.tatsurou@lab.ntt.co.jp
INTRODUCTION materials, in this work, we minimized the temperature sensitivity

Silicon (Si) photonics is one of the most promising technologies of the DL MZI. In the following sections, we show the details of
for overcoming the limitations on integration in commercially the design and fabrication of the DL MZI and present experimen-
available silica-based planar-lightwave circuits. This is because tal results. In addition, as a feasibility demonstration, we show the
it provides ultra-compact waveguides and makes the monolithic thermal stability of the decoding of differential phase-shift eying
integration of active and passive devices possible (Lockwood and (DPSK) signals to on-off keying (OOK) signals at 40 Gbit/s.
Pavesi, 2010; Vivien and Pavesi, 2013). Many compact devices,
such as arrayed-waveguide gratings, Mach–Zehnder interferom- DESIGN AND FABRICATION
eters (MZIs), and ring resonators, have been reported using Si Figure 1 shows a schematic of the DL MZI. The temperature sen-
(Fukazawa et al., 2004; Xia et al., 2007) and silicon nitride (SiN) sitivity could be minimized by balancing the TO effect between the
waveguides (Gondarenko et al., 2009; Chen et al., 2011). One of two arms, while keeping the differential delay between them. The
the issues with these devices is performance degradation with tem- interference condition is expressed as following equation (Guha
perature variations due to the thermo-optic (TO) coefficient’s of et al., 2010)
Si (~1.86 × 10-4 /K) and SiN (4 ~ 5 × 10-5 /K) being higher than
that of the silica (~1.0 × 10-5 /K). To overcome this issue, athermal mλ = neff ,2 L2 − neff ,1 L1
designs of Si-waveguide delay line (DL) MZIs have used different
effective-index changes with temperature (dneff /dT) in the two Here, m is an integer for constructive interference or a half-
arms to balance the TO effects in them (Uenuma and Motooka, integer destructive interference, n eff, 1 and n eff, 2 are the effective
2009; Guha et al., 2010; Hai and Liboiron-Ladouceur, 2011). In indices, and L 1 and L 2 are the physical lengths of arm 1 and 2.
the previous studies, dneff /dT was controlled by means of the dif- Then, the temperature sensitivity of the interference spectrum
ferent optical confinement in the Si cores of narrow and wide could be obtained by differentiating above equation with respect
Si waveguides. However, the dneff /dT of the narrow waveguides to temperature, as expressed by following
significantly depended on the core width; therefore, inevitable
fabrication errors made it difficult to minimize the TO effect. dλ dneff ,2 dneff ,1 dneff ,2 dneff ,1
= L2 − L1 / m − L2 − L1
To prevent the problem, the dneff /dT should be controlled by dT dT dT dλ dλ
changing the TO coefficients of the materials, without using a nar-
row waveguide. In our previous work, we reported control of the Athermal condition is given by the numerator of this equation
refractive indices and TO coefficients of complementary metal- to be 0. Since we have two design parameters L 1 and L 2 , we can
oxide semiconductor (CMOS) compatible materials by changing make dλ/dT to be 0 while keeping the differential delay. The key
the atomic composition of SiOx, SiOxNy, and SiN (Tsuchizawa technique is to control dneff /dT by changing the core materials of
et al., 2011; Nishi et al., 2012; Hiraki et al., 2013). Using these the two arms. In this work, we used Si and SiN waveguides in the
www.frontiersin.org March 2015 | Volume 2 | Article 26 | 23

Hiraki et al. Temperature-insensitive Si-SiN-waveguide MZI
FIGURE 3 | Microscope image of fabricated Si–SiN-waveguide DL MZI.

FIGURE 1 | Schematic of Si–SiN-waveguide DL MZI.
The taper-tip width and the taper length of the fiber-chip inter-
face were 200 nm and 300 µm, respectively. Since the Si and
SiN waveguides were formed in different layers, the interlayer
coupler (ILC) between them was designed using adiabatically
tapers (Huang et al., 2014). We introduced the ILCs into both
arms to cancel out their phase delays. In addition, as refer-
ence samples, we designed a conventional Si-waveguide DL MZI
and a SiN-waveguide DL MZI without any compensation for
thermal sensitivity (dneff, 1 /dT = dneff,2 /dT). In the conventional
DL MZIs, both arms comprised of the same structures, which
were the 440-nm-wide Si waveguide and the 800-nm-wide SiN
waveguide.
The DL MZI was fabricated on an 8-inch silicon-on-insulator
wafer, whose buried-oxide thickness was 3 µm. The Si waveguides
were first patterned; then, a clad film was deposited. After that, the
clad film was flattened, and SiN-waveguide cores were formed. The
interlayer clad thickness between the Si and SiN waveguides was
controlled to be 100 nm. Finally, an overclad film was deposited.
A microscope image of the fabricated Si–SiN-waveguide DL MZI
is shown in Figure 3. The total size of the fabricated DL MZI
FIGURE 2 | Relationships between dneff /dT and core width of Si and is ~0.56 mm2 /ch, which is comparable to that of the conven-
SiN waveguides. tional SiN-waveguide DL MZI. It is still larger than that of the
Si-waveguide DL MZI; however, it is several-hundred times smaller
than one made of the commercially-used silica.
CMOS compatible materials. In the design, the refractive index
and the TO coefficient of the SiN core are 2.0 and 4.0 × 10-5 /K, RESULTS AND DISCUSSIONS
respectively. The core thicknesses of the waveguides were fixed at We measured transmission spectra of the fabricated DL MZI. We
220 and 400 nm, respectively. Figure 2 shows the calculated results used a tunable laser diode (TLD) as a light source and swept the
of relationships between dneff /dT and the core widths of the Si wavelength of the input light, and measured output light power
and SiN waveguides. For little change of the dneff /dT with width from the bar port. The input and output fibers were lensed fibers
variations, we used a 440-nm-wide Si waveguide as arm 1, and an with mode-field diameters of ~3.5 µm, and the polarization of
800-nm-wide SiN-waveguide as arm 2, respectively. We designed the input light was adjusted to the transverse electric (TE) mode.
the DL MZI with a free spectral range (FSR) of 40 GHz. The FSR The chip was set on a temperature-controlled stage by using a
is given by the inverse of the differential delay, or 1-bit delay time heat-dissipation tape. We measured the transmission spectra of
∆t = (n g, 2 L 2 - n g, 1 L 1 )/c, where n g, 1 and n g, 2 are group indices of the DL MZIs, while varying the chip-stage temperature range
the arm 1 and 2, and c is the speed of light in vacuum. Under from 298–302 K so that the wavelength shift should not exceed the
this differential delay condition, the dλ/dT can be 0 by choosing FSR. Figures 4A–C show the transmission spectra of the Si–SiN-
the L 1 and L 2 as 0.95 mm and 5.77 mm, respectively. It is notable waveguide DL MZI, the conventional SiN-waveguide DL MZI,
that if we could use the state-of-the-art fabrication process with and the conventional Si-waveguide DL MZI at 298 and 300 K.
width variations of 3 nm (Shimura et al., 2014), the dλ/dT could The output powers were normalized by the fiber-to-fiber trans-
be less than 0.1 pm/K, which is over 10 times smaller than that mission spectra. It is clear that the Si–SiN-waveguide DL MZI
using a 280-nm-wide (narrow) Si-waveguide as arm 2 (Hai and highly suppresses the wavelength shift with temperature varia-
Liboiron-Ladouceur, 2011) with the same width variations. tions. The measured FSRs and the dλ/dT are listed in Table 1.
As other features to construct the DL MZI structure, we used The TO effects of the fabricated SiN- and Si waveguides are
the inverse taper of the Si waveguide for the fiber-chip interface, almost consistent with their designs. The dλ/dT of the Si-SiN
and 2 × 2 Si-waveguide multimode interference (MMI) couplers. DL MZI is over six times smaller than that of the conventional
Frontiers in Materials | Optics and Photonics March 2015 | Volume 2 | Article 26 | 24

FIGURE 4 | Transmission spectra of (A) Si–SiN-waveguide DL MZI, (B) SiN-waveguide DL MZI, and (C) Si-waveguide DL MZI.
Table 1 | Measured FSRs and dλ/dT.
Sample Arm 1 Arm 2 FSR (GHz) dλ/dT (pm/K)
Si–SiN DL MZI Si SiN 40.5 −2.8

SiN DL MZI (ref.) SiN SiN 40.8 +17.0
Si DL MZI (ref.) Si Si 38.3 +68.5
SiN waveguide, and 24 times smaller than that of the conven-

tional Si-waveguide DL MZI. Although the dλ/dT of the Si–SiN
DL MZI is larger than the expected value (<0.1 pm/K) because
of the large fabrication error over 3 nm, the measured result is FIGURE 5 | Experimental setup for decoding of 40-Gbit/s DPSK signals.
still better than those for DLI MZIs with narrow Si waveguides
(Uenuma and Motooka, 2009; Guha et al., 2010). The FSR of the
Si–SiN DL MZI is 40.5 GHz, which is only about 1% different signals is shown in the inset of Figure 5. The polarization state
from the target value. The insertion loss of the Si–SiN-waveguide of the input light was adjusted to the TE mode. The output
DL MZI is ~13 dB, which includes the fiber-chip coupling loss light was coupled by the lensed fiber and then switched to
of ~2 dB/facet and the interlayer coupling loss of ~0.5 dB/couple. the optical spectrum analyzer or the photodiode. The electri-
The extinction ratio is ~17 dB, which is mainly determined by an cal signals from the photodiode were amplified then fed into
imbalance of propagation loss between two arms of the MZI and the sampling oscilloscope. Figures 6A,B show the constructive-
an unintentional imbalance of the MMI branch. The propaga- and destructive-interference spectra of the bar port at frequency
tion loss of SiN and Si waveguides are ~12 dB/cm and ~4 dB/cm, of 193.611 and 193.592 THz, respectively. They were measured
respectively. The unintentional imbalance of the MMI branch is at 298 K. The decoded 40-Gbit/s signals were observed as 40-
~1.3 dB. Although the imbalance between two arms is large, the GHz-span dips in the destructive-interference spectra. Their
imbalance of the MMI branch improves the extinction ratio. The eye diagrams are shown in the insets of Figures 6A,B, respec-
insertion loss and extinction ratio would be further improved tively. Here, the vertical axis is 20 mV/div and the horizon-
because a well-controlled fabrication environment would reduce tal axis is 10 ps/div. The two output signals carried logically
the SiN-waveguide loss (Huang et al., 2014). We must men- inverted data streams in the DPSK format. The constructive inter-
tion that the device was designed for only TE mode light. For ference carried duobinary, whereas the destructive-interference
polarization diversity circuits (Fukuda et al., 2008), a polarization carries alternate-mark inversion (Gnauck and Winzer, 2005).
rotator using parallel cores has already been demonstrated using Using the destructive interference, we demonstrated a ther-
the same Si and SiN layers as in this work (Fukuda and Wada, mal tolerance to decode the DPSK format to the OOK format
2014). (Winzer and Leuthold, 2001; Lazzeri et al., 2010). Figures 7A,B
As a feasibility demonstration, we applied the Si–SiN- show the eye diagrams of the destructive-interference signals at
waveguide DL MZI to decode 40-Gbit/s DPSK signals. Figure 5 299 and 302 K, respectively. The eye diagrams clearly open at
shows the experimental setup. We input the DPSK signals 302 K, corresponding to a temperature variation of 4 K from
with a non-return-to-zero (NRZ) 40-Gbit/s pseudo-random bit the initial temperature (298 K). From the measured dλ/dT of
sequence (PRBS) of length 231 - 1. The eye diagram of the input -2.8 pm/K, the estimated frequency shift by the temperature

FIGURE 6 | Decoded-signal spectra and eye diagrams (insets) of (A) constructive- and (B) destructive interference at 298 K.
FIGURE 7 | Eye diagrams of Si–SiN-waveguide DL MZI at (A) 299 and FIGURE 8 | Eye diagrams of Si-waveguide DL MZI at (A) 298 and (B)
(B) 302 K. 299 K.
variation of +4 K is -1.4 GHz, which could cause only about

1-dB penalty for a 40-Gbit/s system (Hoon and Winzer, 2003). REFERENCES
By using the state-of-the-art fabrication process, as discussed Chen, L., Doerr, C. R., Dong, P., and Chen, Y. K. (2011). Monolithic silicon chip
with 10 modulator channels at 25 Gbps and 100-GHz spacing. Opt. Express 19,
in the above section, the tolerable temperature variations could B946–B951. doi:10.1364/OE.19.00B946
be over 90 K. As references, the eye diagrams of the conven- Fukazawa, T., Ohno, F., and Baba, T. (2004). Very compact arrayed-waveguide-
tional Si-waveguide DL MZI at 298 and 299 K are shown in grating demultiplexer using Si photonic wire waveguides. Jpn. J. Appl. Phys. 43,
Figures 8A,B, respectively. Here, the vertical axis is 69.7 mV/div L673–L675. doi:10.1143/JJAP.43.L673
Fukuda, H., and Wada, K. (2014). Parallel-core-type polarization rotator for silicon
and the horizontal axis is 10.0 ps/div. The eye diagram was com-
wire waveguide platform. Photonics Res. 2, A14–A18. doi:10.1364/PRJ.2.000A14
pletely closed with a temperature variation of only +1 K. These Fukuda, H., Yamada, K., Tsuchizawa, T., Watanabe, T., Shinojima, H., and Itabashi,
results clearly show that the Si–SiN-waveguide DL MZI actually S. (2008). Silicon photonic circuit with polarization diversity. Opt. Express 16,
improves the thermal-insensitivity of the Si-photonic DL MZI 4872–4880. doi:10.1364/OE.16.004872
and that it has a potential to be used in the telecommunica- Gnauck, A. H., and Winzer, P. J. (2005). Optical phase-shift-keyed transmission. J.
Lightw. Technol. 23, 115–130. doi:10.1109/JLT.2004.840357
tions devices. A well-controlled fabrication environment would
Gondarenko, A., Levy, J. S., and Lipson, M. (2009). High confinement micron-
reduce the insertion loss of the DL MZI and also improves the scale silicon nitride high Q ring resonator. Opt. Express 17, 11366–11370.
signal-to-noise ratio. doi:10.1364/OE.17.011366
Guha, B., Gonarenko, A., and Lipson, M. (2010). Minimizing temperature sen-
CONCLUSION sitivity of silicon Mach-Zehnder interferometers. Opt. Express 18, 1879–1887.
We demonstrated a small sensitivity to temperature variations doi:10.1364/OE.18.001879
of DL MZI on a Si-photonics platform. The key technique Hai, M. S., and Liboiron-Ladouceur, O. (2011). “Temperature compensated 50
Gb/s DPSK demodulator,” in Conference on Lasers and Electro-Optics (Baltimore:
is to balance the TO effect in two arms by using waveguides OSA), JTuI76.
made of different materials, which are Si and SiN. The Si–SiN- Hiraki, T., Nishi, H., Tsuchizawa, T., Kou, R., Fukuda, H., Takeda, K., et al. (2013). Si-
waveguide DL MZI with an FSR of 40 GHz showed dλ/dT of Ge-Silica monolithic integration platform and its application to a 22-Gb/s x 16-
-2.8 pm/K, which is about 24 times smaller than that of the con- ch WDM receiver. Photonics J. 5, 4500407. doi:10.1109/JPHOT.2013.2269676
ventional Si-waveguide DL MZI. The technology has the poten- Hoon, K., and Winzer, P. J. (2003). Robustness to laser frequency offset in
direct-detection DPSK and DQPSK systems. J. Lightw. Technol. 21, 1887–1891.
tial to reduce temperature sensitivities of various Si-photonic doi:10.1109/JLT.2003.816816
devices, such as wavelength filters, phase demodulators, and ring Huang, Y., Song, J., Luo, X., Liow, T. Y., and Lo, G. Q. (2014). CMOS compatible
resonators. monolithic multi-layer Si3 N4 -on-SOI platform for low-loss high performance
Frontiers in Materials | Optics and Photonics March 2015 | Volume 2 | Article 26 | 26

silicon photonics dense integration. Opt. Express 22, 21859–21865. doi:10.1364/ Winzer, P. J., and Leuthold, J. (2001). Return-to-zero modulator using a single NRZ
OE.22.021859 drive signal and an optical delay interferometer. IEEE Photon. Technol. Lett. 13,
Lazzeri, E., Nguyen, A. T., Serafino, G., Kataoka, N., Wada, N., Bogoni, A., et al. 1298–1300. doi:10.1109/68.969887
(2010). “All-optical NRZ-DPSK to RZ-OOK format conversion using optical Xia, F., Rooks, M., Sekaric, L., and Vlasov, Y. (2007). Ultra-compact high order
delay line interferometer and semiconductor optical amplifier,” in Photonics in ring resonator filters using submicron silicon photonic wires for on-chip optical
Switching Conference (Monterey), JTuB45. interconnects. Opt. Express 15, 11934–11941. doi:10.1364/OE.15.011934
Lockwood, D. J., and Pavesi, L. (eds) (2010). Silicon Photonics. Berlin: Springer.
Nishi, H., Tsuchizawa, T., Kou, R., Shinojima, H., Yamada, T., Kimura, H., et al. Conflict of Interest Statement: The authors declare that the research was conducted
(2012). Monolithic integration of a silica AWG and Ge photodiodes on Si in the absence of any commercial or financial relationships that could be construed
photonic platform for one-chip WDM receiver. Opt. Express 20, 9312–9321. as a potential conflict of interest.
doi:10.1364/OE.20.009312
Shimura, D., Horikawa, T., Okayama, H., Jeong, S.-H., Tokushima, M., Sasaki, H., Received: 16 January 2015; accepted: 16 March 2015; published online: 30 March 2015.
et al. (2014). “High precision Si waveguide devices designed for 1.31µm and Citation: Hiraki T, Fukuda H, Yamada K and Yamamoto T (2015) Small sensitivity to
1.55µm wavelengths on 300mm-SOI,” in Conference on Group IV Photonics. temperature variations of Si-photonic Mach–Zehnder interferometer using Si and SiN
Vancouver: IEEE. waveguides. Front. Mater. 2:26. doi: 10.3389/fmats.2015.00026
Tsuchizawa, T., Yamada, K., Watanabe, T., Sungbong, P., Nishi, H., Kou, R., et al. This article was submitted to Optics and Photonics, a section of the journal Frontiers
(2011). Monolithic integration of silicon-, germanium-, and silica-based optical in Materials.
devices for telecommunications applications. IEEE J. Sel. Top. Quant. Electron. Copyright © 2015 Hiraki, Fukuda, Yamada and Yamamoto. This is an open-access
17, 516–525. doi:10.1109/JSTQE.2010.2089430 article distributed under the terms of the Creative Commons Attribution License (CC
Uenuma, M., and Motooka, T. (2009). Temperature-independent silicon waveguide BY). The use, distribution or reproduction in other forums is permitted, provided the
optical filter. Opt. Lett. 34, 599–601. doi:10.1364/OL.34.000599 original author(s) or licensor are credited and that the original publication in this
Vivien, L., and Pavesi, L. (eds). (2013). Handbook of Silicon Photonics. New York: journal is cited, in accordance with accepted academic practice. No use, distribution or
Taylor and Francis. reproduction is permitted which does not comply with these terms.

ORIGINAL RESEARCH
published: 05 May 2015
doi: 10.3389/fmats.2015.00036
Ultrahigh temperature-sensitive
silicon MZI with titania cladding
Jong-Moo Lee 1,2*
1
Electronics and Telecommunications Research Institute, Daejeon, South Korea, 2 School of Advanced Device Technology,
University of Science and Technology, Daejeon, South Korea
We present a possibility of intensifying temperature sensitivity of a silicon Mach-Zehnder

interferometer (MZI) by using a highly negative thermo-optic property of titania (TiO2 ). Tem-
perature sensitivity of an asymmetric silicon MZI with a titania cladding is experimentally
measured from +18 to −340 pm/°C depending on design parameters of MZI.
Keywords: silicon, photonics, temperature, sensor, titania
Introduction
There have been many efforts to adjust temperature-dependent wavelength shift (TDWS) of
a photonic waveguide device using a cladding material with a negative thermo-optic coeffi-
cient (TOC) differently from a core material with a positive TOC (Kokubun et al., 1998; Lee
Edited by: et al., 2007, 2008; Alipour et al., 2010; Guha et al., 2013; Bovington et al., 2014; Lee, 2014).
Jifeng Liu, Polymers have been popularly used as the cladding material with a negative TOC (Kokubun
Thayer School of Engineering, USA et al., 1998; Lee et al., 2007, 2008), and titania (TiO2 ) is recently attracting attention with a
Reviewed by: highly negative TOC (Alipour et al., 2010; Guha et al., 2013; Bovington et al., 2014; Lee, 2014)
Raul J. Martin-Palma, and its merit of complementary-metal-oxide-semiconductor (CMOS) compatibility in fabrica-
Universidad Autonoma de Madrid,
tion when it is used for a silicon photonic waveguide device. Silicon has a very high TOC of
1.8 × 10−4 /°C and there have been many efforts to reduce the high TDWS of silicon photonic
Spain
Venu Gopal Achanta,
Tata Institute of Fundamental devices such as a ring resonator by using polymer (Kokubun et al., 1998; Lee et al., 2007,
Research, India 2008) or titania cladding (Alipour et al., 2010; Guha et al., 2013; Lee, 2014) with a highly
*Correspondence:
negative TOC.
Jong-Moo Lee, In case of silicon photonic Mach-Zehnder interferometer (MZI), there were reports showing the
Electronics and Telecommunications way to reduce the TDWS of MZI without using a cladding with a negative TOC (Uenuma and
Research Institute, 161 Gajong-dong, Moooka, 2009; Guha et al., 2010; Dwivedi et al., 2013). TDWS of silicon MZI was shown to be
Yusong-gu, Daejeon 305-350, Korea reduced by using different widths of waveguide (Uenuma and Moooka, 2009; Guha et al., 2010)
jongmool@etri.re.kr
or by using different polarization (Dwivedi et al., 2013) in each of the MZI arm, respectively. The
difference in each of the MZI arm can induce a different temperature-dependent phase change for
Specialty section: the each arm, resulting reduction in TDWS of MZI.
This article was submitted to Optics The previous efforts of silicon photonic devices using a negative thermo-optic cladding have
and Photonics, a section of the been focused on reducing the TDWS, but there have been demands also on high TDWS in
such applications of low power temperature tuning (Masood et al., 2013) and integrated-photonic
Received: 30 January 2015 temperature sensors (Irace and Breglio, 2003; Kim et al., 2010; Deng et al., 2014). So, it would also be
Accepted: 08 April 2015
attractive if there is a method to intensify the TDWS of the silicon device by using a cladding material
Published: 05 May 2015
such as titania with a very high TOC. There have been MZIs using titania for chemical sensing (Qi
Citation:
et al., 2002; Celo et al., 2009), but no reports on temperature sensors using titania cladding to the
Lee J-M (2015) Ultrahigh
temperature-sensitive silicon MZI with
best of our knowledge.
titania cladding. In this regard, here, we combine the method used to in reducing TDWS of silicon MZI with
Front. Mater. 2:36. different dimension for each arm and the method of adding titania cladding on the silicon MZI to
doi: 10.3389/fmats.2015.00036 show the possibility of ultrahigh temperature-sensitive silicon MZI.
Frontiers in Materials | www.frontiersin.org 28 May 2015 | Volume 2 | Article 36

Lee Ultrahigh temperature-sensitive MZI
Experiment and Results cladding. We deposited 400 nm thickness of titania cladding on

the fabricated device without an upper cladding, using electron-
Design and Fabrication beam evaporation. The initial vacuum level of the electron-beam
Temperature dependence of a silicon MZI can be adjusted by evaporation was 5 × 10−7 Torr, and it was kept at 8 × 10−5 Torr
asymmetric geometry of two waveguide arms with different effec- with O2 during the deposition. The temperature of a plate holding
tive refractive indexes induced by different cross-sectional dimen- SOI chip was maintained at 150°C during the evaporation, and the
sion as in reference Uenuma and Moooka (2009) and Guha et al. speed of deposition was about 3Å/s. The refractive index of titania
(2010). Silicon MZIs in this experiment are designed with varia- was measured using ellipsometry as 2.13 at 1550 nm.
tions in the length of MZI arm and cross-sectional dimension of
each MZI arm as in Figure 1. Figure 1A shows AsyL, which is Measured Results
for asymmetry in the length of each MZI arm of 80 μm, L for the One pair of single-mode fibers is coupled to the silicon devices
common length of MZI arm, which is varied from 110 to 360 μm, for measurement through grating couplers which are with 630-
w0 for the common width of waveguide core, which is 450 nm, nm pitch and 70 nm depth of the shallow etch. Figure 3 shows
w1 for the cross-sectional dimension of a waveguide core which normalized transmission spectra of a silicon MZI with a silica
is 1350 nm and shaped as a rib waveguide shown in Figure 1B or
450 nm and shaped as a channel waveguide for a comparison, and
w2 for the cross-sectional dimension of a waveguide core which is
350 or 450 nm for a comparison. Figures 1B,C show the cross-
sectional structure of the waveguide with a silica cladding and
a titania cladding, respectively. The rib waveguide is formed by
shallow etch of 70 nm for the width w1, and there are tapers at
the both ends of the rib waveguide for an adiabatic transfer to the
channel waveguide with the width of w0.
Figure 2 shows a microscopic image and scanning electron
microscope (SEM) image of the fabricated MZIs. There are many
variations for the length, L, in design, but we limit our discus-
sion here to the two extreme case of L, 110 and 360 μm. Sili-
con waveguide core with the width of 450 nm was patterned by
DUV lithography on a silicon-on-insulator (SOI) wafer with a
220-nm thick silicon layer on a 2-μm thick buried oxide (BOX)
layer. The fabrication of the devices except a deposition of titania
cladding were processed using a standard CMOS fabrication pro-
cess through ePIXfab. There were two types of fabricated device:
one with a silica (SiO2 ) cladding and the other without an upper
FIGURE 2 | A microscope image and a SEM image of silicon MZIs

fabricated in this experiment.
FIGURE 1 | (A) Schematic diagram of asymmetric MZI with arms different in

length and cross-sectional dimension of waveguide, and schematic diagram FIGURE 3 | Normalized transmission spectra of MZI with silica
of the cross-section of the arms in (B) with silica cladding and (C) with titania cladding as the temperature varied from 15 to 55°C when L = 360 μm,
cladding, respectively. w1 = 1350 nm, and w2 = 350 nm.

cladding when the temperature varied from 15 to 55°C. The The input and output fibers are coupled to the silicon waveg-
main design parameters of the silicon MZI are 360 μm for L, uide through grating couplers with the pitch of 630 nm. The fiber
1350 nm for w1, and 350 nm for w2. Figure 3 shows TDWS of the was coupled at the vertical angle of 10° for the waveguide with
silicon MZI is +48 m/°C, which was reduced from +74 pm/°C of silica cladding and 15° for the waveguide with titania cladding.
a ring resonator included in the same chip for a comparison. The There was not a big difference in the coupling loss of the grating
normalized transmissions can be regarded as the insertion loss couplers for the silica cladding and titania cladding. It was about
of the silicon MZI, because they were calculated by subtracting 5 dB/facet for the silica cladding and 5.5 dB/facet for the titania
the amount of fiber-to-fiber transmission of a straight silicon cladding. The slightly excessive loss of the grating coupler in case
waveguide from the amount of fiber-to-fiber transmission of the of titania cladding is expected to be reduced by optimizing the
silicon MZI device. Figure 3 shows that the insertion loss through design of gratings or the thickness of titania if it is required. The
the silicon MZI is negligibly small. normalized transmission in Figures 3 and 4 were calculated by
Figure 4 shows normalized transmission spectra of a silicon subtracting the amount of fiber-to-fiber transmission of a straight
MZI, whose design is the same as in Figure 3 but with a titania silicon waveguide from the amount of fiber-to-fiber transmission
cladding instead of the silica cladding, when the temperature of the silicon MZI device for each case of silica cladding and titania
varied from 25 to 35°C. Figure 4 shows TDWS of the silicon MZI cladding, respectively. So, the normalized transmission spectra
is intensified with opposite sign by the titania cladding as high as show the insertion loss of MZI compared to a straight waveguide.
−340 pm/°C, which is about seven times bigger than the TDWS of The refractive index of titania cladding was measured using
the same design of MZI with a silica cladding and five times bigger ellipsometry as 2.13 at 1550 nm, and TOC of titania film was
than the TDWS of the ring resonator with a silica cladding.
Figure 5 shows the relative wavelength shift of various MZIs
with silica or titania cladding in this experiment compared to the
TDWS of the ring resonator with the silica cladding. The radius of
the ring resonator with the silica cladding was 5 μm and TDWS
of the ring resonator was measured at +74 pm/°C as in Figure 6.
TDWS of a silicon MZI with the same cross-section dimension
of 450 nm and titania cladding was measured as +18 pm/°C as
in Figure 5. TDWS of another titania-covered silicon MZI with
1350 nm for w1, 450 nm for w2, and 110 μm for L was measured
as −70 pm/°C as in Figure 5.
Discussion
The experimental results show that we can adjust TDWS of the
titania-covered silicon MZIs with proper design and can inten-
sify the temperature sensitivity highly enough to be useful in
applications requiring an ultrahigh temperature sensitivity such
FIGURE 5 | Relative wavelength shift depending on temperature of
as thermo-optic tuning devices or photonic temperature sensors. MZI with silica and titania cladding, respectively, in comparison with
+74 pm/°C or a ring resonator with silica cladding. w350 and w450 are
for 350 and 450 nm width, respectively, of the MZI narrow arm. L110 and
L360 are for 110 and 360 μm length, respectively, of the MZI arm.
FIGURE 4 | Normalized transmission spectra of MZI with titania FIGURE 6 | Normalized transmission spectra of the ring resonator with
cladding as the temperature varied from 25 to 35°C when L = 360 μm, silica cladding mentioned in Figure 5, as the temperature varied from
w1 = 1350 nm, and w2 = 350 nm. 15 to 55°C.

not directly measured but estimated from −5 to −7 × 10−4 /°C thermo-optic titania cladding. We experimentally showed tem-
by the measured TDWS of a ring resonator with titania cladding perature sensitivity of an asymmetric silicon MZI with a titania
as in reference Lee (2014). The absolute value of TOC of titania cladding could be adjusted from +18 to −340 pm/°C depending
is several times higher than TOC of silicon or polymer, and on design parameters such as the width and length of MZI. We
that is the reason we used it as highly negative thermo-optic believe these results show the possibility of ultrahigh temperature-
cladding in this experiment. The reason for the highly negative sensitive silicon MZI for new applications requiring ultrahigh
TOC of the titania cladding and the variation of TOC is not fully temperature sensitivity such as thermo-optic tuning devices or
understood yet, and finding the reason remains for our future photonic temperature sensors.
research.
Acknowledgments
Conclusion
We would like to thank ePIXfab (www.epixfab.eu) for the fabrica-
We experimentally showed that TDWS of a silicon MZI can tion of SOI waveguide before our deposition of titania cladding.
be reduced or intensified by proper design of the width and This work was supported by Korean IT R&D program MOTIE
length of arms of MZI when it is used with a highly negative [N019800001] and [10044735].
References Kokubun, Y., Yoneda, S., and Matsuura, S. (1998). Temperature-independent optical
filter at 1.55μm wavelength using a silica-based athermal wavelength. Electron.
Alipour, P., Atabaki, A. H., Eftekhar, A. A., and Adibi, A. (2010). “Athermal Lett. 34, 367–369. doi:10.1049/el:19980245
performance in titania-clad microring resonators on SOI,” in OSA Technical Lee, J. M. (2014). Influence of titania cladding on SOI grating coupler and 5 μm-
Digest. Integrated Photonics Research, Silicon and Nanophotonics and Photonics radius ring resonator. Opt. Commun. 338, 101–105. doi:10.1016/j.optcom.2014.
in Switching, Paper IMC6; 2010 Jul 25-28, (Monterey, CA: Optical Society of 10.039
America). Lee, J. M., Kim, D. J., Ahn, H. Y., Park, S. H., and Kim, G. O. (2007). Temperature
Bovington, J., Wu, R., Cheng, K. T., and Bowers, J. E. (2014). Thermal stress dependence of silicon nanophotonic ring resonator with a polymeric overlayer.
implications in athermal TiO2 waveguides on a silicon substrate. Opt. Express J. Lightwave Technol. 25, 2236–2243. doi:10.1109/JLT.2007.899792
21, 661–666. doi:10.1364/OE.22.000661 Lee, J. M., Kim, D. J., Kim, G. H., Kwon, O. K., Kim, K. J., and Kim, G. O. (2008).
Celo, D., Post, E., Summers, M., Smy, T., Brett, M. J., and Albert, J. (2009). Controlling temperature dependence of silicon waveguide using slot structure.
Interferometric sensing platform with dielectric nanostructured thin films. Opt. Opt. Express 16, 1645–1652. doi:10.1364/OE.16.001645
Express 17, 6655–6664. doi:10.1364/OE.17.006655 Masood, A., Pantouvakiy, M., Lepagey, G., Verheyeny, P., Campenhouty, J. V.,
Deng, Q. Z., Li, X. B., Chen, R. B., and Zhou, Z. (2014). “Low-cost silicon photonic Absily, P., et al. (2013). “Comparison of heater architectures for thermal control
temperature sensor using broadband light source,” in IEEE 11th International of silicon photonic circuits,” in IEEE 10th International Conference, Group IV
Conference, Group IV Photonics (GFP), WP23; 2014 Aug 27-29, (Paris). Photonics (GFP), ThC2; 2013 Aug 28-30, (Seoul).
Dwivedi, S., D’heer, H., and Bogaerts, W. (2013). A compact all-silicon temperature Qi, Z., Matzuda, N., Itoh, K., Murabayashi, M., and Labers, C. R. (2002). A design
insensitive filter for WDM and bio-sensing applications. IEEE Photonics Technol. for improving the sensitivity of Mach-Zehnder interferometer to chemical
Lett. 25, 2167–2170. doi:10.1109/LPT.2013.2282715 and biological measurands. Sens Actuators B Chem 81, 254–258. doi:10.1016/
Guha, B., Cardenas, J., and Lipson, M. (2013). Athermal silicon microring res- S0925-4005(01)00960-1
onators with titanium oxide cladding. Opt. Express 21, 26557–26563. doi:10. Uenuma, M., and Moooka, T. (2009). Temperature-independent silicon waveguide
1364/OE.21.026557 optical filter. Opt. Lett. 34, 599–601. doi:10.1364/OL.34.000599
Guha, B., Gondarenko, A., and Lipson, M. (2010). Minimizing temperature sen-
Conflict of Interest Statement: The author declares that the research was con-
sitivity of silicon Mach-Zehnder interferometers. Opt. Express 18, 1879–1887.
doi:10.1364/OE.18.001879
Irace, A., and Breglio, G. (2003). All-silicon optical temperature sensor based
on multi-mode interference. Opt. Express 11, 2807–2812. doi:10.1364/OE.11. Copyright © 2015 Lee. This is an open-access article distributed under the terms of the
002807 Creative Commons Attribution License (CC BY). The use, distribution or reproduction
Kim, G. D., Lee, H. S., Park, C. H., Lee, S. S., Lim, B. T., Bae, H. K., et al. (2010). in other forums is permitted, provided the original author(s) or licensor are credited
Silicon photonic temperature sensor employing a ring resonator manufactured and that the original publication in this journal is cited, in accordance with accepted
using a standard CMOS process. Opt. Express 18, 22215–22221. doi:10.1364/OE. academic practice. No use, distribution or reproduction is permitted which does not
18.022215 comply with these terms.

ORIGINAL RESEARCH
published: 27 April 2015
doi: 10.3389/fmats.2015.00034
Silicon-nitride-based integrated
optofluidic biochemical sensors
using a coupled-resonator optical
waveguide
Jiawei Wang, Zhanshi Yao and Andrew W. Poon *
Photonic Device Laboratory, Department of Electronic and Computer Engineering, The Hong Kong University of Science and
Technology, Hong Kong, China
Silicon nitride (SiN) is a promising material platform for integrating photonic components
and microfluidic channels on a chip for label-free, optical biochemical sensing applications
Edited by:
in the visible to near-infrared wavelengths. The chip-scale SiN-based optofluidic sensors
Dan-Xia Xu, can be compact due to a relatively high refractive index contrast between SiN and the
National Research Council Canada, fluidic medium, and low-cost due to the complementary metal-oxide-semiconductor
Canada
(CMOS)-compatible fabrication process. Here, we demonstrate SiN-based integrated
Reviewed by:
Koji Yamada, optofluidic biochemical sensors using a coupled-resonator optical waveguide (CROW) in
Nippon Telegraph and Telephone the visible wavelengths. The working principle is based on imaging in the far field the out-
Corporation, Japan
Weidong Zhou,
of-plane elastic-light-scattering patterns of the CROW sensor at a fixed probe wavelength.
University of Texas at Arlington, USA We correlate the imaged pattern with reference patterns at the CROW eigenstates. Our
Robert Halir,
sensing algorithm maps the correlation coefficients of the imaged pattern with a library of
Universidad de Málaga, Spain
calibrated correlation coefficients to extract a minute change in the cladding refractive
*Correspondence:
Andrew W. Poon, index. Given a calibrated CROW, our sensing mechanism in the spatial domain only
Photonic Device Laboratory, requires a fixed-wavelength laser in the visible wavelengths as a light source, with the
Department of Electronic and
Computer Engineering, The Hong
probe wavelength located within the CROW transmission band, and a silicon digital
Kong University of Science and charge-coupled device/CMOS camera for recording the light scattering patterns. This is
Technology, Clear Water Bay,
in sharp contrast with the conventional optical microcavity-based sensing methods that
Kowloon, Hong Kong, China
eeawpoon@ust.hk impose a strict requirement of spectral alignment with a high-quality cavity resonance
using a wavelength-tunable laser. Our experimental results using a SiN CROW sensor
Specialty section: with eight coupled microrings in the 680 nm wavelength reveal a cladding refractive
This article was submitted to Optics
index change of ~1.3 × 10−4 refractive index unit (RIU), with an average sensitivity of
and Photonics, a section of the
journal Frontiers in Materials ~281 ± 271 RIU−1 and a noise-equivalent detection limit of 1.8 × 10−8 ~ 1.0 × 10−4 RIU
Received: 31 January 2015 across the CROW bandwidth of ~1 nm.
Accepted: 01 April 2015
Published: 27 April 2015 Keywords: silicon nitride, biochemical sensor, integrated optofluidics, coupled-resonator optical waveguide,
microring resonators, CMOS-compatible, elastic light scattering, visible wavelengths
Citation:
Wang J, Yao Z and Poon AW (2015)
Silicon-nitride-based integrated
optofluidic biochemical sensors using Introduction
a coupled-resonator optical
waveguide In recent years, the increasing demands of medical diagnostics outside a clinic or a laboratory and
Front. Mater. 2:34. self-monitoring for personal healthcare have highly motivated the rapid research and development
doi: 10.3389/fmats.2015.00034 of portable, low-cost biochemical sensors (Estevez et al., 2012). Particularly, miniaturized, label-free
Frontiers in Materials | www.frontiersin.org 32 April 2015 | Volume 2 | Article 34

Wang et al. Silicon-nitride coupled-microresonator biochemical sensors
biochemical sensors are highly desired in order to be readily choice of the SOI platform and the experimental setup config-
deployed at or carried to the sensing environment and to read- uration (including a 1550 nm laser, an optical amplifier and an
out in real-time, quantitative biochemical information about the InGaAs camera) render our previous work not practical for point-
environment (Vollmer et al., 2008). Among various demonstrated of-care optical biochemical sensing applications. Particularly, in
chip-scale photonic biochemical sensors, optical microresonator- order to leverage the wide availability of smartphones for bio-
based biosensors featuring optical resonances with a high quality chemical sensing (Lakshminarayanan et al., 2015), it would be
(Q) factor (103 ~ 104 ) promise a high sensitivity [few tens to advantageous to switch the operational wavelength of the sensor
hundreds of nanometer resonance shift per refractive index unit from the telecommunication wavelengths to the visible or near-
(RIU)], a low detection limit (10−7 ~ 10−4 RIU) and a compact infrared wavelengths that can be readily recorded using high-
footprint (few to hundreds of micrometer square) (De Vos et al., resolution silicon charge-coupled device (CCD)/complementary
2007; Ciminelli et al., 2013; Sedlmeir et al., 2014). However, such metal-oxide-semiconductor (CMOS) cameras.
high-Q microcavity-based sensors working in the spectral domain In this paper, we report our experimental demonstration of
are constrained by a narrow resonance bandwidth as the sensing the CROW-based biochemical sensors in the visible wavelengths
window, which requires a strict resonance alignment and thus in the silicon-nitride (SiN) platform. The SiN platform is trans-
may compromise the reliability of the sensor system. Besides, the parent to the visible and near-infrared wavelengths (Gorin et al.,
sensing implementation typically requires a precision wavelength- 2008; Subramanian et al., 2013) and its fabrication process is
scanning setup, such as a wavelength-tunable laser, which may CMOS-compatible. After the CROW calibration steps, our sens-
limit the portability of the sensor system. ing scheme in principle only requires a fixed-wavelength, low-
Other than microcavity-based biochemical sensors, integrated output-power, visible laser source, and a silicon CCD/CMOS
interferometric optical biochemical sensors also attract increasing camera for recording out-of-plane light-scattering patterns from
attentions. Various kinds of interferometer structures, including the top-view. This offers a promising opportunity to integrate
Mach–Zehnder interferometers (MZI) (Densmore et al., 2008; the CROW sensor with a smartphone that is equipped with a
Kozma et al., 2009; Duval et al., 2013; Halir et al., 2013; Dante compact laser source and a high-resolution camera with a properly
et al., 2015), Young interferometers (Ymeti et al., 2007), and designed optical interface for future smartphone-based point-of-
Hartman interferometers (Xu et al., 2007) have been adopted as care applications.
integrated interferometric biochemical sensors, demonstrating a
high sensitivity (102 ~ 104 rad/RIU) along with a low detection Principle and Methods
limit (10−7 ~ 10−5 RIU). One key merit of such integrated inter-
ferometric sensors is that they require a relatively simple configu- Principle and the Sensing Algorithm
ration, which typically comprises a fixed-wavelength laser source Figure 1 illustrates the principle of the CROW-based biochemical
and a photodetector. However, these interferometric sensors are sensor following our previous work (Wang et al., 2014). Here,
not tolerant to equipment noises that cause output intensity vari- we outline the key concepts of the principle for understanding
ations, such as laser intensity variations. this work. Figure 1A schematically shows a SiN CROW sen-
Previously, our research group has proposed a coupled- sor comprising eight coupled microring resonators with identi-
resonator optical waveguide (CROW)-based biochemical sensing cal design, coupled to input and output bus waveguides in an
scheme using what we termed “pixelized pattern detection” in add-drop filter configuration. For a perfect CROW compris-
the spatial domain (Lei and Poon, 2011). The scheme employs ing C coupled identical single-mode resonators, the inhomoge-
the discrete transition of the CROW eigenstate excited at a fixed neously broadened transmission spectrum features a combination
laser wavelength upon a small change in the cladding refrac- of split mode resonances, with each mode slightly shifted from the
tive index, Δn, and detects the resulting change in mode-field- original resonance frequency due to inter-cavity-coupling effect.
intensity distribution by far-field measurement of the out-of-plane Therefore, the eigenstate number N within each transmission
elastic-light-scattering intensity patterns. Such a sensing scheme band always equals to the resonator number C. While a perfect
in principle only requires relatively simple optical sources and CROW exhibits distinctive mode-field-amplitude distributions
imaging systems including a fixed-wavelength laser and a cam- at eigenstates, the pair of symmetric and anti-symmetric split-
era. Recently, we have experimentally demonstrated a proof of modes at different eigenfrequencies have non-distinctive mode-
concept of such a chip-scale CROW-based sensor on the silicon- field-intensity distributions. In practice, a CROW inevitably suf-
on-insulator (SOI) platform in the 1550 nm telecommunication fers from fabrication imperfections. The coupled resonators are
wavelengths (Wang et al., 2014). We have extended the scheme no longer identical nor are identically coupled. The symme-
by detecting the continuous modulation of the CROW mode- try breaking between the pair of symmetric and anti-symmetric
field-intensity distribution at a fixed wavelength upon a Δn by split-modes therefore results in distinctive mode-field-intensity
correlating the elastic-light-scattering patterns with reference pat- distributions at all discernable eigenstates. The resulting phase
terns at the CROW eigenstates. Compared with interferometric disorders and coupling disorders can result in the split mode
sensors, the correlation analysis allows our sensing scheme to resonances to be spectrally overlapped. Therefore, in the presence
be more tolerant to equipment noises that are common to all of structural non-uniformity, N could be equal to or smaller than
pixels of the CROW sensor yet do not cause a spectral shift, C (N ≤ C).
including laser intensity variations. Our previous experiment Figure 1B schematically illustrates the inhomogeneously
demonstrated a Δn of ~1.5 × 10−4 RIU and a noise-equivalent broadened transmission bands upon applying cladding refrac-
detection limit (NEDL) of 2 × 10−7 ~ 9 × 10−4 RIU. However, the tive indices n0 and n0 + Δn, for an imperfect eight-microring

FIGURE 1 | Principle of SiN CROW-based biochemical sensors characterizing an imperfect eight-microring CROW, including the
using out-of-plane elastic light scattering at the visible inhomogeneously broadened transmission bands upon a buffer solution
wavelengths. (A) Schematic of a SiN CROW-based sensor integrated (n0 ) and a test solution (n0 + Δn), and pixelized mode-field-intensity
with a microfluidic channel. An objective lens and a CMOS/CCD camera distributions at eigenstate wavelength λ j upon n0 , A(λ j ). Insets: pixelized
are applied on top of the optofluidic chip in order to image the mode-field-intensity distributions at probe wavelength, λ p , (i) upon n0
out-of-plane elastic-light-scattering pattern. (B) Illustration of [B(λ p )]; and (ii) upon n0 + Δn [T(λ p )].
CROW exhibiting a complete set of eight distinctive eigenstate respectively. The bar sign denotes the mean of the entire pixelized
mode-field-intensity distributions. With the mode-field intensity pattern over C pixels.
of each microring integrated as a pixel, we denote the pixelized We adopt the Pearson’s correlation coefficient approach to
one-dimensional pattern at the eigenstate as {Aj }, with j index- describe the linear dependence of the measured and calibrated
ing the eigenstate. Any mode-field-amplitude distribution at an intensity distributions. The Pearson’s correlation approach is
arbitrary wavelength, λp , within the CROW transmission band insensitive to both level and scale variations of the intensity
upon n0 can be expressed by a linear superposition of the com- distributions. Therefore, the approach is tolerant to equipment
plete set of eigenstate mode-field-amplitude distributions upon noise sources, such as uniform background light imaged onto the
n0 . Therefore, we are able to uniquely identify any pixelized camera and the intensity variation of the laser source, which are
mode-field-intensity profile at λp upon n0 , B(λp ), as shown in common to all pixels and do not cause a spectral shift. However,
inset (i), with {Aj } by a correlation analysis. Upon a small Δn this approach still suffers from the noises that cause a spectral
applied homogenously to the cladding, we can uniquely iden- shift, such as a wavelength drift of the laser source and thermal
tify by the correlation analysis any pixelized mode-field-intensity variations in the test environment.
distribution at λp upon n0 + Δn, T(λp ), as shown in inset (ii), Here, we detail our sensing algorithm following our previous
with {Aj }. work (Wang et al., 2014). Figure 2 shows a flow chart illustrating
As in our previous work (Wang et al., 2014), we adopt the Pear- our sensing algorithm including calibration.
{ ( )} We first generate a
son’s correlation coefficient, ρ, in order to analyze the degree of library of correlation coefficients ρ′j λ0 , defined at a fixed
correlation between a pixelized pattern at an arbitrary probe
reference wavelength λ0 centered at the CROW transmission
wavelength λp , B(λp ), and the pixelized patterns at the eigenstate
band. The library is calibrated over a range of Δn values, Δnd ,
wavelengths λj , A(λj ). For a CROW with a number of coupled
given by an integer{multiple}of a minimum refractive index change
single-mode cavities, C, and a number of discernable eigenstates, ( )
N (≤C), we define ρ at λp for A(λj ) as follows: interval Δni . The ρ′j λ0 thus comprises a library of data array
of N (rows) × M (columns), where M is given by Δnd /Δni .
∑
C
For sensing, we first measure the pixelized mode-field-intensity
(A(i, λj ) − A(λj ))(B(i, λp ) − B(λp ))
ρj (λp ) = √ i=1
√ pattern in a buffer solution at a fixed probe wavelength λp (which
∑
C 2 ∑
C 2 is generally offset from λ0 ) as B(λp ) (Figure 1B). We correlate
(A(i, λj ) − A(λj )) (B(i, λp ) − B(λp )) B(λp ) with the eigenstate patterns {Aj } in order to extract
i=1 i=1
{ρj (λp )}. }
{ We look for the closest match of {ρj (λp )} with the library
(1) ( )
where j = 1, 2, . . ., N is the eigenstate number, and i = 1, 2, . . ., ρ′j λ0 , using only the principal (largest) component, ρp ,
C is the cavity (pixel) number. A(i, λj ) and B(i, λp ) are the pixel and the second-principal (second-largest) component, ρs , of
values normalized to the total intensity of the entire patterns, {ρj (λp )} in order to streamline the pattern recognition process

We study the effects of these empirical inputs on the device

parameters, including the waveguide effective refractive index,
neff , and the inter-cavity coupling coefficient, κ. We calculate using
the numerical finite-element method (FEM) (COMSOL RF mod-
ule) the neff of a SiN channel ridge waveguide for the transverse-
magnetic (TM)-polarized mode, as a function of waveguide width
around 427.5 nm at a fixed waveguide height of 300 nm upon a
water upper-cladding. We adopt the measured material refractive
index of the deposited 300 nm-thick SiN film as a function of
wavelength using ellipsometry. The mean value of the calculated
neff is 1.5994 ± 0.0003 at 686 nm. We choose the TM polarization
mode in order to obtain a large evanescent field exposure near
the waveguide top surface for better light–analyte interaction.
We calculate the coupling coefficient in each directional coupling
region as a function of the coupling gap spacing, assuming the
waveguide width is fixed at 427.5 nm. We estimate the waveguide
propagation loss upon a water upper-cladding to be relatively
high at ~17 dB/cm based on our measurements. We attribute this
primarily to surface-roughness-induced scattering losses from the
waveguide sidewall. We apply the designed racetrack arc radius
and interaction length into the modeled CROW. We find from
our FEM calculations a linear relationship between Δn and the
resulting effective refractive index change Δneff , which we apply
to our transfer-matrix modeling (see Supplementary Materials S1
and S2).
Device Fabrication
We fabricate the CROW devices in a 4′′ silicon wafer. The silicon
wafer is first grown with a ~2 μm-thick thermal oxide. We grow
FIGURE 2 | A flow chart showing the sensing algorithm including
calibration of CROW-based sensors.
nitrogen-rich SiN by plasma-enhanced chemical vapor deposi-
tion (PECVD) (SiH4 :NH3 = 25:40 (sccm), 300°C, 13.56 MHz).
The thickness of SiN layer is ~300 nm. We fabricate the CROW
(Wang et al., 2014). We thus obtain a unique equivalent refractive device pattern by electron-beam lithography (JEOL JBX-6300FS)
index change for the buffer solution, ΔnB , which is only due to using a positive electron-beam resist ZEP-520A. We transfer the
the offset between λp and λ0 . We repeat the same procedure for device pattern to the SiN layer by inductively coupled plasma
measuring the pattern at λp upon the test solution, T(λp ), and etching with C4 F8 and SF6 gases (STS ICP DRIE Silicon Etcher).
obtain another unique equivalent refractive index change ΔnT . Figure 3A shows the optical micrograph of the fabricated SiN
Finally, we obtain Δn = ΔnT − ΔnB . eight-microring CROW device. The racetrack microring com-
prises two half circles with a radius of 20 μm and two straight
Transfer-Matrix Modeling of Imperfect CROW waveguides with an interaction length (Lc ) of 4 μm. We design the
Sensors waveguide width to be 450 nm and the coupling gap spacing to
We model imperfect SiN CROWs in 680 nm wavelengths using be 100 nm. Figure 3B shows a zoom-in-view optical microscope
transfer-matrix method with empirical inputs (Wang et al., 2014) image of the CROW. Figure 3C shows a SEM picture of the
(see Supplementary Materials S1 and S2). We measure and accu- coupling region.
mulate statistics of the measured waveguide widths and coupling We fabricate a microfluidic chamber on a polydimethylsiloxane
gap widths from scanning-electron microscope (SEM) character- (PDMS) layer. We pattern a SU8 film by contact photolithogra-
ization of our fabricated devices. We sample six waveguide widths phy as a mold in order to form the PDMS microfluidic channel
and three coupling gap width in one coupling region, and measure by imprinting. The designed dimension of microfluidic channel
a total of eighteen coupling regions in two representative eight- is 8 mm × 2 mm × 50 μm (length, width, and height). We use
microring CROW devices (see Supplementary Material S3). The a puncher to make two holes, each with a diameter of 1 mm,
statistics of the waveguide widths and the coupling gap widths as an inlet and outlet for solution delivery. The diced silicon
approximately follow two Gaussian distributions. We extract the chip and the PDMS microfluidic layer are treated with oxy-
fabricated waveguide width of 427.5 ± 1.1 nm and coupling gap gen plasma and directly bonded, with the microfluidic chan-
spacing of 129.1 ± 1.0 nm. In the modeling, we assume that the nel encompassing the CROW sensor. The bonded PDMS–SiN
two Gaussian distributions are independent, and we generate a set interface is stable enough for repeating the sensing experiments
of varied waveguide widths and coupling gap spacing randomly for many times under a relatively high fluidic pump pressure.
distributed across the CROW using the Gaussian number gener- Figure 3D schematically shows the cross-sectional view of the
ator in Matlab. optofluidic chip.

FIGURE 3 | (A) Optical micrograph of the fabricated eight-microring integrated with a microfluidic channel. (E) Schematic of the
CROW. (B) Zoom-in-view picture of the CROW. experimental setup. HWP, half-wave plate; PBS, polarizing beam
(C) Scanning-electron microscope image of an inter-cavity coupling splitter; LWD OB, long-working-distance objective lens; OB, objective
region of the CROW. (D) A cross-sectional view of the SiN chip lens; PD, photodetector; MMLF, multimode lensed fiber.
Experimental Method spectra with a sum of multiple inverted Lorentzian lineshapes,

Figure 3E schematically shows the experimental setup. The each centered at the resonance (eigenstate) wavelength. The over-
wavelength-tunable laser light in the 680 nm wavelengths is end- all transmission band shift is taken as the average value of the
fired into a tapered 3 μm-wide SiN waveguide through an objec- spectral shifts of all the eigenstates.
tive lens (NA = 0.65). The laser power before coupling into the
chip is ~2 mW. The polarization is controlled by a half-wave Results
plate before a polarizing beam splitter. The output light from the
throughput- or drop-port is collected using a multimode lensed Modeling Results
fiber to a silicon power meter and a lock-in amplifier. Figure 4 shows the modeling results for N = C (see Supple-
For elastic-light-scattering pattern imaging from the top view, mentary Material S4 for modeling results corresponding to the
we use a long-working-distance microscope objective lens (20× case N < C). Figure 4A schematically shows an imperfect SiN
Mitutoyo Plan Apo, NA = 0.42) and a CCD camera (Diagnos- CROW with varied waveguide width and coupling gap width of
tic Instruments, Inc., RT3) with 1600 × 1200 pixels (7.4 μm-sized each microring. Inset shows the numerically calculated waveg-
pixels). The camera has an effective differential cooling of −43°C uide mode-field-amplitude profile in the TM mode at 686 nm
and an 8-bit analog-to-digital conversion in data readout. We fix wavelength. Figure 4B shows the modeled throughput- and drop-
the exposure time as 60 ms and the gain of ~1. For background transmission spectra of an imperfect eight-microring CROW.
subtraction, we set the probe wavelength in between the CROW We define the CROW transmission bandwidth, ΔλBW , as the
spectral range between the first and last discernable eigenstates
transmission bands in order to obtain a background image.
within the transmission band. Figure 4C shows the modeled
In order to acquire the library of calibrated correlation coef-
pixelized patterns at the
{ eight eigenstates. Figure 4D shows
ficients, we scan the laser wavelength in steps of 0.02 nm over ( )}
~2 free spectral ranges (FSRs) of the CROW sensor. We record the calculated library ρ′j λ0 as a function of Δn, with
at each wavelength eight successive images over a time period of Δnd = 2.523 × 10−2 and Δni = 3.6 × 10−4 RIU. Figure 4E shows
4 s (at 2 frames/s). We take average of these successive images in the calculated
( (differential correlation coefficients per unit Δn,
))
order to reduce the systematic equipment noise contribution. In given as |d ρ′j λ0 /d (Δn) |.
the sensing tests, we inject the buffer and test solutions, and start
We define the CROW sensitivity (in units of RIU−1 ) at
recording the images after the scattering pattern is stabilized upon
an(arbitrary λp within the transmission band as the larger
an essentially static fluidic medium. We record over 50 successive ′
( ))
images during a time period of 25 s at a fixed probe wavelength. |d ρj λp /d (Δn) | of the ρp and ρs . Figure 4F shows the
In order to calibrate the spectral sensitivity of the CROW, we modeled sensitivity as a non-linear function of λp . The sensitivity
prepare NaCl solutions with mass concentrations from 1 to 5% in the transmission band spans a range from ~73 to ~1440 RIU−1 ,
(in steps of 1%) and test the transmission band spectral shifts with an average sensitivity of ~553 ± 290 RIU−1 . We quantify
upon a Δn. Between each measurement, we rinse the chip by the non-uniformity of the sensitivity by the ratio of SD value
injecting deionized (DI) water using a fluidic pump. We obtain the to average sensitivity value. A lower ratio value suggests a more
resonance spectral shifts by fitting the throughput-transmission uniform sensitivity. The extracted non-uniformity ratio from

FIGURE 4 | (A) Schematic of an imperfect CROW model with non-uniform respectively. (C) Modeled normalized pixelized intensity patterns at the eight
waveguide widths (W 0 , W 1 , W 2 , . . . W N+1 ) and coupling gap spacing eigenstates, I–VIII. (D) Calculated library of the correlation coefficients
(g1 , g2 , . . . gN+1 ). Inset (i): numerically calculated waveguide mode-field ρ ′1 − ρ ′8 as a function of Δn at λ 0 , with Δnd = 2.523 × 10−2 and
amplitude profile in the TM mode. (B) Modeled throughput- and drop-port Δni = 3.6 × 10−4 RIU. (E) Calculated library of the ( differential
) correlation
transmission spectra of an imperfect eight-microring CROW using coefficients as a function of Δn at λ 0 , given as |d ρ ′j (λ 0 ) /d (Δn) |.
transfer-matrix modeling. Green and red dashed-lines indicate the reference (F) Calculated sensitivity as a function of λ p . The red dashed-line indicates a
wavelength λ 0 of 688.06 nm and the probe wavelength λ p of 688.14 nm, sensitivity of 772 RIU−1 at λ p = 688.14 nm.
Figure 4F is ~0.52. Although such a sensitivity variation is not agrees with the arbitrarily chosen Δn value. We attribute the
ideal, we can obtain a practical sensitivity within a wide enough deviation of 2 × 10−5 RIU to the interpolation error. In principle,
wavelength window without fine-tuning the probe wavelength. the maximum error upon the sampling interval in the library is
As an example, we can set a practical sensitivity of ~100 RIU−1 given by ± Δni /2, which is ~1.8 × 10−4 RIU given the assumed
Δni value.
in order to sense a Δn down to 10−5 RIU (assuming a noise-
induced uncertainty of correlation coefficients of ~±10−3 ). From Calibrating the CROW Sensor in a Buffer Solution
Figure 4F, the width of the probe wavelength window with a sen- Figure 6 summarizes the characterization results upon a buffer
sitivity >100 RIU−1 is 1.1 nm. We consider this sufficiently wide solution (DI water). Figure 6A shows the measured TM-polarized
for sensing with a practical sensitivity at an arbitrarily set probe transmission spectra with DI water upper-cladding. The measured
wavelength. If a higher practical sensitivity of, say, 200 RIU−1 FSR of ~1.80 nm is consistent with the microring circumference.
is desired, the width of the probe wavelength window with a The CROW exhibits an inhomogeneously broadened transmis-
sion band, with a ΔλBW of ~1.10 nm. We discern eight eigenstates
sensitivity >200 RIU−1 narrows to ~1.06 nm.
within each transmission band (labeled by I to VIII for the first
Here, we arbitrarily choose λp at 688.14 nm near the center
transmission band, and I′ –VIII′ for the second transmission band
of the CROW transmission band (Figure 4B) in order to model
in Figure 6A).
the sensing test. The sensitivity at λp is ~772 RIU−1 . Figure 5
Figure 6B shows the measured elastic-light-scattering images
illustrates the modeled sensing results. Figure 5A shows the
at eigenstates I–VIII. We observe a non-uniform scattering image
modeled pixelized patterns at λp , B(λp ) and T(λp ), assuming a
profile across each microring. We attribute this to the extra
water buffer (n0 = 1.331) and an arbitrarily chosen Δn value of
modulation of the surface roughness and local defects to the
2.50 × 10−3 RIU, respectively. Figure 5B shows the two sets of
intrinsic mode-field-intensity distributions. We notice an obvious
correlation coefficients extracted from the two modeled pixelized
“local hotspot” in the coupling region between microring 3 and
patterns without and with Δn. The ρp and ρs without Δn are ρ4 and
microring 4 in all the light-scattering images. We attribute that
ρ5 , respectively. The ρp and ρs with Δn are ρ5 and ρ3 , respectively.
to the larger surface roughness localized in the coupling region
Figures
{ ( )5C,D } show the zoom-in view of the calculated library
between microring 3 and microring 4. We integrate within a cer-
ρ′j λ0 as a function of Δn. Insets show the detailed mappings tain window the elastic-light-scattering intensity of each micror-
of ρp and ρs with the library. We extract using linear interpola- ing to form a single pixel. The window excludes the coupling
tion from the library Δn = ΔnT − ΔnB = 2.52 × 10−3 RIU, which region in order to avoid scattering-induced crosstalks between

FIGURE 5 | (A) Modeled normalized pixelized patterns at λ p (688.14 nm) the calculated library of ρ ′j as a function of Δn. Dashed-lines indicate the
upon n0 and n0 + Δn. (B) Calculated correlation coefficients at mapping of ΔnB , ΔnT for buffer solution and test solution, respectively.
λ p = 688.14 nm upon n0 and n0 + Δn. The dashed-line and the Insets (i)–(iv): Mapping of ρ p and ρ s with the library to extract ΔnB
dotted-line boxes indicate ρ p and ρ s , respectively. (C,D) Zoom-in view of and ΔnT .
the coupled waveguides and local hotspots. Here we normalize still relatively tolerant to set a probe wavelength. In conventional
the patterns with the estimated contributions of the surface- microcavity-based sensing methods, the sensitivity is only appli-
roughness-induced scattering as a step for pattern correction (see cable within the high-Q transmission band (~0.1 nm in De Vos
Supplementary Material S5). Figure 6C shows the corrected pix- et al., 2007), which is generally much narrower than our probe
elized mode-field-intensity patterns at the eight eigenstates. We wavelength window.
use the corrected pixelized patterns for sensing. We define the NEDL at λp as the uncertainty of extracted Δn.
Figure 6D shows the measured library of the calibrated cor- We repeat the extraction of Δn values based on ρp and ρs at each
relation coefficients as a function of Δn. Here, we calibrate the λp for eight times and calculate the SD of the eight extracted
sensor by scanning the laser wavelength over ±Δλ (Δλ = 0.7 nm) Δn values. Figure 6G shows the extracted NEDL values as a
about the center of the CROW transmission band spanning a function of λp, which shows a high dependence on the choice of
FSR upon a fixed buffer solution (DI water), with a minimum λp. The NEDL values range from ~2 × 10−8 to ~1 × 10−4 RIU.
wavelength step of 0.02 nm. This interval corresponds to a Δni of We observe particularly low NEDL values (~10−8 RIU) at λp
~3.5 × 10−4 RIU, based on the calibrated linear spectral sensitiv- aligning with the eigenstate wavelengths. We attribute the low
ity of ~57.30 nm/RIU of the CROW sensor (see Supplementary NEDL at each eigenstate to the particularly low uncertainty of
Material S6). We also convert Δλ back to Δn using the calibrated ρp (~±10−6 − 10−4 ) close to 1 at each eigenstate. Upon eight
linear spectral sensitivity. The corresponding range of Δnd is repeated tests at a fixed probe wavelength at each eigenstate, the
~±1.2 × 10−2 RIU. measured pixelized patterns only slightly deviate from the cali-
Figure 6E shows the calculated |dρ′j /d (Δn) | as a function of brated eigenstate distributions due to the low noise in the cooled
Δn. Figure 6F shows the calculated sensitivity as a function of silicon CCD camera and the low thermo-optic coefficient of SiN.
λp over the λ0 ± Δλ range. The calculated sensitivity value shows The low uncertainties of ρp at each eigenstate are converted into
highly non-uniform profiles. The sensitivity ranges from ~15 to particularly low NEDL values.
~1420 RIU−1 , with an average value of ~281 ± 271 RIU−1 . The In order to quantify the sensing resolution, here we define
extracted non-uniformity ratio from Figure 6F is ~0.96. The width the resolution of the CROW sensor as the lowest refractive
of the probe wavelength window with a sensitivity >100 RIU−1 index change that can be sensed reliably and repeatedly. In
is 0.88 nm. Whereas, the width of the probe wavelength window practice, there are two main limiting factors to the resolution.
with a sensitivity >200 RIU−1 narrows to ~0.48 nm, which is One is the interpolation error in extracting Δn. The other is

FIGURE 6 | (A) Measured TM-polarized throughput- and drop-port coefficients ρ ′1 − ρ ′8 as a function of Δn. White dashed-lines indicate the
transmission spectra of the eight-microring SiN CROW with DI water ΔnB values at λ p1 , λ p2 , and λ p3 . (E) Calculated
( differential
) correlation
upper-cladding. Green and red dashed-lines indicate the reference coefficients as a function of Δn, given as |d ρ ′j (λ 0 ) /d (Δn) |.
wavelength λ 0 (686.86 nm) and three probe wavelengths, λ p1 (687.06 nm), (F) Calculated sensitivity as a function of λ p . Red dashed-lines indicate
λ p2 (687.38 nm), and λ p3 (686.42 nm). (B) Measured elastic-light-scattering sensitivities of 214, 279, and 541 RIU−1 at probe wavelengths λ p1 , λ p2 , and
images with DI water upper-cladding at the eight eigenstates I–VIII. The λ p3 , respectively. (G) Extracted noise-equivalent detection limit (NEDL) as a
white-line box indicates the integration window for pixelization. function of λ p . Red dashed-lines indicate NEDL values of ~4 × 10−6 ,
(C) Normalized pixelized mode-field intensity patterns at the eight CROW ~2 × 10−8 , and ~1 × 10−6 RIU at λ p1 , λ p2 , and λ p3 , respectively. Green
eigenstates I–VIII. (D) Calculated library of the calibrated correlation dashed-lines indicate the eight eigenstate wavelengths λ j .
the NEDL taking into account all the noise sources that our are both close to the calibrated results of the first transmission
correlation approach is not tolerant to. Therefore, given a cali- band.
bration interval of Δni (3.5 × 10−4 RIU), the worst resolution is
~1.8 × 10−4 RIU given ± Δni /2 (1.8 × 10−4 RIU) and the NEDL Blind Sensing Test Results
(~1.8 × 10−8 − 1.0 × 10−4 RIU in Figure 6G). The interpolation- We implement blind sensing tests at three different probe wave-
error-limited resolution (±Δni /2) suggests that a Δn below Δni /2 lengths (λp1 , λp2 , and λp3 ) within the CROW transmission band.
may not be tested reliably or repeatedly. The resolution can be We prepare one buffer solution (DI water) and three NaCl solu-
improved by adopting a finer Δni . tions, X, Y, and Z, with different mass concentration values
We also calibrate the CROW sensor in the adjacent trans- unknown to the researcher conducting the sensing tests. We study
mission band (see Supplementary Material S7). The pixelized the images upon the buffer solution at the initial stage and upon
mode-field intensity patterns at eigenstates I′ –VIII′ show a rinsing after each sensing test. We confirm that the pixelized pat-
high similarity with the corresponding patterns at eigenstates tern returns to the baseline pattern (see Supplementary Material
I–VIII, respectively. The extracted sensitivity and NEDL range S8). Table 1 summarizes the experimental sensing results.

TABLE 1 | Sensing results at the three probe wavelengths upon the buffer solution and the three test solutions.
λp Solution ρp ρs ΔnB or ΔnT Δn (× 10−3 RIU) Sensed

(× 10−3 RIU) concentration (%)
λ p1 (687.06 nm) Buffer (DI water) ρ3 (0.927 ± 0.003) ρ6 (0.866 ± 0.002) ~−3.54 ± 0.02 – –
X (NaCl) ρ6 (0.924 ± 0.003) ρ2 (0.838 ± 0.003) ~4.21 ± 0.01 ~7.75 ± 0.02 ~4.35 ± 0.01
Y (NaCl) ρ4 (0.946 ± 0.002) ρ5 (0.854 ± 0.003) ~−2.46 ± 0.03 ~1.08 ± 0.04 ~0.61 ± 0.03
Z (NaCl) ρ3 (0.903 ± 0.002) ρ6 (0.854 ± 0.008) ~−3.41 ± 0.01 ~0.13 ± 0.03 ~0.073 ± 0.014
λ p2 (687.38 nm) Buffer (DI water) ρ 1 (0.99996 ± 0.00003) ρ7 (0.622 ± 0.001) ~−9.1800 ± 0.0001 – –
X (NaCl) ρ 4 (0.941 ± 0.007) ρ5 (0.875 ± 0.005) ~−0.79 ± 0.02 ~8.39 ± 0.02 ~4.70 ± 0.01
Y (NaCl) ρ 2 (0.841 ± 0.012) ρ6 (0.807 ± 0.012) ~−8.16 ± 0.02 ~1.03 ± 0.02 ~0.58 ± 0.01
Z (NaCl) ρ 1 (0.989 ± 0.001) ρ7 (0.622 ± 0.003) ~−9.14 ± 0.01 ~0.05 ± 0.01 ~0.03 ± 0.01
λ p3 (686.42 nm) Buffer (DI water) ρ7 (0.851 ± 0.001) ρ 2 (0.485 ± 0.002) ~7.775 ± 0.003 – –
X (NaCl) ρ1 (0.205 ± 0.015) ρ 7 (−0.013 ± 0.014) – – –
Y (NaCl) ρ2 (0.519 ± 0.006) ρ 8 (0.305 ± 0.009) ~8.81 ± 0.01 ~1.04 ± 0.01 ~0.58 ± 0.01
Z (NaCl) ρ7 (0.783 ± 0.005) ρ 2 (0.477 ± 0.005) ~7.90 ± 0.01 ~0.13 ± 0.01 ~0.069 ± 0.005
Prepared concentration values: X: (4.5 ± 0.1)%, Y: (0.60 ± 0.02)%, Z: (0.070 ± 0.002)%.
Sensing at an Arbitrarily Set Probe Wavelength λp1 attribute this deviation to a not sufficiently fine calibration of the
Figure 7 shows the sensing results at an arbitrarily set probe wave- library and the error from linear interpolation. The calibrated
length λp1 (687.06 nm) near the center of the CROW transmission response of ρp around the eigenstate is in the proximity to the
band. The sensitivity at λp1 is ~214 RIU−1 (see Figure 6F). The maximum (unity). The limited sampling resolution of Δni may not
NEDL at λp1 is ~4 × 10−6 RIU (see Figure 6G). Figure 7A shows be sufficient to describe the response around an extremum.
the measured elastic-light-scattering images of the CROW upon
the buffer solution and the three test solutions at λp1 . Figure 7B Sensing at λp3 Near Eigenstate VII
shows the corresponding pixelized patterns. Figure 7C shows the Figure 9 shows the sensing results at another specifically chosen
corresponding calculated correlation coefficients. Figures 7D–G probe wavelength λp3 (686.42 nm). We specifically set λp3 at the
show the mapping of ρp and ρs in the buffer solution and the blue-edge of the transmission band near eigenstate VII. The sen-
three test solutions with the library. Insets (i)–(viii) show the sitivity at λp3 is ~541 RIU−1 (see Figure 6F). The NEDL at λp3 is
mapping ρp and ρs to the corresponding ΔnB or ΔnT using linear ~1 × 10−6 RIU (see Figure 6G). We consider λp3 as a near opti-
interpolations in between Δni . mized choice with a relatively high sensitivity and a low NEDL.
We acquire for solution X a ΔnX of ~(7.75 ± 0.02) × 10−3 RIU Figure 9A shows the measured elastic-light-scattering images
and for solution Y a ΔnY of ~(1.08 ± 0.04) × 10−3 RIU, both cor- upon the buffer solution and the three test solutions. Figure 9B
responding to a relatively large Δn but still within Δnd . We acquire shows the corresponding pixelized patterns. Figure 9C shows the
for solution Z (Figure 7G) a ΔnZ of ~(1.3 ± 0.3) × 10−4 RIU. For corresponding calculated correlation coefficients. Figures 9D–F
all three solutions, we convert from the measured Δn values the show the mapping of ρp and ρs values with the library (see
sensed concentration values (see Table 1), which show a good Supplementary Material S9 for detailed mappings).
agreement with the prepared values. For solution X, however, we observe an almost dark scattering
pattern, which suggests that λp3 upon solution X is relatively
shifted out of the transmission band. Both the extracted ρp and
Sensing at λp2 Aligned with Eigenstate I ρs values out of ρj (λp3 ) upon solution X are particularly low. By
Figure 8 shows the sensing results at a specifically chosen probe mapping the extracted ρj (λp3 ) values with the library, we find no
wavelength λp2 (687.38 nm) aligned with eigenstate I. The sensi- match to indicate the corresponding Δnx. Therefore, in the case
tivity at λp2 is ~279 RIU−1 (see Figure 6F). The NEDL at λp2 is that there is a chance to measure a large Δn near Δnd (in the order
~2 × 10−8 RIU (see Figure 6G), which is much lower compared of 10−2 ~ 10−3 RIU in this case), it is better to position λp close
with that at λp2 . Figure 8A shows the measured elastic-light- to the red-side of the transmission band in order to leverage the
scattering images upon the buffer solution and the three test dynamic range given by ΔλBW in full.
solutions. Figure 8B shows the corresponding pixelized patterns. For solution Y, we acquire a ΔnY of ~(1.04 ± 0.02) × 10−3 RIU.
Figure 8C shows the corresponding calculated correlation coef- For solution Z, we acquire a ΔnZ of ~(1.24 ± 0.1) × 10−4 RIU.
ficients. Figures 8D–G show the mapping of ρp and ρs values in Both sensing results agree with the prepared concentrations of
the buffer solution and the three test solutions with the library (see solutions Y and Z. Compared with the sensing result of solution
Supplementary Material S9 for detailed mappings). Z at λp1 , we obtain a more accurate value of ΔnZ with a much
We acquire for solution X a ΔnX of ~(8.40 ± 0.02) × 10−3 RIU improved uncertainty. We attribute this to a higher sensitivity and
and for solution Y a ΔnY of ~(1.03 ± 0.02) × 10−3 RIU. Both a lower NEDL at λp3 than those at λp1.
sensing results agree with the prepared concentrations of solu-
tions X and Y. For solution Z (Figure 8G), we acquire a ΔnZ of Discussion
~(0.5 ± 0.1) × 10−4 RIU, corresponding to a mass concentration
of ~(0.03 ± 0.01)%. This, however, shows a significant devia- Here, we benchmark our work with other silicon- and SiN-based
tion from the prepared concentration [~(0.070 ± 0.002)%]. We on-chip optical biochemical sensors that have been demonstrated

FIGURE 7 | (A) Measured elastic-light-scattering images of CROW upon the buffer solution and solutions X, Y and Z at λ p1 . Dashed-line and dotted-line
buffer solution and the three blind-test solutions X, Y, and Z at an arbitrarily set boxes indicate ρ p and ρ s , respectively. (D–G) Zoom-in view of the library to
probe wavelength λ p1 . The white-line box indicates the integration window for extract Δn. (D) Upon the buffer solution. (E) Upon solution X. (F) Upon solution
pixelization. (B) Normalized pixelized patterns upon the buffer solution and Y. (G) Upon solution Z. Insets (i)–(viii): Mapping of ρ p and ρ s upon the buffer
solutions X, Y, and Z at λ p1 . (C) Calculated correlation coefficients upon the solution and solutions X, Y, and Z.
in recent years, including our previous work (Wang et al., 2014), De Vos et al., calibrating the spectral sensitivity of a microring sen-
as summarized in Table 2. All of the work including this work sor of Q ~ 20,000 involved measuring the microring transmission
have attained a detection limit of 10−7 ~ 10−4 RIU. Two of the spectrum three times for each of the four given NaCl solutions
microcavity-based sensors (Ghasemi et al., 2013; Doolin et al., with different concentrations (De Vos et al., 2007). In contrast,
2015) and three of the MZI-based sensors (Duval et al., 2013; our library preparation involves scanning laser wavelength across
Misiakos et al., 2014; Dante et al., 2015) operate on the SiN-based the CROW transmission band, recording the pixelized patterns at
platform in the visible wavelengths. each wavelength step corresponding to the refractive index inter-
Most of the reported microcavity-based sensors in the liter- val Δni and deriving the corresponding correlation coefficients
ature (except Ghasemi et al., 2013; Doolin et al., 2015) operate with the eigenstate patterns. The pattern recording and additional
in the telecommunication wavelengths (1.3/1.55 μm) and require computation of the correlation coefficients render our library
a wavelength-tunable laser and a non-silicon photodetector. preparation more reliable and tolerant to the equipment noises
Whereas, our CROW sensor operating in the visible wavelengths that are common to all pixels compared with recording single
only requires in principle a fixed-wavelength visible laser diode intensity values multiple times.
and a silicon CCD/CMOS camera after the library preparation. A major issue requiring further developments is the signifi-
In terms of the sensor calibration, the main difference between cant variation of sensitivity values upon different probe wave-
our library preparation and the conventional calibration process lengths. We can modify the CROW design in order to attain
for a microcavity-based sensor is the recording of the pixelized a more uniform sensitivity (see Supplementary Material S10).
patterns instead of single intensity values. A typical calibration Our modeling results suggest that an imperfect CROW with a
for a conventional microcavity-based sensor [e.g., De Vos et al. reduced cavity size along with an enhanced inter-cavity coupling
(2007) and Iqbal et al. (2010)] involves scanning laser wavelength coefficient offers a more uniform sensitivity. Upon a small cavity
across a narrow transmission band. As an example, in the work of radius R = 10 μm and a strong inter-cavity coupling coefficient

FIGURE 8 | (A) Measured elastic-light-scattering images of CROW upon λ p2 . (C) Calculated correlation coefficients upon the buffer solution and
the buffer solution and the three blind-test solutions X, Y, and Z at a solutions X, Y, and Z at λ p2 . Dashed-line and dotted-line boxes indicate
specifically chosen probe wavelength λ p2 at eigenstate I. The white-line ρ p and ρ s , respectively. (D–G) Zoom-in view of the library to extract Δn.
box indicates the integration window for pixelization. (B) Normalized (D) Upon the buffer solution. (E) Upon solution X. (F) Upon solution Y.
pixelized patterns upon the buffer solution and solutions X, Y, and Z at (G) Upon solution Z.
FIGURE 9 | (A) Measured elastic-light-scattering images of CROW upon Y and Z at λ p3 . (C) Calculated correlation coefficients upon the buffer
the buffer solution and the three blind-test solutions X, Y, and Z at a solution and solutions X, Y, and Z at λ p3 . Dashed-line and dotted-line
specifically chosen probe wavelength λ p3 near eigenstate VII. The boxes indicate ρ p and ρ s , respectively. (D–F) Zoom-in view of the library
white-line box indicates the integration window for pixelization. to extract Δn. (D) Upon the buffer solution. (E) Upon solution Y. (F) Upon
(B) Normalized pixelized patterns upon the buffer solution and solutions X, solution Z.

TABLE 2 | Summary of silicon- and silicon-nitride-based on-chip optical biochemical sensors.
Device config. Reference Material Operational Footprint Q-factor Sensitivity Detected Detection
platform wavelength (nm) (μm2 ) (RIU−1 ) Δn (RIU) limit (RIU)
MZI Densmore et al. (2008) SOI ~1550 ~40,000 N/A 920 π rad ~7 × 10−4 ~1 × 10−5
Duval et al. (2013) Si3 N4 658 ~108 N/A 4950 π rad ~3 × 10−4 ~2 × 10−7
Misiakos et al. (2014) SiN ~600–900 ~107 N/A 581 rad ~4 × 10−5 ~1 × 10−5
Dante et al. (2015) Si3 N4 660 ~60,000 N/A 6000 π rad ~2 × 10−4 ~4 × 10−7
Microdisk Wang et al. (2013) SOI ~1550 ~3 ~100 130 nm ~9 × 10−3 ~8 × 10−4
Doolin et al. (2015) Si3 N4 ~770 ~900 10000 200 nm ~4 × 10−4 ~10− 6
Microring with Barrios et al. (2007) Si3 N4 ~1300 ~20,000 1800 212 nm ~10−3 ~2 × 10−4
slot-waveguide Claes et al. (2009) SOI ~1550 ~240 ~450 298 nm ~4 × 10−3 ~4.2 × 10−5
Carlborg et al. (2010) Si3 N4 ~1300 ~20,000 – 248 nm ~3 × 10−4 ~5 × 10−6
Microring De Vos et al. (2007) SOI ~1550 ~110 20,000 70 nm ~9 × 10−4 ~10−5
Iqbal et al. (2010) SOI ~ 1550 ~900 43,000 163 nm ~10−6 –
Ghasemi et al. (2013) SiN ~656 ~400 – 48 nm – –
Liu et al. (2014) SOI ~1550 ~1600 15000 6000 rad ~4 × 10−4 ~2.5 × 10−6
Eight-microring Wang et al. (2014) SOI ~1550 ~1716 N/A ~199 ~1.5 × 10−4 2 × 10−7 ~ 9 × 10−4
CROW in the (This work) SiN ~680 ~14080 N/A ~281 ± 271 ~1.3 × 10−4 2 × 10−8 ~ 1 × 10−4
spatial domain
κ ~ 0.9, we obtain for an imperfect eight-microring CROW a ~281 ± 271 RIU−1 and a NEDL of 2 × 10−8 ~ 1 × 10−4 RIU. Our
modeled sensitivity of ~384 ± 153 RIU−1 , with an improved non- blind sensing tests using NaCl solutions showed a detection
uniformity ratio of ~0.40 compared to the modeled ratio of of ~1.26 × 10−4 RIU. Therefore, we have shown that such a
~0.52 following our experimental device parameters. Assuming chip-scale, microresonator-based SiN CROW sensor operating
a practical sensitivity of ~100 RIU−1 , the width of the mod- in the visible wavelengths is promising as a potentially high-
eled probe wavelength window with a sensitivity >100 RIU−1 performance, portable, and low-cost optical biochemical sensor
is 2.2 nm, which is much improved compared to the modeled for applications such as point-of-care biochemical analyses and
width of 1.1 nm following the experimental device parameters. If self-monitoring of personal healthcare using smartphones.
a higher practical sensitivity of 300 RIU−1 is desired, the modeled
probe wavelength window width with a sensitivity >300 RIU−1 Acknowledgments
is ~1.56 nm, which is still sufficiently wide for practical appli-
cations. Based on our current imperfect CROW model, we can This work is supported by grants from the Research Grants
further design the CROW with tailored non-uniform parameters Council of the Hong Kong Special Administrative Region, China
to optimize the sensitivity and sensitivity variation. (Project No. 618010, 617612, and 617913). The authors acknowl-
In summary, we demonstrated a SiN CROW-based sensing edge the HKUST Nanoelectronics Fabrication Facility for fabri-
scheme in the spatial domain in the visible wavelengths. Given cating the optofluidic chip.
a calibrated CROW sensor, this sensing scheme in principle
only requires a low-power, fixed-wavelength laser source in the Supplementary Material
visible wavelengths and a silicon CCD or CMOS camera to
image the elastic-light-scattering patterns in the far field. Our The Supplementary Material for this article can be found
proof-of-concept experiment using an eight-microring CROW online at http://www.frontiersin.org/Journal/10.3389/fmats.2015.
on the SiN-on-silica platform showed an average sensitivity of 00034/abstract
References Dante, S., Duval, D., Fariña, D., González-Guerrero, A. B., and Lechuga, L. M.
(2015). Linear readout of integrated interferometric biosensors using a peri-
Barrios, C. A., Gylfason, K. B., Sánchez, B., Griol, A., Sohlström, H., Holgado, odic wavelength modulation. Laser Photon. Rev. 9, 248–255. doi:10.1002/lpor.
M., et al. (2007). Slot-waveguide biochemical sensor. Opt. Lett. 32, 3080–3082. 201400216
doi:10.1364/ol.32.003080 De Vos, K., Bartolozzi, I., Schacht, E., Bienstman, P., and Baets, R. (2007). Silicon-
Carlborg, C. F., Gylfason, K. B., Kazmierczak, A., Dortu, F., Polo, M. B., Catala, on-insulator microring resonator for sensitive and label-free biosensing. Opt.
A. M., et al. (2010). A packaged optical slot-waveguide ring resonator sensor Express 15, 7610–7615. doi:10.1364/oe.15.007610
array for multiplex label-free assays in labs-on-chips. Lab. Chip 10, 281–290. Densmore, A., Xu, D.-X., Janz, S., Waldron, P., Mischki, T., Lopinski, G.,
doi:10.1039/b914183a et al. (2008). Spiral-path high-sensitivity silicon photonic wire molecular sensor
Ciminelli, C., Campanella, C. M., Dell’olio, F., Campanella, C. E., and Armenise, with temperature-independent response. Opt. Lett. 33, 596–598. doi:10.1364/ol.
M. N. (2013). Label-free optical resonant sensors for biochemical applications. 33.000596
Prog. Quant. Electron. 37, 51–107. doi:10.1016/j.pquantelec.2013.02.001 Doolin, C., Doolin, P., Lewis, B. C., and Davis, J. P. (2015). Refractometric sensing
Claes, T., Molera, J. G., De Vos, K., Schachtb, E., Baets, R., and Bienstman, P. of Li salt with visible-light Si3N4 microdisk resonators. Appl. Phys. Lett. 106,
(2009). Label-free biosensing with a slot-waveguide-based ring resonator 081104. doi:10.1063/1.4913618
in silicon on insulator. Photonics J. IEEE 1, 197–204. doi:10.1109/jphot.2009. Duval, D., Osmond, J., Dante, S., Dominguez, C., and Lechuga, L. M. (2013). Grat-
2031596 ing couplers integrated on Mach-Zehnder interferometric biosensors operating

in the visible range. Photonics J. IEEE 5, 3700108. doi:10.1109/jphot.2013. index and bio-chemical sensor. Opt. Express 22, 26803–26813. doi:10.1364/OE.
2251873 22.026803
Estevez, M.-C., Alvarez, M., and Lechuga, L. M. (2012). Integrated optical devices Sedlmeir, F., Zeltner, R., Leuchs, G., and Schwefel, H. G. (2014). High-Q MgF2
for lab-on-a-chip biosensing applications. Laser Photon. Rev. 6, 463–487. doi:10. whispering gallery mode resonators for refractometric sensing in aqueous
1002/lpor.201100025 environment. Opt. Express 22, 30934–30942. doi:10.1364/oe.22.030934
Ghasemi, F., Eftekhar, A. A., Gottfried, D. S., Song, X., Cummings, R. D., and Adibi, Subramanian, A., Dhakal, A., Peyskens, F., Selvaraja, S., Baets, R., Neutens, P., et al.
A. (2013). “Self-referenced silicon nitride array microring biosensor for toxin (2013). Low-loss singlemode PECVD silicon nitride photonic wire waveguides
detection using glycans at visible wavelength,” in SPIE BiOS (San Francisco, CA: for 532–900 nm wavelength window fabricated within a CMOS pilot line.
International Society for Optics and Photonics). Photonics J. IEEE 5, 2202809. doi:10.1109/jphot.2013.2292698
Gorin, A., Jaouad, A., Grondin, E., Aimez, V., and Charette, P. (2008). Fabrication Vollmer, F., Arnold, S., and Keng, D. (2008). Single virus detection from the
of silicon nitride waveguides for visible-light using PECVD: a study of the reactive shift of a whispering-gallery mode. Proc. Natl. Acad. Sci. U.S.A. 105,
effect of plasma frequency on optical properties. Opt. Express 16, 13509–13516. 20701–20704. doi:10.1073/pnas.0808988106
doi:10.1364/oe.16.013509 Wang, J., Yao, Z., Lei, T., and Poon, A. W. (2014). Silicon coupled-resonator
Halir, R., Vivien, L., Le Roux, X., Xu, D.-X., and Cheben, P. (2013). Direct and optical-waveguide-based biosensors using light-scattering pattern recognition
sensitive phase readout for integrated waveguide sensors. Photonics J. IEEE 5, with pixelized mode-field-intensity distributions. Sci. Rep. 4, 07528. doi:10.1038/
6800906. doi:10.1109/jphot.2013.2276747 srep07528
Iqbal, M., Gleeson, M. A., Spaugh, B., Tybor, F., Gunn, W. G., Hochberg, M., et al. Wang, X., Guan, X., Huang, Q., Zheng, J., Shi, Y., and Dai, D. (2013). Suspended
(2010). Label-free biosensor arrays based on silicon ring resonators and high- ultra-small disk resonator on silicon for optical sensing. Opt. Lett. 38, 5405–5408.
speed optical scanning instrumentation. IEEE J. Sel. Top. Quantum Electron. 16, doi:10.1364/ol.38.005405
654–661. doi:10.1109/jstqe.2009.2032510 Xu, J., Suarez, D., and Gottfried, D. S. (2007). Detection of avian influenza virus
Kozma, P., Hamori, A., Cottier, K., Kurunczi, S., and Horvath, R. (2009). Grating using an interferometric biosensor. Anal. Bioanal. Chem. 389, 1193–1199.
coupled interferometry for optical sensing. Appl. Phys. B 97, 5–8. doi:10.1007/ doi:10.1007/s00216-007-1525-3
s00340-009-3719-1 Ymeti, A., Greve, J., Lambeck, P. V., Wink, T., Van Hövell, S. W., Beumer, T. A., et al.
Lakshminarayanan, V., Zelek, J., and Mcbride, A. (2015). “Smartphone Science” in (2007). Fast, ultrasensitive virus detection using a Young interferometer sensor.
eye care and medicine. Opt. Photonics News 26, 44–51. doi:10.1364/opn.26.1. Nano Lett. 7, 394–397. doi:10.1021/nl062595n
000044
Conflict of Interest Statement: The authors declare that the research was con-
Lei, T., and Poon, A. W. (2011). Modeling of coupled-resonator optical
waveguide (CROW) based refractive index sensors using pixelized spatial detec-
tion at a single wavelength. Opt. Express 19, 22227–22241. doi:10.1364/oe.19.
022227 Copyright © 2015 Wang, Yao and Poon. This is an open-access article distributed
Liu, J., Zhou, X., Qiao, Z., Zhang, J., Zhang, C., Xiang, T., et al. (2014). Inte- under the terms of the Creative Commons Attribution License (CC BY). The use, dis-
grated optical chemical sensor based on an SOI ring resonator using phase- tribution or reproduction in other forums is permitted, provided the original author(s)
interrogation. Photonics J. IEEE 6, 1–7. doi:10.1109/jphot.2014.2352973 or licensor are credited and that the original publication in this journal is cited, in
Misiakos, K., Raptis, I., Makarona, E., Botsialas, A., Salapatas, A., Oikonomou, P., accordance with accepted academic practice. No use, distribution or reproduction is
et al. (2014). All-silicon monolithic Mach-Zehnder interferometer as a refractive permitted which does not comply with these terms.

MATERIALS
doi: 10.3389/fmats.2015.00028
High-throughput multiple dies-to-wafer bonding

technology and III/V-on-Si hybrid lasers for heterogeneous
integration of optoelectronic integrated circuits
Xianshu Luo 1 , Yulian Cao 2 , Junfeng Song 1 , Xiaonan Hu 2 , Yuanbing Cheng 2 , Chengming Li 2 ,
Chongyang Liu 2 , Tsung-Yang Liow 1 , Mingbin Yu 1 , Hong Wang 2 , Qi Jie Wang 2 and Patrick Guo-Qiang Lo 1 *
1
Institute of Microelectronics, Agency for Science, Technology and Research (A*STAR), Singapore, Singapore
2
Photonics Center of Excellence (OPTIMUS), School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore, Singapore
Edited by: Integrated optical light source on silicon is one of the key building blocks for optical intercon-
Laurent Vivien, Université Paris-Sud,
nect technology. Great research efforts have been devoting worldwide to explore various
France
approaches to integrate optical light source onto the silicon substrate. The achievements
Reviewed by:
Junichi Fujikata, Photonics Electronics so far include the successful demonstration of III/V-on-Si hybrid lasers through III/V gain
Technology Research Association, material to silicon wafer bonding technology. However, for potential large-scale integration,
Japan leveraging on mature silicon complementary metal oxide semiconductor (CMOS) fabrica-
Yasuhiko Ishikawa, The University of
tion technology and infrastructure, more effective bonding scheme with high bonding yield
Tokyo, Japan
is in great demand considering manufacturing needs. In this paper, we propose and demon-
*Correspondence:
Patrick Guo-Qiang Lo, Institute of strate a high-throughput multiple dies-to-wafer (D2W) bonding technology, which is then
Microelectronics, Agency for Science, applied for the demonstration of hybrid silicon lasers. By temporarily bonding III/V dies to a
Technology and Research (A*STAR), handle silicon wafer for simultaneous batch processing, it is expected to bond unlimited III/V
Singapore Science Park II, 11 Science
dies to silicon device wafer with high yield. As proof-of-concept, more than 100 III/V dies
Park Road, 117685 Singapore
e-mail: logq@ime.a-star.edu.sg bonding to 200 mm silicon wafer is demonstrated. The high performance of the bonding
interface is examined with various characterization techniques. Repeatable demonstrations
of 16-III/V die bonding to pre-patterned 200 mm silicon wafers have been performed for
various hybrid silicon lasers, in which device library including Fabry–Perot (FP) laser, lateral-
coupled distributed-feedback laser with side wall grating, and mode-locked laser (MLL).
From these results, the presented multiple D2W bonding technology can be a key enabler
toward the large-scale heterogeneous integration of optoelectronic integrated circuits.
Keywords: silicon photonics, hybrid lasers, heterogeneous integration, die-to-wafer bonding, optoelectronic
integrated circuits
INTRODUCTION 2004; Lipson, 2005; Guillot and Pavesi, 2006; Jalali and Fathpour,
In the future generation of datacom and computercom, which 2006; Soref, 2006; Poon et al., 2009a,b; Vivien and Pavesi, 2013;
demand ever higher bandwidth and lower power, the conventional Xu et al., 2014), which utilizes low-cost silicon material along with
electrical interconnection routing the electronic signals becomes leveraging on the advancement of silicon complementary metal
bandwidth-limited along with prohibitively high power consump- oxide semiconductor (CMOS) process, integration, and mature
tion (Beausoleil et al., 2008). One solution to the challenge is the infrastructure. Envisioned by Soref and Lorenzo (1985), silicon
optical interconnect technology (Goodman et al., 1984; Miller, photonics has emerged and progressed steadily. Especially in the
2000, 2009; Ohashi et al., 2009), in which high bandwidth optical past decade, we have been witnessing rapid growth in research and
signals are routed by low-loss optical fiber and waveguides. In con- development activities along with product development efforts
trast to the electrical interconnection (i.e., the copper wire), optical exploiting silicon photonics technology for the optical intercon-
interconnect has many merits, e.g., high speed, low crosstalk, nect (Pavesi and Lockwood, 2004; Reed and Knights, 2004; Lipson,
immunity to electromagnetic interference, low overall power con- 2005; Guillot and Pavesi, 2006; Jalali and Fathpour, 2006; Soref,
sumption (Alduino and Paniccia, 2007). Most importantly, with 2006; Fedeli et al., 2008; Poon et al., 2009a,b; Michel et al., 2010;
the up scaling potential, optical interconnect is expected to provide Reed et al., 2010; Feng et al., 2012; Liow et al., 2013; Vivien and
much higher transmission capacity and longer signal transmission Pavesi, 2013; Dong et al., 2014a,b; Lim et al., 2014; Xu et al., 2014).
distance than the electrical interconnect. For instance, to minimize the small core silicon waveguide prop-
Although it was proposed initially 30 years ago (Goodman et al., agation losses, considerable research work has been devoted to
1984), there was no significant development progress with solid minimize the waveguide sidewall roughness by using the deep
demonstrations of optical interconnect for very-large-scale inte- ultra-violet (DUV) photolithography and optimized patterning
gration (VLSI). The situation has changed since the concept of technique (Dumon et al., 2004; Bogaerts et al., 2005) and sidewall
silicon photonics (Pavesi and Lockwood, 2004; Reed and Knights, smoothing technique [e.g., double thermal oxidations (Sparacin
www.frontiersin.org April 2015 | Volume 2 | Article 28 | 45

Luo et al. III/V-on-Si bonding and hybrid lasers
et al., 2005; Xia et al., 2006)]. Indeed, submicrometer-scale silicon technology, which comprised a semiconductor optical ampli-
wire waveguides have shown a propagation loss of 2 dB/cm and fier (SOA) chip and a silicon photonic chip, and were hybrid-
less (Xia et al., 2006). Furthermore, owing to the enabling CMOS integrated by using passive alignment technology. An SiON mode-
fabrication technologies, we have seen the establishment and uti- size converter was adopted between the silicon waveguide and III/V
lization of a myriad of essential silicon photonic passive and active SOA for low coupling loss. Later on, silicon photonic-based optical
components including optical filters (Xiao et al., 2007; Zhou and interconnects were also demonstrated by integrating lasers, sili-
Poon, 2007; Guha et al., 2010; Fang et al., 2012), optical switches con modulators, and Ge photodetectors on single silicon substrate
(Poon et al., 2009a,b; Van Campenhout et al., 2009; Luo et al., (Urino et al., 2011). While such demonstrations have shown the
2012; Song et al., 2013), low-power-consuming modulators with advantage in principal of being capable to integrate various build-
up to 50 Gb/s-speed operation (Dong et al., 2009; Reed et al., 2010; ing blocks together for optical interconnection, the main issue
Tu et al., 2013, 2014), and Ge-on-Si photodetectors with band- is the complicated fabrication process. It typically requires pre-
width larger than 40 GHz (Michel et al., 2010; Liow et al., 2013). cise alignment between the SOA and the silicon waveguide, even
It is these demonstrated silicon photonic devices and technolo- with assistance of the mode-size converter. Considering the III/V
gies that make ultimate optical interconnection a viable solution gain region of <200 nm in thickness, for instance, it became a
to address the distance/bandwidth/cost and power-consumption difficult challenge for alignment with acceptable coupling loss.
challenges. To this end, silicon photonics provides nearly all key Such complicated fabrication process is a potential show stop-
building blocks for optical interconnection. Furthermore, the per for future massive production demanding high yield, thus
CMOS-compatible fabrication processes make it possible to inte- significantly increases the product cost.
grate both electronics and photonics either through monolithic Heterogeneous integration of III/V gain materials on silicon
or heterogeneous approach. Such significant progress has led the through wafer bonding technology is another major directional
optical interconnect to become a much more practical technology. strategy for silicon lasers. UMR-CNRS and LETI initiated the
However, silicon-based on-chip optical light source, which is research work of III/V laser on silicon wafers for photonic inte-
one of the key components for the light generation for carrying gration by using wafer bonding technology. In 2001, they demon-
information, has been the missing piece for optical interconnect. strated the first InP-based microdisk laser integrated on a silicon
This is mainly because silicon is transparent in the telecommunica- wafer through SiO2 –SiO2 molecular bonding (Seassal et al., 2001).
tion wavelengths (i.e., 1310 and 1550 nm wavelengths) due to the Although this work did not show complete integration of III/V
indirect bandgap, which prohibits efficient light emission from sil- optoelectronics with silicon photonics waveguide structures, it
icon. Thus, to solve the challenge, numerous research efforts have showed the potential of such wafer bonding technology for future
been devoted to explore various technologies for light source on heterogeneous integrated optoelectronic circuit. Following such
silicon chips. demonstration, Hattori et al. (2006) demonstrated an integra-
tion scheme of III/V microdisk laser with silicon waveguide in
REVIEW ON RESEARCH FOR LASERS ON SILICON 2006. By aligning the microdisk laser atop silicon waveguide, the
Historically, researchers worldwide have devoted many research laser emissions can be vertically coupled into the underneath sili-
efforts by exploring various possibilities for the development of con waveguide with 35% coupling efficiency. Such demonstration
lasers on silicon, which mainly focused in the following directions: showed the capability of the hybrid photonic integration of III/V
laser with silicon waveguide for photonic links application.
(1) silicon material engineering by introducing emissive centers The so-called hybrid silicon laser was proposed and first
to assist the efficient light emission (Pavesi et al., 2000; Han demonstrated by Park et al. (2005) with optical injection. In
et al., 2001; Rotem et al., 2007a,b; Shainline and Xu, 2007), this work, the III/V wafer with AlGaInAs quantum well struc-
(2) strained Ge (Liu et al., 2007, 2009, 2010; Cheng et al., 2009; ture is directly bonded to pre-patterned silicon wafer using low-
Sun et al., 2009b,c; Camacho-Aguilera et al., 2012), temperature oxygen plasma-assisted wafer bonding. The laser
(3) silicon Raman laser (Boyraz and Jalali, 2004, 2005; Rong et al., cavity was defined by endface-polished silicon waveguide struc-
2005a,b, 2007), and ture, while the III/V provides the optical gain. As the III/V
(4) heterogeneous integration of III/V gain materials through optoelectronic structures are fabricated after the wafer bonding
packaging (Chu et al., 2009; Fujioka et al., 2010; Urino et al., with best precise, only possibly achieved via lithographic process,
2011) or wafer bonding (Park et al., 2005; Fang et al., 2007a,b; alignment to the silicon device layer, thus there is no stringent
Liang et al., 2009a,b, 2010; Stanković et al., 2010; Grenouillet alignment requirement to the bonding process, which signifi-
et al., 2012). cantly simplifies the fabrication process and makes the possibility
of wafer-level-oriented manufacturing ability. Subsequently, Fang
Here, we will limit our review to the heterogeneous inte- et al. (2006) demonstrated an electrically pumped AlGaInAs-
grated silicon lasers. With regard to the silicon laser through silicon evanescent laser with continuous-wave (CW) operation
heterogeneous integration of III/V gain materials on silicon, there in 2006. Subsequently, various hybrid lasers with different struc-
are two major types of integration strategies, namely the pack- tures and also enhanced laser performances are demonstrated by
aging scheme and the bonding scheme. Research groups from various research groups using molecular wafer bonding technol-
Japan devoted many efforts for the development of silicon lasers ogy, including Fabry–Pérot lasers (FP) (Ben Bakir et al., 2011;
using packaging methods. Chu et al. (2009) demonstrated the Dong et al., 2013), racetrack lasers (Fang et al., 2007a,b), dis-
first wavelength-tunable-laser fabricated with silicon photonic tributed Bragg reflector (DBR) lasers (Fang et al., 2008a,b,c),
Frontiers in Materials | Optics and Photonics April 2015 | Volume 2 | Article 28 | 46

distributed-feedback (DFB) lasers (Fang et al., 2008a,b,c), micror- From these analyses, it shows that among various approaches,
ing lasers (Liang et al., 2009a,b, 2012), wavelength tunable lasers the hybrid silicon laser through wafer bonding technology can
(Keyvaninia et al., 2013a,b,c), multiple-wavelength lasers (Van be considered as the most successful and promising one for sili-
Campenhout et al., 2008; Kurczveil et al., 2011), and mode-locked con photonic heterogeneous integration circuits due to the ever-
lasers (MLL) (Fang et al., 2008a,b,c). demonstrated advanced performances and the fabrication process
Besides such direct bonding method, wafer bonding can also compatibility with silicon photonics. Table 1 summarizes some of
be realized through an adhesive material as the bonding interlayer. the representative demonstrations of hybrid silicon lasers through
Among all kinds of adhesive bonding materials, divinylsiloxane- wafer bonding technology.
bisbenzocyclobutene (DVS-BCB or BCB) is the most popular one
for hybrid silicon lasers due to the merits such as the high bond- WAFER BONDING TECHNOLOGIES FOR ON-CHIP SILICON LASERS
ing strength and the sustainability in the subsequent III/V process. In general, there are two mainstreams of wafer bonding methods
IMEC has used BCB-assisted adhesive bonding method for hetero- applying to heterogeneous integrated silicon photonics, namely
geneous integration (Roelkens et al., 2005). In 2006, Roelkens et al. the molecular bonding through interfacial bonds, and the adhe-
(2006) demonstrated the first electrically injected InP/InGaAsP sive bonding assisted with another adhesive material as bonding
laser integrated on silicon waveguide circuit using BCB-assisted interface such as polymer or metal. Such wafer bonding tech-
adhesive bonding technology. Similar to Seassal et al. (2001), the nology is a mature process, which is widely applicable for SOI
optical laser is purely made with III/V layer with the laser facets wafer fabrication, MEMS technology, and optoelectronic device
being defined by dry etching. With optimized mode-size converter, fabrication. As a lot of review papers already exist (Lasky, 1986;
the optical light can be vertically coupled down to the underneath Maszara, 1991; Tong and Goesele, 1999; Alexe and Gösele, 2004;
silicon waveguide with high efficiency. By designing hybrid mode Christiansen et al., 2006), we thus only focus the discussion on
waveguide comprising silicon waveguide and III/V gain medium, the application of hybrid silicon lasers. According to the existing
they also demonstrated a hybrid FP laser (Stanković et al., 2011) demonstrations, we further summarize here the major bonding
and a DFB laser (Stanković et al., 2012; Keyvaninia et al., 2013a,b,c), technologies as below:
and multiple-wavelength laser (Keyvaninia et al., 2013a,b,c) using
such adhesive bonding technology. (1) wafer-to-wafer (W2W) molecular bonding,
Apart from BCB, some kinds of metal can also be adopted as (2) die-to-wafer (D2W) molecular bonding,
the bonding interlayer for adhesive bonding. AuGeNi is one of the (3) BCB-assisted D2W adhesive bonding, and
most popular metals for metal bonding as it not only functions (4) metal-assisted adhesive bonding.
as a bonding media but can also be used for the Ohmic contact
to the InGaAsP structure. Tanabe et al. (2010) demonstrated a The W2W molecular bonding for hybrid silicon lasers is mainly
InAs/GaAs quantum-dot laser on Si substrate by metal-assisted driven by the UCSB group. Through such plasma-activated low-
wafer bonding with room temperature operation at 1.3 µm wave- temperature W2W molecular bonding (Pasquariello and Hjort,
length. Meanwhile, Hong et al. (2010) also demonstrated an FP 2002), they, together with their collaborators, have demonstrated
laser through selective-area metal bonding using AuGeNi. The sil- various hybrid silicon lasers, starting from the first-hybrid FP
icon waveguide in such demonstration is with 5 µm and 800 nm laser (Fang et al., 2006), followed by racetrack-shaped laser (Fang
thickness. The demonstrated FP laser is with threshold current et al., 2007a,b), DBR lasers (Fang et al., 2008a,b,c), DFB lasers
density of 1.7 kA/cm and a maximum output power of 3 mW. (Fang et al., 2008a,b,c), MLL (Fang et al., 2008a,b,c), and multi-
However, the drawback of the AuGeNi-assisted bonding is the Au wavelength arbitrary waveform generation (AWG) laser (Kurczveil
contamination. Thus, Tatsumi et al. (2012) further demonstrated et al., 2011). However, for the conventional III/V-to-Si W2W bond-
an Au-free metal-assisted wafer bonding for lasers on silicon chip. ing without thick oxide interlayer, the generated gas by-products
Besides, Creazzo et al. (2013) also demonstrated another type of H2 and H2 O are easily trapped inside the bonding interface and
of silicon laser by using metal-assisted bonding of III/V epitax- form the interfacial voids, which subsequently affect the bonding
ial material directly onto the silicon substrate. The demonstrated quality. In order to effectively remove such trapped gases, some
that silicon laser had a threshold of ~50 mA and maximum optical proper outgas channels are designed, such as in-plane outgassing
power of ~8 mW. The benefit of such metal-assisted bonding is the channels (IPOC) (Kissinger and Kissinger, 1993) or vertical out-
advantage of effective thermal dissipation, which shows a thermal gassing channels (VOC) (Liang and Bowers, 2008). IPOS is formed
resistant of only 21°C/W. by etching some lateral channels extended to the chip edges, so
Beyond these two major heterogeneous integration schemes, that the by-product gases can be directed to outside the bonding
there are also other methods for III/V-on-Si lasers, including direct interface to the chip edge during post bonding annealing. How-
III/V epitaxy on silicon substrate (Liu et al., 2011; Lee et al., 2012) ever, for some close-loop structures, such as microrings, there is
and III/V epitaxial layer transfer-printing to silicon wafers (Justice no way to design such IPOS. In order to solve such issue, VOC
et al., 2012; Yang et al., 2012). However, while the direct epitaxy is proposed by etching some array of holes down to the BOX
method faces major challenges of high-density dislocations due to layer. The generated by-product gases can migrate to the closest
the lattice mismatch between III/V material and silicon after many VOC and are absorbed by SOI BOX. As both IPOS and VOC can
years of research, the transfer-printing method for hybrid silicon be formed during the waveguide etching, there is no particular
laser needs further demonstrations to show the repeatability and design requirement from the fabrication point of view. However,
reliability. as the formation of such outgas channels affects the silicon layer

Table 1 | Representative demonstrations of hybrid silicon lasers through wafer bonding technology.
Laser types Bonding type Waveguide scheme Performances Reference
λ (nm) T (°C) I th (mA) P out (mW) SE (mW/mA) Z t (°C/W)
Fabry–Perot Molecular Hybrid WG (75 vs. 3% mode 1577 CW @ 15 65 1.8 0.013a 40 Fang et al.
laser bonding confinement within Si WG (2006)
and QW)
DBR laser Molecular Hybrid waveguide with 1569 CW @ 15 65 11 0.088a 40 Fang et al.
bonding inverse taper (66 vs. 4.4% (2008a,b,c)
mode confinement within Si
WG and QW)
D2W molecular Hybrid waveguide with 1570 Pulse @ 20 100 7.2 0.021a – Ben Bakir
bonding with adiabatic mode transformer et al. (2011)
oxide interlayer
W2W molecular Hybrid waveguide with 1553 CW @ 20 40 4 0.025a – Keyvaninia
bonding with inverse taper, thermal tunable et al.
oxide interlayer microring for wavelength (2013a,b,c)
tuning
Metal-assisted III/V gain material 1562 CW @ 20 41 8 0.038 21 Creazzo
D2W bonding butt-coupling with Si et al. (2013)
waveguide through a
waveguide coupler
DFB laser Molecular Hybrid waveguide with 1600 CW @ 10 25 5.4 0.072a 132 Fang et al.
bonding inverse taper (59.2 vs. 5.2% (2008a,b,c)
mode confinement within Si
WG and QW)
D2D BCB Hybrid waveguide (70 vs. 3% 1308 CW @ 20 20 2.1 0.026 – Stanković
adhesive mode confinement within Si et al. (2011)
bonding WG and QW)
Selective-area Hybrid WG (94% mode 1554 Pulse @ RT 35 3 0.05 – Hong et al.
metal bonding confinement within Si) (2010)
Microdisk D2W molecular InP microdisk laser light 1590 CW @ 20 0.9 0.012 0.008 – Van Camp-
laser bonding vertically coupling to Si WG enhout
et al. (2008)
Microring Molecular Racetrack microring, hybrid 1590 CW @ 15 175 29 0.089 – Fang et al.
laser bonding waveguide (2007a,b)
Molecular Hybrid microring with side – CW @ 10 7.5 2.5 0.2 – Liang et al.
bonding coupled Si WG (2011)
a
Data are extracted from the power–current curves.
pattern density, which will finally affect the bonding strength, it silicon wafers were covered with oxide layer. Pick-and-place tech-
is desirable to take into account the design tradeoff between the nology was used in order to align the InP dies to specific spots
bonding strength and the gas removal effectiveness. in silicon wafer, as well as to supply mechanical force to the
Alternatively, plasma-assisted D2W molecular bonding has also dies through pick-and-place head. Using such D2W bonding,
been investigated for hybrid silicon lasers mainly by LETI. For they demonstrated electrically pumped microdisk lasers integrated
large-scale manufacturability for potential massive production, with a silicon waveguide circuit (Van Campenhout et al., 2007).
the key enabling capability is the multiple dies to wafer bond- However, in such D2W bonding, as the cleaning of the dies is per-
ing with high yield. Kostrzewa et al. (2006) first demonstrated formed ahead of the pick-and-place process, the bonding surface
a molecular bonding of multiple InP dies to a 200 mm silicon could be contaminated, and subsequently affecting the bonding
CMOS wafer with only 1 mm × 1 mm die size (Kostrzewa et al., quality and bonding yield. Furthermore, with the consideration
2006). For strong hydrophilic molecular bonding, both InP and of the pick-and-place time of 30 s/die, it takes approximately an

hour for bonding 100 dies. Such long bonding time through In Table 2, we summarize and compare these four different
pick-and-place process for individual die would cause the bond- bonding methods.
ing surface deactivation for the molecular bonding with plasma
activation. OUTLINE OF THE MANUSCRIPT
BCB-assisted D2W adhesive bonding can address such issue The rest of the submission is organized as follows. In the Section
with potential capability of bonding unlimited number of dies. “III/V-to-Si Wafer-to-Wafer (W2W) Bonding Technology,” we
Stanković et al. (2010) demonstrated such D2D adhesive bond- show a demonstration of the wafer-to-wafer bonding by using low-
ing technology using BCB. The BCB is first spin-coated on silicon temperature plasma-activated molecular bonding method with
wafer with controlled thickness of <100 nm in order to ensure oxide as the bonding interlayer. In the Section “High-Throughput
the vertical light coupling efficiency, followed with die attaching Multiple Dies-to-Wafer Bonding Technology,” we propose and
and subsequent curing at 240°C for 1 h in a nitrogen atmosphere show the demonstration of an alternative bonding technology
at 1000 mbar. With the assistant of BCB adhesion, the stringent that can perform high-throughput D2W bonding for potential
requirements of contamination-free and smooth bonding surfaces massive production of silicon hybrid lasers, which is based on
for molecular bonding are relieved significantly. Furthermore, a batch process to simultaneously bonding all the dies to the
there is, in principle, no limitation for multiple dies bonding by the silicon wafer. In the Section “Design of III/V-on-Si Lasers,” we
assistance of BCB adhesive layer (Keyvaninia et al., 2012). Through provide some design guidelines of hybrid silicon laser, includ-
such bonding method, various hybrid lasers, including FP laser ing the design of III/V multiple quantum wells (MQW) struc-
(Stanković et al., 2011), DFB laser (Stanković et al., 2012), micror- tures, the silicon waveguide thickness selection for hybrid laser,
ing and AWG integrated multi-wavelength DBR lasers (Keyvaninia and the design of the vertical coupling structures. The Section
et al., 2013a,b,c), and microdisk laser (Mechet et al., 2013) have “Demonstration of III/V-on-Si Hybrid Lasers” shows some hybrid
been demonstrated. However, although such adhesive bonding is silicon laser demonstrations using the bonded wafers from the
with good robustness and bonding strength, the thermal dissipa- proposed high-throughput D2W bonding, including FP laser,
tion could be a major problem due to the low thermal conductivity lateral-coupled distributed-feedback (LC-DFB) laser with side
of the BCB layer. Besides, robust polymer coating process ensuring wall grating, and MLL. The Section “Summary and Future Out-
the controllable BCB thickness is also very important. look” summarizes this paper and addresses some of the future
Apart from these three major bonding methods, metal-assisted challenges.
adhesive bonding (Hong et al., 2010; Tanabe et al., 2010; Creazzo
et al., 2013) is another one that can be used for hybrid laser integra- III/V-TO-Si WAFER-TO-WAFER BONDING TECHNOLOGY
tion. However, due to the potential metal contamination and the We have started the development of wafer bonding technology
non-compatibility with the subsequent fabrication process, such for hybrid silicon photonics integration in 2011. Considering the
as acid etching for substrate removal, the metal-assisted adhe- complete integration with existing silicon photonic integrated cir-
sive bonding method might not be an optimal choice for silicon cuit, which consisting various silicon passive waveguide devices,
heterogeneous optoelectronic integrated circuits. high-speed modulator, and photodetectors, and are normally
Table 2 | The major bonding technology for hybrid integration.
Bonding methods Process description Fabrication tolerance Manufacturing scalability Comments
W2W molecular O2 plasma-assisted Small tolerance of contamination-free, Difficult due to wafer size Low utilization of both
bonding direct bonding with smooth, and flat bonding surfaces mismatch III/V and silicon wafers
12 h annealing at 300°C Sensitive to the wafer
bowing
D2W molecular O2 plasma-assisted Small tolerance of contamination-free Difficult due to contamination Difficult to ensure high
bonding bonding with oxide and smooth bonding surfaces and surface deactivation during yield with large number
interlayer and 3 h pick-and-place process for large of dies bonding
annealing at 250°C number of dies bonding
BCB-assisted BCB adhesive bonding Large tolerance with low requirement Easy to be scalable with multiple Thermal dissipation
D2W adhesive with post curing of 1 h on the bonding surface. Yet it requires dies and large-sized wafers problem due to the low
bonding at 240°C controllable polymer coating regarding thermal conductivity of
the thickness and flatness the BCB layer
Metal-assisted Metal-assisted bonding Large tolerance with low requirement on Easy to be scalable with multiple Enhanced thermal
adhesive bonding with annealing the bonding surface. However, potential dies and large-sized wafers resistant due to the
metal contamination, and process metal utilization
incompatibility. Potential coupling
problem due to the metal absorption

with thick oxide cladding, we adopt the low-temperature plasma- The bonding process starts with the wafer cleaning, separately
activated molecular bonding method (Pasquariello and Hjort, for silicon wafer and III/V wafer. First, standard SPM clean for
2002) with oxide as the bonding interlayer. Furthermore, such 10 min is performed to the silicon wafer in order to remove any
thick cladding oxide also serves as the diffusion and absorption organic contaminants, followed with 5 min SC1 clean with mega-
medium for the bonding by-products gases, thus with enhanced sonic to remove any particle on the surface. The III/V wafer is sepa-
bonding quality and bonding yield. rately cleaned in the NH4 OH solution (NH4 OH:DI water = 1:15)
For the initial development, we deposit 1.1 µm PECVD oxide for 1 min. Second, O2 plasma activation in a RIE chamber is
on top of silicon wafers, followed with chemical mechanic pol- performed for both silicon wafer and III/V wafer, subsequently
ishing (CMP) to remove 100 nm oxide in order to ensure the followed with DI water rinse. These two bonded wafers are then
smooth bonding interface. For all the bonding process described physically contacted with each other immediately after drying and
hereafter, we will use the similar PECVD oxide as cladding fol- placed inside to the EVG 520 bonder for pre-bonding under N2
lowed with CMP to smooth the bonding surface. Thus, we char- for 3 min with 1000 N mechanic force. After that, post bonding
acterize and compare the oxide properties in terms of wafer- annealing at 300°C in vacuum is applied to the bonded pair for
level uniformity and surface roughness before and after CMP. 12 h. Figure 2A shows the optical image of a 50 mm InP wafer
Figures 1A,B show the wafer-level oxide thickness before and bonding to a 200 mm silicon wafer before unloading from the
after CMP. The non-uniformity is only ~1% after CMP, which bonder track.
suggests a very flat surface. Figures 1C,D show the AFM results The bonded wafers are first characterized by scanning acoustic
before and after CMP. As deposited, the surface is relatively microscope (SAM) using Sonix HS3000. Figure 2B shows the typi-
rough, with RMS of ~2.5 nm, while after CMP, the surface rough- cal CSAM image for the bonded wafer. We observe that larger than
ness is reduced significantly with RMS of ~0.4 nm, which is 98% of the 50 mm InP area is bonded to the silicon wafer, with
more suitable for wafer molecular bonding (Christiansen et al., only limited voids, which are attributed to the particles remain-
2006). ing on bonding surface. Besides, the bonding quality in the wafer
FIGURE 1 | The oxide thickness (A) before and (B) after CMP. The AFM results of the wafer surface roughness (C) before and (D) after CMP.

FIGURE 2 | (A) The bonded wafer before unloading from the EVG bonder. (B) The CSAM result. (C) The shear testing result. (D) The TEM results show the
high-quality bonding interface.
periphery is also not good enough, which is due to the ring-shaped increased device cost and waste, which in turn makes it ineffec-
imperfection of the InP wafer. tual to use the silicon photonics though it is of low cost. Second,
The whole wafer is then diced into 5 mm × 5 mm dies for shear the main stream silicon photonics has already adopted 200 mm
testing by using a Die Shear Tester (Dage Series 4000). Figure 2C silicon wafers. However, due to the brittleness of the InP wafers,
shows the extracted bonding strength, with maximum bonding it is very difficult to make large-sized wafers to match with sili-
strength of ~30 MPa in the wafer center region, and the averaged con wafers. Although the largest available III/V epitaxial wafer can
value of 15 MPa. We believe such bonding strength is high enough go with 150 mm, the commercially available largest-sized III/V
for any of the post optoelectronic fabrication process. Figure 2D epitaxial wafer is only 75 mm. Thus, such wafer size mismatch-
shows the TEM results of the bonded wafer, which indicate a ing definitely results in the insufficient utilization of the silicon
very tight bonding between InP and oxide, again suggesting a device wafer, which in turn increased the cost. Furthermore, InP
high-quality bonding. wafers with multiple quantum well structures are normally with
However, although such W2W bonding has been demonstrated high global stress, which induces the wafer bowing. Such stress-
with high quality, there are still existing big challenges, including: induced wafer bowing will easily trap the air between the bond-
ing interfaces with remained voids, thus reducing the bonding
(1) insufficient III/V wafer utilization, quality.
(2) insufficient silicon wafer utilization due to wafer size mis-
match, HIGH-THROUGHPUT MULTIPLE DIES-TO-WAFER BONDING
(3) III/V wafer global stress-induced bonding voids. TECHNOLOGY
Based on the aforementioned W2W bonding method, we pro-
First of all, for practical application of optical interconnec- pose an alternative proprietary high-throughput multiple D2W
tion, only very small portion of the silicon waveguide area needs bonding method, which is based on temporarily bonding III/V
to be bonded with III/V material for optoelectronic fabrication dies to a handle silicon wafer through pick-and-place process for
to form optical lasers. With whole III/V wafer bonding, most of simultaneous batch processing. Such high-throughput multiple
the III/V material will be subsequently etched away during post D2W bonding method is the key enabling technique for potential
optoelectronic fabrication. Giving such precious III/V wafers, the manufacturability of large-scale hybrid optoelectronic integrated
insufficient utilization of the III/V material results in significantly circuit (H-OEIC).

FIGURE 3 | Illustration of the key processing steps of the multiple D2W bonding technology.
Figure 3 shows the key process steps of the proposed multiple misalignment of only ±5 µm from the programmable reconfigu-
D2W bonding technology, which includes: ration by pick-and-place process. However, as the alignment of the
III/V devices to the silicon waveguide device is determined through
(a) the programmable reconfiguration of III/V dies onto a handle photolithography during optoelectronic fabrication process after
wafer via pick-and-place process, wafer bonding, such misalignment can easily be compensated by
(b) the D2W bonding through the notch alignment between the adopting relatively large-sized III/V dies.
two 800 wafers, after batch processing of wafer cleaning and Figure 4 schematically illustrates the detailed bonding process
plasma activation, and flow starting from the preparations of silicon and III/V wafer. For
(c) the dies releasing from the handle wafer and transferring to either blanket silicon wafer or patterned silicon wafer with pho-
silicon device wafer. tonic devices, the wafers are cladded with PECVD oxide, followed
with CMP process to smooth the bonding surface. As the hybrid
The most critical step here is the choice of the adhesion layer laser performance is largely dependent on the vertical coupling
for the temporary III/V dies bonding to the handle wafer, which efficiency, which is determined by the inter-layer oxide thickness,
includes the following two trade-off considerations. it is of very importance to control the oxide thickness by CMP
process.
(1) The adhesion should be strong enough to stick the III/V dies The preparation of the III/V dies includes the III/V wafer dicing
on the handle wafer without peeling off during the subsequent into certain sized dies, the preparation of the adhesion layer to the
III/V dies batch processing, including InGaAs cap layer wet handle wafer, and the programmable reconfiguration of the III/V
etching, pre-clean, wafer drying, and plasma activation, etc. dies temporary bonding to the handle wafer through pick-and-
(2) The adhesion should not be excessively strong so that the place process. Typically, the mechanical wafer dicing will result in
III/V dies can be successfully released and transferred to the edge roughness along the dicing lane, thus subsequently cause the
Si device wafer after pre-bonding. low quality bonding near the die periphery. Besides, such dicing
process may also introduce particles to the wafer surface cause
The programmable reconfiguration of the dies onto the han- contamination. Thus, a sacrificial InGaAs cap layer in order to
dle wafer is realized through pick-and-place process by pre- protect the III/V bonding surface is designed in our demonstra-
determining the position coordinates of each die with considering tion. The applied mechanic force by the pick-and-place head also
the wafer-level silicon device die distribution. Unlike the pick-and- needs to be well controlled in order to ensure the successful die
place process in flip-chip bonding, which directly bonds the dies releasing from the handle wafer after pre-bonding. Due to the
to the actual wafer (Kostrzewa et al., 2006), the pick-and-place in direct contact of the pick-and-place head, the die surface could be
our proposed method only helps to distribute the dies onto a han- contaminated. However, owing to the sacrificial InGaAs layer, the
dle wafer without flipping the chips. Thus, all the dies attached to bonding surface can be well protected without contamination or
the handle wafer can be simultaneously performed with different surface damage. We have checked and compared the surface con-
process steps for wafer bonding, such as InGaAs cap layer etching, dition of the III/V dies before and after the etching of the sacrificial
wafer clean, and plasma activation. InGaAs layer. We observe the particles on the chip surface after the
The D2W bonding alignment accuracy is mainly determined wafer dicing and pick-and-place process, which is with relatively
by the notch alignment, which is performed manually and induces high RMS of 0.198 nm. In comparison, after the etching of InGaAs
a relatively large misalignment of ±500 µm, compared to the cap layer, the surface roughness is improved with reduced RMS of

FIGURE 4 | The fabrication process flow of the multiple D2W bonding batch process and the silicon device wafer fabrication, and D2W bonding
technology. The process includes two different folds, i.e., the bonding through dies releasing from handle wafer and transferring to silicon device
wafers preparation including the III/V dies adhesion to handle wafer for wafer.
FIGURE 5 | Demonstration of 104 III/V dies bonding to silicon wafer. (A) Photo image of the bonded wafer, (B) CSAM results, (C) shear testing results.
0.182 nm, which is far below the required RMS of <1 nm for wafer of the bonded wafers with nearly all InP dies are successfully
direct bonding (Christiansen et al., 2006). bonded to the silicon wafer. The only missing piece is peeled off
Prior the physical contact of the wafers for molecular bonding, during pick-and-place process. The CSAM shown in Figure 5B
the silicon wafer is performed with standard SPM clean for 10 min suggests a successful bonding. The dark areas, which suggest less
and SC1 clean with mega sonic for 5 min, while III/V die-attached strong bonding, come from the dies located in the InP wafer edge.
handle wafer is first performed with sacrificial InGaAs cap layer Figure 5C shows the sheer testing results. The maximum bonding
etching in H3 PO4 solution for 1 min, followed with standard clean strength is larger than 20 MPa, with an averaged bonding strength
in NH4 OH solution for 2 min. After that, O2 plasma activation is of ~13 MPa, which is comparable with that of W2W bonding
applied to both silicon wafer and III/V dies in a RIE chamber for under the same process.
1 min, followed with DI water rinsing and wafer drying. The III/V All in all, we believe that there are at least two significant
dies and the silicon wafer are then physically contacted with each implications of the proposed multiple D2W bonding technology:
other by notch alignment between two 800 wafers, followed with
pre-bonding in the 200 mm EVG bonder for 2 min with 1000 N (1) The significantly increased bonding efficiency owing to the
mechanical force applied. The III/V dies are released from the han- simultaneous batch process. Through the batch process of the
dle wafer after pre-bonding, and all III/V dies are now transferred III/V dies (pre-clean, plasma activation, etc.), it is possible to
to the silicon device wafer. Finally, the bonded pairs are placed back bond unlimited number of dies. It also helps to avoid poten-
to the EVG bonder for post-bonding annealing at 300°C for 12 h. tial contamination by performing the pick-and-place before
As a proof-of-concept demonstration, we show here the bond- cleaning process, and eliminates the time link constraint of
ing of 104 InP dies to an 800 silicon wafer. The silicon wafer is the bonding surface deactivation. This is the most signif-
covered with 1 µm PECVD oxide after CMP. The InP dies are icant processing advantage comparing to the conventional
diced into 5 mm × 5 mm in size. Figure 5A shows the photo image pick-and-place method.

(2) The scalability to whatever-sized silicon wafers. Such multi- which means that the optical mode needs to be well confined in
ple D2W bonding technology can easily be adopted for even the III/V waveguide. However, on the other hand, the light has
larger-sized silicon wafers, such as 300 mm wafer. This is the to be confined sufficiently inside the silicon output waveguide
most critical step toward the potential manufacturability of for the efficient light extraction. In view of this, there are mainly
H-OEIC. two kinds of waveguide structures considering the optical power
distribution for the hybrid laser with optical mode predominantly
DESIGN OF III/V-on-Si LASERS confined either in the silicon waveguide or in the III/V overlay. This
A hybrid III/V-on-silicon laser consists of a III/V epitaxial-layered leads to two different optical cavity designs. In the first design, the
structure and a silicon waveguide. It is a device that emits laser optical cavity comprises both III/V and silicon waveguides and the
beams from silicon waveguides by electrical/optical injection to mode is mainly guided within the Si waveguide and evanescently
the III/V region. In this section, we will discuss the design of coupled with the III/V waveguide. Such structure is also called as
hybrid III/V-on-silicon lasers with regard to two fundamental overlapped structure with hybrid mode (Fang et al., 2006, 2007a,b,
laser elements, namely, optical gain medium and optical waveguide 2008a,b,c, 2009). It has the advantage of making the coupling to a
cavity. passive silicon waveguide straightforward and wavelength selective
features can easily be defined in the silicon waveguide layer using
DESIGN OF III/V MQW STRUCTURES CMOS fabrication techniques, which provides an accurate mech-
There are two main material systems for the fabrication of long- anism to control the emission wavelength of the laser. However, it
wavelength lasers emitting at 1.55 µm, which are InGaAsP/InP requires a controllable thin bonding layer (<50 nm) for efficient
and InGaAlAs/InP systems. Both kinds of materials can be used optical coupling, which may increase the difficulty of bonding
to fabricate hybrid lasers. InGaAlAs MQWs exhibit a larger con- process. Furthermore, due to the weak interaction between the
duction band discontinuity (E c = 0.72E g ), and smaller valence optical mode and gain material, it usually requires longer laser
band discontinuity compared with InGaAsP MQW. This leads cavity, and thus resulting in high power consumption. In the sec-
to an improved electron confinement, which can improve the ond design, the mode in the hybrid section is mainly guided by the
temperature characteristics of semiconductor laser diodes. Thus, III/V waveguide, and the light is coupled from the III/V waveguide
InGaAlAs/InP material system is more suitable for high speed and to the silicon waveguide through waveguide mode transformer,
uncooled operation of semiconductor laser diode. In this study, such as inverse tapers (Yariv and Sun, 2007; Sun et al., 2009a; Ben
we select this material system for the hybrid silicon lasers demon- Bakir et al., 2011). In such design, the bonding interface can be
stration. The MQW region includes eight Al0.055 Ga0.292 In0.653 As relatively thick (typically from 30 to 150 nm) due to the released
quantum wells separated by nine Al0.055 Ga0.292 In0.653 As barriers. coupling constrain for the bonding interface. The advantage is
The gain spectrum of the MQW is calculated and the wavelength of that the optical mode experiences a high optical gain in the central
peak gain is designed at 1550 nm when the carrier injection density region of the laser structure. However, the challenge of this struc-
increases from 5 × 1017 to 5 × 1018 /cm3 as shown in Figure 6A. ture is the fabrication of low-loss tapered waveguides. Hereafter,
Figure 6B shows the measured photoluminescence (PL) spec- we name such design as adiabatic tapered coupling structure.
trum for III/V epitaxial wafer at room temperature with the peak
wavelength at about 1550 nm. Silicon waveguide thickness selection
The selection of the silicon waveguide thickness depends on the
DESIGN OF HYBRID LASER VERTICAL WAVEGUIDE STRUCTURE detailed device dimensions/structures and the fabrication process.
As mentioned, the optical gain comes from overlying III/V stack For indium phosphide (InP)-based gain waveguides, the effective
layer, which needs to be structured to efficiently inject electrons refractive index is typically larger than 3.2 if the waveguide width
or holes into the MQW regions. A high overlap between the opti- and height are larger than 1 µm. In order to achieve this index for
cal mode and the MQW benefits to achieve a high optical gain, silicon waveguides for effective coupling with the InP-based gain
FIGURE 6 | (A) The calculated gain spectrum under different carrier injection concentration. (B) The measured photoluminescence (PL) spectrum for III/V
epitaxial wafer.

region, the corresponding silicon waveguide thickness needs to be structure. The bonded III/V-Si structure forms the hybrid wave-
sufficiently large. Figure 7 shows the calculated effective refractive guide cavity. The effective refractive index of III/V active and Si
index of the fundamental mode in silicon waveguide depending regions are critical parameters for the hybrid waveguides, which,
on the waveguide thickness. It indicates that the required silicon respectively, determine the light confinement factors in III/V and
thickness needs to be larger than 450 nm to achieve an effective Si region. In our design, the confinement factors over the silicon
index of 3.2 for the waveguide with 2 µm in width. Such thick and the quantum well regions are modified by altering the silicon
silicon layer does not match with the current mainstream silicon waveguide thickness and the separate confinement heterostruc-
photonics. However, on the other hand, it is still possible to couple ture (SCH) thickness in order to ensure sufficiently low-threshold
light from 220 nm silicon to InP waveguides by using very narrow gain for lasing. While the thicker silicon waveguide pulls the opti-
InP waveguides (~200 nm) to push down the value of effective cal mode into the silicon layer, the larger SCH thickness can drag
index, although the fabrication is difficult to form these narrow back the optical mode into the III/V region.
InP waveguides by conventional photolithography. Figure 8A shows the calculated optical confinement factors in
MQW and Si depending on the Si waveguide width under differ-
ent Si thickness, with assuming the III/V ridge width and the SCH
Overlapped structure with hybrid mode
thickness of 6 µm and 250 nm, respectively. It shows that when the
As mentioned above, there is a tradeoff between the optical mode
waveguide widths of III/V and Si are fixed, the optical confinement
confinement in the III/V and silicon regions for the overlapped
in Si waveguide can be increased by using a thick Si layer. With the
silicon waveguide thickness of 700 nm, large confinement of up
to 70% in silicon waveguide is achieved. However, the device per-
formance is very sensitive to the bonding interface quality due to
the overlapping of the optical mode with the bonding interface
between the III/V and the silicon. Based on the analysis, we adopt
silicon thickness of 500 nm for the demonstration of hybrid Si
lasers.
Figure 8B shows the simulated optical confinement factor in
MQW and in silicon waveguide with different SCH thickness, with
the fixed III/V and Si waveguide widths of 4 and 2 µm, and Si
waveguide thickness of 500 nm. As the SCH thickness increases,
the optical mode confinement in III/V region increase, which in
turn significantly decreases the optical mode confinement in sil-
icon waveguide. Inserts in Figure 8B show the simulated field
distributions with SCH thicknesses of 0.1 and 0.5 µm. It shows
obviously that for the small SCH thicknesses, the optical mode lies
primarily in the silicon region, while the optical mode is dragged
into III/V region with increased SCH thickness. The ability to con-
FIGURE 7 | Effective refractive index of the silicon waveguide trol the optical mode with the SCH thickness is a key feature of this
fundamental mode as the function of silicon waveguide thicknesses.
platform. For hybrid lasers, higher optical confinement is needed
Both top and bottom claddings are oxide (n = 1.45) and silicon index is
chosen as 3.48 at the wavelength of 1550 nm.
to achieve lower threshold current. Thus, we choose an optimized
SCH thickness of 0.18 µm for the hybrid lasers.
FIGURE 8 | Confinement factor of optical mode in multiple quantum wells (MQW) and Si waveguide as a function of (A) the Si waveguide width
under different Si waveguide thicknesses, and (B) the SCH thickness. Insets: the simulated field distributions of the fundamental TE mode with different
SCH thicknesses. It shows that by increasing the thickness of SCH layer up to 500 nm, the optical mode is more confined in the III/V active layer.

For such hybrid III/V-on-silicon lasers, another challenge arises 1.3 W/m/K. The modal gain of laser is dependent on the tem-
from the control of the bonding layer thickness. Generally, a thin perature of the active region. As the temperature of active region
bonding layer (<50 nm) is needed for efficient optical coupling increases, the modal gain decreases due to the increased carrier
between III/V and silicon regions, while the thicker bonding layer leakage out or not reaching the active region, and/or increased
benefits to the bonding quality of III/V layer and the bonding non-radiative recombination. The decrease of modal gain leads to
yield improvement. For direct bonding without oxide interlayer, high threshold current and low output optical power. In order
it is easy to achieve such thin thickness, which is usually only the to investigate the effect of interlayer on the thermal charac-
native oxide. However, this process is particularly sensitive to sur- teristics of the hybrid lasers, a two-dimensional model of the
face roughness and particles contamination, which would limit device structure is conducted using COMSOL by mapping out
the bonding quality and bonding yield. DVS-BCB bonding can be the heat dissipation of each layer. Figure 9B shows the simula-
used for the heterogeneous integration of III/V material on silicon tion structure. In the simulation, the structure parameters are
to improve the yield. However, it is difficult to obtain a control- as follows: III/V ridge width, Si ridge width and thickness, and
lable thin bonding interlayer of <50 nm. In our case, we choose the laser cavity length are assumed to be 6 µm, 1 µm, 500 nm,
silicon oxide as interlayer between III/V and silicon, which is also and 1000 µm, respectively. Injection current is 500 mA and the
compatible with the mainstream silicon phonics, in which all the corresponding voltage is 4 V. Figure 9C shows the calculated
devices are with oxide cladding. working temperatures in the III/V active region with different
Figure 9A shows the calculated optical confinement factor in thicknesses of the oxide interlayer. The increase of temperature
MQW and silicon waveguide as the function of interlayer oxide in the III/V active region versus interlayer thickness is about
thickness. In the simulation, we assume the fixed silicon thickness 0.02 K/nm. For illustration purpose, Figure 9D shows as an
of 500 nm, the silicon waveguide width of 3 µm, and III/V ridge example the thermal distribution within the layered structure
waveguide with of 6 µm. We observe from the results that the Si for the oxide interlayer thickness of 0 nm. The thermal distri-
confinement factor largely decreases when the interlayer thickness butions for the other thicknesses are similar. Actually, we can
increases from 10 to 100 nm. Only approximately 5% optical light conclude from the study that the main hurdle for the heat dis-
is confined in the silicon waveguide when the interlayer thickness sipation is the SOI BOX layer, which can be seen from the results
is 100 nm. without oxide bonding interlayer (thickness = 0 nm). Thus, for
Additionally, the interlayer of oxide at the bonding interface enhanced thermal management, novel designs such as thermal
also affects the characteristics temperature of hybrid III/V-on- shunt (Liang et al., 2012) are required to effectively remove the
silicon laser due to the poor thermal conductivity of as low as generated heat.
FIGURE 9 | (A) The confinement factor of the optical mode in the MQW changes in the III/V active region with regard to the oxide interlayer
and silicon layers, respectively, as a function of the interlayer thickness. thickness. (D) The simulated thermal distribution with the interlayer
(B) The simulation structure of the thermal distribution. (C) The temperature thickness is 0 nm.

Adiabatic tapered coupling structure direction. Here, we adopt a three-dimensional approximated

For the adiabatic tapered coupling structure, the mode in the model based on beam propagation method (BPM) in order to
hybrid section is mainly guided by the III/V waveguide, and the optimize the tapering structure of the silicon waveguide and III/V
light is coupled from the III/V waveguide to the silicon wave- waveguide for an efficient coupling. Figure 10A schematically
guide through a tapered waveguide. It shows that a tapering length illustrates the design of such waveguide tapering structure. The
~100 µm is required for a sufficient light coupling with mini- coupling efficiency largely depends on the tapering design, espe-
mized optical loss. By using such tapered coupling, it eliminates the cially the III/V waveguide taper width and taper length. Here, we
tricky tradeoff between the modal gain and vertical coupling effi- simulate such dependency by varying the taper width and taper
ciency, which is inherent in the overlapped waveguide structures. length, while fixing the III/V waveguide width of 5 µm, the silicon
Therefore, hybrid lasers with a short cavity as pure III/V laser are waveguide width of 1 µm, and the silicon taper length of 100 µm.
possible. Up to now, the hybrid lasers with the high performances Figures 10B,C, respectively, show the simulated coupling effi-
are achieved using such tapered coupling scheme (Levaufre et al., ciency from III/V-Si hybrid waveguide to silicon waveguide as
2014; Zhang et al., 2014). functions of the III/V waveguide taper width and tape length.
In order to efficiently couple the light between the Si-III/V It suggests that the coupling efficiency from III/V-Si hybrid wave-
hybrid waveguide and the silicon waveguide, the III/V waveguide guide to Si waveguide can be as high as 85% by using an 80-µm-
and silicon waveguides are tapered simultaneously in the same long III/V waveguide taper and a 100-µm-long silicon waveguide
FIGURE 10 | (A) Schematics of the tapering structure for the vertical waveguide taper length. The whole structure is cladded with oxide.
coupling between III/V waveguide and silicon waveguide. WInP : III/V Coupling efficiency of III/V waveguide to Si waveguide as functions of
waveguide width, Wend : III/V waveguide taper width, Wd : silicon (B) the III/V waveguide taper width, and (C) the III/V waveguide taper
waveguide width, LTaper1 : silicon waveguide taper length, LTaper2 : III/V length.

tape. Through optimizing the III/V waveguide taper width and planarization step, which includes oxide etch-back with 500 nm in
III/V waveguide taper length, the maximum coupling efficiency depth and CMP process. Such planarization steps with CMP also
can be as high as 99%. However, due to the optoelectronic fabri- help to smooth the bonding surface with very small surface rough-
cation limitation, we are not able to demonstrate the hybrid laser ness for molecular bonding. The interlayer oxide thickness can be
using such adiabatic tapered coupling structure. well controlled by the CMP process, with only ~50 nm oxide left
atop the silicon waveguide in our demonstration. For enhanced
DEMONSTRATION OF III/V-on-Si HYBRID LASERS flatness and uniformity of the bonding surface, we only etch away
Using the proposed multiple D2W bonding technology, we have the silicon surrounding the devices, remaining most of the silicon
demonstrated various hybrid silicon lasers, including FP lasers, areas forming silicon plateaus.
DBR lasers, sidewall-grating lasers, racetrack-shaped microring The multiple D2W bonding is then performed after the prepa-
lasers, and MLL. In this section, we will first introduce the hybrid ration of III/V dies, followed with the process described in Section
silicon laser fabrication process, leveraging on IME’s CMOS- “III/V-to-Si Wafer-to-Wafer (W2W) Bonding Technology.” As the
compatible silicon photonic fabrication facilities and NTU’s designed devices are all within an area of 8 mm × 8 mm, the InP
expertise in optoelectronics fabrication capability, followed with dies are all diced with 9 mm × 9 mm with the consideration of
showing some hybrid silicon laser demonstrations as the examples. bonding misalignment of ±500 µm for the notch alignment, thus
ensuring the full covering of all the silicon photonic devices within
III/V-on-Si HYBRID LASER FABRICATION PROCESS the III/V die area. For a 50 mm InP wafer, there are only 16 full
The III/V-on-Si hybrid laser fabrication in our demonstrations dies with 9 mm × 9 mm in size. Thus, for the purpose of hybrid
includes two parts, namely silicon passive device fabrication using silicon laser demonstration using such D2W bonding technol-
IME’s CMOS line and multiple D2W bonding in IME’s MEMS ogy, we only perform 16-dies bonding to 200 mm silicon wafers,
line, and III/V optoelectronics fabrication in NTU. Figure 11 with some of the silicon photonics device area being wasted.
shows the fabrication process flow. We adopt commercially avail- Figure 12A shows the photo image of the 16-dies bonded silicon
able SOI wafer with 340 nm silicon layer sitting on a 2 µm buried wafer. Figure 12B shows the CSAM results. Except some particle-
oxide (BOX) layer. The fabrication starts with the blanket silicon induced bonding defects, all the dies are bonded very well. How-
epitaxy to ~500 nm for refractive index matching between silicon ever, we clearly observe that some of the die edge periphery regions
waveguide device and InP gain medium. After the deposition of are not bonded well due to the wafer dicing induced damage.
70 nm oxide as the hard mask, the waveguide structures, including Figures 12C,D, respectively, show the TEM and cross-SEM of the
both grating coupler and inverse taper coupler are patterned by bonding structures, both suggesting very reliable bonding quality.
deep UV photolithography and transferred onto the silicon layer The III/V optoelectronic fabrication starts from the InP sub-
by using deep RIE etching. For the grating coupler, the silicon strate removal using HCl solution. After photolithography, the
etching thickness is 377 nm. While for other waveguide devices, InP mesa structures are formed by using H3 PO4 , H2 O2 , and
second silicon etch is applied down to the BOX layer by covering HCl mixed solution to etch the InGaAs contact layer and p-InP
the surface grating coupler area with additional photo resist. Oxide cladding layer. The SCH layer and QW layer are also etched using
cladding of 650 nm in thickness is deposited followed by a surface H3 PO4 :H2 O2 :H2 O solution, stopping on the n-InP cladding layer.
FIGURE 11 | The fabrication process of the hybrid silicon laser in a ridge waveguide structure.

Then, an SiO2 insulator layer with the thickness of 300 nm is and increases significantly to ~100 mA at room temperature. We
deposited, followed with the contact opening for p-type and n-type attribute such fast increase of the threshold to the thick oxide
injection by one-time photolithography and the oxide is etched by interlayer, which prohibits the heat dissipation efficiently. The
HF solution. After that, Ti/Au metal contacts are formed by sput- measured output power from a single facet without any reflec-
tering, with wet etching to form a Ti/Au slot between the n-type tion coating is more than 1 mW. This includes the coupling loss
and p-type contacts using diluted HF and KI solution, respectively. due to the un-optimized testing setup for light collection, which is
estimated only with ~20% light collection efficiency. The thermal
FABRY–PEROT LASERS dissipation is very important for CW lasing. Figure 13B shows
The CW operation of the optical laser requires good thermal the measured lasing spectra under different temperatures. The
management to remove the generated heat. In the case of FP wavelength shift with temperature is about 0.75 nm/K.
lasers, another way is to design narrow ridge waveguide to gen-
erate less heat. In our demonstration, we design a FP laser with LATERAL-COUPLED DISTRIBUTED-FEEDBACK LASERS WITH SIDE WALL
ridge waveguide width of 6 µm. The laser facets are formed by GRATING
lapping down the Si substrate to around 60 µm, followed with The FP laser is fabricated by lapping down the silicon substrate and
mechanical cleaving. The length of the FP laser is ~720 µm. The mechanical cleaving to form the facets. From the optical commu-
demonstrated FP laser is able to work at room temperature with nication and optical interconnect applications point of view, such
CW operation. Figure 13A shows the measured P–I curves under FP lasers are not practical for photonic integration. Furthermore,
different temperatures. The threshold current at 264 K is ~45 mA, how to achieve good facet is still a main challenge and a key limit-
ing factor for high-performance hybrid lasers as reflection coating
is always required in order to optimize the cavity transmission
and reflection. In view of this, optical resonators, Bragg grating
structures that form the cavities through fabrication are the good
candidates for on-chip hybrid laser. We here show as an example
of a hybrid laser using LC-DFB structure as the laser cavity.
Figure 14A schematically shows the perspective view of the LC-
DFB hybrid laser with illustration of the key parameters, including
the Bragg grating period Λ, the silicon ridge width D, and the grat-
ing teeth width, W 1 . Considering the fabrication limitation, we
design third-order later Bragg grating in order to achieve single-
mode operation. With regard to the silicon thickness of 500 nm,
the grating period Λ is 670 nm with filling factor of 0.5. The ridge
width D and the teeth width W1 are, respectively, designed with
2 and 1 µm. The Bragg grating is centered beneath the III/V gain
region, which is with the width of 12 µm. Both LC-DFB struc-
ture and III/V gain region are designed with the same lengths. In
order to extract the output laser light for easy characterization, the
vertical grating couplers are adopted. For the vertical grating cou-
FIGURE 12 | (A) The optical image of the 16-dies bonded silicon wafer.
pler, the period is designed to be 640 nm with filling factor of 0.5,
(B) CSAM results. (C) TEM of the bonding structure. (D) Cross-SEM of the
bonding structures. and the silicon etch depth is 377 nm. Such grating coupler design
is purely based on theoretical calculation, without any process
FIGURE 13 | (A) The power–current characteristic curves and (B) the lasing spectra of a typical hybrid silicon FP laser measured under different temperature.

FIGURE 14 | (A) The perspective view of the hybrid LC-DFB laser the LC-DFB structures and the grating coupler. (F) The cross SEM of
integrating with surface grating coupler, with illustration of the key the vertical layered structure. (G) The light power output and (H) the
design parameters. Λ: grating period, D: the ridge width, and W1 : the laser spectrum of the hybrid silicon laser with sidewall Bragg grating
grating teeth width. (B) Top-view optical microscope image of a LC-DFB structure. The output power is directly measured from the surface
hybrid laser integrated with vertical surface coupler. (C–E) The SEM of grating coupler.
verification and optimization. For this demonstration, we did not power is only ~10 µW upon 250 mA current injection, which also
design any mode transformer between III/V layer and silicon wave- can be observed from the spectrum measurement. We attribute the
guide layer, thus expecting some transition loss. Figure 14B shows relatively low output power to the following two reasons, namely,
the optical microscope image of the fabricated hybrid LC-DFB the accumulated optical loss, and the inefficient vertical light cou-
laser with integrated vertical grating couplers. The LC-DFB struc- pling. First of all, the optical loss, which mainly includes the surface
ture and the III/V gain region are designed with same length of grating coupler coupling loss, the Bragg grating scattering loss, and
700 µm, while the silicon device including two grating couplers is the non-radiative recombination loss from the bonding interface,
~2750 µm. Due to the optoelectronic fabrication limitation, there affects the light output significantly. From the reference measure-
is no designed taper between III/V waveguide and silicon wave- ment for the device only with passive silicon waveguide yet bonded
guide, thus expecting relatively high transition loss. Figures 14C–E III/V layer, the accumulated total loss is >40 dB, which is mainly
show the SEM images of the fabricated LC-DFB and vertical grat- due to the unoptimized surface grating coupler. Second, the oxide
ing coupler, while Figure 14F shows the cross-sectional SEM of the interlayer in our design, which might not be able to control pre-
vertical structures, illustrating the Si waveguide, the III/V layer, and cisely, will affect the light coupling efficiently from III/V layer to
the Ti/Au layer. silicon waveguide. Furthermore, the polarization sensitivity of the
Figures 14G,H show the measured P–I curve and the spec- surface grating coupler can also induce additional optical loss.
trum of the LC-DFB hybrid laser under pulse operation. The Thus, the optimized grating coupler design for the light extraction
threshold current is ~120 mA, corresponding to a threshold cur- from the silicon waveguide and the vertical light transition struc-
rent density of ~1.42 kA/cm2 . From the spectrum, we see clearly ture for light coupling from III/V layer to silicon waveguide can
single-mode operation with the peak wavelength at 1559.8 nm and significantly increase the laser output power. Besides, by optimiza-
a side-mode-suppression ratio (SMSR) larger than 20 dB. This is tion of the Bragg grating period and silicon waveguide thickness,
expected from the LC-DFB design. However, the maximum output the SMSR can also be enhanced.

PASSIVELY MODE-LOCKED LASERS Passive mode locking of the device is obtained by forward bias-
Semiconductor MLLs are excellent candidates for generating stable ing the gain section (I gain ) and reverse biasing (V sa ) or un-biasing
ultra-short optical pulses, which have a corresponding wide opti- the SA section. The mode locking behavior of the device is charac-
cal spectrum of phase correlated modes and high repetition rate. terized by measuring the radio frequency (RF) spectrum using the
Optical frequency combs emitted by MLLs can have high extinc- spectrum analyzer (Agilent E4448A). Figure 15B shows the mea-
tion ratios, low jitter, and low chirp, which can be utilized in a vari- sured RF spectrum of the III/V-on-Si MLL at the injection current
ety of applications including AWG, optical clock generation and to the gain section (I gain ) of 110 mA and reverse biasing the SA
recovery, coherent communications systems, high-speed analog- section at −0.9 V. The resolution bandwidth (RBW) during mea-
to-digital conversion (ADC), and optical time-division multiplex- surement is 1 MHz. The repetition frequency is about 30.0 GHz
ing (OTDM), etc. Integration of MLLs on silicon is very promising with signal-to-noise ratio above 30 dB. By changing I gain , it can
as it combines the low-loss and low-dispersion characteristics be tuned to more than 30 GHz, giving clear evidence of passively
of silicon material with high gain III/V material, thus ensuring mode locking of light signal. The measured RF linewidth of the
improved performance. Furthermore, it will be possible for semi- injection locked laser is about 7 MHz by Lorentzian fitting the RF
conductor MLLs to generate ultra-short optical pulses with low spectrum.
repetition rate on the silicon platform owing to the long cavity
length. Here, we show our preliminary demonstration of a passive SUMMARY AND FUTURE OUTLOOK
MLL using the developed heterogeneous integration platform. KEY ACHIEVEMENTS
The optical cavity of the MLL is defined by a 1250-µm- In summary, we reviewed in this paper the recent demonstra-
long gain section, a saturable absorber (SA) with the length of tions of optical light source in silicon for the application of H-
30-100µm, and cleaved facet at the waveguide end. The gain OEIC, with major focus on hybrid silicon lasers through wafer
section and SA are separated by a 20 µm electrical isolation region bonding technology. Furthermore, we proposed a proprietary
with isolation resistance >1 kΩ. The SA is made up of the same high-throughput multiple dies-to-wafer (D2W) bonding tech-
active material as the gain section. The difference between the SA nology by temporarily bonding III/V dies to a handle silicon
and gain section is that SA absorbs the light in the cavity upon wafer through pick-and-place process for subsequent simulta-
applying a reverse bias, while the gain section amplifying light neous batch processing. Such high-throughput multiple D2W
upon forward current injections. bonding technology features the merits of high bonding yield
The laser optical output is collected by a photodiode located with unlimited III/V dies and scalability to whatever-size sili-
in front of the cleaved facet. The typical threshold current with con wafers, thus is the key enabling technique toward potential
an unbiased 50-µm-long SA section is 88 mA. The device has a manufacturability of large-scale H-OEIC. As proof-of-concept
maximum single facet CW output power of 1 mW at room temper- demonstration, we showed the III/V dies to silicon wafer bond-
ature when the injection current is 140 mA. The series resistance ing with up to 104 dies. Repeatable demonstrations of 16-III/V
is about 8.5 Ω, while the slope efficiency is about 0.02 mW/mA. dies bonding to pre-patterned 200 mm silicon wafers are per-
Figure 15A shows the measured optical spectra at different injec- formed for the fabrication of hybrid silicon lasers with various
tion currents. It shows that the widest optical emission is centered laser cavities, including FP lasers, LC-DFB laser with side wall
at about 1605 nm with a full-width at half-maximum (FWHM) of grating, and MLL.
5.4 nm at the injection current of 110 mA measured by an optical
spectrum analyzer (OSA). Assuming the generated optical pulse is CHALLENGES AND FUTURE OUTLOOK
chirp-free and the shape of the pulse is with a Sech-function, the However, there are still many key issues need to be addressed before
width of the optical pulse is calculated to 0.5 ps. the hybrid silicon laser applied to optical interconnects system.
FIGURE 15 | (A) The measured optical spectra of the III/V-on-silicon mode-locked laser at different injection currents to the gain section while the SA section is
floating. (B) Measured RF spectrum of the III/V-on-Si MLL at 110 mA injection current and saturation voltage of −0.9V. The RBW during measurement is 1 MHz.

Here, we will only discuss three of the most important issues, et al., 2011). Thus, III/V-SiN platform for hybrid lasers is another
including: interest research area, which can address the loss issue. The inte-
gration between SiN waveguide and other SOI-based devices is
(1) the thermal management; also CMOS-compatible and ready for further application (Huang
(2) the integration with other silicon photonic devices with full et al., 2014).
wafer processing capability; and
(3) the new platform beyond silicon for high-performance
ACKNOWLEDGMENTS
advanced hybrid lasers.
This work was supported by A*STAR SERC Future Data Cen-
ter Technologies Thematic Strategic Research Programme under
Thermal management is one of the major obstacles for achiev-
Grant No. 112 280 4038, and A*STAR – MINDEF Science
ing high-performance hybrid silicon lasers for practical applica-
and Technology Joint Funding Programme under Grant No.122
tions (Sysak et al., 2011). Due to the poor thermal conductivity of
331 0076.
the SOI BOX layer as well as the inter oxide layer between bonding
surface, such layers would prevent the efficient heat dissipation to
the silicon substrate, thus resulting in the poor laser performances, REFERENCES
such as lower laser power. One of the simple ways to increase the Alduino, A., and Paniccia, M. (2007). Interconnects: wiring electronics with light.
Nat. Photonics 1, 153–155. doi:10.1038/nphoton.2007.17
thermal dissipation efficiency is to design the contact electrodes Alexe, M., and Gösele, U. (2004). Wafer Bonding: Applications and Technology. Berlin:
with thick and large area metal, serving as the top surface heat Springer-Verlag.
sink. Another efficient way for thermal dissipation is to remove the Bauters, J. F., Heck, M. J., John, D. D., Barton, J. S., Blumenthal, D. J., Bowers, J. E.,
BOX layer in some areas and refill it with high thermal conductive et al. (2011). “Ultra-low-loss (< 0.1 dB/m) planar silica waveguide technology,”
in Paper Presented at the IEEE Photonics Conference (Arlington, VA: Marriott
materials such as polycrystalline silicon or metals, serving as ther-
Crystal Gateway).
mal shunt (Liang et al., 2012). However, although such approach Beausoleil, R. G., Kuekes, P. J., Snider, G. S., Wang, S.-Y., and Williams, R. S.
has been demonstrated with enhanced thermal management and (2008). Nanoelectronic and nanophotonic interconnect. Proc. IEEE 96, 230–247.
increased laser performance, it still requires further development doi:10.1109/JPROC.2007.911057
in order to further improve the performance. Ben Bakir, B., Descos, A., Olivier, N., Bordel, D., Grosse, P., Augendre, E., et al. (2011).
Electrically driven hybrid Si/III-V Fabry-Pérot lasers based on adiabatic mode
The integration of such hybrid laser with the existing silicon
transformers. Opt. Express 19, 10317–10325. doi:10.1364/OE.19.010317
photonics building blocks is another key issue before it is applied Bogaerts, W., Baets, R., Dumon, P., Wiaux, V., Beckx, S., Taillaert, D., et al. (2005).
for H-OEIC. For most of the demonstrated hybrid silicon lasers, Nanophotonic waveguides in silicon-on-insulator fabricated with CMOS tech-
the silicon waveguide is normally with more than 500 nm thick- nology. J. Lightwave Technol. 23, 401–412. doi:10.1109/JLT.2004.834471
ness in order to ensure the optical index matching with III/V Bovington, J., Heck, M., and Bowers, J. (2014). Heterogeneous lasers and coupling
to Si 3 N 4 near 1060 nm. Opt. Lett. 39, 6017–6020. doi:10.1364/OL.39.006017
material for efficient light coupling to silicon waveguide. Such Boyraz, O., and Jalali, B. (2004). Demonstration of a silicon Raman laser. Opt. Express
thick waveguide design is actually not compatible with the main- 12, 5269–5273. doi:10.1364/OPEX.12.005269
stream silicon photonics, with most of the key building blocks Boyraz, O., and Jalali, B. (2005). Demonstration of directly modulated silicon Raman
are demonstrated in 220 nm silicon wafers (Xu et al., 2014). Thus, laser. Opt. Express 13, 796–800. doi:10.1364/OPEX.13.000796
novel designs taking care of both of these design considerations Camacho-Aguilera, R. E., Cai, Y., Patel, N., Bessette, J. T., Romagnoli, M., Kimerling,
L. C., et al. (2012). An electrically pumped germanium laser. Opt. Express 20,
are required. Recently, Dong et al. (2014a,b) demonstrated novel 11316–11320. doi:10.1364/OE.20.011316
integration scheme with associated transition structure via epitax- Cheng, S.-L., Lu, J., Shambat, G., Yu, H.-Y., Saraswat, K., Vuckovic, J., et al. (2009).
ial growth of silicon in a pre-defined trench. Such epitaxial-grown Room temperature 1.6 µm electroluminescence from Ge light emitting diode
silicon mesa also serves as the bonding interface with III/V gain on Si substrate. Opt. Express 17, 10019–10024. doi:10.1364/OE.17.010019
Christiansen, S. H., Singh, R., and Gosele, U. (2006). Wafer direct bonding: from
material. Thus, the rest of the device area leaves with 220 nm silicon
advanced substrate engineering to future applications in micro/nanoelectronics.
for other existing silicon photonic devices. Such novel demonstra- Proc. IEEE 94, 2060–2106. doi:10.1109/JPROC.2006.886026
tion sets a path toward the integration of hybrid silicon laser with Chu, T., Fujioka, N., and Ishizaka, M. (2009). Compact, lower-power-consumption
existing silicon photonics building blocks. However, for practical wavelength tunable laser fabricated with silicon photonic-wire waveguide micro-
integration with high-speed modulator and photodetector, which ring resonators. Opt. Express 17, 14063–14068. doi:10.1364/OE.17.014063
Creazzo, T., Marchena, E., Krasulick, S. B., Yu, P. K., Van Orden, D., Spann, J. Y.,
involves even complicated integration process with multiple oxide
et al. (2013). Integrated tunable CMOS laser. Opt. Express 21, 28048–28053.
etch-back and CMP, it is still very challenging on how to ensure doi:10.1364/OE.21.028048
the flatness and smoothness of the bonding surface. More sophis- Dong, P., Hu, T.-C., Liow, T.-Y., Chen,Y.-K., Xie, C., Luo, X., et al. (2014a). Novel inte-
ticated design and further demonstration with integration of such gration technique for silicon/III-V hybrid laser. Opt. Express 22, 26854–26861.
are highly demanded. doi:10.1364/OE.22.026854
Dong, P., Liu, X., Chandrasekhar, S., Buhl, L. L., Aroca, R., and Chen, Y.-K. (2014b).
The third issue is associated with current new demonstration Monolithic silicon photonic integrated circuits for compact 100 Gb/s coherent
trend that utilizes extremely low-loss SiN or SiON waveguide as optical receivers and transmitters. IEEE J. Sel. Top. Quantum Electron. 20, 1–8.
the passive waveguide layer (Bovington et al., 2014; Luo et al., doi:10.1109/JSTQE.2013.2295181
2014). As we know, for some advanced type of hybrid lasers, such Dong, P., Hu, T.-C., Zhang, L., Dinu, M., Kopf, R., Tate, A., et al. (2013). 1.9 µm
as the MLLs, extremely low optical loss is required for achieving hybrid silicon/iii-v semiconductor laser. Electron. Lett. 49:664. doi:10.1049/el.
2013.0674
high performances. The state-of-the-art demonstration of the sil- Dong, P., Liao, S., Feng, D., Liang, H., Zheng, D., Shafiiha, R., et al. (2009). Low V
icon waveguide is still with propagation loss of ~2 dB/cm, which pp, ultralow-energy, compact, high-speed silicon electro-optic modulator. Opt.
is higher comparing that of SiN waveguide of 0.1 dB/m (Bauters Express 17, 22484–22490. doi:10.1364/OE.17.022484

Dumon, P., Bogaerts, W., Wiaux, V., Wouters, J., Beckx, S., Van Campenhout, J., et al. Keyvaninia, S., Muneeb, M., Stanković, S., Roelkens, G., Van Thourhout, D., and
(2004). Low-loss SOI photonic wires and ring resonators fabricated with deep Fedeli, J. (2012). “Multiple die-to-wafer adhesive bonding for heterogeneous
UV lithography. IEEE Photonics Technol. Lett. 16, 1328–1330. doi:10.1109/LPT. integration,” in Paper Presented at the 16th European Conference on Integrated
2004.826025 Optics (ECIO-2012). Sitges, Barcelona, Spain.
Fang, A. W., Jones, R., Park, H., Cohen, O., Raday, O., Paniccia, M. J., et al. (2007a). Keyvaninia, S., Roelkens, G., Van Thourhout, D., Jany, C., Lamponi, M., Le Liepvre,
Integrated AlGaInAs-silicon evanescent racetrack laser and photodetector. Opt. A., et al. (2013a). Demonstration of a heterogeneously integrated III-V/SOI sin-
Express 15, 2315–2322. doi:10.1364/OE.15.002315 gle wavelength tunable laser. Opt. Express 21, 3784–3792. doi:10.1364/OE.21.
Fang, A. W., Park, H., Kuo, Y.-H., Jones, R., Cohen, O., Liang, D., et al. (2007b). 003784
Hybrid silicon evanescent devices. Mater. Today 10, 28–35. doi:10.1016/S1369- Keyvaninia, S., Verstuyft, S., Pathak, S., Lelarge, F., Duan, G.-H., Bordel, D., et al.
7021(07)70177-3 (2013b). III-V-on-silicon multi-frequency lasers. Opt. Express 21, 13675–13683.
Fang, A. W., Koch, B. R., Gan, K.-G., Park, H., Jones, R., Cohen, O., et al. (2008a). doi:10.1364/OE.21.013675
A racetrack mode-locked silicon evanescent laser. Opt. Express 16, 1393–1398. Keyvaninia, S., Verstuyft, S., Van Landschoot, L., Lelarge, F., Duan, G.-H., Mes-
doi:10.1364/OE.16.001393 saoudene, S., et al. (2013c). Heterogeneously integrated III-V/silicon distributed
Fang, A. W., Koch, B. R., Jones, R., Lively, E., Liang, D., Kuo, Y.-H., et al. (2008b). A feedback lasers. Opt. Lett. 38, 5434–5437. doi:10.1364/OL.38.005434
distributed Bragg reflector silicon evanescent laser. IEEE Photonics Technol. Lett Kissinger, G., and Kissinger, W. (1993). Void-free silicon-wafer-bond strengthen-
20, 1667–1669. doi:10.1109/LPT.2008.2003382 ing in the 200–400 C range. Sens. Actuators A 36, 149–156. doi:10.1016/0924-
Fang, A. W., Lively, E., Kuo, Y.-H., Liang, D., and Bowers, J. E. (2008c). A distributed 4247(93)85009-5
feedback silicon evanescent laser. Opt. Express 16, 4413–4419. doi:10.1364/OE. Kostrzewa, M., Di Cioccio, L., Zussy, M., Roussin, J., Fedeli, J., Kernevez, N., et al.
16.001393 (2006). InP dies transferred onto silicon substrate for optical interconnects appli-
Fang, A. W., Park, H., Cohen, O., Jones, R., Paniccia, M. J., and Bowers, J. E. (2006). cation. Sens. Actuators A 125, 411–414. doi:10.1016/j.sna.2005.07.023
Electrically pumped hybrid AlGaInAs-silicon evanescent laser. Opt. Express 14, Kurczveil, G., Heck, M. J., Peters, J. D., Garcia, J. M., Spencer, D., and Bowers, J. E.
9203–9210. doi:10.1364/OE.14.009203 (2011). An integrated hybrid silicon multiwavelength AWG laser. IEEE J. Sel. Top.
Fang, A. W., Sysak, M. N., Koch, B. R., Jones, R., Lively, E., Kuo, Y.-H., et al. (2009). Quantum Electron. 17, 1521–1527. doi:10.1109/JSTQE.2011.2112639
Single-wavelength silicon evanescent lasers. IEEE J. Sel. Top. Quantum Electron. Lasky, J. (1986). Wafer bonding for silicon-on-insulator technologies. Appl. Phys.
15, 535. doi:10.1364/OE.22.005448 Lett. 48, 78–80. doi:10.1063/1.96768
Fang, Q., Song, J., Luo, X., Jia, L., Yu, M., Lo, G., et al. (2012). High efficiency Lee, A., Jiang, Q., Tang, M., Seeds, A., and Liu, H. (2012). Continuous-wave
ring-resonator filter with NiSi heater. IEEE Photonics Technol. Lett. 24, 350–352. InAs/GaAs quantum-dot laser diodes monolithically grown on Si substrate with
doi:10.1109/LPT.2011.2177816 low threshold current densities. Opt. Express 20, 22181–22187. doi:10.1364/OE.
Fedeli, J., Di Cioccio, L., Marris-Morini, D., Vivien, L., Orobtchouk, R., Rojo-Romeo, 20.022181
P., et al. (2008). Development of silicon photonics devices using microelec- Levaufre, G., Le Liepvre, A., Jany, C., Accard, A., Kaspar, P., Brenot, R., et al.
tronic tools for the integration on top of a CMOS wafer. Adv. Opt. Tech. 2008. (2014). “Hybrid III-V/silicon tunable laser directly modulated at 10Gbit/s for
doi:10.1155/2008/412518 short reach/access networks,” in Paper Presented at the Optical Communication
Feng, S., Lei, T., Chen, H., Cai, H., Luo, X., and Poon, A. W. (2012). Silicon pho- (ECOC), 2014 European Conference on, Cannes France.
tonics: from a microresonator perspective. Laser Photonics Rev. 6, 145–177. Liang, D., and Bowers, J. (2008). Highly efficient vertical outgassing channels for
doi:10.1002/lpor.201100020 low-temperature InP-to-silicon direct wafer bonding on the silicon-on-insulator
Fujioka, N., Chu, T., and Ishizaka, M. (2010). Compact and low power consump- substrate. J. Vac. Sci Technol. B 26, 1560–1568. doi:10.1116/1.2943667
tion hybrid integrated wavelength tunable laser module using silicon waveguide Liang, D., Fang, A. W., Chen, H.-W., Sysak, M. N., Koch, B. R., Lively, E., et al. (2009a).
resonators. J. Lightwave Technol. 28, 3115–3120. doi:10.1109/JLT.2010.2073445 Hybrid silicon evanescent approach to optical interconnects. Appl. Phys. A 95,
Goodman, J. W., Leonberger, F. J., Kung, S.-Y., and Athale, R. A. (1984). Optical 1045–1057. doi:10.1007/s00339-009-5118-1
interconnections for VLSI systems. Proc. IEEE 72, 850–866. doi:10.1109/PROC. Liang, D., Fiorentino, M., Okumura, T., Chang, H.-H., Spencer, D. T., Kuo, Y.-
1984.12943 H., et al. (2009b). Electrically-pumped compact hybrid silicon microring lasers
Grenouillet, L., Dupont, T., Philippe, P., Harduin, J., Olivier, N., Bordel, D., et al. for optical interconnects. Opt. Express 17, 20355–20364. doi:10.1364/OE.17.
(2012). Hybrid integration for silicon photonics applications. Opt. Quant. Elec- 020355
tron. 44, 527–534. doi:10.1038/nnano.2014.215 Liang, D., Fiorentino, M., Srinivasan, S., Bowers, J. E., and Beausoleil, R. G. (2011).
Guha, B., Kyotoku, B. B., and Lipson, M. (2010). CMOS-compatible athermal silicon Low threshold electrically-pumped hybrid silicon microring lasers. IEEE J. Sel.
microring resonators. Opt. Express 18, 3487–3493. doi:10.1364/OE.18.003487 Top. Quantum Electron. 17, 1528–1533. doi:10.1109/JSTQE.2010.2103552
Guillot, G., and Pavesi, L. (2006). Optical Interconnects: The Silicon Approach. Berlin: Liang, D., Roelkens, G., Baets, R., and Bowers, J. E. (2010). Hybrid integrated plat-
Springer-Verlag. forms for silicon photonics. Materials 3, 1782–1802. doi:10.1038/nnano.2014.
Han, H.-S., Seo, S.-Y., and Shin, J. H. (2001). Optical gain at 1.54 µm in erbium- 215
doped silicon nanocluster sensitized waveguide. Appl. Phys. Lett. 79, 4568–4570. Liang, D., Srinivasan, S., Fiorentino, M., Kurczveil, G., Bowers, J., and Beausoleil,
doi:10.1063/1.1419035 R. (2012). “A metal thermal shunt design for hybrid silicon microring laser,”
Hattori, H. T., Seassal, C., Touraille, E., Rojo-Romeo, P., Letartre, X., Hollinger, in Paper Presented at the Optical Interconnects Conference, 2012 (Santa Fe,
G., et al. (2006). Heterogeneous integration of microdisk lasers on silicon strip NM: IEEE).
waveguides for optical interconnects. IEEE Photonics Technol. Lett. 18, 223–225. Lim, A. E.-J., Song, J., Fang, Q., Li, C., Tu, X., Duan, N., et al. (2014). Review of sili-
doi:10.1109/LPT.2005.861542 con photonics foundry efforts. IEEE J. Sel. Top. Quantum Electron. 20, 8300112.
Hong, T., Ran, G.-Z., Chen, T., Pan, J.-Q., Chen, W.-X., Wang, Y., et al. (2010). A doi:10.1109/JSTQE.2013.2293274
selective-area metal bonding InGaAsP–Si laser. IEEE Photonics Technol. Lett. 22, Liow, T.-Y., Song, J., Tu, X., Lim, A.-J., Fang, Q., Duan, N., et al. (2013). Silicon
1141–1143. doi:10.1364/OE.22.005448 optical interconnect device technologies for 40 Gb/s and beyond. IEEE J.
Huang, Y., Song, J., Luo, X., Liow, T.-Y., and Lo, G.-Q. (2014). CMOS com- Sel. Top. Quantum Electron. 19, 8200312–8200312. doi:10.1109/JSTQE.2012.
patible monolithic multi-layer Si3N4-on-SOI platform for low-loss high per- 2218580
formance silicon photonics dense integration. Opt. Express 22, 21859–21865. Lipson, M. (2005). Guiding, modulating, and emitting light on silicon-challenges
doi:10.1364/OE.22.021859 and opportunities. J. Lightwave Technol. 23, 4222. doi:10.1109/JLT.2005.858225
Jalali, B., and Fathpour, S. (2006). Silicon photonics. J. Lightwave Technol. 24, Liu, H.,Wang, T., Jiang, Q., Hogg, R., Tutu, F., Pozzi, F., et al. (2011). Long-wavelength
4600–4615. doi:10.1109/JLT.2006.885782 InAs/GaAs quantum-dot laser diode monolithically grown on Ge substrate. Nat.
Justice, J., Bower, C., Meitl, M., Mooney, M. B., Gubbins, M. A., and Corbett, B. Photonics 5, 416–419. doi:10.1364/OE.20.022181
(2012). Wafer-scale integration of group III-V lasers on silicon using transfer Liu, J., Sun, X., Camacho-Aguilera, R., Kimerling, L. C., and Michel, J. (2010). Ge-
printing of epitaxial layers. Nat. Photonics 6, 610–614. doi:10.1038/nphoton. on-Si laser operating at room temperature. Opt. Lett. 35, 679–681. doi:10.1364/
2012.204 OL.35.000679

Liu, J., Sun, X., Kimerling, L. C., and Michel, J. (2009). Direct-gap optical gain Seassal, C., Rojo-Romeo, P., Letartre, X., Viktorovitch, P., Hollinger, G., Jalaguier,
of Ge on Si at room temperature. Opt. Lett. 34, 1738–1740. doi:10.1364/OL.34. E., et al. (2001). InP microdisk lasers on silicon wafer: CW room temperature
001738 operation at 1.6 µm. Electron. Lett. 37, 222–223. doi:10.1049/el:20010173
Liu, J., Sun, X., Pan, D., Wang, X., Kimerling, L. C., Koch, T. L., et al. (2007). Tensile- Shainline, J. M., and Xu, J. (2007). Silicon as an emissive optical medium. Laser
strained, n-type Ge as a gain medium for monolithic laser integration on Si. Opt. Photonics Rev. 1, 334–348. doi:10.1002/lpor.200710021
Express 15, 11272–11277. doi:10.1364/OE.15.011272 Song, J., Luo, X., Tu, X., Jia, L., Fang, Q., Liow, T.-Y., et al. (2013). On-chip quasi-
Luo, X., Song, J., Feng, S., Poon, A. W., Liow, T.-Y., Yu, M., et al. (2012). Silicon digital optical switch using silicon microring resonator-coupled Mach-Zehnder
high-order coupled-microring-based electro-optical switches for on-chip opti- interferometer. Opt. Express 21, 12767–12775. doi:10.1364/OE.21.012767
cal interconnects. IEEE Photonics Technol. Lett. 24, 821–823. doi:10.1109/LPT. Soref, R. (2006). The past, present, and future of silicon photonics. IEEE J. Sel. Top.
2012.2188829 Quantum Electron. 12, 1678–1687. doi:10.1021/ar900141y
Luo, X., Song, J., Zhou, H., Liow, T. Y., Yu, M., and Lo, P. G. Q. (2014). United States Soref, R., and Lorenzo, J. (1985). Single-crystal silicon: a new material for 1.3 and
Patent No. US 2014/0153600 A1. 1.6 µm integrated-optical components. Electron. Lett. 21, 953–954. doi:10.1049/
Maszara, W. (1991). Silicon-on-insulator by wafer bonding: a review. J. Electrochem. el:19850673
Soc. 138, 341–347. doi:10.1149/1.2085575 Sparacin, D. K., Spector, S. J., and Kimerling, L. C. (2005). Silicon waveguide
Mechet, P., Verstuyft, S., De Vries, T., Spuesens, T., Regreny, P., Van sidewall smoothing by wet chemical oxidation. J. Lightwave Technol. 23, 2455.
Thourhout, D., et al. (2013). Unidirectional III-V microdisk lasers heteroge- doi:10.1109/JLT.2005.851328
neously integrated on SOI. Opt. Express 21, 19339–19352. doi:10.1364/OE.21. Stanković, S., Jones, R., Sysak, M. N., Heck, J. M., Roelkens, G., and Van Thourhout,
019339 D. (2011). 1310-nm hybrid III–V/Si Fabry–Pérot laser based on adhesive
Michel, J., Liu, J., and Kimerling, L. C. (2010). High-performance Ge-on-Si pho- bonding. IEEE Photonics Technol. Lett. 23, 1781–1783. doi:10.1109/LPT.2011.
todetectors. Nat. Photonics 4, 527–534. doi:10.1038/nphoton.2010.157 2169397
Miller, D. A. (2000). Optical interconnects to silicon. IEEE J. Sel. Top. Quantum Stanković, S., Jones, R., Sysak, M. N., Heck, J. M., Roelkens, G., and Van Thourhout,
Electron. 6, 1312–1317. doi:10.1109/2944.902184 D. (2012). Hybrid III–V/Si distributed-feedback laser based on adhesive
Miller, D. A. (2009). Device requirements for optical interconnects to silicon chips. bonding. IEEE Photonics Technol. Lett. 24, 2155–2158. doi:10.1109/LPT.2012.
Proc. IEEE 97, 1166–1185. doi:10.1109/JPROC.2009.2014298 2223666
Ohashi, K., Nishi, K., Shimizu, T., Nakada, M., Fujikata, J., Ushida, J., et al. (2009). Stanković, S., Van Thourhout, D., Roelkens, G., Jones, R., Heck, J., and Sysak, M.
On-chip optical interconnect. Proc. IEEE 97, 1186–1198. doi:10.1109/JPROC. (2010). Die-to-die adhesive bonding for evanescently-coupled photonic devices.
2009.2020331 ECS Trans. 33, 411–420. doi:10.1149/1.3483531
Park, H., Fang, A., Kodama, S., and Bowers, J. (2005). Hybrid silicon evanescent laser Sun, X., Liu, H.-C., and Yariv, A. (2009a). Adiabaticity criterion and the shortest adi-
fabricated with a silicon waveguide and III-V offset quantum wells. Opt. Express abatic mode transformer in a coupled-waveguide system. Opt. Lett. 34, 280–282.
13, 9460–9464. doi:10.1364/OPEX.13.009460 doi:10.1364/OL.34.000280
Pasquariello, D., and Hjort, K. (2002). Plasma-assisted InP-to-Si low temperature Sun, X., Liu, J., Kimerling, L. C., and Michel, J. (2009b). Direct gap pho-
wafer bonding. IEEE J. Sel. Top. Quantum Electron. 8, 118–131. doi:10.1109/2944. toluminescence of n-type tensile-strained Ge-on-Si. Appl. Phys. Lett. 95,
991407 011911–011911–011913. doi:10.1364/OL.38.000652
Pavesi, L., Dal Negro, L., Mazzoleni, C., Franzo, G., and Priolo, F. (2000). Optical Sun, X., Liu, J., Kimerling, L. C., and Michel, J. (2009c). Room-temperature direct
gain in silicon nanocrystals. Nature 408, 440–444. doi:10.1038/35044012 bandgap electroluminesence from Ge-on-Si light-emitting diodes. Opt. Lett. 34,
Pavesi, L., and Lockwood, D. J. (2004). Silicon Photonics. Berlin: Springer-Verlag. 1198–1200. doi:10.1364/OL.34.001198
Poon, A. W., Luo, X., Xu, F., and Chen, H. (2009a). Cascaded microresonator- Sysak, M. N., Liang, D., Jones, R., Kurczveil, G., Piels, M., Fiorentino, M., et al. (2011).
based matrix switch for silicon on-chip optical interconnection. Proc. IEEE 97, Hybrid silicon laser technology: a thermal perspective. IEEE J. Sel. Top. Quantum
1216–1238. doi:10.1109/JPROC.2009.2014884 Electron. 17, 1490–1498. doi:10.1109/JSTQE.2011.2109940
Poon, A. W., Luo, X., Zhou, L., Li, C., Lee, J. Y., Xu, F., et al. (2009b). “Microresonator- Tanabe, K., Guimard, D., Bordel, D., Iwamoto, S., and Arakawa, Y. (2010). Electri-
based devices on a silicon chip: novel shaped cavities and resonance coherent cally pumped 1.3 µm room-temperature InAs/GaAs quantum dot lasers on Si
interference,” in Practical Applications of Microresonators in Optics and Photonics, substrates by metal-mediated wafer bonding and layer transfer. Opt. Express 18,
ed. A. B. Matsko (Boca Raton, FL: CRC Press), 211–263. 10604–10608. doi:10.1364/OE.18.010604
Reed, G. T., and Knights, A. P. (2004). Silicon Photonics: An Introduction. England: Tatsumi, T., Tanabe, K., Watanabe, K., Iwamoto, S., and Arakawa, Y. (2012). 1.3 µm
John Wiley & Sons. InAs/GaAs quantum dot lasers on Si substrates by low-resistivity, Au-free metal-
Reed, G. T., Mashanovich, G., Gardes, F., and Thomson, D. (2010). Silicon optical mediated wafer bonding. J. Appl. Phys. 112, 033107. doi:10.1063/1.4742198
modulators. Nat. Photonics 4, 518–526. doi:10.1038/nphoton.2010.219 Tong, Q.-Y., and Goesele, U. (1999). Semiconductor Wafer Bonding: Science and Tech-
Roelkens, G., Van Thourhout, D., and Baets, R. (2005). Ultra-thin benzocy- nology. Wiley, 1998.
clobutene bonding of III–V dies onto SOI substrate. Electron. Lett. 41, 561–562. Tu, X., Chang, K.-F., Liow, T.-Y., Song, J., Luo, X., Jia, L., et al. (2014). Silicon
doi:10.1049/el:20050807 optical modulator with shield coplanar waveguide electrodes. Opt. Express 22,
Roelkens, G., Van Thourhout, D., Baets, R., and Smit, M. (2006). Laser emis- 23724–23731. doi:10.1364/OE.22.023724
sion and photodetection in an InP/InGaAsP layer integrated on and cou- Tu, X., Liow, T.-Y., Song, J., Luo, X., Fang, Q., Yu, M., et al. (2013). 50-Gb/s silicon
pled to a Silicon-on-Insulator waveguide circuit. Opt. Express 14, 8154–8159. optical modulator with traveling-wave electrodes. Opt. Express 21, 12776–12782.
doi:10.1364/OE.14.008154 doi:10.1364/OE.21.012776
Rong, H., Jones, R., Liu, A., Cohen, O., Hak, D., Fang, A., et al. (2005a). A Urino, Y., Shimizu, T., Okano, M., Hatori, N., Ishizaka, M., Yamamoto, T., et al.
continuous-wave Raman silicon laser. Nature 433, 725–728. doi:10.1038/ (2011). First demonstration of high density optical interconnects integrated with
nature03273 lasers, optical modulators, and photodetectors on single silicon substrate. Opt.
Rong, H., Liu, A., Jones, R., Cohen, O., Hak, D., Nicolaescu, R., et al. (2005b). An Express 19, B159–B165. doi:10.1364/OE.19.00B159
all-silicon Raman laser. Nature 433, 292–294. doi:10.1038/nature03273 Van Campenhout, J., Green, W. M., Assefa, S., and Vlasov, Y. A. (2009). Low-power,
Rong, H., Xu, S., Kuo, Y.-H., Sih, V., Cohen, O., Raday, O., et al. (2007). Low- 2x2 silicon electro-optic switch with 110-nm bandwidth for broadband recon-
threshold continuous-wave Raman silicon laser. Nat. Photonics 1, 232–237. figurable optical networks. Opt. Express 17, 24020–24029. doi:10.1364/OE.17.
doi:10.1038/nphoton.2007.29 024020
Rotem, E., Shainline, J. M., and Xu, J. M. (2007a). Electroluminescence of nanopat- Van Campenhout, J., Liu, L., Romeo, P. R., Van Thourhout, D., Seassal, C., Regreny,
terned silicon with carbon implantation and solid phase epitaxial regrowth. Opt. P., et al. (2008). A compact SOI-integrated multiwavelength laser source based
Express 15, 14099–14106. doi:10.1364/OE.15.014099 on cascaded InP microdisks. IEEE Photonics Technol. Lett. 20, 1345–1347.
Rotem, E., Shainline, J. M., and Xu, J. M. (2007b). Enhanced photoluminescence doi:10.1109/LPT.2008.926857
from nanopatterned carbon-rich silicon grown by solid-phase epitaxy. Appl. Van Campenhout, J., Rojo Romeo, P., Regreny, P., Seassal, C.,Van Thourhout, D.,Ver-
Phys. Lett. 91, 051127. doi:10.1063/1.2766843 stuyft, S., et al. (2007). Electrically pumped InP-based microdisk lasers integrated

with a nanophotonic silicon-on-insulator waveguide circuit. Opt. Express 15, Zhou, L., and Poon, A. W. (2007). Electrically reconfigurable silicon micror-
6744–6749. doi:10.1364/OE.15.006744 ing resonator-based filter with waveguide-coupled feedback. Opt. Express 15,
Vivien, L., and Pavesi, L. (2013). Handbook of Silicon Photonics. Boca Raton, FL: CRC 9194–9204. doi:10.1364/OE.15.009194
Press.
Xia, F., Sekaric, L., and Vlasov, Y. (2006). Ultracompact optical buffers on a silicon Conflict of Interest Statement: The authors declare that the research was conducted
chip. Nat. Photonics 1, 65–71. doi:10.1038/nphoton.2006.42 in the absence of any commercial or financial relationships that could be construed
Xiao, S., Khan, M. H., Shen, H., and Qi, M. (2007). Multiple-channel silicon micro- as a potential conflict of interest.
resonator based filters for WDM applications. Opt. Express 15, 7489–7498.
doi:10.1364/OE.15.009386 Received: 08 January 2015; paper pending published: 13 February 2015; accepted: 17
Xu, D.-X., Schmid, J. H., Reed, G. T., Mashanovich, G. Z., Thomson, D. J., March 2015; published online: 07 April 2015.
Nedeljkovic, M., et al. (2014). Silicon photonic integration platform-have we Citation: Luo X, Cao Y, Song J, Hu X, Cheng Y, Li C, Liu C, Liow T-Y, Yu M, Wang
found the sweet spot? IEEE J. Sel. Top. Quantum Electron. 20, 189–205. H, Wang QJ and Lo PG-Q (2015) High-throughput multiple dies-to-wafer bonding
doi:10.1109/JSTQE.2014.2299634 technology and III/V-on-Si hybrid lasers for heterogeneous integration of optoelectronic
Yang, H., Zhao, D., Chuwongin, S., Seo, J.-H., Yang, W., Shuai, Y., et al. (2012). integrated circuits. Front. Mater. 2:28. doi: 10.3389/fmats.2015.00028
Transfer-printed stacked nanomembrane lasers on silicon. Nat. Photonics 6, This article was submitted to Optics and Photonics, a section of the journal Frontiers
615–620. doi:10.1038/nphoton.2012.160 in Materials.
Yariv, A., and Sun, X. (2007). Supermode Si/III-V hybrid lasers, optical ampli- Copyright © 2015 Luo, Cao, Song , Hu, Cheng , Li, Liu, Liow, Yu, Wang , Wang and
fiers and modulators: a proposal and analysis. Opt. Express 15, 9147–9151. Lo. This is an open-access article distributed under the terms of the Creative Commons
doi:10.1364/OE.15.009147 Attribution License (CC BY). The use, distribution or reproduction in other forums is
Zhang, C., Srinivasan, S., Tang, Y., Heck, M. J., Davenport, M. L., and Bowers, J. E. permitted, provided the original author(s) or licensor are credited and that the original
(2014). Low threshold and high speed short cavity distributed feedback hybrid publication in this journal is cited, in accordance with accepted academic practice. No
silicon lasers. Opt. Express 22, 10202–10209. doi:10.1364/OE.22.010202 use, distribution or reproduction is permitted which does not comply with these terms.

REVIEW ARTICLE
MATERIALS
published: 17 September 2014
doi: 10.3389/fmats.2014.00015
Group IV light sources to enable the convergence of

photonics and electronics
Shinichi Saito 1 *, Frederic Yannick Gardes 1 , Abdelrahman Zaher Al-Attili 1 , Kazuki Tani 2,3,4 , Katsuya Oda 2,3,4 ,
Yuji Suwa 2,3,4 , Tatemi Ido 2,3,4 , Yasuhiko Ishikawa 5 , Satoshi Kako 3,6 , Satoshi Iwamoto 3,6 and
Yasuhiko Arakawa 3,6
1
Faculty of Physical Sciences and Engineering, University of Southampton, Southampton, UK
2
Photonics Electronics Technology Research Association (PETRA), Tokyo, Japan
3
Institute for Photonics-Electronics Convergence System Technology (PECST), Tokyo, Japan
4
Central Research Laboratory, Hitachi Ltd., Tokyo, Japan
5
Department of Materials Engineering, Graduate School of Engineering, The University of Tokyo, Tokyo, Japan
6
Institute of Industrial Science, The University of Tokyo, Tokyo, Japan
Edited by: Group IV lasers are expected to revolutionize chip-to-chip optical communications in terms
Jifeng Liu, Dartmouth College, USA
of cost, scalability, yield, and compatibility to the existing infrastructure of silicon industries
Reviewed by:
for mass production. Here, we review the current state-of-the-art developments of silicon
Androula Galiouna Nassiopoulou,
National Centre for Scientific and germanium light sources toward monolithic integration. Quantum confinement of elec-
Research Demokritos, Greece trons and holes in nanostructures has been the primary route for light emission from silicon,
Raul J. Martin-Palma, Universidad and we can use advanced silicon technologies using top-down patterning processes to fab-
Autonoma de Madrid, Spain
ricate these nanostructures, including fin-type vertical multiple-quantum-wells. Moreover,
the electromagnetic environment can also be manipulated in a photonic crystal nanocavity
*Correspondence:
Shinichi Saito, Nano Research Group, to enhance the efficiency of light extraction and emission by the Purcell effect. Germanium
Electronics and Computer Science, is also widely investigated as an active material in Group IV photonics, and novel epitaxial
Faculty of Physical Sciences and growth technologies are being developed to make a high quality germanium layer on a sil-
Engineering, Highfield Campus,
icon substrate. To develop a practical germanium laser, various technologies are employed
University of Southampton,
Southampton SO17 1BJ, UK for tensile-stress engineering and high electron doping to compensate the indirect valleys
e-mail: s.saito@soton.ac.uk in the conduction band. These challenges are aiming to contribute toward the convergence
of electronics and photonics on a silicon chip.
Keywords: silicon, photonics, CMOS, germanium, epitaxy, luminescence, quantum, strain
1. INTRODUCTION complementary-metal-oxide-semiconductor (CMOS) technolo-

As the integration of transistors in a chip increases, the demands gies for fabrication. The hybrid integration of III-V devices on
of the interconnections are expanding, since more information an Si substrate (Fang et al., 2006) or feeding of an optical fiber to
will be transferred between chips optically (Miller, 2009). The an Si waveguide coupled with a grating from a III-V laser diode
advantage of optical interconnection over electrical wiring is (Gunn, 2006) would be the near-term solution, but it is desirable to
fundamentally coming from the elementary particles, photons, realize monolithic light sources for the long term. Comprehensive
used for signal transmission. We can transmit photons with- reviews on developing practical lasers on Si have been published by
out an electrical connection throughout an optical fiber, since various authors (Cullis et al., 1997; Ossicini et al., 2006; Daldosso
photons do not have charge. Of course, optical loss exists, but and Pavesi, 2009; Liang and Bowers, 2010; Steger et al., 2011; Liu
still the total energy consumption of the optical interconnec- et al., 2012; Michel and Romagnoli, 2012; Boucaud et al., 2013;
tion can be much lower than that of the electrical connection, Shakoor et al., 2013; Liu, 2014). Here, we review this active field
especially for the long-distance communications at higher data focusing on the progress of Si and germanium (Ge) light sources
rate, even including the energy required to convert electrons to fabricated by standard CMOS processes.
photons and vice versa (Miller, 2009). Si photonics is revolu- Photoluminescence (PL) (Canham, 1990) and electrolumi-
tionizing optical interconnections in terms of cost, power, band- nescence (EL) (Koshida and Koyama, 1992) from porous-Si are
width, and scalability (Zimmermann, 2000; Pavesi and Lockwood, the most famous achievements to overcome the fundamental
2004; Reed and Knights, 2004; Pavesi and Guillot, 2006; Reed, limitations of the indirect band-gap character of Si. The max-
2008; Deen and Basu, 2012; Fathpour and Jalali, 2012; Vivien imum PL (Gelloz and Koshida, 2000) and EL (Gelloz et al.,
and Pavesi, 2013). III-V (Wale, 2008; Evans et al., 2011) and Si- 2005) quantum efficiency exceeded 23 and 1%, respectively. The
based platform technologies (Reed and Knights, 2004; Gunn, 2006; mechanism of light emission from porous-Si is considered to
Rylyakov et al., 2011; Arakawa et al., 2013; Urino et al., 2013) are originate from quantum confinement effects (Canham, 1990;
competing for the next generation of optical interconnections. Koshida and Koyama, 1992; Cullis et al., 1997; Nassiopoulou,
The critical missing component for Si photonics is a mono- 2004; Ossicini et al., 2006; Daldosso and Pavesi, 2009) in the
lithic light source compatible with the existing infrastructure of self-organized nanostructure. The typical length scale to expect
Frontiers in Materials | Optics and Photonics September 2014 | Volume 1 | Article 15 | 66

Saito et al. Group IV light sources
quantum confinement would be comparable to the exciton Bohr

radius, which is about 5 nm for Si and 18 nm for Ge (Cullis et al.,
1997). On the other hand, the gate length fabricated by CMOS
technologies is comparable to the exciton Bohr radius so that we
can fabricate various quantum structures, including quantum dots
(Arakawa and Sakaki, 1982), nano-wires, and quantum-wells, by
lithographically controlled top-down processes. In addition, novel
cavity structures (Iwamoto and Arakawa, 2012) can be fabricated
to enhance the internal quantum efficiency by the Purcell effect
(Purcell, 1946) as well as the extraction efficiency by improved
coupling to a lens. Ge is also intensively studied, since the direct
band-gap energy is closer to the indirect transition energy than
that of Si. Highly, n-type doping and strain engineering are effec-
tive to enhance the light emissions from Ge (Liu et al., 2012; Michel
and Romagnoli, 2012; Boucaud et al., 2013; Liu, 2014), and some
of these recent advances are reviewed in this paper.
FIGURE 1 | Optical gain of Si(001) thin films calculated from direct
transitions across the energy gap only.
2. STRATEGIES TO ENHANCE LIGHT EMISSION FROM
GROUP IV MATERIALS
2.1. THEORETICAL STUDY OF LIGHT EMISSION FROM SILICON
Both Si and Ge are known to be poor light emitters because of their near the X -point. The other is that Ge has a local minimum of
indirect band-gap structures. Even so, there are some methods for the LCB at the 0-point (0-valley), while Si does not. An L-valley
making direct transitions to occur in these materials. These possi- is projected onto the 0-point in the two-dimensional k-space for
bilities were examined theoretically by first-principles calculations Ge(111) QW. For Ge(001) QW, no L-valley is projected onto the
based on density functional theory using plane-wave-based ultra- 0-point. The small 0-valley, which is not occupied unless a large
soft pseudo-potentials (Vanderbilt, 1990; Laasonen et al., 1993). number of electrons are injected, is always projected onto the
Generalized gradient approximation (Perdew et al., 1996) is used 0-point independently of the direction of the QW. While there
for the calculation of Si, and hybrid functional (Perdew et al., 1996) are two approaches to obtain efficient light emission from Ge
is used for Ge. The optical matrix elements are calculated with the by direct transitions, using L-valleys of a Ge(111) QW or using
aid of core-repair terms (Kageshima and Shiraishi, 1997). the 0-valley of bulk Ge, we think the latter is more promis-
The lowest conduction band (LCB) of bulk Si has a minimum ing. This is due to the fact that the calculated optical matrix
near the X -point, and six electron valleys exist near X -points. element for the 0-valley is very large compared to that for the
Two valleys among the six are projected onto 0-point in two- L-valleys.
dimensional momentum (k)-space when an Si quantum-well To enhance light emission from bulk Ge, applying tensile strain
(QW) with (001) surfaces is fabricated; this is called a valley- is known to be effective (Liu et al., 2010). Tensile strain makes
projection. Because the top of the valence band is also projected the energy difference between the 0 and L-valleys small, and that
onto 0, direct transitions are possible in an Si(001) QW. makes electron injection into the 0-valley easier. Also heavy n-type
Optical gain of Si(001) QWs is shown in Figure 1 as a function doping is known to be effective, because electrons can be injected
of the thickness (Suwa and Saito, 2009). Here, losses due to tran- into the 0-valley if the L-valleys are already occupied by doped
sitions within conduction bands and those within valence bands electrons.
are not taken into account. The thinner QW shows the larger gain, In order to predict required strength of strain and amount
since the surface of the QW plays an important role in this direct of doping, we calculated optical gains of bulk Ge with and
transition and it dominates if the QW is thin. Figure 1 also shows without strain. Figure 2 shows calculated optical gain as func-
that the surface structure of the QW affects the efficiency of light tions of injected electron density and hole densities. Here, the
emission strongly. applied strain is assumed to be 0.25% biaxial tensile strain parallel
Experimentally, optical gain from the Si quantum dots (Pavesi to (001) surface and optical losses due to free carrier absorp-
et al., 2000) embedded in an insulating matrix (Pavesi et al., 2000; tions (Wang et al., 2013) are taken into account. This result
Nassiopoulou, 2004; Ossicini et al., 2006; Pavesi and Guillot, 2006) shows that even bulk Ge without strain can have a positive
has been reported. It was confirmed that the interface states asso- optical gain, but number of electrons required for that is very
ciated with oxygen atoms were important to explain the positive large (1020 cm−3 ). Despite the relatively small amount of the
optical gain (Pavesi et al., 2000; Nassiopoulou, 2004; Ossicini et al., strain (0.25%), the impact on the gain is clear. Owing to this
2006; Pavesi and Guillot, 2006). It will be interesting to make these enhancement, only half the electron density (5 × 1019 cm−3 ) is
structures by top-down CMOS processes. needed to have positive gain. In experiment, applying 0.25%
strain is rather easy, and making higher strain will be possi-
2.2. THEORETICAL STUDY OF LIGHT EMISSION FROM GERMANIUM ble, as we see in the following sections. Therefore, Ge lasers
Ge has two important differences from Si. One is that Ge has the will be realized when an appropriate strain and carrier injection
minimum of the LCB at the L-point (L-valley), while Si has it are achieved.
www.frontiersin.org September 2014 | Volume 1 | Article 15 | 67

3. ELECTRO-LUMINESCENCE FROM SILICON p-i-n diodes (Saito et al., 2006a,b, 2008, 2009; Noborisaka et al.,
QUANTUM-WELL 2011; Saito, 2011). Another advantage of these device structures
As we reviewed in Section 1, it is well established that efficient is the fabrication of the Si QW through the LOCal-Oxidation of
recombination is observed in Si nanostructures by quantum con- Si (LOCOS) process. The LOCOS process was originally devel-
finement effects (Canham, 1990; Koshida and Koyama, 1992; oped for isolation of CMOS transistors (Sze and Lee, 2012; Taur
Cullis et al., 1997; Ossicini et al., 2006; Daldosso and Pavesi, 2009). and Ning, 2013). It was also used to evaluate the carrier mobility
The nanostructures include quantum dots (Arakawa and Sakaki, in the ultra-thin Si QW (Uchida and Takagi, 2003). Oxidation is
1982), nano-wires (Canham, 1990; Koshida and Koyama, 1992), one of the most precisely controlled processes in CMOS technolo-
quantum-well (QW) (Saito et al., 2006a,b, 2008, 2009; Saito, 2011), gies, and we can routinely oxidize a large Si wafer (typically 8–1000
and fins (Saito et al., 2011a,b). One of the difficulties in develop- in diameter) within the local variation of <0.1 nm. Besides, the
ing an efficient light-emitting diode (LED) made of Si comes from interface between Si and SiO2 is excellent with low interface trap
the trade-off between quantum confinement and carrier injection. density (<1011 cm−2 ) (Sze and Lee, 2012; Taur and Ning, 2013).
The surface of these Si nanostructures is easily oxidized to SiO2 , The excellent interfacial quality and strong quantum confinement
and the band offsets between Si and SiO2 are too high to expect in Si nanostructures are critical to ensure high quantum efficiency
efficient current injection except for tunneling. In order to over- (Gelloz et al., 2005). As shown in Figure 3E, EL is observed exclu-
come this trade-off, lateral carrier injection into the Si QW was sively from the thin Si QW and EL from thick Si electrodes is
proposed (Saito et al., 2006a,b, 2008, 2009; Hoang et al., 2007; negligible (Saito et al., 2006a). This supports the mechanism of EL
Noborisaka et al., 2011; Saito, 2011). As shown in Figures 3A–C, based on quantum confinement (Ossicini et al., 2006; Suwa and
the Si QW LEDs were fabricated by local thinning of a silicon-on- Saito, 2009). The high carrier density in the thin Si QW also con-
insulator (SOI) substrate, and the Si QW was directly connected to tributes to enhance the emissions (Saito et al., 2006a). By applying
the thick Si diffusion electrodes (Saito et al., 2006a,b). Both elec- the back gate to the Si substrate, we can modulate the intensities of
trons and holes are laterally injected to the Si QW in these planar light emission, and the device can be called as an Si light-emitting
transistor (Saito et al., 2006b).
The next step toward the practical light source for Si photonics
is to couple the light from Si to a cavity and a waveguide (WG).
An Si-based WG cannot be used for emission from Si QW due
to the absorption. An Si3 N4 WG was fabricated on top of the Si
QW by conventional lithography and dry etching (Saito et al., 2008,
2009). To enhance the optical confinement in the WG of the Si Res-
onant Cavity LED (RCLED), part of the supporting substrate was
removed by using double sided aligner and anisotropic wet etch-
ing (Saito et al., 2008, 2009), as shown in Figure 3B. Evanescent
coupling between the propagating optical mode and Si QW was
expected, and the enhanced EL from the edge of the waveguide was
observed (Figure 3F). More recently, SOI substrates with superior
uniformities with thick Buried-OXide (BOX) (>2 µm) became
available, and by using these wafers, strong optical confinement
within the Si3 N4 WG was ensured without removing the support-
FIGURE 2 | Optical gain of germanium as a function of injected ing Si substrate (Saito, 2011), as shown in Figure 3C. In fact, the
electron density and hole density is shown. Those without strain and
with 0.25% biaxial tensile strain parallel to (001) surface are shown.
near-field image of the propagating optical mode was taken at the
edge of the WG (Figure 3G).
FIGURE 3 | Development of an Si light source. (A) Si QW LED, (B) Si RCLED, (C) Si QW LED with thick BOX, (D) Si FinLED, and (E–H) EL images from these
devices. (E,F) are plan views. (G,H) are near-field images at the edge of WG.

refractive index but also on the electromagnetic environment sur-

rounding the material. In the previous sections, engineering group
IV materials themselves such as quantum confinement, doping,
and strain engineering have been discussed. Here, we discuss
another approach, i.e., tailoring the electromagnetic environment
by photonic nanostructures for improving light emission proper-
ties. The total efficiency of light-emitting devices can be expressed
as a product of three factors: light emission efficiency ηemission ,
extraction efficiency ηextraction , and collection efficiency ηcollection .
ηemission denotes how efficiently injected carriers recombine by
emitting photons. ηextraction takes into account the fact that only
a part of emitted photons can be extracted from the material.
ηcollection expresses how much extracted photons can be collected
by the first lens of the setup. All of them can be improved by
photonic nanostructures. Photonic crystal (PhC) (Jannopoulos
and Winn, 1995), which has a wavelength-scale periodic variation
of refractive index, is an important photonic nanostructure for
this application (see discussions in Iwamoto and Arakawa, 2012).
Figure 5A shows a scanning electron microscope (SEM) image of a
FIGURE 4 | EL from Si FinLED taken from edge of WG. (A) Spectra and
two-dimensional (2D) PhC slab, which is the most widely studied
(B) integrated intensity. PhC structure. The structure can be fabricated by forming air holes
in a thin semiconductor plate using conventional lithography and
etching processes. In the structure, owing to the periodic modula-
The obvious disadvantage of using the planar Si QW is the tion in refractive index, in-plane light propagation is governed by
small confinement factor of the optical mode in the Si QW due the photonic band structure. Strikingly, propagation is forbidden
to the thin single QW layer. It is not straightforward to make Si in photonic bandgaps (PBGs). Photonic band structures and PBGs
Multiple QWs (MQWs) (Fukatsu et al., 1992), if the surface of the can play roles to improve mainly ηextraction and ηcollection . Another
Si QW is covered with the amorphous SiO2 . As an alternative to important structure is the PhC nanocavity (Figure 5B), which is
the stacking of the Si MQWs, the Si FinLED has been proposed created by omitting air holes from the regular array. Photons are
(Saito et al., 2011b), as shown in Figure 3D. Si fin is a vertical QW confined in in-plane and out-of-plane directions due to the PBG
located perpendicular to an Si substrate, and it was proposed for effect and total internal reflection, respectively. PhC nanocavities
a self-aligned double-gate CMOS field-effect-transistor, called a have a high quality factor Q and small mode volume V c (~1 cubic
FinFET (Hisamoto et al., 2000). FinFETs are already used for mass wavelength or less). These two quantities are key parameters to
production and more than one billion of FinFETs are integrated enhance the spontaneous emission rate through the Purcell effect
in the most recent MPU (INTEL, 20131 ; ITRS, 20122 ). Therefore, (Purcell, 1946) and improve ηemission . Particularly, for light emit-
we can fabricate thousands of Si fins as MQWs at the same time ters with broad linewidth such as bulk Si, V c has a stronger impact
simply by conventional photolithography and dry etching (Saito (Ujihara, 1995). Such high-Q PhC nanocavities can uncover the
et al., 2011b). By applying forward bias to the Si FinLED, we can quantum nature of light-matter interaction. Cavity quantum elec-
observe edge emission from the Si3 N4 WG (Figure 3H). The EL trodynamics in a high-Q PhC nanocavity coupled with a single
spectra from the edge of the Si FinLED are shown in Figure 4A. semiconductor quantum dot is a hot topic in the field (see, for
The enhanced peaks from the edge of the stop band were observed example, Arakawa et al., 2012). Purcell enhancement factors of as
due to the distributed-feedback structure of the periodic fins large as 12 (Lo Savio et al., 2011) and 30 (Sumikura et al., 2014)
(Saito et al., 2011b). The non-linear increase of the EL intensity were reported, which would be limited by the emission linewidth
against the current is considered to come from stimulated emis- and the Q factor, respectively.
sion (Figure 4B), but the estimated gain of <1 cm−1 was too low to
overcome the threshold for a laser operation (Saito et al., 2011b). 4.2. ENHANCED LIGHT EMISSION FROM SILICON PHOTONIC CRYSTAL
STRUCTURES
PhC structures without cavities have been firstly applied to con-
4. APPLICATION OF PHOTONIC NANOSTRUCTURES TO
trol the light emission from crystalline Si. In 2003, Zelsmann et al.
GROUP IV MATERIALS
(2003) reported enhanced PL extraction from a 2D PhC slab fab-
4.1. CONTROL OF LIGHT EMISSION BY PHOTONIC CRYSTALS
ricated into the top Si layer of a SOI substrate at low temperature
The light emission properties of materials depend not only
(Zelsmann et al., 2003). Similar enhancements at room tempera-
on material characteristics such as the dipole moment and the
ture have been observed from arrays of Si nanoboxes (Cluzel et al.,
2006a) and rods (Cluzel et al., 2006b) formed on SOI substrates.
1 http://www.intel.com/content/www/us/en/history/museum-transistors-to- Strong light emission was observed at wavelengths corresponding
transformations-brochure.html to photonic band edges at the 0 point. Increasing the number of
2 http://www.itrs.net band edge within the emission spectrum of Si can lead to higher

luminescence intensity. This is experimentally verified by increas- an L3-type PhC nanocavity compared to a non-patterned region
ing the lattice constant of PhC so that normalized frequencies (see the inset). The L3 PhC nanocavity was also fabricated into
corresponding to the Si emission wavelengths are increased (Fujita an SOI substrate. In this sample, the BOX layer was etched out
et al., 2008). Si light-emitting diodes (LEDs) with PhC patterns in order to confine the photons strongly in the vertical direction.
have also been demonstrated (Nakayama et al., 2010a; Iwamoto The PL intensity from the cavity was much larger than that from
and Arakawa, 2012). The device schematically shown in Figure 6A the non-patterned region. In addition, sharp peaks are observed
was fabricated using a SOI substrate. Firstly, a lateral p-i-n junction only in the spectrum from the cavity. These peaks originate from
was formed into the top 200-nm-thick Si layer by area-selective the cavity resonant modes. For this particular sample with the air
implantations of boron and phosphorous ions. Then, a PhC struc- hole radius r = 0.37a, large enhancement of PL over 300 times was
ture was patterned. To keep mechanical stability and better thermal obtained for a cavity mode at 1,191 nm. As discussed in Section
conductivity, the buried-oxide (BOX) layer was not removed. An 1, this enhancement can be attributed to three factors. Detailed
SEM image of the central part of a device is shown in Figure 6A. analysis including numerical simulation indicated that ηemission is
The i-region is 5 µm in length and 250 µm in width. EL spectra improved by ~5 times (Iwamoto et al., 2007). The enhancement
from devices with different PhC periods and from a device without factors in ηemission ranging ~5−10 have been reported for Si inter-
PhC are shown in the inset of Figure 6B. EL emission increased band transition (Fujita et al., 2008) and for light emission from
as the period a increased. Figure 6B shows the integrated inten- optically active defects in Si (Lo Savio et al., 2011). The temper-
sities from these devices as a function of injected current. The ature dependence of cavity mode emission (Hauke et al., 2010;
integrated intensity from the device with a = 750 nm is ~14 times Lo Savio et al., 2011) and the dependence of PL on cavity mode
stronger than that from an unpatterned LED. This enhancement volume V c (Nakayama et al., 2012) suggest that the Purcell effect
is mainly caused by the improvement of ηextraction and ηcollection plays a role in this enhancement. The enhancement in ηemission
due to the photonic band structures as discussed above. ηemission reported so far is still too small for practical applications. How-
is also expected to be enhanced in PhC nanocavities. Figure 7 ever, this research would provide important insights for further
shows room-temperature µ-PL spectra measured at the center of development of light-emitting devices using group IV materials.
Indeed, these pioneering works have stimulated theoretical inves-
tigations, which discuss the possibility of lasing oscillation in Si
(Escalante and Martínez, 2012, 2013). Recent advances in this field
are developments of Si LEDs with PhC nanocavities (Nakayama
et al., 2011; Shakoor et al., 2013). Shakoor et al. (2013) recently
reported Si LEDs using L3-type nanocavity structure, in which
optically active defects created by hydrogen bombardment are used
as light emission centers. They carefully designed the cavity struc-
ture to improving ηcollection and obtained sharp light emission at
around 1.5 µm with a power density of 0.4 mW/cm2 . The strain-
induced dislocations (Ng et al., 2001; Kittler et al., 2013) will also
be compatible to PhC nanocavities, since the emission energies
FIGURE 5 | SEM images of a regular PhC structure with a triangular are smaller than the band gap of Si. The combination of PhC
lattice (A) and a L3-type PhC nanocavity, in which three air holes along
a 0-K direction are omitted (B).
nanocavities and defect engineering is very promising, and a wall
plug efficiency of 0.7 × 10−8 was reported (Shakoor et al., 2013).
FIGURE 6 | (A) Schematic representation of a silicon PhC LED is shown. The SEM image shows the center area of a device. (B) Integrated EL intensities for
silicon PhC LED with various periods a and for an SOI LED with a flat surface. The inset shows corresponding EL spectra at 10 mA.

used in the optical fiber communications (1.3–1.6 µm). A fur-

ther attractive feature of the tensile strain in Ge is the reduction
of energy difference in the conduction band between the direct
0 valley and indirect L-valley (e.g., Fischetti and Laux, 1996;
Wada et al., 2006; Camacho-Aguilera et al., 2012; Nama et al.,
2013; Süess et al., 2013). This feature stimulates researchers to
obtain efficient NIR light emission from tensile-strained Ge due
to the enhanced direct transition around the 0 point (e.g., Liu
et al., 2007; Lim et al., 2009). In this section, the grown-in ten-
sile strain in Ge on Si, generated due to the thermal expansion
mismatch, is described. Figure 8A shows typical ω − 2θ x-ray
diffraction (XRD) curves taken for 0.6-µm-thick Ge grown on
a 525-µm-thick Si(001) substrate with the Cu Kα radiation as
the x-ray source (0.15406 nm in wavelength). The samples were
grown by ultrahigh-vacuum chemical vapor deposition with a
source gas of GeH4 (9%) diluted in Ar. The growth tempera-
ture was 600°C, while a lower temperature of 370°C was used
at the initial stage of Ge growth (~50 nm) in order to prevent
FIGURE 7 | Room-temperature µ-PL spectra measured at the center of
the islanding, leading to Ge layers uniform in thickness (Luan
a PhC nanocavity and at a non-patterned area. The spectrum for the
latter is magnified by ten times for better viewing. et al., 1999; Ishikawa and Wada, 2010). After the growth, high-
temperature annealing was carried out for one of the samples at
800°C for 20 min. Such annealing is often performed in order to
4.3. APPLICATION OF PHOTONIC CRYSTAL STRUCTURES TO OTHER reduce the threading-dislocation density (Luan et al., 1999). In our
EMITTERS IN GROUP IV MATERIALS case, the density was reduced from 1 × 109 to 1 − 2 × 108 cm−2 . In
Erbium ions have been investigated as one of the promising Figure 8A, the peaks due to the (004) diffraction are clearly seen at
light emitters in Si. PhC nanocavities have been also applied to around 2θ ~ 66° for both of the as-grown and annealed samples.
enhance the light emission from Er ions (Wang et al., 2012; Savio It is important that the peaks were located at larger diffraction
et al., 2013). Narrowing the cavity linewidth in Er-doped silicon angles than that for unstrained Ge, indicating the reduction of
nitride PhC nanocavities has been also demonstrated under opti- out-of-plane lattice constant, i.e., the increase of in-plane lattice
cal pumping condition (Gong et al., 2010). As discussed in the constant due to the generation of tensile strain. According to the
previous sections, Ge is, at present, the most important material peak positions, the in-plane biaxial tensile strain was estimated
for future light-emitting devices in Si photonics. PhC (Nakayama to be 0.11 and 0.22% for the as-grown and annealed samples,
et al., 2010b) and PhC nanocavities (Kurdi et al., 2008; Ngo et al., respectively.
2008) have been applied to increase the light emission from bulk As mentioned above, such a tensile strain is generated in Ge
Ge. Applying advanced strain/doping engineering technologies to due to the mismatch of thermal expansion coefficient with Si.
photonic nanostructures would open a new route for boosting the As schematically shown in Figure 8B, the compressive strain in
light emission efficiency of Ge. Ge due to the 4.2% lattice mismatch should be relaxed at the
growth/annealing temperature, while the shrinkage in the Ge
lattice during the cooling should be prevented by the thick Si sub-
5. GENERATION OF TENSILE STRAIN IN Ge LAYERS strate, since Si has a smaller thermal expansion coefficient than
EPITAXIALLY GROWN ON Si SUBSTRATE that of Ge. This means that a tensile (compressive) stress/strain is
In epitaxial growth of Ge on an Si substrate, a compressive strain generated in Ge (Si), as in the bottom of Figure 8B. Taking into
in Ge, derived from the 4.2% lattice mismatch with Si, should be account the balance of forces together with the balance of moments
relaxed after growth beyond the critical thickness, while it has been in the stacked structure of Ge and Si, the tensile (compressive)
reported by one of the authors that, during the cooling from the strain in Ge (Si) is theoretically expressed as:
growth temperature to room temperature, a biaxial tensile strain
as large as 0.2% is built-in due to the thermal expansion mis- 1

Y1 t13 + Y2 t23

t1

match (Ishikawa et al., 2003, 2005; Cannon et al., 2004; Liu et al., ∈|| (Ge) = + − z1 (1)
R 6Y1 t1 (t1 + t2 ) 2
2005). It is known that the strain in semiconductors causes shifts
in band edge energies, e.g., de Walle, 1989, modifying the gap ener-
Y1 t13 + Y2 t23

1 t2
gies, i.e., properties of optical transitions. The 0.2% tensile strain ∈|| (Si) = & − − − z2 , (2)
in Ge reduces the direct bandgap energy from 0.80 to ~0.77 eV, R 6Y2 t2 (t1 + t2 ) 2
and as a result, the optical absorption edge (or the longer limit of
detection wavelength) shifts from 1.55 to >1.60 µm, causing the where, α i , Yi , ti , and zi represent the thermal expansion coefficient,
increase of optical absorption coefficient at 1.55 µm (Ishikawa the Young’s modulus, the layer thickness, and the location in the
et al., 2003, 2005; Cannon et al., 2004; Liu et al., 2005). This layer measured from the bottom of the layer for the i-th layer (1
property is effective for the detection of near-infrared (NIR) light for Ge and 2 for Si), respectively. The radius of curvature R is

FIGURE 8 | (A) ω − 2θ XRD curves for 0.6-µm-thick Ge on Si (001) substrate, (B) schematic illustration showing the generation of tensile stress/strain in Ge, and
(C) theoretical curves and experimental data for biaxial tensile strain in Ge.
represented from the Ge layer as well as deposition of dielectric films embedding a

strain could intentionally modify the strain in the Ge.
R TRT
1 6 (t1 + t2 ) Y1 Y2 t1 t2 TGR/AN (α1 − α2 ) dT
= , (3) 6. DIRECT GERMANIUM EPITAXIAL GROWTH PROCESS ON
3(t1 + t2 ) Y1 Y2 t1 t2 + Y1 t1 + Y2 t23 (Y1 t1 + Y2 t2 )
2 3

R
SILICON
The Ge was epitaxially grown by using a cold-wall rapid ther-
where T GR/AN and T RT represent the growth/annealing tempera- mal chemical vapor deposition system. Germane (GeH4 ) was
ture before cooling and the room temperature (after the cooling), used as a source gas, which was supplied with H2 carrier gas.
respectively. Since the first term is dominant in the right side of As the starting point of improving the crystallinity and control-
equation (1), the strains are almost independent of z i , the location ling the lattice strain, Ge layers with good surface morphology
within the layer. Therefore, equations (1) and (2) are simplified to: were grown at 420°C under relatively high pressure of 7,000 Pa.
Then, the Ge layers were annealed in the same H2 atmosphere
to improve the crystallinity. Figure 9 shows a reciprocal space
1 Y1 t13 + Y2 t23 map (RSM) of XRD (XRD-RSM) from the 130-nm-thick Ge layer
∈|| (Ge) ∼ (4)
R 6Y1 t1 (t1 + t2 ) directly grown on the Si substrate before and after H2 anneal-
ing. An intense Si (-1-13) peak was observed, which represented
1 Y1 t13 + Y2 t23 the diffraction from the Si substrate under the Ge layer. Since the
∈|| (Si) ∼ − . (5)
R 6Y2 t2 (t1 + t2 ) XRD-RSM was measured by using semiconductor array detec-
tors, errors in the counts occur if the diffraction intensity is
The lines in Figure 8C represent the strains calculated for the Ge very high; therefore, the streak line observed around the Si (-1-
thickness of 0.6 µm and the Si thickness of 525 µm. Note that 13) peak does not represent any actual diffraction. Since a Ge
almost identical results can be obtained when the thickness of Si (-1-13) diffraction peak was observed from the Ge layer with-
substrate t 2 is much larger (more than ~100 times) than the Ge out annealing (Figure 9A), it could be confirmed that a single
thickness t 1 . The parameters used in the calculation can be found crystalline Ge layer was obtained by using low-temperature epi-
in Ishikawa et al. (2005). It is found that a tensile strain on the taxial growth. The displacement of the diffraction peak shows that
order of 0.1% is generated in Ge at room temperature, while the the as-grown Ge layer still contained a compressive strain just
compressive strain in Si is negligible. It is also found that higher after the low-temperature epitaxial growth at 420°C due to the
growth/annealing temperature generates larger tensile strain after larger lattice constant of Ge compared to that of the Si substrate.
the cooling. These properties are qualitatively in good agreement It has been reported that cyclic annealing at a relatively higher
with the XRD results in Figure 8A. However, quantitatively, the temperature can reduce the threading-dislocation density (Luan
tensile strain observed by XRD was smaller than the theoretical et al., 1999) in Ge layers. This has led to studies on the effect
one. This is probably ascribed to the residual compressive strain of annealing on the crystallinity and lattice strain of Ge layers.
in Ge at the growth/annealing temperature (Ishikawa et al., 2005). After low-temperature epitaxial growth of Ge layers at 420°C, the
From the viewpoint of optoelectronic integration of Ge devices temperature was increased to the annealing temperature in the
on an Si platform, Si-on-insulator (SOI) wafers have been widely same H2 atmosphere as that during the epitaxial growth, and the
used. For Ge layers grown on SOI wafers, a similar amount of Ge layers were then annealed at various temperatures for 10 min.
tensile strain should be generated, since the elastic deformation, XRD-RSMs of Ge layers annealed at a temperature (T GR/AN ) of
derived from the thermal expansion mismatch, is governed by the 700°C after the low-temperature epitaxial growth are shown in
thick Si substrate, rather than the buried SiO2 and the top Si layers Figure 9B. The Ge (-1-13) diffraction peaks became much steeper
with the thicknesses on the order of 1 µm or below. Patterning of and the peak intensity increased when the annealing temperature

for this experiment, so that <001> is perpendicular to the sur-

face of the Ge film, while <110> is in the plane of Ge film.
The lattice strain in the <110> crystal orientation increased as
T GR/AN increased, and the strain in the <001> crystal orien-
tation showed an opposite dependence. Although the Ge layer
contained a compressive strain in the <110> crystal orienta-
tion at T GR/AN = 420°C, i.e., without annealing, this strain started
decreasing when T GR/AN was increased, and the Ge was completely
un-strainedat T GR/AN = 530°C. Furthermore, the sign of the lat-
tice strain changed from compressive to tensile after annealing at
T GR/AN > 530°C, and the tensile strain at T GR/AN = 700°C reached
0.19%. This result is consistent with previous studies (Cannon
et al., 2004). Normally, a grown layer with a larger lattice constant
compared to a substrate contains a compressive strain within the
growth plane. However, since the Ge layers grown on the Si sub-
strate were almost completely relaxed even after low-temperature
growth, the Ge lattice could be dislocated at the Ge/Si interface by
post-annealing, and the lattice strain of the Ge layer was relaxed
FIGURE 9 | XRD-RSM of (-1-13) diffraction from Ge layer grown on Si during annealing at the relatively higher temperature with the
substrate, (A) after low-temperature epitaxial growth and (B) after volumes of Ge and Si determined by the thermal expansion coeffi-
post-annealing. cients (Singh, 1968; Okada and Tokumaru, 1984). After annealing,
the volume of the Ge layer and the Si substrate both shrunk as
the temperature decreased, and there was barely any change to the
lattice alignment at low-temperature. The volume of the Si sub-
strate returned to its original value because it was thick enough.
However, the volume of the Ge layer could not return due to its
larger thermal expansion coefficients. Therefore, the tensile lattice
strain remained only in the Ge layers after cooling (Cerdeira et al.,
1972). The ideal lattice strain in <110> crystal orientation was also
plotted in the inset of Figure 10, which was calculated with only
the difference of the thermal expansion coefficients between Si
and Ge, so these values indicate the maximum lattice strain. Since
there are large discrepancies between calculation and measured
values, it seems that relaxation ratio has a large effect on the lattice
strain even at the lower temperatures. PL spectra from the post-
annealed Ge layers with various annealing temperatures are shown
in Figure 10. Although Ge is an indirect bandgap material and the
L-valley has the lowest energy level in the conduction band, we
were able to observe recombination between electrons and holes
at the 0-valley as luminescence at a wavelength of 1,550 nm, even
from the bulk Ge (dashed line in Figure 10). A comparison with
the post-annealed Ge layers shows that although the spectrum was
very weak and broad for the as-grown Ge layer, an obvious peak
FIGURE 10 | Photoluminescence spectra from Ge layers annealed with could be observed from annealed samples at T GR/AN > 530°C.
different temperatures are shown. Peak wavelength of Moreover, the PL intensity increased and the peak shape became
photoluminescence from Ge layers red-shifted as annealing temperature sharper as the annealing temperature was increased. The PL spec-
increased, consistent with temperature induced tensile strain. Inset shows
trum is strongly affected by crystallinity, because non-radiative
lattice strain of Ge layers grown on Si substrate along <001> and <110>
crystal orientations as a function of annealing temperature. Dotted line
recombination was significantly increased with defects such as
indicates lattice strain calculated with difference between thermal dislocation and stacking faults. Therefore, these results suggest
expansion coefficients of Si and Ge. that the crystallinity of the Ge layers was improved by the post-
annealing. The peak was observed at a shorter wavelength from
the Ge layer annealed at 500°C compared with that from bulk Ge,
was increased, indicating that the crystallinity of the Ge layers was and a red shift of the PL peaks occurred after post-annealing at
increased by the post-annealing. a higher temperature. In addition, the peak wavelength from the
The inset of Figure 10 shows the lattice strain in the Ge unstrained Ge was 1,550 nm, which is almost the same value as that
layers in the <001> and <110> crystal orientations as a func- of the bulk Ge. These results show that the bandgap energy at the
tion of the annealing temperature. We used standard Si wafers 0-point was varied by the lattice strain in the Ge layers (Cerdeira

et al., 1972; de Walle and Martin, 1986; de Walle, 1989), which the possibility of using RMG to obtain high quality Ge crystalline
is consistent with the XRD measurements. These results indicate layers to create a bridge between electronic components and pho-
that, in the range of this study, the most favorable PL characteristic tonic components. This vision is clearly demonstrated by Going
can be obtained from the Ge layer after post-annealing at higher et al. (2014) in a Ge Gate PhotoMOSFET (Carroll et al., 2012)
temperatures. where a Ge-gated NMOS phototransistor is integrated on an Si
photonics platform on SOI substrate. The resulting device, with
7. GERMANIUM LIQUID-PHASE EPITAXY AND DEVICES 1-µm channel length, and 8-µm channel width, demonstrates a
FOR PHOTONIC APPLICATION responsivity of over 18 A/W at 1550 nm with 583 nW of incident
Liquid-phase epitaxy (LPE) is a technique that was invented in light. By increasing the incident power to 912 µW, the device oper-
the 1960s (Nelson, 1963) and developed in the 1970s (Wieder ates at 2.5 GHz. Ge RMG or LPE on Si is therefore a promising
et al., 1977) for the fabrication of detectors, solar cells, LEDs technology for the fabrication of heterogeneous devices requiring
(Saul and Roccasecca, 1973), and laser diodes (Panish et al., 1970). high quality Ge layers such as MOSFETs, near-infrared detectors
Originally used for III-V crystal growth, it has been adapted for but also Ge-based lasers that are still to be demonstrated using this
SiGe-on-insulator (SGOI) and Ge-on-Insulator (GOI) growth by specific process technique. In fact, a highly tensile strain of 0.4%
various groups (Liu et al., 2004; Tweet et al., 2005; Feng et al., 2008; has successfully been applied to a Ge film grown by RMG process
Hashimoto et al., 2009; Miyao et al., 2009; Ohta et al., 2011) and is (Matsue et al., 2014), which is quite promising for light emission.
also referred to as rapid melt growth (RMG). The GOI technique
was pioneered by Liu et al. (2004) for Ge-on-insulator fabrication. 8. TIME-RESOLVED PHOTOLUMINESCENCE STUDY OF
In this technique, a thin insulating layer is deposited on an Si sub- GERMANIUM ON SILICON
strate and patterned to open up seed windows. The target material, The use of n-type tensile-strained Ge grown on Si substrates is one
in this case Ge, is deposited using a non-selective method and pat- promising way to realize an efficient light source for Si photon-
terned to form the desired features. This is then encapsulated using ics through the enhanced direct recombination from the 0 valley.
an insulating layer and heated up in a rapid-thermal-annealer However, the large lattice mismatch between Ge and Si inherently
(RTA) in order to melt the Ge. The micro-crucible holds the melt in causes misfit dislocations at the interface, and threading disloca-
place until the liquid epitaxial growth is complete. Upon cooling, tions during the growth. Besides, epitaxially grown Ge is usually a
liquid-phase epitaxial growth starts from the seed and propagates thin layer, so that both the interface and the surface become impor-
to the extremities of the strip structure. For the realization of sin- tant. Therefore, investigation of the excess carrier lifetime is crucial
gle crystal Ge, epitaxial growth must proceed faster than unseeded for the realization of efficient light-emitting devices. Recently,
random nucleation, so that the crystal regrowth starting from the the excess carrier dynamics of thin Ge film grown on either Si
seed is uninterrupted. Misfit dislocations arising at the SiGe inter- or SOI substrates have been investigated by time-resolved pho-
face in the seed area are necked down to the seed window as shown toluminescence (Kako et al., 2012), microwave photoconductive
in Figure 11. The RMG is limited to the growth of structures of the decay (Sheng et al., 2013), and pump-probe transmission (Geiger
order of around 3 µm in width and with a length of above 100 µm. et al., 2014) methods. Here, we present the time-resolved photo-
The limitation is largely due to the surface tension of the insulator luminescence study of both non-doped and n-type Ge samples
causing the Ge to form ball shapes while in the liquid phase. grown on Si.
RMG is very attractive for the heterogeneous integration of Ge- The Ge samples were epitaxially grown on (100) Si substrates by
based devices on insulator for electronics and photonics and has using a cold-wall rapid thermal chemical vapor deposition system
been demonstrated for Gate all around P-MOSFET (Feng et al., (Oda et al., 2014). There were two primary growth steps. The first
2008), P-Channel FinFET (Feng et al., 2007), waveguide integrated step was the growth of an intrinsic Ge thin layer (≈100 nm) at low
Ge/Si heterojunction photodiodes (Tseng et al., 2013), or Ge Gate temperature followed by an annealing process. The second step
PhotoMOSFET (Going et al., 2014). These devices demonstrate was the regrowth of Ge on the first layer with another annealing
FIGURE 11 | Transmission-electron-microscope (TEM) image of a high quality single crystalline Ge-on-insulator obtained using RMG. It can clearly be
seen that the misfit dislocations from the lattice mismatch are confined to the seed region and that the crystalline Germanium lateral overgrowth is free from
defects.

process. In situ n-type doping was carried out during the second thickness. The excess carrier lifetimes obtained for undoped Ge
growth step by supplying phosphine. The Ge becomes biaxially layers with different thicknesses (filled black circles) are shown in
strained (≈0.15%) due to the difference of the thermal expansion the inset of Figure 12A together with the black curve, which is a
coefficients between Si and Ge. Time-resolved photoluminescence fit to the data using equation (6) with parameters of τ B = 3.5 ns,
measurements were performed using a time-correlated single- S = 5.5 × 103 cm/s, and D = 30 cm2 /s (The ambipolar diffusion
photon counting method employing a superconducting single- constant Da could be estimated by changing the spot size and mea-
photon detector (SSPD) with a time resolution of about 50 ps. A suring the photoluminescence decay time). Both SRH bulk recom-
Ti:Sapphire pulsed laser was used as the excitation source (wave- bination and the surface recombination processes determine the
length 710 nm, repetition rate 80 MHz, and pulse-duration 100 fs). excess carrier dynamics in our undoped Ge samples.
The laser beam was focused on the sample surface using an objec- Figure 12B shows time-resolved photoluminescence decay
tive lens. The photoluminescence from the samples was collected curves measured at two different excitation power densities from
by the same objective and focused on to an optical fiber connected an n-type Ge sample (thickness 500 nm, doping concentration
to the SSPD. Photoluminescence ranging from 1.2 to 1.8 µm was 7 × 1019 cm−3 ). The measured decay depends on both the exci-
detected. tation power density and time (in contrast to those measured
Figure 12A shows a time-resolved photoluminescence decay from undoped samples, which are independent of the excita-
curve measured from a nominally undoped Ge sample (thickness tion power). The instantaneous lifetime (that measured at a
500 nm). In order to limit the effects of lateral diffusion, the laser particular point during the decay) depends on the photolumi-
spot size was set to ≈10 µm. The decay is a single exponential with nescence intensity, and thus the excess carrier density. Based
a lifetime of 1 ns, which corresponds to the excess carrier lifetime on the SRH non-radiative recombination model, the lifetime of
of 2 ns. Germanium has an indirect bandgap, and as such, its excess excess carriers depends on their density (Linnros, 1998). This
carrier dynamics are determined by non-radiative recombination dependence can be simplified to τ hl = τ n + τ p (τ ll = τ p for n-
processes, such as Shockley-Read Hall (SRH) recombination and type doping) in the two extreme conditions where the carrier
surface recombination processes. The photoluminescence decay density is high (low) when compared to the doping concentra-
lifetime, τ PL , of undoped Ge is then related to the excess carrier tion (τ n and τ p are the inverse capture rates of the electrons
lifetime, τ ex , as 2τ PL = τ ex . The lifetime of excess carriers τ ex of and holes, respectively). The photoluminescence lifetime can be
an indirect semiconductor film depends on the thickness and can expressed as 2τ PL = τ hl (high excess carrier density) and τ PL = τ ll
be represented by Sproul (1994) and Gaubas and Vanhellemont (low excess carrier density). Therefore, from our measurements,
(2006) as: we estimate τ ll = 0.14 ns, τ hl = 0.8 ns based on SRH theory. The
estimated τ hl value is shorter than those found from the undoped
samples. This difference might be attributed to an increased
1 1 1 dislocation density introduced by the doping, but the estima-
= + , (6)
τex τB d
+ d2 tion of τ hl could be underestimation because the Auger process
2S π 2D
becomes important for doped samples (Gaubas and Vanhelle-
where τ B is the bulk lifetime, S is the surface recombination veloc- mont, 2006). Further investigation is needed in order to obtain
ity, D is the ambipolar diffusion constant, and d is the layer a better understanding.
9. ELECTRO-LUMINESCENCE FROM GERMANIUM

Realization of monolithic light sources compatible with the exist-
ing Si photonics platform is one of the most difficult challenges.
Ge has attracted much attention as for possible future monolithic
light sources owing to its emission wavelengths of ~1.6 µm suitable
for an Si-based WG, in addition to the CMOS compatibility and
the pseudo-direct band-gap character (Menéndez and Kouvetakis,
2004; Liu et al., 2007, 2012; Liang and Bowers, 2010; Michel et al.,
2010; Boucaud et al., 2013; Liu, 2014). Recently, laser operation
from Ge pumped optically (Liu et al., 2010) and electrically (Cheng
et al., 2007; Camacho-Aguilera et al., 2012) has been reported.
However, there is no report so far to reproduce their results. The
optical gain from Ge is also achieved by the tensile-stress engineer-
ing (de Kersauson et al., 2011). The precise nature of the optical
gain in Ge is still controversial (Carroll et al., 2012), but the high
crystalline quality of Ge is one of the most critical factor to avoid
FIGURE 12 | (A) Time-resolved PL curve for an undoped Ge sample. The non-radiative recombinations at dislocations. It is confirmed by
inset shows the measured excess carrier lifetimes for two Ge thicknesses several groups (Michel et al., 2010; Liu et al., 2012; Boucaud et al.,
with simulated lifetimes using the equations shown in the text. 2013; Liu, 2014) that the primary challenges for engineering Ge
(B) Time-resolved PL curves of an n-type sample for two excitation powers,
150 kW/cm2 (black line) and 15 kW/cm2 (red line).
as an active layer are: (i) crystallinity, (ii) high n-type doping, (iii)
tensile strain, as confirmed theoretically (Suwa and Saito, 2010,

2011; Virgilio et al., 2013a,b). Here, we review some of the Ge light loss due to free carrier absorption in the diffusion electrodes. The
sources developed on SOI substrates. Ge waveguide was doped with 1 × 1019 cm−3 of phosphorus, and
the surface of the Ge waveguide was then passivated with GeO2
9.1. DEVICE STRUCTURE AND FABRICATION PROCESS formed by low-temperature oxidation to reduce interfacial traps
As we discussed in section for Si light sources, lateral carrier injec- (Tani et al., 2012, 2013a). Then, metal electrodes were made on
tion is a natural choice for electrical pumping, since fabrication both diffusion regions.
processes are based on planar CMOS technologies. We show sev- To enhance light emission efficiency from Ge by tensile stress,
eral candidates for Ge light sources suitable for lateral carrier several techniques have been developed, e.g., the use of the thermal
injection in Figure 13. expansion of relaxed Ge grown on Si (Ishikawa et al., 2003), the
Figures 13A,E,I show schematic views and a transmission- growth on buffer layers with larger lattice parameter (Huo et al.,
electron-microscope (TEM) image of a Ge FinLED (Saito et al., 2011), the mechanical deformation using membrane structures
2011a), which uses Ge fins as MQWs embedded in Si3 N4 WG. (Kurdi et al., 2010), the stress concentration in a membrane struc-
Ge fins were fabricated by the oxidation condensation technique ture (Nama et al., 2013), and using external stressors (Ortolland
(Tezuka et al., 2009) applied to SiGe fins (Saito et al., 2011a). Rel- et al., 2009; Ghrib et al., 2013). Considering the process compati-
atively, high crystallinity is expected in Ge fins, since the lattice bility to the lateral carrier injection, the Si3 N4 film with the tensile
mismatch between Si and Ge would be relaxed by stretching the stress of 250-MPa was employed (Tani et al., 2013a), as shown in
fins during the oxidation (Saito et al., 2011a). In fact, the low dark Figures 13D,H,L.
current density of 1.86 × 10−5 A/cm−2 at a reverse bias of 1 − V
and the strong breakdown current density of >1 MA/cm−2 were
confirmed (Saito et al., 2011a). 9.2. IMPACT OF STRESS ENGINEERING FOR LATERAL GERMANIUM
In order to enhance the overlap between an optical mode ON SILICON DIODE
and fins, Ge fins with (111) orientation at the sidewall were also Figure 14A shows EL spectra of the Ge waveguide with 500-nm
developed (Tani et al., 2012), as shown in Figures 13B,F,J. To width and 500-µm length taken from the top of the substrate
improve the patterning accuracy, Si (111) fins were fabricated under continuous current injection of 60-mA. EL peak wave-
by anisotropic wet etching, and n-Ge was re-grown after the length of the device with an SiN stressor is slightly longer than
condensation oxidation of SiGe fins (Tani et al., 2011). that without the SiN stressor due to the tensile strain-induced
Further increase of the coupling is realized by using a bulk band-gap shrinkage, although the exact band-gap energy can-
Ge WG (Liu et al., 2007; Camacho-Aguilera et al., 2012; Tani et al., not be quantitatively estimated due to the additional peak shifts
2013a,b), as shown in Figures 13C,G,K for schematic views and the caused by heating under high currents. Moreover, as shown in
scanning electron microscope (SEM) image, rather than using Ge Figure 14B, the peak intensity of the EL of the device with
QW or Ge fins. The p- and n-type diffusion regions were formed SiN stressor is 1.65 times larger than that without SiN stressors.
in the 40 nm-thick SOI layer, and the Ge waveguide with 500- Figure 14C shows two-dimensional stress mapping calculated by
nm width and 500-µm length was directly grown on the SOI a finite element modeling of the Ge waveguide on the Si sub-
diode. The SOI thickness was designed to minimize the optical strate covered by Si3 N4 stressor. The tensile stress of 100 MPa
FIGURE 13 | Development of a Ge light source. (A) i -Ge FinLED, (B) n-Ge FinLED, (C) n-Ge-WG-on-Si LED without SiN, and (D) n-Ge-WG-on-Si LED with SiN.
(A–D) Cross section, (E–H) plan views, and (I–L) microscope images.

FIGURE 14 | Strain engineering for n-Ge-WG-on-Si LED. (A) Spectra and (B) integrated intensity from experiments. (C) Stress mapping simulation.
is localized on the side wall of the Ge waveguide, while the in- ACKNOWLEDGMENTS
plane compressive stress of 40 MPa exists on the top part of the We would like to thank research collaborators, engineers, and
Ge waveguide. The increase of the light emission efficiency was line managers in Hitachi, the University of Tokyo, and Uni-
22% caused by the tensile stress, after subtracting of the addi- versity of Southampton for supporting this project. We are
tional increase of 35% caused by the light extraction efficiency also grateful to Prof. H. N. Rutt for his careful reading of
due to the reduced reflectance at the surface of the Ge waveguide the manuscript and constructive comments. Funding: parts of
by the 500 nm-thick Si3 N4 layer (Tani et al., 2013a). Therefore, the the studied discussed here was supported by Japan Society for
stress engineering by Si3 N4 is an appropriate option to improve the Promotion of Science (JSPS) through its “Funding Pro-
the performance of Ge light sources. Recently, there are significant gram for World-Leading Innovation R&D on Science and Tech-
advances in stress engineering by manipulating free-standing Ge nology (FIRST Program),” the Project for Developing Innova-
structures (Jain et al., 2012; Boztug et al., 2013; Süess et al., 2013; tion Systems, and Kakenhi 216860312, MEXT, Japan. This work
Sukhdeo et al., 2014), and enhanced direct recombination has been is also supported by EU, FP7, Marie-Curie, Carrier Integra-
achieved. tion Grant (CIG), PCIG13-GA-2013-618116, and University of
Southampton, Zepler Institute, Research Collaboration Stimu-
10. CONCLUSION AND FUTURE OUTLOOK lus Fund.
In this paper, we reviewed the recent progress on the developments
of silicon and germanium light sources. There are many process REFERENCES
options to fabricate silicon- and germanium-based nanostructures Arakawa, Y., Iwamoto, S., Nomura, M., Tandaechanurat, A., and Ota, Y. (2012).
Cavity quantum electrodynamics and lasing oscillation in single quantum dot-
by using modern silicon technologies. For active materials, pla- photonic crystal nanocavity coupled systems. IEEE J. Selec. Top. Quant. Elec. 18,
nar silicon single-quantum-well (Saito et al., 2006a,b, 2008, 2009; 1818–1829. doi:10.1109/JSTQE.2012.2199088
Hoang et al., 2007; Noborisaka et al., 2011; Saito, 2011) or multiple- Arakawa, Y., Nakamura, T., and Urino, Y. (2013). Silicon photonics for next genera-
quantum-wells made of silicon or germanium fins (Saito et al., tion system integration platform. Commun. Mag. IEEE 51, 72–77. doi:10.1109/
MCOM.2013.6476868
2011a,b) can be used. To enhance the recombination rates and
Arakawa, Y., and Sakaki, H. (1982). Multidimensional quantum well laser and tem-
the extraction efficiencies, photonic crystal structures have been perature dependence of its threshold current. Appl. Phys. Lett. 40, 939–941.
introduced (Fujita et al., 2008; Nakayama et al., 2010a; Iwamoto doi:10.1063/1.92959
and Arakawa, 2012). The further increase of the efficiency can Boucaud, P., Kurdi, M. E., Ghrib, A., Prost, M., de Kersauson, M., Sauvage, S.,
be achieved by introducing tensile strain and n-type doping of et al. (2013). Recent advances in germanium emission. Photon. Res. 1, 102–109.
doi:10.1364/PRJ.1.000102
the germanium (Ishikawa et al., 2003; Menéndez and Kouvetakis,
Boztug, C., Sanchez-Perez, J. R., Yin, J., Lagally, M. G., and Paiella, R. (2013).
2004; Liu et al., 2007, 2012; Kurdi et al., 2010; Michel et al., 2010; Graiting-coupled mid-infrared light emission from tensile strained germanium
Huo et al., 2011; Boucaud et al., 2013; Ghrib et al., 2013; Nama nanomembranes. Appl. Phys. Lett. 103, 201114. doi:10.1002/smll.201201090
et al., 2013; Tani et al., 2013a; Liu, 2014). Camacho-Aguilera, R. E., Cai, Y., Patel, N., Bessette, J. T., Romagnoli, M., Kimerling,
Considering the success of the laser operation using the bulk L. C., et al. (2012). An electrically pumped germanium laser. Opt. Express 20,
11316–11320. doi:10.1364/OE.20.011316
germanium waveguides (Liu et al., 2010; Camacho-Aguilera et al., Canham, L. T. (1990). Silicon quantum wire array fabrication by electrochem-
2012), the next step will be to reduce the threshold current for ical and chemical dissolution of wafers. Appl. Phys. Lett. 57, 1046–1048.
pumping. It is critical to develop a process technology to fabricate doi:10.1063/1.103561
a high crystalline quality germanium quantum-well compatible Cannon, D. D., Liu, J., Ishikawa, Y., Wada, K., Danielson, D. T., Jongthammanu-
with the silicon photonics platform. If practical silicon or ger- rak, S., et al. (2004). Tensile strained epitaxial Ge film on Si(100) substrate with
potential application to L-band telecommunications. Appl. Phys. Lett. 84, 906.
manium laser diodes are available in the future, these group IV doi:10.1063/1.1645677
lasers will realize the convergence of electronics and photonics on Carroll, L., Friedli, P., Neunschwander, S., Sigg, H., Cecchi, S., Isa, F., et al. (2012).
a silicon chip. Direct-gap gain and optical absorption in germanium correlated to the density

of photoexcited carriers, doping, and strain. Phys. Rev. Lett. 109, 057402. Gelloz, B., Kojima, A., and Koshida, N. (2005). Highly efficient and stable lumi-
doi:10.1103/PhysRevLett.109.057402 nescence of nanocrystalline porous silicon treated by high-pressure water vapor
Cerdeira, F., Buchenauer, C. J., Pollak, F. H., and Cardona, M. (1972). Stress-induced annealing. Appl. Phys. Lett. 87, 031107. doi:10.1063/1.2001136
shifts of first-order Raman frequencies of diamond- and zinc-blende-type semi- Gelloz, B., and Koshida, N. (2000). Electroluminescence with high and stable quan-
conductors. Phys. Rev. B 5, 580–593. doi:10.1103/PhysRevB.5.580 tum efficiency and low threshold voltage from anodically oxidized thin porous
Cheng, T. H., Kuo, P. S., Lee, C. T., Liao, M., Hung, T. A., and Liu, C. W. (2007). silicon diode. J. Appl. Phys. 88, 4319–4324. doi:10.1063/1.1290458
“Electrically pumped Ge laser at room temperature,” in IEEE Int. Conf. Electron Ghrib, A., Kurdi, M. E., de Kersauson, M., Prost, M., Sauvage, S., Checoury, X., et al.
Devices Meeting (IEDM), Washington, 659–662. (2013). Tensile-strained germanium microdisks. Appl. Phys. Lett. 102, 221112.
Cluzel, B., Pauc, N., Calvo, V., Charvolin, T., and Hadji, E. (2006a). Nanobox array doi:10.1063/1.4809832
for silicon-on-insulator luminescence enhancement at room temperature. Appl. Going, R. W., Loo, J., Liu, T.-J. K., and Wu, M. C. (2014). Germanium gate Pho-
Phys. Lett. 88, 133120. doi:10.1063/1.2191089 toMOSFET integrated to silicon photonics. IEEE J. Selec. Top. Quant. Elec. 20,
Cluzel, B., Calvo, V., Charvolin, T., Picard, E., Noé, P., and Hadji, E. (2006b). Single- 8201607. doi:10.1109/JSTQE.2013.2294470
mode room-temperature emission with a silicon rod lattice. Appl. Phys. Lett. 89, Gong, Y., Makarova, M., Selçuk Yerci, R. L., Stevens, M. J., Baek, B., Nam, S. W.,
201111. doi:10.1063/1.2364876 et al. (2010). Linewidth narrowing and Purcell enhancement in photonic crystal
Cullis, A. G., Canham, L. T., and Calcott, P. D. J. (1997). The structural and lumi- cavities on an Er-doped silicon nitride platform. Opt. Express 18, 2601–2612.
nescence properties of porous silicon. J. Appl. Phys. 82, 909–965. doi:10.1063/1. doi:10.1364/OE.18.002601
366536 Gunn, C. (2006). Cmos photonics for high-speed interconnects. Micro IEEE 26,
Daldosso, N., and Pavesi, L. (2009). Nanosilicon photonics. Laser Photon. Rev. 3, 58–66. doi:10.1109/MM.2006.32
508–534. doi:10.1002/lpor.200810045 Hashimoto, T., Yoshimoto, C., Hosoi, T., Shimura, T., and Watanabe, H. (2009). Fab-
de Kersauson, M., Kurdi, M. E., David, S., Checoury, X., Fishman, G., Sauvage, S., rication of local Ge-on-insulator structures by lateral liquid-phase epitaxy: effect
et al. (2011). Optical gain in single tensile-strained germanium photonic wire. of controlling interface energy between Ge and insulators on lateral epitaxial
Opt. Express 19, 17925–17934. doi:10.1364/OE.19.017925 growth. Appl. Phys. Express 2, 066502. doi:10.1143/APEX.2.066502
de Walle, C. G. V. (1989). Band lineups and deformation potentials in the model- Hauke, N., Zabel, T., Müller, K., Kaniber, M., Laucht, A., Bougeard, D., et al.
solid theory. Phys. Rev. B 39, 1871. doi:10.1103/PhysRevB.39.1871 (2010). Enhanced photoluminescence emission from two-dimensional silicon
de Walle, C. G. V., and Martin, R. M. (1986). Theoretical calculations of het- photonic crystal nanocavities. New J. Phys. 98, 053005. doi:10.1088/1367-2630/
erojunction discontinuities in the Si/Ge system. Phys. Rev. B 34, 5621–5634. 12/5/053005
doi:10.1103/PhysRevB.34.5621 Hisamoto, D., Lee, W., Kedzierski, J., Takeuchi, H., Asano, K., Kuo, C., et al. (2000).
Deen, M. J., and Basu, P. K. (2012). Silicon Photonics Fundamentals and Devices. West FinFET-a self-aligned double-gate MOSFET scalable to 20 nm. IEEE Trans. Elec-
Sussex: Wiley. tron Devices 47, 2320–2325. doi:10.1109/16.887014
Escalante, J. M., and Martínez, A. (2012). Theoretical study about the gain in indi- Hoang, T., LeMinh, P., Holleman, J., and Schmitz, J. (2007). Strong efficiency
rect bandgap semiconductor optical cavities. Physica B Condens. Matter 407, improvement of SOI-LEDs through carrier confinement. IEEE Electron Device
2044–2049. doi:10.1016/j.physb.2012.02.002 Lett. 28, 383–385. doi:10.1109/LED.2007.895415
Escalante, J. M., and Martínez, A. (2013). Optical gain by simultaneous photon Huo, Y., Lin, H., Chen, R., Makarova, M., Rong, Y., Li, M., et al. (2011). Strong
and phonon confinement in indirect bandgap semiconductor acousto-optical enhancement of direct transition photoluminescence with highly tensile-
cavities. Opt. Quant. Electron. 45, 1045–1056. doi:10.1007/s11082-013-9715-z strained Ge grown by molecular beam epitaxy. Appl. Phys. Lett. 98, 011111.
Evans, P., Fisher, M., Malendevich, R., James, A., Studenkov, P., Goldfarb, G., et al. doi:10.1063/1.3534785
(2011). “Multi-channel coherent PM-QPSK InP transmitter photonic integrated Ishikawa, Y., and Wada, K. (2010). Germanium for silicon photonics. Thin Solid
circuit (PIC) operating at 112 gb/s per wavelength,” in Optical Fiber Communi- Films 518, S83–S87. doi:10.1016/j.tsf.2009.10.062
cation Conference (OFC) (Los Angeles: (IEEE) PDPC, OSA). Ishikawa, Y., Wada, K., Cannon, D. D., Liu, J., Luan, H.-C., and Kimerling, L. C.
Fang, A. W., Park, H., Cohen, O., Jones, R., Paniccia, M. J., and Bowers, J. E. (2006). (2003). Strain-induced band gap shrinkage in Ge grown on Si substrate. Appl.
Electrically pumped hybrid AlGaInAs-silicon evanescent laser. Opt. Express 14, Phys. Lett. 82, 2044. doi:10.1063/1.1564868
9203–9210. doi:10.1364/OE.14.009203 Ishikawa, Y., Wada, K., Liu, J., Cannon, D. D., Luan, H.-C., Michel, J., et al. (2005).
Fathpour, S., and Jalali, B. (eds) (2012). Silicon Photonics for Telecommunications and Strain-induced enhancement of near-infrared absorption in Ge epitaxial layers
Biomedicine. Boca Raton: CRC Press. grown on Si substrate. J. Appl. Phys. 98, 013501. doi:10.1063/1.1943507
Feng, J., Thareja, G., Kobayashi, M., Chen, S., Poon, A., Bai, Y., et al. (2008). High- Iwamoto, S., and Arakawa, Y. (2012). Enhancement of light emission from silicon
performance gate-all-around GeOI p-MOSFETs fabricated by rapid melt growth by utilizing photonic nanostructures. IEICE Trans. Electron E95-C, 206–212.
using plasma nitridation and ALD Al2 O3 gate dielectric and self-aligned NiGe doi:10.1587/transele.E95.C.206
contacts. IEEE Electron Device Lett. 29, 805–807. doi:10.1109/LED.2008.2000613 Iwamoto, S., Arakawa, Y., and Gomyo, A. (2007). Observation of enhanced photolu-
Feng, J., Woo, R., Chen, S., Liu, Y., Griffin, P. B., and Plummer, J. D. (2007). P-channel minescence from silicon photonic crystal nanocavity at room temperature. Appl.
germanium FinFET based on rapid melt growth. IEEE Electron Device Lett. 28, Phys. Lett. 91, 211104. doi:10.1063/1.2816892
637–639. doi:10.1109/LED.2007.899329 Jain, J. R., Hryciw, A., Baer, T. M., Miller, D. A. B., Brongersma, M. L., and Howe, R.
Fischetti, M. V., and Laux, S. E. (1996). Band structure, deformation potentials, T. (2012). A micromachining-based technology for enhancing germanium light-
and carrier mobility in strained Si, Ge, and SiGe alloys. J. Appl. Phys. 80, 2234. emission via tensile strain. Nat. Photonics 6, 398–405. doi:10.1038/nphoton.2012.
doi:10.1063/1.363052 111
Fujita, M., Tanaka, Y., and Noda, S. (2008). Light emission from silicon in Jannopoulos, J. D., and Winn, R. D. M. J. N. (1995). Photonic Crystals. Princeton,
photonic crystal nanocavity. IEEE J. Selec. Top. Quant. Elec. 14, 1090–1097. NJ: Princeton University Press.
doi:10.1109/JSTQE.2008.918941 Kageshima, H., and Shiraishi, K. (1997). Momentum-matrix-element calculation
Fukatsu, S., Usami, N., Chinzei, T., Shiraki, Y., Nishida, A., and Nakagawa, K. (1992). using pseudopotentials. Phys. Rev. B 56, 14985. doi:10.1103/PhysRevB.56.14985
Electroluminescence from strained SiGe/Si quantum well structures grown by Kako, S., Okumura, T., Oda, K., Suwa, Y., Saito, S., Ido, T., et al. (2012). “Time-
solid source si molecular beam epitaxy. Jpn. J. Appl. Phys. 31, L1015–L1017. resolved photoluminescence study of highly n-doped germanium grown on sil-
doi:10.1143/JJAP.31.L1015 icon,” in The 9th International Conference on Group IV Photonics (GFP), Vol. 7,
Gaubas, E., and Vanhellemont, J. (2006). Dependence of carrier lifetime in ger- San Diego, 340–342.
manium on resistivity and carrier injection level. Appl. Phys. Lett. 89, 142106. Kittler, M., Reiche, M., and Arguirov, T. (2013). “1.55 µm light emitter based on
doi:10.1063/1.2358967 dislocation d1-emission in silicon,” in Microelectronics Technology and Devices
Geiger, R., Frigerio, J., Süess, M. J., Chrastina, D., Isella, G., Spolenak, R., et al. (SBMicro), Symposium on IEEE, Curitiba.
(2014). Excess carrier lifetimes in Ge layers on Si. Appl. Phys. Lett. 104, 062106. Koshida, N., and Koyama, H. (1992). Visible electroluminescence from porous sili-
doi:10.1063/1.4865237 con. Appl. Phys. Lett. 60, 347–349. doi:10.1063/1.106652

Kurdi, M. E., Bertin, H., Martincic, E., de Kersauson, M., Fishman, G., Sauvage, Nakayama, S., Iwamoto, S., Kako, S., Ishida, S., and Arakawa, Y. (2011). “Demon-
S., et al. (2010). Control of direct band gap emission of bulk germanium by stration of silicon nanocavity led with enhanced luminescence,” in International
mechanical tensile strain. Appl. Phys. Lett. 96, 041909. doi:10.1063/1.3297883 Conference on Solid State Devices and Materials, Nagoya, I–8–2.
Kurdi, M. E., Davida, S., Checourya, X., Fishmana, G., Boucauda, P., Kermarrecb, Nama, D., Sukhdeo, D. S., Kang, J.-H., Petykiewicz, J., Lee, J. H., Jung, W. S.,
O., et al. (2008). Two-dimensional photonic crystals with pure germanium-on- et al. (2013). Strain-induced pseudoheterostructure nanowires confining car-
insulator. Opt. Commun. 281, 846–850. doi:10.1016/j.optcom.2007.10.008 riers at room temperature with nanoscale-tunable band profiles. Nano Lett. 13,
Laasonen, K., Pasquarello, A., Car, R., Lee, C., and Vanderbilt, D. (1993). Car- 3118–3123. doi:10.1021/nl401042n
Parrinello molecular dynamics with Vanderbilt ultrasoft pseudopotentials. Phys. Nassiopoulou, A. G. (2004). Silicon Nanocrystals in SiO 2 Thin Layers,Vol. 9. Valencia:
Rev. B 47, 10142. doi:10.1103/PhysRevB.47.10142 American Scientific Publishers, 793–813.
Liang, D., and Bowers, J. E. (2010). Recent progress in lasers on silicon recent progress Nelson, H. (1963). Epitaxial growth from the liquid state and its application to the
in lasers on silicon recent progress in lasers on silicon recent progress in lasers fabrication of tunnel and laser diodes. RCA Rev. 24, 603–615.
on silicon. Nat. Photonics 4, 511–517. doi:10.1088/0034-4885/76/3/034501 Ng, W. L., Lourenço, M. A., Gwilliam, R. M., Ledain, S., Shao, G., and Homewood,
Lim, P. H., Park, S., Ishikawa, Y., and Wada, K. (2009). Enhanced direct bandgap K. P. (2001). An efficient room-temperature silicon-based light-emitting diode.
emission in germanium by micromechanical strain engineering. Opt. Express Nature 410, 192–194. doi:10.1038/35069092
17, 16358–16365. doi:10.1364/OE.17.016358 Ngo, T.-P., Kurdi, M. E., Checoury, X., Boucaud, P., Damlencourt, J. F., Kermar-
Linnros, J. (1998). Carrier lifetime measurements using free carrier absorp- rec, O., et al. (2008). Two-dimensional photonic crystals with germanium on
tion transients. I. Principle and injection dependence. J. Appl. Phys. 84, 275. insulator obtained by a condensation method. Appl. Phys. Lett. 93, 241112.
doi:10.1063/1.368024 doi:10.1063/1.3054332
Liu, J. (2014). Monolithically integrated Ge-on-Si active photonics. Photonics 1, Noborisaka, J., Nishiguchi, K., Ono, Y., Kageshima, K., and Fujiwara, A. (2011).
162–197. doi:10.3390/photonics1030162 Strong stark effect in electroluminescence from phosphorous-doped silicon-on-
Liu, J., Camacho-Aguilera, R., Bessette, J. T., Sun, X., Wang, X., Cai, Y., et al. (2012). insulator metal-oxide-semiconductor field-effect transistors. Appl. Phys. Lett. 98,
Ge-on-Si optoelectronics. Thin Solid Films 520, 3354–3360. doi:10.1016/j.tsf. 033503. doi:10.1063/1.3543849
2011.10.121 Oda, K., Tani, K., Saito, S.-I., and Ido, T. (2014). Improvement of crystallinity by
Liu, J., Cannon, D. D., Wada, K., Ishikawa, Y., Jongthammanurak, S., Danielson, post-annealing and regrowth of Ge layers on Si substrates. Thin Solid Films 550,
D. T., et al. (2005). Tensile strained Ge p-i-n photodetectors on Si platform for 509–514. doi:10.1016/j.tsf.2013.10.136
C and L band telecommunications. Appl. Phys. Lett. 87, 011110. doi:10.1063/1. Ohta, Y., Tanaka, T., Toko, K., Sadoh, T., and Miyao, M. (2011). Growth-direction-
2037200 dependent characteristics of Ge-on-insulator by Si-Ge mixing triggered melting
Liu, J., Sun, X., Camacho-Aguilera, R., Kimerling, L. C., and Michel, J. (2010). Ge- growth. Solid State Electron. 60, 18–21. doi:10.1016/j.sse.2011.01.039
on-Si laser operating at room temperature. Opt. Lett. 35, 679–681. doi:10.1364/ Okada, Y., and Tokumaru, Y. (1984). Precise determination of lattice parameter and
OL.35.000679 thermal expansion coefficient of silicon between 300 and 1500 K. J. Appl. Phys.
Liu, J., Sun, X., Pan, D., Wang, X., Kimerling, L. C., Koch, T. L., et al. (2007). Tensile- 56, 314. doi:10.1063/1.333965
strained, n-type Ge as a gain medium for monolithic laser integration on Si. Opt. Ortolland, C., Okuno,Y.,Veheyen, P., Kerner, C., Stapelmann, C., Aoulaiche, M., et al.
Express 15, 11272–11277. doi:10.1364/OE.15.011272 (2009). Stress memorization technique-fundamental understanding and low-
Liu, Y., Deal, M. D., and Plummer, J. D. (2004). High-quality single-crystal Ge on cost integration for advanced CMOS technology using a nonselective process.
insulator by liquid-phase epitaxy on Si substrates. Appl. Phys. Lett. 84, 2563. IEEE Trans. Electron Devices 56, 1690–1697. doi:10.1109/TED.2009.2024021
doi:10.1063/1.1691175 Ossicini, S., Pavesi, L., and Priolo, F. (2006). Light Emitting Silicon for Microphotonics.
Lo Savio, R., Portalupi, S. L., Gerace, D., Shakoor, A., Krauss, T. F., O’Faolain, L., et al. Berlin: Springer.
(2011). Room-temperature emission at telecom wavelengths from silicon pho- Panish, M. B., Hayashi, I., and Sumski, S. (1970). Double heterostructure injection
tonic crystal nanocavities. Appl. Phys. Lett. 98, 201106. doi:10.1063/1.3591174 lasers with room temperature thresholds as low as 2300 a/cm2 . Appl. Phys. Lett.
Luan, H.-C., Lim, D. R., Lee, K. K., Chen, K. M., Sandland, J. G., Wada, K., et al. 16, 326. doi:10.1063/1.1653213
(1999). High-quality Ge epilayers on Si with low threading-dislocation densities. Pavesi, L., and Guillot, G. (eds) (2006). Optical Interconnects The Silicon Approach.
Appl. Phys. Lett. 75, 2909. doi:10.1063/1.125187 Berlin: Springer.
Matsue, M., Yasutake, Y., Fukatsu, S., Hosoi, T., Shimura, T., and Watanabe, H. Pavesi, L., and Lockwood, D. J. (eds) (2004). Silicon Photonics. Berlin: Springer.
(2014). Strain-induced direct band gap shrinkage in local Ge-on-insulator struc- Pavesi, L., Negro, L. D., Mazzoleni, C., Franzò, G., and Priolo, F. (2000). Optical gain
tures fabricated by lateral liquid-phase epitaxy. Appl. Phys. Lett. 104, 031106. in silicon nanocrystals. Nature 408, 440–444.
doi:10.1063/1.4862890 Perdew, J. P., Burke, K., and Wang, Y. (1996). Generalized gradient approximation
Menéndez, J., and Kouvetakis, J. (2004). Type-I Ge/Ge1−x−v Six Snv strained- for the exchange-correlation hole of a many-electron system. Phys. Rev. B 54,
layer heterostructure with a direct Ge bandgap. Appl. Phys. Lett. 85, 1175. 16533. doi:10.1103/PhysRevB.54.16533
doi:10.1063/1.1784032 Purcell, E. M. (1946). Spontaneous emission probabilities at radio frequencies. Phys.
Michel, J., Liu, J., and Kimerling, L. C. (2010). High-performance ge-on-si photode- Rev 69, 681.
tectors. Nat. Photonics 4, 527–534. doi:10.1038/nphoton.2010.157 Reed, G. (2008). Silicon Photonics the State of the Art. West Sussex: Wiley.
Michel, J., and Romagnoli, M. (2012). Germanium as the Unifying Material for Silicon Reed, G. T., and Knights, A. P. (2004). Silicon Photonics. West Sussex: Wiley.
Photonics. Bellingham: SPIE. doi:10.1117/2.1201206.004285 Rylyakov, A., Schow, C., Lee, B., Green, W., Campenhout, J. V., Yang, M., et al. (2011).
Miller, D. A. B. (2009). Device requirements for optical interconnections to silicon “A 3.9ns 8.9mW 4×4 silicon photonic switch hybrid integrated with CMOS
chips. Proc. IEEE 97, 1166–1185. doi:10.1109/JPROC.2009.2014298 driver,” in Int. Solid-State Circuits Conference (ISSCC) (San Francisco: IEEE),
Miyao, M., Tanaka, T., Toko, K., and Tanaka, M. (2009). Giant Ge-on-insulator for- 222–224.
mation by Si-Ge mixing-triggered liquid-phase epitaxy. Appl. Phys. Express 2, Saito, S. (2011). “Silicon and germanium quantum well light-emitting diode,” in
045503. doi:10.1143/APEX.2.045503 IEEE 8th Int. Conf. Group IV Photonics, London, 166–168.
Nakayama, S., Ishida, S., Iwamoto, S., and Arakawa, Y. (2012). Effect of cavity mode Saito, S., Hisamoto, D., Shimizu, H., Hamamura, H., Tsuchiya, R., Matsui, Y., et al.
volume on photoluminescence from silicon photonic crystal nanocavities. Appl. (2006a). Electro-luminescence from ultra-thin silicon. Jpn. J. Appl. Phys. 45,
Phys. Lett. 98, 171102. L679–L682. doi:10.1143/JJAP.45.L679
Nakayama, S., Iwamoto, S., Ishida, S., and Arakawa, Y. (2010a). “Demonstration of Saito, S., Hisamoto, D., Shimizu, H., Hamamura, H., Tsuchiya, R., Matsui, Y., et al.
a silicon photonic crystal slab led with efficient electroluminescence,” in Inter- (2006b). Silicon light-emitting transistor for on-chip optical interconnection.
national Conference on Solid State Devices and Materials (SSDM 2010) D-4-3, Appl. Phys. Lett. 89, 163504. doi:10.1063/1.2360783
Tokyo. Saito, S., Oda, K., Takahama, T., Tani, K., and Mine, T. (2011a). Germanium fin
Nakayama, S., Iwamoto, S., Ishida, S., Bordel, D., Augedre, E., Calvelier, L., et al. light-emitting diode. Appl. Phys. Lett. 99, 241105. doi:10.1364/OE.22.005927
(2010b). Enhancement pf photoluminescence from germanium by utilizing air- Saito, S., Takahama, T., Tani, K., Takahashi, M., Mine, T., Suwa, Y., et al. (2011b).
bridge-type photonic crytsal slab. Physica E 42, 2556–2559. doi:10.1016/j.physe. Stimulated emission of near-infrared radiation in silicon fin light-emitting diode.
2010.05.026 Appl. Phys. Lett. 98, 261104. doi:10.1063/1.3605255

Saito, S., Sakuma, N., Suwa, Y., Arimoto, H., Hisamoto, D., Uchiyama, H., et al. thermal etching and Ge-condensation processes. Appl. Phys. Lett. 94, 081910.
(2008). “Observation of optical gain in ultra-thin silicon resonant cavity light- doi:10.1063/1.3086884
emitting diode,” in IEEE Int. Conf. Electron Devices Meeting (IEDM), San Fran- Tseng, C.-K., Chen, W.-T., Chen, K.-H., Liu, H.-D., Kang, Y., Na, N., et al. (2013).
cisco, 19.5. A self-assembled microbonded germanium/silicon heterojunction photodiode
Saito, S., Suwa, Y., Arimoto, H., Sakuma, N., Hisamoto, D., Uchiyama, H., et al. for 25 Gb/s high-speed optical interconnects. Sci. Rep. 3, 3225. doi:10.1038/
(2009). Stimulated emission of near-infrared radiation by current injection srep03225
into silicon (100) quantum well. Appl. Phys. Lett. 95, 241101. doi:10.1063/1. Tweet, D. J., Lee, J. J., Maa, J.-S., and Hsu, S. T. (2005). Characterization and reduc-
3273367 tion of twist in Ge on insulator produced by localized liquid phase epitaxy. Appl.
Saul, R. H., and Roccasecca, D. D. (1973). Vapor-doped multislice LPE for efficient Phys. Lett. 87, 141908. doi:10.1063/1.2077860
GaP green LED’s. J. Electrochem. Soc. 120, 1128–1131. doi:10.1149/1.2403644 Uchida, K., and Takagi, S. (2003). Carrier scattering induced by thickness fluctuation
Savio, R. L., Miritello, M., Shakoor, A., Cardile, P., Welna, K., Andreani, L. C., et al. of silicon-on-insulator film in ultrathin-body metal–oxide–semiconductor field-
(2013). Enhanced 1.54 µm emission in Y-Er disilicate thin films on silicon pho- effect transistors. Appl. Phys. Lett. 82, 2916. doi:10.1063/1.1571227
tonic crystal cavities. Opt. Express 21, 10278–10288. doi:10.1364/OE.21.010278 Ujihara, K. (1995). “Effects of atomic broadening on spontaneous emission in an
Shakoor, A., Lo Savio, R., Cardile, P., Portalupi, S. L., Gerace, D., Welna, K., optical microcavity,” in Spontaneous Emission and Laser Oscillation in Microcav-
et al. (2013). Room temperature all-silicon photonic crystal nanocavity light ities, eds H. Yokoyama and K. Ujihara (Boca Raton, FL: CRC Press), 81–107.
emitting diode at sub-bandgap wavelengths. Laser Photon. Rev. 7, 114–121. Urino, Y., Horikawa, T., Nakamura, T., and Arakawa, Y. (2013). “High density optical
doi:10.1002/lpor.201200043 interconnects integrated with lasers, optical modulators and photodetectors on
Sheng, J. J., Leonhardt, D., Han, S. M., Johnston, S. W., Cederberg, J. G., and Carroll, a single silicon chip,” in Optical Fiber Communication Conference (OFC), (Ana-
M. S. (2013). Empirical correlation for minority carrier lifetime to defect density heim: OSA), OM2J.
profile in germanium on silicon grown by nanoscale interfacial engineering. J. Vanderbilt, D. (1990). Soft self-consistent pseudopotentials in a generalized eigen-
Vac. Sci. Technol. B Microelectron. 31, 051201. doi:10.1116/1.4816488 value formalism. Phys. Rev. B 41, 7892(R). doi:10.1103/PhysRevB.41.7892
Singh, H. P. (1968). Determination of thermal expansion of germanium, Virgilio, M., Manganelli, C. L., Grosso, G., Pizzi, G., and Capellini, G. (2013a).
rhodium and iridium by x-rays. Acta Crystallogr. A 24, 469–471. doi:10.1107/ Radiative recombination and optical gain spectra in biaxially strained n-type
S056773946800094X germanium. Phys. Rev. B 87, 235313. doi:10.1103/PhysRevB.87.235313
Sproul, A. B. (1994). Dimensionless solution of the equation describing the effect Virgilio, M., Manganelli, C. L., Grosso, G., Schroeder, T., and Capellini, G. (2013b).
of surface recombination on carrier decay in semiconductors. J. Appl. Phys. 76, Photoluminescence, recombination rate, and gain spectra in optically excited
2851. doi:10.1063/1.357521 n-type and tensile strained germanium layers. J. Appl. Phys. 114, 243102.
Steger, M., Yang, A., Sekiguchi, T., Saeedi, K., Thewalt, M. L. W., Henry, M. O., doi:10.1063/1.4849855
et al. (2011). Photoluminescence of deep defects involving transition met- Vivien, L., and Pavesi, L. (eds) (2013). Handbook of Silicon Photonics. Taylor and
als i Si: new insights from highly enriched 28 Si. J. Appl. Phys. 110, 081301. Francis.
doi:10.1063/1.3651774 Wada, K., Liu, J., Jongthammanurak, S., Cannon, D. D., Danielson, D. T., Ahn, D.,
Süess, M. J., Geiger, R., Minamisawa, R. A., Schiefler, G., Frigerio, J., Chrastina, D., et al. (2006). Si Microphotonics for Optical Interconnection. Berlin: Springer Ver-
et al. (2013). Analysis of enhanced light emission from highly strained germa- lag.
nium microbridges. Nat. Photonics 7, 466–472. doi:10.1038/nphoton.2013.67 Wale, M. (2008). “Photonic integration – an industrial perspective,” in European
Sukhdeo, D. S., Nam, D., Kang, J. H., Brongersma, M. L., and Saraswat, K. C. (2014). Photonic Integration Forum (Brussels: European Photonics Integration Forum).
Direct bandgap germanium-on-silicon inferred from 5.7% <100> uniaxial ten- Wang, X., Li, H., Camacho-Aguilera, R., Cai, Y., Kimerling, L. C., Michel, J., et al.
sile strain. Photon. Res. 2, A8–A13. doi:10.1364/PRJ.2.0000A8 (2013). Infrared absorption of n-type tensile-strained Ge-on-Si. Opt. Lett. 38,
Sumikura, H., Kuramochi, E., Taniyama, H., and Notomi, M. (2014). Ultrafast spon- 652–654. doi:10.1364/OL.38.000652
taneous emission of copper-doped silicon enhanced by an optical nanocavity. Wang, Y., Zhang, J., Wu, Y., An, J., Li, J., Wang, H., et al. (2012). Light emission
Sci. Rep. 4, 5040. doi:10.1038/srep05040 enhancement from Er-doped silicon photonic crystal double-heterostructure
Suwa, Y., and Saito, S. (2009). Intrinsic optical gain of ultrathin silicon quantum microcavity. IEEE Photon. Technol. Lett. 24, 1041–1135. doi:10.1109/LPT.2011.
wells from first-principles calculations. Phys. Rev. B 79, 233308. doi:10.1103/ 2173183
PhysRevB.79.233308 Wieder, H. H., Clawson, A. R., and McWilliams, G. E. (1977). Inx Ga1−x Asv P1−v /InP
Suwa, Y., and Saito, S. (2010). “First-principles study of light emisson from germa- heterojunction photodiodes. Appl. Phys. Lett. 31, 468. doi:10.1063/1.89718
nium guantum-well,” in IEEE 22nd Int. Semiconductor Laser Conference (ISLC), Zelsmann, M., Picard, E., Charvolin, T., Hadji, E., Heitzmann, M., Dalzotto, B.,
Kyoto, 131–132. et al. (2003). Seventy-fold enhancement of light extraction from a defectless
Suwa, Y., and Saito, S. (2011). “First-principles study of light emission from sili- photonic crystal made on silicon-on-insulator. Appl. Phys. Lett. 83, 2542–2544.
con and germanium due to direct transitions,” in IEEE 8th Int. Conf. Group IV doi:10.1063/1.1614832
Photonics, London, 222–224. Zimmermann, H. (2000). Integrated Silicon Opto-Electronics. Berlin: Springer.
Sze, S. M., and Lee, M. K. (2012). Semiconductor Devices: Physics and Technology,
3rd Edn. Singapore: John Wiley and Sons. Conflict of Interest Statement: The authors declare that the research was conducted
Tani, K., Oda, K., Kasai, J., Okumura, T., Mine, T., Saito, S., et al. (2013a). “Ger- in the absence of any commercial or financial relationships that could be construed
manium waveguides on lateral silicon-on-insulator diodes for monolithic light as a potential conflict of interest.
emitters and photo detectors,” in Group IV Photonics (GFP), IEEE 10th Interna-
tional Conference (Seoul: IEEE), 134–135. Received: 12 June 2014; paper pending published: 28 July 2014; accepted: 29 August
Tani, K., Oda, K., Okumura, T., Takezaki, T., Kasai, J., Mine, T., et al. (2013b). 2014; published online: 17 September 2014.
“Enhanced electroluminescence from germanium waveguides by local tensile Citation: Saito S, Gardes FY, Al-Attili AZ, Tani K, Oda K, Suwa Y, Ido T,
strain with silicon nitride stressors,” in International Conference on Solid State Ishikawa Y, Kako S, Iwamoto S and Arakawa Y (2014) Group IV light sources
Devices and Materials (SSDM) (Fukuoka: SSDM), K–6–3. to enable the convergence of photonics and electronics. Front. Mater. 1:15. doi:
Tani, K., Saito, S., Oda, K., Miura, M., Mine, T., Sugawara, T., et al. (2011). “Ge(111)- 10.3389/fmats.2014.00015
fin light-emitting diodes,” in IEEE 8th Int. Conf. Group IV Photonics, London, This article was submitted to Optics and Photonics, a section of the journal Frontiers
217–219. in Materials.
Tani, K., Saito, S., Oda, S., Okumura, T., Mine, T., and Ido, T. (2012). “Lateral car- Copyright © 2014 Saito, Gardes, Al-Attili, Tani, Oda, Suwa, Ido, Ishikawa, Kako,
rier injection to germanium for monolithic light sources,” in Group IV Photonics Iwamoto and Arakawa. This is an open-access article distributed under the terms of the
(GFP), IEEE 9th Int. Conf, San Diego, 328–330. Creative Commons Attribution License (CC BY). The use, distribution or reproduction
Taur, Y., and Ning, T. H. (2013). Fundamentals of Modern VLSI Devices, 2nd Edn. in other forums is permitted, provided the original author(s) or licensor are credited
Cambridge: Cambridge University Press. and that the original publication in this journal is cited, in accordance with accepted
Tezuka, T., Toyoda, E., Irisawa, T., Hirashita, N., Moriyama, Y., Sugiyama, N., academic practice. No use, distribution or reproduction is permitted which does not
et al. (2009). Structural analyses of strained SiGe wires formed by hydrogen comply with these terms.

Review
published: 15 July 2015
doi: 10.3389/fmats.2015.00052
Group iv direct band gap photonics:

methods, challenges, and opportunities
Richard Geiger, Thomas Zabel and Hans Sigg*
Laboratory for Micro- and Nanotechnology, Paul Scherrer Institut, Villigen, Switzerland
The concept of direct band gap group IV materials may offer a paradigm change for
Si-photonics concerning the monolithic implementation of light emitters: the idea is to
integrate fully compatible group IV materials with equally favorable optical properties
as the chemically incompatible group III–V-based systems. The concept involves either
mechanically applied strain on Ge or alloying of Ge with Sn, which permits to drastically
improve the radiative efficiency of Ge. The favorable optical properties result from a
modified band structure transformed from an indirect to a direct one. The first demon-
stration of such a direct band gap laser has recently been accomplished in GeSn. This
demonstration proves the capability of this new concept, which may permit a qualitative
as well as a quantitative expansion of Si-photonics in not only traditional but also new
Edited by: areas of applications. This review aims to discuss the challenges along this path in terms
Koji Yamada,
of fabrication, characterization, and fundamental understanding, and will elaborate on
National Institute of Advanced
Industrial Science and Technology, evoking opportunities of this new class of group IV-based laser materials.
Japan
Keywords: Si photonics, germanium, strain, GeSn, direct band gap, laser
Reviewed by:
Shinichi Saito,
University of Southampton, UK
Krishna C. Saraswat,
Introduction
Stanford University, USA
The Si-based optical platform is rapidly changing the landscape of photonics by offering powerful
*Correspondence: solutions, for example, for data links (Miller, 2010) and sensing (Passaro et al., 2012) to name only
Hans Sigg,
two out of many. This development has taken place in spite of the fact that Si itself is a poor emitter of
Laboratory for Micro- and
Nanotechnology, Paul Scherrer
light. This is without a doubt due to the fact that Si technology as used in very large-scale integration
Institut, Villigen PSI, CH 5232, (VLSI) and complementary metal-oxide-semiconductor (CMOS) technology is extremely mature
Switzerland and advanced. This fact seemingly compensates for the shortfalls in concepts for Si to generate light.
hans.sigg@psi.ch Nowadays, group III–V materials are implemented to integrate active light sources onto the Si
platform by using involved coupling schemes and/or heterogeneous integration (Fang et al., 2013).
Specialty section: However, because these materials are chemically intolerant to Si, their integration bears a lot of
This article was submitted to Optics burdens, which raises the fabrication costs. Strongly preferred are materials that are compatible to Si,
and Photonics, a section of the tolerated by the technology (preferentially CMOS), and capable of producing light similar in efficiency
to traditional group III–V semiconductor systems.
Received: 13 May 2015 In direct band gap systems, light generation is based on radiative recombination of electrons and
Accepted: 29 June 2015 holes, both with practically the same momentum as schematically shown in Figure 1A. In unstrained,
Published: 15 July 2015
i.e., “regular” bulk Ge, however, the excited electrons will preferentially occupy the lower conduction
Citation: band energy states of the L-valley. In Ge, the momentum of the electrons does, thus, not match those
Geiger R, Zabel T and Sigg H (2015)
of the holes, which occupy the degenerated heavy- and light-mass valence bands at the Γ-point (c.f.
Group IV direct band gap photonics:
methods, challenges,
Figure 1B). The appearing momentum mismatch requires a phonon for the recombination. But note
and opportunities. that except the position of the indirect L-valley, which is in Ge 140 meV below the Γ valley minimum,
Front. Mater. 2:52. the band alignments near the Γ-point in system A (say InGaAs, one of the most prominent group
doi: 10.3389/fmats.2015.00052 III–V systems used for lasing) and B are very similar. To achieve the favorable direct recombination
Frontiers in Materials | www.frontiersin.org 81 July 2015 | Volume 2 | Article 52

Geiger et al. Group IV direct band gap photonics
A B C
direct direct direct
valley valley indirect valley indirect
CONDUCTION Γ Γ valley Γ valley
BAND Offset L L
0.14 eV Offset
> kT
momentum
transfer
heavy hole
heavy hole
VALENCE
heavy hole
BANDS light light
hole hole
light
hole
momentum space
FIGURE 1 | Band structure in momentum space (A) direct band gap L-valleys. Radiative recombination is hindered by the momentum mismatch.
semiconductor. Electron/hole recombination occurs at the Γ-point. (C) Due to strain (or alloying with Sn), the band gap of Ge (GeSn) shrinks and
(B) Unstrained Ge. Electrons occupy the conduction band states of the the population of electrons at the Γ-point increases.
condition also for Ge, we need to find a way to inject electrons

into that conduction band valley with its energetic minimum at uniaxial
the Γ-point. This is realized in the most straightforward fashion [100] 0 1 2 3 4 5
(%)
when all unwanted electron levels are energetically shifted above biaxial
the Γ-states, which is equivalent to transfer the system from a (001)
0 0.5 1.0 1.5 2.0
(%)
fundamentally indirect to a fundamentally direct one.
We will discuss the two methods that make this conversion GexSn1-x
(Sn%) 0 2 4 6 8 10
possible. One involves the application of tensile strain, while the
second approach relies on alloying Ge with Sn. Thus, the obtained
band alignments are depicted in Figure 1C. With either one of the 100 80 60 40 20 0 %
methods, the Γ-valley can be reduced below the indirect one at L E/Eoffset (%)
enabling efficient carrier injection into the Γ-valley. Moreover, the
VB degeneracy is lifted depending on the strain state and its load- FIGURE 2 | Relative offset of the Γ- and L-conduction band minima
ing, biaxial or uniaxial, c.f. Section “Modeling” for more details. realized in Ge either by uniaxial tensile strain along [100] direction
[dark blue: (Capellini et al., 2013), light blue: (Nam et al., 2013;
In Figure 2, we show the state-of-the-art of the strain and alloy- Sukhdeo et al., 2014), olive: (Süess et al., 2013; Geiger et al., 2014c)],
ing approach toward the realization of a direct band gap group IV biaxial tensile strain on (001) oriented substrate [black: (Liu et al.,
material. For our discussion, we selected those approaches that are 2010); violet: (Ghrib et al., 2013, 2015)] or by alloying with Sn [yellow:
potentially compatible with CMOS fabrication and are suited for (Chen et al., 2013a); orange (Gupta et al., 2013b); red: (Wirths et al.,
2015)]. The shown offset versus Sn concentration relates to the unstrained
optical applications. Very thin membranes (Sánchez-Pérez et al.,
case. About 100% (0%) offset refers to 140 meV (vanishing energy offset).
2011), nanowires clamped in bulky mechanical strain apparatus Hence, the dash-dotted line marks the transition from an indirect to a direct
(Greil et al., 2012), Ge bulk layers on III–V substrates (Huo et al., band gap semiconductor.
2011), etc., are not considered here because they are unpractical
for integration on Si. Not considered either is light emission from
Si-based quantum wells and defects; for a recent review, see Saito is introduced to fill the parasitic indirect states (Xiaochen et al.,
et al. (2014). In our compilation, Figure 2, we benchmark the two 2010). Such doping does not transform the material into a direct
strain loadings (uniaxial and biaxial) and Sn alloy composition gap system, but it appeared that under optical excitation and
against the achieved relative band offset, ΔE/E0, where an offset electrical pumping the light emission shows an intensity threshold
ΔE of 100% is equal to E0 ~ 140 meV for the case of unstrained Ge. as well as linewidth narrowing (Liu et al., 2010; Camacho-Aguilera
An offset parameter of 0 meV (0%) corresponds, thus, to Γ- and et al., 2012). These results became widely known as the optically
L-valleys having their band edges at the same energy. and electrically pumped Ge-laser. However, since the time when
The black arrow on the left hand side of the second line in these announcements were made in 2010 and 2012, only one other
Figure 2 marks the case of highly n-doped Ge (Liu et al., 2010), demonstration of these effects has been reported so far (Koerner
where a maximum of 0.25% biaxial strain is accomplished. This et al., 2015). This very recent and only result concerns a Ge diode
value of 0.25% is the one typically obtained from direct epitaxy of structure with an unstrained active region doped at 3 × 1019 cm−3.
Ge on Si. It arises due to the difference between the thermal expan- The obtained emission spectra are similar to the one from the
sion coefficients of Si and Ge (Michel et al., 2010). High n-doping original work from the MIT group. However, as we will show

below, these spectra significantly differ in several aspects – the similarities of the physics and the characterization methods used
intensity, the linewidth, and the Fabry–Perot (FP) multi-mode for investigations, we hope to provide a comprehensive overview
behavior – from those obtained with the here discussed direct band that will support and interest many scientists to enter this highly
gap lasers. Moreover, as will be discussed in Section “Lifetime, relevant field of research.
Gain, and Loss,” these Ge-lasing observations are contradicted In Section “Direct Band Gap Group IV Materials,” the band
by gain experiments (Carroll et al., 2012) as well as by theoretical structure in Ge is given in dependence of strain. We then sum-
analysis performed by several groups (Liu et al., 2007; Chow, marize the fabrication steps for strain engineering and Sn alloying.
2012; Dutt et al., 2012; Peschka et al., 2015) when the lasing cur- In Section “Characterization Methods: Optical Properties,” several
rent density threshold is calculated using realistic non-radiative optical characterization methods are introduced, such as pump and
lifetimes (Geiger et al., 2014a). As the experimental foundation probe spectroscopy developed for this very purpose at the infrared
for understanding this peculiar threshold behavior of low strained beamline of the Swiss light source (SLS). Gain and loss studies
(and in one case even unstrained) Ge is ambiguous, we will focus performed on Ge layers as well as carrier lifetime measurements are
here on reports concerning direct band gap group IV systems, shown in Section “Lifetime, Gain, and Loss.” These results impact
which includes the first unmistakable proof of interband lasing in the discussion on lasing in n-doped Ge, which is briefly repeated
a group IV system. Without doubt, with the advent of direct band to exemplify the capability of these experimental methods. The
gap systems showing unambiguous lasing, an excellent opportunity analysis of temperature-dependent photoluminescence (PL) is
is created, which will help to unravel in the very near future above found to deliver a quantitative measure for the directness of GeSn
raised questions regarding the lasing in highly n-doped Ge. layers, shown in Section “Photoluminescence – Direct Band Gap,”
Coming back to Figure 2: the arrows colored in violet depict and narrow emission spectra together with an intensity versus exci-
1.0 and 1.5% biaxially tensilely strained structures that have been tation-threshold represent the first observation of lasing in a direct
achieved via deposition of Si–Nitride (SiN) stressor layers (Ghrib band gap group IV system, shown in Section “Optically Pumped
et al., 2013, 2015). This strain is equivalent to a band offset of ~70 Laser.” Investigation challenges, such as the quantitative analysis of
and 30 meV, which corresponds to 50 and 30% of the unstrained the Auger recombination and the carrier transport, are appointed
band offset value, respectively. So far, the highest strain values in Section “Challenges” together with other fundamental device-
are obtained in suspended microbridges under uniaxial loading related issues, such as cavity design, band gap renormalization, and
as is shown on the top stroke. There, the ~0.25% biaxial prestrain thermal budgets for alloys. We speculate about the opportunities
is enhanced and transformed into uniaxial strain. The arrows in for Si photonics offered by an efficient monolithically integrated
olive (Süess et al., 2013; Geiger et al., 2014c) and blue (Nam et al., laser source in Section “Opportunities,” and furthermore discuss
2013; Sukhdeo et al., 2014) mark recent achievements from the two the prospect of a Ge and/or GeSn electro-optical data processing
leading groups. The latest result (Sukhdeo et al., 2014) indicates platform. We conclude in Section “Conclusion and Outlook” and
that the bridge technology can indeed provide direct band gap give a short outlook.
strained Ge. SiN stressor layers on suspended microbridges or
FP cavities deliver far less strain and offset reductions (Capellini
et al., 2013, 2014). As shown by the red arrows on the third stroke, Direct Band Gap Group IV Materials
alloying Ge with Sn also provides optical group IV material with
a fundamental direct band gap. The transition from fundamental Modeling
indirect to direct occurs at a Sn concentration of ~9% for relaxed Band Structure
GeSn. Depending on the strain loading, i.e., tensile or compressive, The effect of tensile strain on Ge’s band edges shown in Figure 3
the crossover shifts to a higher or lower Sn concentration. Hence, illustrates the path of the transitions’ energies going from an indi-
a 20-nm thick GeSn layer with 8% Sn sandwiched between Ge rect to a direct band gap system. The energies for interband- (solid
claddings and processed into microdisks is not as close to the lines) and intervalence-band transitions (broken lines) between
direct transition as a relaxed layer with 6% Sn because of the −1% the respective conduction- and valence-band edges are calculated
biaxial compressive loading (Chen et al., 2013a). The GeSn alloy via deformation potential theory as implemented in the nextnano®
above the crossover in Figure 2 exhibits 0.7% in-plane strain at a Sn modeling software (Birner et al., 2007). Due to the fact that the
concentration of 13%. This system shows lasing at low temperature Γ-valley energy reduces faster than the one of the L-valley, Ge
(Wirths et al., 2015). We will present this recent result and will, transforms into a direct band gap semiconductor at ~4.7% uniaxial
thereby, clarify the characteristic of the experimental observation strain along [100] when the direct transition (black line) decreases
of lasing. below the energy of the indirect recombination (green line). For
The availability of direct band gap group IV semiconductors as Ge under biaxial tensile strain or GeSn alloys, the band edges
compiled in Figure 2, together with the rise of promising results, behave similarly with an indirect-to-direct band gap crossover at
in particular, the demonstration of lasing in the GeSn system, has ~1.6–2.0% strain (El Kurdi et al., 2010; Virgilio et al., 2013; Wen
motivated the writing of this review. It is meant to present the and Bellotti, 2015) or and at a Sn-content of ~9% (Low et al., 2012;
current understanding evoked from the research undertaken at Gupta et al., 2013b; Wirths et al., 2015) for a fully relaxed layer.
many places worldwide. Although some of the following descrip- In the valence band, strain lifts the degeneracy of light hole and
tions are exemplified for only one of the systems (strain or Sn heavy hole bands and introduces a mixing such that this distinc-
alloying), we will argue that the physics of this two direct gap tion becomes meaningless, especially under high strain. For low
systems can be understood by analogy. By merely emphasizing the strain, VB1 and VB2 in Figure 3 are mostly “heavy hole”- and

is nearly reached at a low injection of 1 × 1018 cm−3 as depicted in

Figure 4A. Applying a moderate doping of 1 × 1019 cm−3 results in a
gain of >500 cm−1, which is sufficient to overcome typical resonator
losses. When the doping level is increased to 2 × 1019 cm−3, the
gain approximately triples to ~1500 cm−1 according to our model.
This suggests that n-doping is a very effective method to promote
high gain for correspondingly low excitation in direct band gap
systems. This is due to the fact that as long as the offset is not much
larger than kT (Sukhdeo et al., 2014), electrons will nevertheless
spread into the L-band from where they cannot contribute to gain.
Transparency can also be achieved for an undoped and still
indirect, strained Ge system under higher excitation. We obtain
transparency at an injection of 1 × 1019 cm−3 for a system with a
remaining offset of 25% (35 meV) and an n-doping level below
1 × 1018cm−3 (see Figure 4B). When the direct band gap is reached
(0% offset), the net gain amounts to 1000 cm−1, which can be
increased up to 2000 cm−1 at a doping of 2 × 1019 cm−3. In contrast
to the indirect band gap Ge, a reduction in temperature helps to
increase the gain as soon as the Γ valley constitutes the lowest
conduction band energy due to condensation of the carriers into
the direct gap states. For example, an intrinsic direct band gap Ge
FIGURE 3 | Transition energies between the direct and indirect
conduction band valleys and the two top valence bands and within
system with 25 meV band offset exhibits a net gain of the order of
the valence band of uniaxially, tensilely stressed Ge. For a strain of 4500 cm−1 at a temperature of 20 K and an injection of 1 × 1019 cm−3
~4.7%, the conduction band minimum at the Γ-point reaches a lower energy compared to 1700 cm−1 at RT.
than the indirect L-valleys. The upper x-axis denotes the offset between the When comparing gain predictions in literature, we experi-
Γ- and L-band edges in relative units.
ence larger differences than between predictions of energy levels
and their relative positions. The reason for this stems from
the uncertainty in the loss. For weakly strained and relaxed
“light hole”-like, respectively. VB3 refers to the split-off band. The Ge, experimental values are available as discussed in Section
energetic order of the heavy and light hole bands is reverted when “Lifetime, Gain, and Loss.” Hence, the overall agreement of
moving from the uniaxial to the biaxial case. the predictions is largely coherent. For example, calculations
Most of the theoretical work concerning Ge light emission uti- consistently predict gain for Ge with a large offset (80%) only
lizes k·p theory including 6 bands (Aldaghri et al., 2012; Chang and for the case of very high doping of >5 × 1019 cm−3. For strained
Cheng, 2013; Virgilio et al., 2013), 8 bands (Zhu et al., 2010; Wirths and alloyed systems, however, the interband energies approach
et al., 2013b), or 30 bands (El Kurdi et al., 2010). The latter is not the one of the intervalence band transitions. The energies may
restricted to the Brillouin-zone center but describes the full energy even cross, as shown in Figure 3. Hence, loss processes related
dispersion. In other works, the empirical pseudopotential method to these transitions will become critical. Furthermore, the gain
(Dutt et al., 2013; Wen and Bellotti, 2015), density functional as predicted by a Green’s functional approach (Wen and Bellotti,
theory (Tahini et al., 2012), and the tight-binding model (Dutt 2015) tend to be smaller than the commonly used joint density
et al., 2012) are employed. The agreement between the models is of state formalism as applied for Figure 4.
generally found to be satisfactory.
Gain Fabrication
In Figure 4, we show gain calculations for uniaxially stressed Ge Microbridges
in dependence of n-type doping and conduction band offset (c.f. Strain engineering is nowadays a standard tool in microelec-
the scale of the upper x-axis in Figure 3). The band structure was tronics to improve device performance, where the lattice mis-
computed with an 8-band k·p approach (Birner et al., 2007). The match between Si and Ge is used to generate strain via epitaxy.
gain was calculated via Fermi’s golden rule, assuming cylindri- However, the pseudomorphic deposition of Ge on Si leads to
cal symmetry for the valence bands to simplify the calculation compressive strain, which deteriorates the light emission effi-
of the joint density of states (JDOS), c.f. Virgilio et al. (2013). ciency and is, furthermore, limited to small layer thicknesses.
More details of the calculation can be found in Süess et al. (2013), Therefore, the main method used to introduce strain is the
supplementary information. The peak gain at room temperature application of external stressor layers, such as silicon nitride
(RT) is plotted after subtraction of the loss following the experi- (SiN), which is compatible with CMOS processing. Some work
mentally determined electron- and hole-absorption cross-sections following this approach includes the deposition of stressors on
from Carroll et al. (2012) (Süess et al., 2013). The black, broken the back side of Ge membranes (Nam et al., 2011, 2012), on
line indicates when transparency is reached. As an example, for a micropillars (Velha et al., 2013), or on selectively grown Ge
system at the crossover to a direct band gap system, transparency (Oda et al., 2013).

FIGURE 4 | Maximum net gain at room temperature for uniaxially parameters. The map is calculated for a carrier injection of (A) 1 × 1018 cm−3
stressed Ge in dependence of n-type doping and conduction band and (B) 1 × 1019 cm−3. The black, broken line depicts the transparency
offset. An 8-band k·p model was employed to calculate the band structure condition when the gain equals the losses.
An advantage of using external stressor layers is the simplicity from 0.15% biaxial strain in Ge on Si (blue squares) or Ge on
to combine the strain transfer with standard cavity structures like silicon-on-insulator (SOI) (green circles and red triangles) for
FP waveguides (Capellini et al., 2014). However, the achieved strain bridges with varying geometrical dimensions. The agreement of
is so far limited to a predominantly uniaxial strain of 1.5%. In the experimental values with the ones predicted by finite element
other efforts, SiN layers were deposited on Ge microdisks, resulting modeling (red triangles) is excellent. For Ge on SOI, the highest
in a biaxial strain of 1.0% (Ghrib et al., 2013) and 1.5% (Ghrib strain achieved in 6 μm × 2 μm constrictions is 3.1%. When
et al., 2015). However, these stressor layer approaches suffer from starting from 200 nm thick germanium-on-insulator substrates
a large strain inhomogeneity across the Ge layer, and elaborated (GOI), which feature a significantly reduced dislocation density
all-around stressor techniques using wafer transfer and bonding. (Akatsu et al., 2006; Hartmann et al., 2010), a strain of 5.7% was
These results are included in Figure 2. observed in a 5.0 μm × 0.2 μm constriction (Sukhdeo et al., 2014).
Following a different route, it was shown that high levels of According to Figures 2 and 3, such a strain is by far sufficient to
tensile strain can be locally induced without the use of any external transform Ge into a direct band gap material showing the prospect
stressor layers (Minamisawa et al., 2012; Süess et al., 2013). In of the strain-enhancement technique given a starting material
the approach by Süess et al., the starting substrate is the com- with high-crystal quality.
monly used tensilely strained Ge layer with a biaxial strain of
~0.2%. Subsequently, the layer is patterned into a microbridge GeSn Alloying
with a narrow central cross-section (the “constriction”) and larger The epitaxial growth of GeSn alloys poses several challenges, such
outer cross-sections (the “pads”) as shown in Figures 5A,B. As as a large lattice mismatch between α-Sn and Si (17%) or Ge (15%),
last processing step, the structure is underetched by selectively and a low solid-solubility of <1%. Therefore, the fabrication of
removing the underlying buried oxide with hydrofluoric acid, c.f. high quality and smooth epilayers was a demanding task for many
Figure 5C. Releasing the structure leads to a relaxation of the strain years and the development of new growth processes to deposit
in the pads, which in turn increases the strain in the constriction. GeSn under non-equilibrium conditions at low temperatures was
Due to Hooke’s law and force balance, strain accumulated in the required. Whereas the first attempts to grow GeSn alloys were
constriction will depend on the ratio of pad and constriction based on molecular beam epitaxy (MBE) in 1980s and 1990s
widths as well as the ratio between their lengths (Minamisawa (Pukite et al., 1989; Harwit et al., 1990; Wegscheider et al., 1990;
et al., 2012; Süess et al., 2013). Hence, following this principle, Fitzgerald et al., 1991; He and Atwater, 1997), device-grade GeSn
any strain can be generated in the constriction by solely varying epilayers could be synthesized since the early 2000s when the
the geometrical parameters independent of the actual dimen- first chemical vapor deposition (CVD) processes were developed
sions of the structure. In contrast to external stressors where the (Bauer et al., 2003).
achievable strain is limited by the efficiency of strain transfer, Nowadays, several groups established growth processes for
this strain enhancement is only limited by the material strength. GeSn utilizing either MBE (Bratland et al., 2003; Chen et al.,
Figure 5A shows enhancement factors of more than 20× realized 2011a; Bhargava et al., 2013; Oehme et al., 2013) or CVD

A B
Exp. strain (%) ε (%)
0 1 2 3 4 4
4
25 Ge/Si: analytical 3.5
Ge/SOI: analytical
3
Model enhancement 20 Ge/SOI: FEM 3
Model strain (%)

2.5
15
2 2
10 1.5
1
5 1
0.5
0 0
0 5 10 15 20 25 0
Exp. enhancement
C
Electron-beam RIE Wet

lithography etching etching
FIGURE 5 | Suspended microbridges from thermally pre-strained Ge. (B) Strain profile of a suspended bridge structure as obtained by finite element
(A) Experimental and modeled strain for Ge microbridges fabricated on Si and modeling (FEM). Due to the relaxation in the pads, the strain in the central
SOI. The analytical strain-enhancement model is given in Süess et al. (2013). The constriction is enhanced. (C) Process flow for the fabrication of suspended
enhancement of 22× corresponds to 3.1% uniaxial strain (Süess et al., 2013). microbridges.
(Vincent et al., 2011; Chen et al., 2013b; Wirths et al., 2013a; the epitaxial GeSn layers. The authors envision their method
Xu et al., 2013; Du et al., 2014) for a variety of applications, to enable the fabrication of direct band gap GeSn micro disks.
e.g., photodiodes, photodetectors, or MOSFETs. Here, due to Figures 6A–C show transmission electron microscopy images
the reduced lattice mismatch compared to Si, Ge is preferred as of a GeSn layer with 13% Sn grown via reduced pressure CVD on
virtual substrate (VS) in order to ensure layers of high monocrys- a Ge VS (Wirths et al., 2015). The advantageous relaxation mecha-
talline quality. Regarding the epitaxial growth of direct band nism mentioned above can be seen here with dislocation-loops
gap GeSn alloys, nearly strain relaxed or even tensilely strained (blue arrows) emitted into the Ge VS. Despite the high-Sn content,
layers are highly desired, since for compressively strained GeSn the thickness of the GeSn layer could be increased up to 560 nm
layers, i.e., GeSn coherently grown on Ge VS, higher Sn contents without deteriorating the high-crystalline quality. Owing to the
are necessary for the indirect to direct transition (Gupta et al., large thickness, a relaxation of 60% could be achieved such that only
2013b). Owing to an advantageous relaxation mechanism for a mild compressive strain of −0.6% was present. As will be shown
GeSn layers on Ge VS, dislocations seem to mostly protrude into in Section “Photoluminescence – Direct Band Gap,” this epilayer
the Ge VS rather than into the GeSn layer, which is beneficial was proven to be a direct band gap group IV semiconductor that
for optical properties as the density of non-radiative recombina- provides net gain and, hence, shows lasing under optical pumping.
tion centers is reduced (Takeuchi et al., 2006; Senaratne et al., We conclude this section on the fabrication of GeSn alloys by
2014; Wirths et al., 2015). Although relaxation takes place, a summarizing the list of beneficial assets GeSn epitaxy brings to the
certain level of compressive biaxial strain (typically between current Si technology facilitating future developments and integra-
−0.6 to −0.8%) remains nevertheless, which, as already said tion. Apart from the prospect to fabricate a fundamental direct
in connection with Figure 2, shifts the indirect-to-direct band band gap group IV material, GeSn alloys are attractive because of
gap crossover to higher Sn concentrations with respect to fully (i) low-temperature deposition on Si(001) compatible with existing
relaxed GeSn. Therefore, several approaches are being followed to CMOS processes; (ii) strain relaxation with reasonably low thread-
reduce the compressive strain, such as growth on lattice-matched ing dislocation density; (iii) available option for selective growth
InGaAs VS (Chen et al., 2011a), which is not acceptable within on silicon, which is attractive for photonic integration; (iv) GeSn/
a CMOS processing line, or deposition of ever thicker layers to SiGeSn heterojunction layers to generate carrier confinement in
enforce further strain relaxation (Senaratne et al., 2014; Wirths quantum wells; (v) and therefore, tunability of the lattice constant
et al., 2015). Gupta et al. (2013b) introduced a robust etching offering opportunities to combine the alloys with the strained
approach enabling to selectively dry etch the Ge VS underneath membrane method.

FIGURE 6 | Cross-sectional transmission electron microscopy (TEM) GeSn/Ge-interface (orange arrows) penetrating only into the Ge virtual
image of a Ge0.87Sn0.13 alloy. (A) Expanded view showing the high-crystalline substrate. (C) High-resolution TEM image of the interface used for Burgers
quality of the GeSn epilayer. The defects are located near the interface to the vector calculations. Lomer dislocations with b = a/2[110] are observed (Wirths
Ge virtual substrate. (B) Dislocation-loops (blue arrows) emitted below the et al., 2015).
Characterization Methods: Optical interpretation of the Ge-lasing observations. We renarrate this

Properties discussion at the end of this chapter.
Figure 7A shows the mid-infrared reflection of a Ge layer
Lifetime, Gain, and Loss grown on Si plotted as the ratio of pumped (RP) and unpumped
When describing a material with regards to its suitability as an (RU) reflection signal. The different colors depict different opti-
efficient laser source, key properties that decide upon adequacy cal excitation strengths between 1 and 160 MW cm−2. All of
are the gain and loss, i.e., the material’s ability to amplify light, as the spectra were taken for a pump–probe delay time of 250 ps.
well as the non-radiative lifetime, which determines the internal The distinct minimum observed in the spectra is attributed
quantum efficiency as well as the achievable steady-state carrier to the carriers’ plasma frequency. For an increasing excitation
density. These characteristics can be extracted in a direct way power, the minimum shifts to higher energy and becomes at the
using broadband, time-resolved pump–probe transmission, and same time more pronounced. As the plasma frequency shifts in
reflection spectroscopy. Possible ways for performing such experi- first order proportional to the square root of the total amount
ments could be via tunable lasers or supercontinuum sources. of charge carriers in the system, such reflection measurements
However, particularly synchrotron-based infrared pump–probe facilitate a convenient method for the quantitative determination
spectroscopy has been shown to offer advantageous conditions for of the carrier density. Thus, the extracted carrier concentration in
measuring the carrier density, their lifetime as well as gain and loss dependence of the optical pump power for delay times of 0 and
due to its extended bandwidth and suitable pulse lengths (Carroll 250 ps is shown in the inset of Figure 7A. Moreover, by analyzing
et al., 2012; Geiger et al., 2014a,b). At the infrared beamline of the carrier density at a fixed pump power for varying delay times,
the SLS, 100 ps long pulses of infrared light are supplied from the the reflection spectra can be used to extract the carrier decay times.
synchrotron and serve as broadband probe pulses, whereas the In the case shown here, the carrier density drops to ~4 × 1019 cm−3
excess charge carriers are optically excited by a 100 ps Nd:YAG within 250 ps for all generated carrier concentrations larger than
laser at 1064 nm (Carroll et al., 2011). The delay time between 4 × 1019 cm−3. This behavior indicates an increasingly faster decay
pump and probe pulses can be varied electronically, which offers time at high-carrier concentrations, which is attributed to Auger
the possibility to follow the dynamics of a system over a long time recombination (Carroll et al., 2012).
period by probing at different times after excitation. In the follow- While the analysis of mid-infrared reflection spectra enables
ing, we review some of the pump–probe measurements performed to directly access charge carrier concentration and decay time,
at the SLS and give the most important results that challenge the the latter can also be extracted from near-infrared transmission

FIGURE 7 | Time-resolved infrared reflection and transmission the total amount of charge carriers. The inset shows the carrier concentration for
spectroscopy: (A) mid-infrared reflection spectra of Ge on Si expressed 0 and 250 ps delay time in dependence of the excitation power. (B) Normal-
as the ratio of pumped (RP) and unpumped reflection (RU) for varying incidence pump–probe transmission spectra for Ge on SOI for varying delay
excitation power at a pump–probe delay of 250 ps. The resonance in the times. Strong Fabry–Perot oscillations are observed from the thin film
spectra is attributed to the carrier plasma frequency, which enables to extract interference. Analyzing the peak-shifts facilitates the extraction of the decay time.
measurements. In Figure 7B, normal-incidence transmission

spectra of intrinsic Ge are plotted, while the delay time between
pump and probe is varied. As SOI is used as substrate, distinct
FP oscillations are observed due to standing wave interferences
between the Ge/air and Si/SiO2 interfaces. For short delay times,
the transmission is significantly reduced due to absorption. Above
the direct band gap of ~0.8 eV, there is an increase compared
to the unpumped transmission due to gain or bleaching. By fol-
lowing the shifts of the minima or maxima, the dynamics of the
refractive index is obtained, which enables the extraction of the
carriers’ decay time. Compared to the decay time analysis from
the mid-infrared reflection, the sensitivity to detect small carrier
densities is higher in such a measurement because the refractive
index – and, hence, the oscillation extrema – follows the carrier
densities linearly, which enables to follow the decay processes
within an extended time window.
In Figure 8, the time-dependent FP peak shifts are shown for
differently prepared Ge layers. The shifts were normalized to unity
at t = 0 ns and the decay fitted to an exponential curve (Geiger
et al., 2014a). The defective Ge/Si interface was identified as the
main non-radiative loss channel, as (i) Ge selectively grown via FIGURE 8 | Normalized peak shifts taken from normal-incidence
transmission spectra (as, e.g., in Figure 6B) for a series of differently
ultrahigh vacuum CVD (selGe in Figure 8) and a full epilayer prepared Ge layers: iGe, nGe, selGe refers to intrinsic, n-doped, and
grown via low-energy plasma-enhanced CVD (iGe in Figure 8) selectively grown Ge, respectively. GOI refers to Ge on insulator. The
feature the same surface recombination velocity (SRV) – i.e., the non- radiative lifetime is obtained through an exponential fit to the data
carrier lifetime normalized to the layer thickness – of ~800 m s−1, (Geiger et al., 2014a).
(ii) a built-in field introduced by modulation doping (nGe/iGe)
increases the lifetime compared to iGe by keeping electrons away efficiency and, thus, a low-threshold laser. Furthermore, similar
from the interface, and (iii) the longest lifetime was observed for pump–probe transmission studies on strained microbridges
an overgrown GOI wafer, where the defective Ge/Si interface is showed that neither strain, at least up to ~2%, nor processing
removed (SRV = 490 m s−1). These results demonstrate the impor- affects the lifetime (Geiger et al., 2014b), indicating that a high-
tance of engineering the material- and, in the case of Ge on Si, crystal quality can be maintained using the microbridge strain-
especially the interface quality to obtain a high-internal quantum enhancement technology.

For the analysis of gain and loss, the transmission spectra From the preceding analysis, it is clear that a solid understand-
should be recorded under the Brewster angle such that the ing and consideration of the loss processes is required for an
obstructing FP resonances do not occur, c.f. Figure 9A for the accurate description of gain in Ge. For illustration, the absorp-
case of unstrained Ge on Si for different pump–probe delay tion cross-sections for three Ge samples (Ge#1: Nd = 0, εxx = 0;
times. Δt = 0 ps refers to the maximum overlap between pump Ge#2: Nd = 2.5 × 1019 cm−3, εxx = 0; Ge#3: Nd = 0, εxx = 0.25%) are
and probe and, hence, to the highest carrier density. The thick plotted in Figure 10 in dependence of the total carrier density
lines in blue and red show modeled transmission spectra for the NT = Nd + NP, where Nd refers to the doping concentration and NP
unpumped and pumped case. Under excitation, a strong absorp- to the pump-induced carrier density. As a comparison, the cross-
tion occurs with a linear dependence on energy. At the direct sections of three InGaAs layers (InGaAs#1: Nd = 0, InGaAs#2:
band gap, the absorption gets reduced due to gain, but the gain Nd = 5.3 × 1018 cm−3; InGaAs#3: Nd = 2.1 × 1019 cm−3) are plotted
is too small to generate a negative absorption and, hence, light as well. Therefore, Figure 10 reveals that the absorption scales pre-
amplification. This situation holds true for all other delay times, dominantly with NP indicating that the absorption cross-section
i.e., carrier concentrations, as well. from holes σh is much larger than the cross-section for electrons
In Figure 9B, the situation for Δt = 0 ps is shown again in terms σe. Indeed, describing the absorption via a linearly dependent
of the absorption coefficient with the modeled functions being cross-section as α = σeNe + σhNh (where subscripts e and h refer
plotted separately for (i) direct gap absorption before pumping to electrons and holes, respectively) offers a good representation
(blue, thick line), (ii) direct gap absorption under excitation (red, of the experimental data with σh/σe > 10. The absorption cross-
thick line), and (iii) the featureless pump-induced absorption section for holes is significantly larger than for electrons, because
decreasing linearly in energy (red, thin line). Even though a gain in addition to the non-momentum conserving intraband or
of ~850 cm−1 is observed as displayed by a negative absorption, the Drude-type free carrier absorption, the holes can undergo vertical
loss from that spectrally distributed absorption at the same energy intervalence band transitions (Newman and Tyler, 1957), which
is >6000 cm−1 making light amplification impossible. To show the are hereby identified as the main loss channel in Ge. A similar
contrast to an established laser material featuring a direct band gap, conclusion concerning the cross-section ratio can be deduced from
the same absorption properties are plotted in Figure 9C for the the InGaAs data shown in Figure 10 in agreement with common
case of InGaAs. Here, the pump-induced losses are independent knowledge for direct band gap lasing materials (Adams et al., 1980;
on energy and amount to ~1000 cm−1, which is compensated by Childs et al., 1986). Furthermore, the absolute values of the hole
a direct gap gain of ~1700 cm−1 such that a net gain of 700 cm−1 cross-section are in a similar range but larger for InGaAs than for
is revealed. We should mention here that the theoretical analysis Ge, c.f. Figure 10. However, the absorption is much higher in Ge
of Carroll et al. (2012) has been questioned (Dutt et al., 2012) due to a much larger total carrier density needed to achieve a gain
concerning the strength of the gain (red line, Figure 9B) but not like in the InGaAs sample.
the experiments, which clearly show that the loss is by far larger Finally, we would like to relate the above presented data of gain
than the gain. and loss as well as of lifetimes in Ge layers on Si to the observation
A B C
FIGURE 9 | Analysis of pump and probe measurements. absorption (red, thin line) obtained from the spectra in (A) at 0 time delay.
(A) Transmission spectra for Ge on Si measured under Brewster angle for (C) Similar extraction of direct gap gain and losses for undoped InGaAs
different delay times. (B) Modeled direct gap absorption unpumped showing light amplification, as the gain surpasses the pump-induced losses
(blue, thick line) and pumped (red, thick line) as well as pump-induced (Carroll et al., 2012).

PL, as has been demonstrated recently in Wirths et al. (2015). In

Pump-induced Absorption ( cm -1 ) 5000 Figure 11B, the temperature-dependent PL intensity for a set of
Linear Fits
InGaAs#1 samples with Sn content from 8 to 13% is shown. The data have
4000 InGaAs#2 been normalized to unity at 300 K. For the sample with the lowest
InGaAs#3
Sn content, a rapid drop in intensity on lowering the temperature is
3000 observed, whereas for the three other samples a steady increase in
intensity can be seen with the intensity increase being dependent
2000
on the Sn concentration. Qualitatively, the increase from sample
Ge#1 to sample can be explained by the reduced conduction band offset
1000 Ge#2 with increasing Sn, whereas cooling down leads to a condensation
Ge#3
of the carriers into the lowest energy states such that the direct
0
0.0 0.4 0.8 1.2 1.6 0.0 0.6 gap emission either vanishes for indirect band gap materials as
20
NT x 10 ( cm )
-3 sample Ge0.92Sn0.08 because the Γ valley is not populated anymore or
increases strongly when the electrons condense at the minimum of
FIGURE 10 | Absorption coefficients for differently doped and strained the Γ valley. To quantify the band offsets for the set of GeSn samples,
Ge and InGaAs layers. The data are well described with a linear absorption the emission efficiency is calculated via a similar JDOS model as the
cross-section model. one used for calculating the gain in Figure 4. Therein, the offset
ΔE between Γ- and L-valleys as well as the injected charge carrier
density Ni represent the free fitting parameters. Furthermore, the
of lasing in highly n-doped and weakly strained Ge (Liu et al., 2010; temperature dependence of the lifetime is assumed to be identi-
Camacho-Aguilera et al., 2012), which was recently repeated at the cal for all samples and follows the Shockley–Read–Hall (SRH)
University of Stuttgart with an unstrained, highly n-doped light recombination characteristics (Shockley and Read, 1952; Schubert,
emitting diode (LED) (Koerner et al., 2015). First, as is shown 2006) that describes non-radiative recombination via trap states.
above by the gain and loss experiment (Carroll et al., 2011), the Using this model, an excellent agreement with the experimental
loss in Ge layers strongly exceeds the gain at all the investigated data is obtained (Wirths et al., 2015). The GeSn sample with 13%
carrier densities up to 1020 cm−3 and in all investigated cases, i.e., Sn content is, hence, identified as a true direct band gap group IV
Ge with and without weak strain and/or n-doping. Hence, Carroll’s semiconductor with its Γ-valley being 25 meV below the indirect
results are apparently in conflict with the observation of lasing L-valleys. From the second fit parameter, a RT carrier density Ni
(Liu et al., 2010) in similar but not identical material. Second, the of 4 × 1017 cm−3 is deduced consistent with a carrier lifetime of
non-radiative lifetimes, which have been determined for such Ge 0.35 ns, which corresponds to a SRV of 570 m s−1. This value is
layers only recently by Geiger et al. (2014a), c.f Figure 8, shine close to the ones reported for elemental Ge on Si (Geiger et al.,
a new light on previous and recent gain and threshold current 2014a), c.f. Figure 8, which is an indication of the high-crystalline
density calculations (Dutt et al., 2012; Peschka et al., 2015). Using quality of the investigated GeSn epilayers.
the obtained carrier lifetime of the order of 1–2 ns for thresh- The temperature dependence of the non-radiative carrier
old current estimates, the calculated threshold of the order of lifetime is obtained as:
100 kA cm−2 – obtained by assuming a lifetime of 100 ns – needs
to be rescaled by a factor of 50–100. Surely, such a current density is t = (1 / t0 + 1 / tSRH + 1 / t Auger )−1
(1)
above the material’s limit and also exceeds the observed threshold
values by ~2 orders of magnitude. Therefore, not only the gain/loss where τ0 describes the lifetime at low temperature, and τSRH
experiments but also the theory (when fed with properly valued describes the decay due to the capture of charge carriers by mid-gap
parameters) shows that more research is needed to understand states, i.e., τSRH = A × (1 + cosh(ET/kT)). τAuger describes the Auger
the MIT results. The recent paper by the Stuttgart group (Koerner recombination time, which can be neglected here due to the low-
et al., 2015) may give the directions for further thinking: a “lasing” carrier densities. Furthermore, ET is the difference between the trap
threshold was reached only shortly before their devices failed, level energy and the intrinsic Fermi-level, k is the Boltzmann con-
hinting at a carrier breakthrough. The heat pulse related to the stant, and A is to normalize τ to 0.35 ns at 300 K as obtained from
breakthrough may have caused the peaked emission signal. the temperature-dependent PL. For ΔE = 19 meV and τ0 = 2.1 ns,
a good agreement between the extracted lifetimes and the lifetime
model is obtained (Wirths et al., 2015). For temperatures >50 K,
Photoluminescence – Direct Band Gap there is a drastic decrease in carrier lifetime from ~2 ns to 350 ps
Photoluminescence spectroscopy offers a convenient tool for prob- for higher temperatures. As the temperature dependence of this
ing the changes of the electronic band structure induced via strain process is well described via the SRH model, the lifetime decay is
or Sn alloying. As shown in Figure 11A, the reduced offset between attributed to the capture of carriers via mid-gap states originating
Γ- and L-valleys manifests in an increased emission intensity of from defects (Wirths et al., 2015). These defects could potentially
the PL signal (Süess et al., 2013). A similar effect has also been be related to defects located at the GeSn/Ge-interface (Geiger
observed by Chen et al. (2011a). et al., 2013; Wirths et al., 2015), but further studies are needed
A quantitative analysis of the relative alignment between Γ- to unambiguously identify the origin of this deterioration and,
and L can be obtained from the temperature dependence of the subsequently, improve the material quality.

A B
FIGURE 11 | Photoluminescence investigation of strained Ge and Sn-concentration leads to a more pronounced increase in intensity.
GeSn alloys. (A) Room-temperature PL spectra for Ge samples with The offset between Γ- and L-valleys is extracted from JDOS modeling
increasing uniaxial strain up to 3.1%. The inset shows the good agreement (solid lines), which reveals Ge0.87Sn0.13 to have a fundamental, direct band
between the experimental and modeled band edges. (B) Temperature- gap. The inset shows the experimentally extracted non-radiative lifetime
dependent integrated PL intensity normalized to unity at 300 K for a series modeled with a Shockley–Read–Hall-like temperature dependence
of GeSn layers with Sn-content ranging from 8 to 13%. An increase in (red line).
Optically Pumped Laser temperature is equivalent to the temperature range where the
According to the modeling results shown in Section “Modeling,” lifetime was found to drop substantially from ~2 ns to 350 ps.
a direct band gap Ge-based system should feature a net gain and, Hence, it is tempting to attribute the limitation of lasing to
hence, enable light amplification at low excitation. In the previ- temperatures <100 K to the carrier capture by defect-induced
ous analysis of low-temperature PL on GeSn alloys in Section mid-gap states, as has appealed from the analysis shown in
“Photoluminescence – Direct Band Gap,” a fundamental direct Figure 11B, inset. However, carrier transfer to the L-valleys and
band gap could be identified for a Sn content of 13% in a strain- carrier out diffusion into the Ge may be a determining factor,
relaxed layer with 0.7% compressive biaxial strain. To show lasing, as well.
a 560-nm epilayer of such Ge0.87Sn0.13 material was grown providing Despite the breakthrough of presenting for the first time a direct
an overlap of 60% for the fundamental transverse electric mode in band gap group IV material that is lasing under optical pumping,
a 5 μm wide FP cavity (Wirths et al., 2015). For this layer, modal there still remain open questions. For example, with an excitation
gain could be observed at 20 K via the variable-stripe-length (VSL) power of 325 kW cm−2, a non-radiative lifetime of 2 ns, as shown
method under pulsed optical excitation at 1064 nm with a dif- in Figure 11, and a typical absorbance of 1 × 104 cm−1 at 1064 nm,
ferential gain of ≈0.4 cm kW−1 and a threshold excitation density of a steady-state carrier density of ~3.5 × 1019 cm−3 is estimated.
≈325 kW cm−2 (c.f. Figure 12). Above threshold, the gain increases With this number, the gain at low temperature from our model
linearly with excitation and can easily pass 100 cm−1. The stripe is found to be >5000 cm−1. And, interpolating from Figure 4, at
length-dependent PL analysis is a widely applied technique to excitation density of 0.6 × 1018 cm−3 we would expect for a system
measure net modal gain, but it does not allow to resolve the gain with positive offset of about 15%, a material gain of ~300 cm−1
and loss as by pump and probe spectroscopy. More evidentially, a at RT. We assign this large discrepancy from what is observed at
gain statement, such as provided by Figure 12, becomes respected low temperature and what a RT calculation predicts to resonant
only after showing lasing. intervalence band absorption. As mentioned above, due to the
Indeed, when pumping a FP cavity over its full length, a lack of experimental data for direct gap Ge or GeSn, the energy
strongly enhanced emission and narrowing of the line spectra dependence of the loss as measured by pump–probe experiments
is observed as soon as the modal gain surpasses the cavity losses. for Ge (Carroll et al., 2012) has been used for Figure 4. Its proper
This behavior is shown in Figure 13 where the edge-emission inclusion possibly adds significant contribution to the loss (Wen
spectra from a 1-mm long FP cavity at 20 K are plotted for vary- and Bellotti, 2015).
ing optical excitation powers. The curves are offset for clarity. In order to improve such gain calculations, which critically
The threshold obtained from the lasing experiments matches well depend on the knowledge of the band structure, mappings of the
with the one obtained from the VSL method. For an excitation entire valence, and conduction band in reciprocal space would cer-
density of 1 MW cm−2, lasing could be observed up to 90 K. This tainly be highly valuable. This could be possible via angle-resolved

A B
FIGURE 12 | Gain extraction via the variable-stripe-length method extracted by an exponential fit to the data. (B) Differential gain extracted from
(VSL). (A) Edge-emitted intensity of Ge0.87Sn0.13 at 20 K in dependence of the the spectra in (A) indicating a linear dependence on excitation density with a
pumped waveguide length for varying excitation densities. The modal gain is threshold at 325 kW cm−2.
In fact, similar lifetimes for Ge as obtained from synchrotron

measurements (Geiger et al., 2014a) were found (Nam et al., 2014).
Challenges
In the previous section, we reviewed experiments and results
related to the dependence of the optical properties on strain and
alloying of Ge with Sn. Furthermore, we summarized investiga-
tions concerning the first lasing of a direct band gap group IV
semiconductor and expounded on the temperature dependence
of the PL as a powerful tool to determine the directness of a
group IV material. We illustrated optical methods based on pump
and probe spectroscopy using synchrotron light to determine the
carrier lifetime, gain, and loss under optical pumping related to
the injected carrier density.
Future experiments along these lines on both, the strained Ge
system and GeSn alloys at various strain and Sn concentration,
respectively, will allow to establish the fundamentals of lasing in
direct band gap group IV systems. The impact of doping on gain,
loss, and carrier lifetime should also be addressed in dependence
FIGURE 13 | Lasing emission spectra measured from the facet of a of the directness of the respective system to verify the picture
5-μm wide and 1 mm long FP waveguide cavity under optical elucidated by Figure 4 of Section “Gain.”
pumping at 20 K. A clear threshold behavior can be observed in the spectra As an example, intervalence band absorption, Auger recom-
with respect to output intensity and linewidth, c.f. inset to the right and left
bination, and the electrical injection are some of the many
hand side, respectively.
fundamental aspects of group IV direct band gap lasing pending
to be understood and quantified.
photoelectron spectroscopy (ARPES) at high energy (Gray et al.,
2011) or the soft x-ray regime (Strocov et al., 2014). Other promis- Intervalence Band Absorption
ing experimental techniques not covered because of lack of space One of the most essential parameters determining the efficiency
include lifetime measurements via time-resolved PL measure- of a laser is associated to the parasitic absorption due to the
ments (He and Atwater, 1997; Nam et al., 2014; Saito et al., 2014). injected holes (Adams et al., 1980; Childs et al., 1986). As shown

experimentally by Carroll et al. (2012) for Ge, this absorption (μe = 3900 cm2/Vs and μh = 1900 cm2/Vs, respectively) (Golikova
depends linearly on the excitation and inclines with decreasing et al., 1962; Jacoboni et al., 1981). For optical devices, this is
energy. This can be understood from the Drude dependence of appealing because it results in diffusion lengths of several
the free carrier absorption modified by dipole allowed intervalence 100 μm, more than sufficient for, e.g., typical detector absorber
band transitions. Because the emission wavelength increases when sizes and laser cavity lengths. However, at the same time, Ge
approaching the direct band gap configuration, and the initially suffers from its low band gap, which causes large leakage currents
degenerate heavy and light hole bands split due to strain, the para- in Ge pn junctions (Metzger et al., 2001; Satta et al., 2006), thus,
sitic absorption will strongly increase in direct band gap systems requiring extensive work on surface passivation due to the lack
and may, thus, obstruct the efficiency of lasing (Wen and Bellotti, of a native oxide.
2015). Applying the above introduced optical characterization The already beneficial mobility properties can be further
methods should allow to investigate these effects in detail, which, improved by employing tensile strain as has been shown for both,
together with evolving theoretical results, will enable to complete uniaxially and biaxially strained Ge (Schetzina and McKelvey,
our understanding. 1969; Chu et al., 2009; Chen et al., 2011b). Here, we would like to
highlight that all these studies have been performed on indirect
band gap Ge where the increase in electron mobility is mediated
Auger Recombination
by a reduction of the effective mass in the L valley. Even without
On the material level, the performance of optical devices depends
strain, the Γ-valley already offers an ~8 times smaller effective
strongly on the charge carrier recombination lifetime similarly
mass and correspondingly higher electron mobility. Similarly,
as described by rate Eq. 1. Here, both the radiative and the
an increase in hole mobility is expected due to the lifting of the
Auger recombination lifetime depend on the carrier density
valence band degeneracy (Beattie and Landsberg, 1959; Fischetti
n: 1/τrad = B × n, 1/τAuger = C × n2. The quadratic carrier density
and Laux, 1996). This is advantageous as a high mobility strongly
dependence implicates that Auger recombination becomes a
reduces the resistivity of the device allowing an efficient injection
dominant loss mechanism at high-charge carrier densities, which
and extraction of charge carriers.
can be the order of >1019 cm−3 for typical laser devices.
GeSn emerged as a material of interest in electronics only
Extensive theoretical work based on perturbation theory has
recently; therefore, less transport data are available. However,
shown that, despite its indirect band gap, the band structure of
theoretical studies predict very large electron mobilities as well
both, Si and Ge, is favorable for direct Auger recombination (Huldt,
as hole mobilities of the order of 4500 cm2/Vs for direct band
1974; Lochmann, 1978) with Auger recombination coefficients of
gap GeSn (Sau and Cohen, 2007). The first reported experimental
the order of 10−30 cm6s−1. This is comparable to direct band gap
mobility study has been done on low Sn content (<6%) indirect
materials like GaAs or GaN used for optical devices in the visible
band gap GeSn layers yielding a Hall mobility of the order of
part of the spectrum but still much smaller than for low band gap
~200–300 cm2/Vs (Nakatsuka et al., 2010). Slightly better results
materials like InAs (Metzger et al., 2001).
have been obtained thereafter investigating p-MOSFETs hole
Significantly, less is known about direct band gap group IV
channel mobility (Gupta et al., 2013a; Wang et al., 2013).
Auger recombination. For direct band gap Ge0.9Sn0.1/Ge0.75Si0.1Sn0.15
In summary, we see that a vast amount of knowledge concerning
multi-quantum-well structures, Sun et al. recently showed
the mobility exists leaving a good base for further studies. Moreover,
theoretically that the RT Auger recombination lifetime is of the
many electrical devices and the corresponding fabrication tech-
order of 50 ns compared to a radiative lifetime of 10 ns (Sun et al.,
niques, e.g., passivation, contacting, or annealing, have been
2010). Other work on GeSn (Dutt et al., 2013) and n-doped or
conceived allowing for a fast implementation in optical devices.
tensile strained Ge (Liu et al., 2007; Jain et al., 2012) only refers
However, besides the tremendous changes in the carrier mobil-
to unstrained bulk Ge recombination coefficients to include in
ity, there are additional effects coming into play with electrical
their gain models.
injection of charge carriers from indirect to direct band gap Ge.
We believe, however, that this is unjustified considering that
Exemplarily in Figure 14, such an injection scheme in form of a
there is an exponential dependence of the Auger lifetime on the
pin diode is discussed for the case of tensile strained Ge bridges
band gap and effective masses (Beattie and Landsberg, 1959; Huldt,
where the strain profile is shown in Figure 5B.
1971; Adams et al., 1980), both being strongly altered in direct
Far from the strained constriction, electrons can be injected
band gap Ge. Using the simple exponential dependence derived
into the L-valleys of the conduction band as in standard Ge diodes.
by Beattie and Landsberg (1959) to scale the experimentally
However, close to the center the strain profile alters the band
determined Auger recombination coefficient of CGe = 10−30 cm6s−1
structure with L- and γ-valley starting to cross, which allows for
(Carroll et al., 2012) via the effective masses and band gaps of direct
intervalley scattering (Boucaud et al., 2013) from a high- into a
band gap Ge, an Auger coefficient of the order of 10−26–10−27 cm6s−1
low-effective mass valley with a higher mobility, a process inverse
is obtained. Despite the overly strong simplicity of this comparison,
to the Gunn effect (Gunn, 1963). This may support current extrac-
it shows that Auger coefficients will most probably increase and
tion and injection in optical devices. However, an actual impact
need to be addressed and investigated in the future.
still needs to be proven.
Carrier Injection Cavity Design

From an electronic point of view, Ge is one of the most interest- For the usage of direct band gap materials in lasing structures,
ing materials as it offers both, high electron and hole mobilities high-quality factor optical resonators are necessary confining the

Finding the intrinsic stress limits of Ge is another item of

interest in this context.
By saying this, we conclude our listing of fundamental and
materials-related challenges. This list may be incomplete. However,
it confirms that the research and development of a laser source
from a group IV material will involve many disciplines from
fundamental to device physics and from wave optics to material
and transport properties. To progress fast, a collaborative effort
FIGURE 14 | Schematic of the band structure of a forward biased is demanded.
pin strained Ge bridge diode where the intrinsic layer and strained
region overlap.
Opportunities
light in the gain region and, thus, allowing for stimulated emission.
Photonics
Direct band gap group IV laser systems may permit a qualita-
Helpful in this regard is the high-refractive index contrast of the
tive as well as a quantitative expansion of Si-photonics (group
Ge–air and GeSn–air interfaces (Kasper et al., 2013), which should
IV photonics) into traditional but also new areas of applications.
allow for good confinement properties. The design of suitable
However, it is requested that such lasers can be operated energy
cavities seems to be straightforward for GeSn lasers where the
efficiently, under ambient conditions and can be fully integrated
wafer-scale direct band gap on Si or Ge already facilitated the
with current Si technology. An answer to whether this is possible
implementation of well-known laser cavities, such as FP cavities
cannot be given yet as the research is at an early stage. We can
(Wirths et al., 2015) or microdisks (Cho et al., 2011). At the same
only speculate about the specifications of such a laser and, thus,
time, optical microdisk cavities with quality factors of ~1400
have to guess which of the applications would profit most from
have been demonstrated in tensile strained Ge using external
a successful implementation of group IV lasers. Hence, for the
SiN stressors (Ghrib et al., 2013) as well as waveguide cavities
following discussion, let us assume that this all-group-IV laser does
(Capellini et al., 2014).
indeed exist and it operates (i) under electrical injection, (ii) at RT
The situation is much more complex for uniaxially strained
or above, and (iii) with reasonable power conversion efficiency.
Ge bridges where patterning of the bridge inevitably relaxes the
What could we do with such a device, where is the highest impact,
strain and, hence, prohibits a fundamental direct band gap. This
and what is the platform of choice?
excludes many popular cavity designs, in particular, microdisks,
A high return is merely achievable when this laser device will be
photonic crystals, and FP cavities. Hence, distributed feedback
combined with the current Si photonics by using the same platform.
structures, which do not rely on patterning of the strained region,
Most advanced are photonic elements fabricated on SOI, except
are currently under investigation (Marin et al., 2015).
for applications in the visible part of the spectrum – not covered
here – where SiN-based structures are often used. SOI for photon-
Band Gap Renormalization and Material Stability ics typically consists of Si layers with a thickness of ~200–250 nm
Relevant for both here discussed direct band gap systems is a quan- and a several micrometer thick buried oxide to avoid leakage of
titative analysis of the band gap renormalization of the involved Γ the propagating modes into the Si substrate. For strain engineer-
and L valleys in dependence of their respective carrier population. ing, the compatibility with SOI has already been shown (Süess
So far, experiments suggest that the renormalization corrections et al., 2013), c.f. Section “Microbridges,” and, as mentioned above,
are comparable for the two valleys. Hence, the offset between the bridges with even higher mechanical strength are fabricated from
Γ and L states would not depend on the injection density, which GOI using wafer transfer (Sukhdeo et al., 2014). In fact, wafer-scale
is essential for a stable injection. fabrication of GOI using the SmartCut® process has already been
Moreover, material specific investigation concerns, for example, established several years ago for electronics (Augendre et al., 2009).
the thermal stability of GeSn and SiGeSn metastable alloys with GOI for photonic applications, where thicker layers and a thicker
regards to Sn diffusion and segregation where extensive segrega- BOX are required, has been presented recently by Reboud et al.
tion can result in changes of the emission wavelength and/or (2015). A photonic platform based on GOI, in comparison to SOI,
emission efficiency. Recently, investigations have been pursued has the advantage that all photonic elements, such as waveguides,
to examine the temperature budget a GeSn or SiGeSn device would bends, and the resonant structures, can be reduced in size because
be able to withstand, e.g., by in situ studies (Fournier-Lupien et al., of the larger refractive index contrast. This allows for the potential
2014) or annealing experiments (Wirths et al., 2014). First, in situ fabrication of more dense optical circuits and, hence, for an easier
results indicate phase separation of a 12% Sn containing ternary integration with electronics. Furthermore, Ge provides coverage
SiGeSn and binary GeSn alloys at ~420 and 460°C, respectively, of the longer wavelengths toward 10 μm and more. Moreover,
which is surprising considering that the higher mixing entropy by using processes that are selective for either Ge or Si, the GOI
usually results in a higher thermal stability of ternary alloys platform may provide additional fabrication opportunities. The
(Fournier-Lupien et al., 2014). Annealing experiments revealed high quality (Si)GeSn presented by Wirths et al. (2013b) has been
distinct Sn diffusion at 300°C for GeSn with approximately the deposited on a Ge VS on Si(001) indicating that the growth on SOI
same composition (Wirths et al., 2014). and certainly GOI is possible, as well.

With a laser device implemented on the currently used SOI or physics, and Si electronics and photonics will cooperate and
similarly on GOI, various new applications will emerge. Before define the routes to opto-electronics for fast and energy-efficient
speculating, we may picture the many already existing photonic data processing.
elements. To select a few: Low-loss (<1 dB cm−1) single mode
waveguides in various designs, tapers to adiabatically match the
waveguide modes to fibers, and low-loss grating couplers (<1 dB) Conclusion and Outlook
in 1D or 2D providing polarization splitting. Other standard
elements include directional couplers, Mach Zehnder interfer- We reviewed the methods for achieving a direct band gap in
ometers, ring resonators, and light modulators based on free group IV semiconductors in the most promising material system
carrier injections or quantum confined Stark effect. Detectors for the prospect of a Si compatible laser, namely, Ge modified
based on slightly tensile strained (0.2%) Ge provide more than either via tensile strain or by alloying with Sn. We expanded on
10 GHz speed for high-data rate transmission. Other device the methods to characterize these systems and gave examples
elements include add-drop filters, buffers, and switches, which on their optical properties. The recent advances in numerous
can be integrated with fluidic channels for (bio-) sensing. The approaches to achieve a direct band gap have finally concluded
short wavelength infrared at which the here discussed group IV in the first demonstration of lasing in a direct band gap GeSn
direct band gap lasers emit (see Figure 3) is certainly a clear asset alloy (Wirths et al., 2015).
for sensing applications (Soref, 2010; Nedeljkovic et al., 2013). With this demonstration, we are at the beginning of an excit-
Silicon-on-insulator is (and GOI could become) a very conveni- ing journey in the field of silicon photonics. As shown in great
ent platform to realize high-performing photonic crystal structures detail, the many optical characterization tools at hand allow us
enabling unique photonic circuits, such as compact high-Q cavi- to address a large amount of fundamental questions, including
ties, which can operate stably at single and dual-wavelength, and band gap renormalization, various recombination processes, and
as wavelength division multiplexer as desired for optical signal doping level-dependent lasing performance, but also material- and
processing. This listing can be extended almost indefinitely, nam- technology-related issues, such as high Q-factor cavity design,
ing, e.g., switching and steering of optical signals, slow light, pulse diffusive carrier transport, stress, and thermal diffusion limits.
compression, customized reflectors, and filters. Together with the We hope that with outlining these challenges, we can motivate
expectation that such photonic circuits will be very cost effective, a vast amount of new researchers from various backgrounds in
compact, reliable, and efficient, a monolithically integrated laser optics, material science, and device physics to join this interesting
source will certainly bring new functionality, in particular when research field. We believe that combined efforts will converge in
optics can be merged with electronics. a reasonable time to a demonstration of a practical laser source
being electrically pumped, highly efficient, and fully integrated on
an electro-optical CMOS platform. This building block will finally
CMOS Integration pave the way for true monolithic on-chip integration of photonics
The combination of optics with CMOS electronics to realize an and CMOS electronics for new sensors in the long wavelength
on-chip data distribution network (Heck and Bowers, 2014) infrared, and will eventually enable to build an on-chip or off-chip
is – without any doubt – one of the most advanced and chal- electro-optical data distribution network for high-performance
lenging applications for direct band gap group IV lasers. The computing.
requirements are so complex (Miller, 2009) that before the start
of such a development, many fundamental questions have to be
answered, such as the efficiency issues among other challenges, Acknowledgments
which have been addresses in the previous section. However,
once these hurdles are taken, we expect to arise a highly competi- We would like to acknowledge the many scientific collaborators
tive and attractive platform solution for future data processing we were fortunate to work with over the last few years. They
applications. In fact, the extension of CMOS by integration of Ge supported us in building up a strong portfolio in the investiga-
and (Si)GeSn may not just resolve the demands for a monolithic tion and understanding of lasing in group IV systems, and the
laser gain medium, but, as discussed widely elsewhere (Kao et al., fabrication of direct band gap group IV materials. In particular,
2014), (Si)GeSn would already advance the performance of the we thank our previous group members Gustav Schiefler, Martin
electronic circuits. This appealing double benefit, together with J. Süess, and Renato Minamisawa for their contributions,
the potential compatibility to CMOS of such an all-group-IV which led to this appealing strain concept, and the group of
solution, bears an essential advantage in comparison to other Dan Buca (FZ Jülich), who contacted us for investigating their
emerging technologies, such as spin- and/or valley-based elec- high quality material and thus gave us the opportunity to learn
tronics, which rely in part on non-conform chemical elements also about GeSn alloys. The tremendous progress achieved in a
and non-CMOS fabrication processes. short time is a shining example of our good collaboration. We
Hence, we expect that as soon as the fundamental lessons also thank Jérôme Faist and Ralph Spolenak (ETHZ) for their
of direct band gap lasing are learnt and a gain medium well- whole-hearted support to this subject and their many essential
qualified for injection pumping at RT is defined, research and contributions. Finally, we acknowledge the Swiss Science
development of a new opto-electronic platform will quickly Foundation (SNF) for supporting part of the here reviewed
advance. Experts in CMOS technology, group IV epitaxy, laser studies over several years.

References Chen, Y. T., Lan, H. S., Hsu, W., Fu, Y. C., Lin, J. Y., and Liu, C. W. (2011b). Strain
response of high mobility germanium n-channel metal-oxide-semiconductor
Adams, A. R., Asada, M., Suematsu, Y., and Arai, S. (1980). The temperature dependence field-effect transistors on (001) substrates. Appl. Phys. Lett. 99, 022106.
of the efficiency and threshold current of In 1- xGa xAs yP 1- y lasers related doi:10.1063/1.3604417
to intervalence band absorption. Jpn. J. Appl. Phys. 19, L621–L624. doi:10.1143/ Childs, G. N., Brand, S., and Abram, R. A. (1986). Intervalence band absorp-
JJAP.19.L621 tion in semiconductor laser materials. Semicond. Sci. Technol. 1, 116–120.
Akatsu, T., Deguet, C., Sanchez, L., Allibert, F., Rouchon, D., Signamarcheix, T., et al. doi:10.1088/0268-1242/1/2/004
(2006). Germanium-on-insulator (GeOI) substrates – a novel engineered substrate Cho, S., Chen, R., Koo, S., Shambat, G., Lin, H., Park, N., et al. (2011). Fabrication
for future high performance devices. Mater. Sci. Semicond. Process. 9, 444–448. and analysis of epitaxially grown Ge1-xSnx microdisk resonator with 20-nm
doi:10.1016/j.mssp.2006.08.077 free-spectral range. IEEE Photon. Technol. Lett. 23, 1535–1537. doi:10.1109/
Aldaghri, O., Ikonić, Z., and Kelsall, R. W. (2012). Optimum strain configurations LPT.2011.2163929
for carrier injection in near infrared Ge lasers. J. Appl. Phys. 111, 053106. Chow, W. W. (2012). Model for direct-transition gain in a Ge-on-Si laser. Appl. Phys.
doi:10.1063/1.3691790 Lett. 100, 191113. doi:10.1063/1.4714540
Augendre, E., Sanchez, L., Benaissa, L., Signamarcheix, T., Hartmann, J.-M., Chu, M., Sun, Y., Aghoram, U., and Thompson, S. E. (2009). Strain: a solution for
Le Royer, C., et al. (2009). Challenges and progress in germanium-on-insulator higher carrier mobility in nanoscale MOSFETs. Annu. Rev. Mater. Res. 39, 203–229.
materials and device development towards ULSI integration. ECS Trans. 25, doi:10.1146/annurev-matsci-082908-145312
351–362. doi:10.1149/1.3203972 Du, W., Ghetmiri, S. A., Conley, B. R., Mosleh, A., Nazzal, A., Soref, R. A., et al.
Bauer, M. R., Tolle, J., Bungay, C., Chizmeshya, A. V. G., Smith, D. J., Menéndez, J., (2014). Competition of optical transitions between direct and indirect bandgaps
et al. (2003). Tunable band structure in diamond-cubic tin-germanium alloys in Ge1−xSnx. Appl. Phys. Lett. 105, 051104. doi:10.1063/1.4892302
grown on silicon substrates. Solid State Commun. 127, 355–359. doi:10.1016/ Dutt, B., Lin, H., Sukhdeo, D. S., Vulovic, B. M., Gupta, S., Nam, D., et al. (2013).
S0038-1098(03)00446-0 Theoretical analysis of GeSn alloys as a gain medium for a Si-compatible laser. IEEE J.
Beattie, A. R., and Landsberg, P. T. (1959). Auger effect in semiconductors. Proc. R. Sel. Top. Quantum Electron. 19, 1502706–1502706. doi:10.1109/JSTQE.2013.2241397
Soc. A Math. Phys. Eng. Sci. 249, 16–29. doi:10.1098/rspa.1959.0003 Dutt, B., Sukhdeo, D. S., Nam, D., Vulovic, B. M., Ze, Y., and Saraswat, K. C. (2012).
Bhargava, N., Coppinger, M., Gupta, J. P., Wielunski, L., and Kolodzey, J. (2013). Lattice Roadmap to an efficient germanium-on-silicon laser: strain vs. n-type doping.
constant and substitutional composition of GeSn alloys grown by molecular beam IEEE Photon. J. 4, 2002–2009. doi:10.1109/JPHOT.2012.2221692
epitaxy. Appl. Phys. Lett. 103, 041908. doi:10.1063/1.4816660 El Kurdi, M., Fishman, G., Sauvage, S., and Boucaud, P. (2010). Band structure and
Birner, S., Zibold, T., Andlauer, T., Kubis, T., Sabathil, M., Trellakis, A., et al. (2007). optical gain of tensile-strained germanium based on a 30 band k·p formalism.
Nextnano: general purpose 3-D simulations. IEEE Trans. Electron Devices 54, J. Appl. Phys. 107, 013710. doi:10.1063/1.3279307
2137–2142. doi:10.1109/TED.2007.902871 Fang, Z., Chen, Q. Y., and Zhao, C. Z. (2013). A review of recent progress in lasers
Boucaud, P., Kurdi El, M., Ghrib, A., Prost, M., de Kersauson, M., Sauvage, S., et al. on silicon. Opt. Laser Technol. 46, 103–110. doi:10.1088/0034-4885/76/3/034501
(2013). Recent advances in germanium emission. Photon. Res. 1, 102. doi:10.1364/ Fischetti, M. V., and Laux, S. E. (1996). Band structure, deformation potentials, and
PRJ.1.000102 carrier mobility in strained Si, Ge, and SiGe alloys. J. Appl. Phys. 80, 2234–2252.
Bratland, K., Foo, Y., Soares, J., Spila, T., Desjardins, P., and Greene, J. (2003). Mechanism doi:10.1063/1.363052
for epitaxial breakdown during low-temperature Ge(001) molecular beam epitaxy. Fitzgerald, E. A., Xie, Y. H., Green, M. L., Brasen, D., Kortan, A. R., Michel, J., et al.
Phys. Rev. B 67, 125322. doi:10.1103/PhysRevB.67.125322 (1991). Totally relaxed GexSi1−x layers with low threading dislocation densities
Camacho-Aguilera, R. E., Cai, Y., Patel, N., Bessette, J. T., Romagnoli, M., Kimerling, L. grown on Si substrates. Appl. Phys. Lett. 59, 811. doi:10.1063/1.105351
C., et al. (2012). An electrically pumped germanium laser. Opt. Express 20, 11316. Fournier-Lupien, J. H., Chagnon, D., Levesque, P., AlMutairi, A. A., Wirths, S., Pippel,
doi:10.1364/OE.20.011316 E., et al. (2014). In situ studies of germanium-tin and silicon-germanium-tin
Capellini, G., Kozlowski, G., Yamamoto, Y., Lisker, M., Wenger, C., Niu, G., et al. thermal stability. ECS Trans. 64, 903–911. doi:10.1149/06406.0903ecst
(2013). Strain analysis in SiN/Ge microstructures obtained via Si-complementary Geiger, R., Frigerio, J., Süess, M. J., Chrastina, D., Isella, G., Spolenak, R., et al. (2014a).
metal oxide semiconductor compatible approach. J. Appl. Phys. 113, 013513. Excess carrier lifetimes in Ge layers on Si. Appl. Phys. Lett. 104, 062106–062106.
doi:10.1063/1.4772781 doi:10.1063/1.4865237
Capellini, G., Reich, C., Guha, S., Yamamoto, Y., Lisker, M., Virgilio, M., et al. (2014). Geiger, R., Süess, M. J., Bonzon, C., Frigerio, J., Chrastina, D., Isella, G., et al. (2014b).
Tensile Ge microstructures for lasing fabricated by means of a silicon complemen- Carrier Lifetimes in Uniaxially Strained Ge Micro Bridges. IEEE, 227–228.
tary metal-oxide-semiconductor process. Opt. Express 22, 399–410. doi:10.1364/ Geiger, R., Süess, M. J., Bonzon, C., Spolenak, R., Faist, J., and Sigg, H. (2014c). Strained
OE.22.000399 Ge Microbridges to Obtain a Direct Bandgap Laser. IEEE, 7–8.
Carroll, L., Friedli, P., Lerch, P., Schneider, J., Treyer, D., Hunziker, S., et al. (2011). Ultra- Geiger, R., Frigerio, J., Süess, M. J., Minamisawa, R. A., Chrastina, D., Isella, G., et al.
broadband infrared pump-probe spectroscopy using synchrotron radiation and (2013). Excess Carrier Lifetimes in Ge Layers on Si. IEEE, 103–104.
a tuneable pump. Rev. Sci. Instrum. 82, 063101–063101–9. doi:10.1063/1.3592332 Ghrib, A., Kurdi El, M., de Kersauson, M., Prost, M., Sauvage, S., Checoury, X., et al.
Carroll, L., Friedli, P., Neuenschwander, S., Sigg, H., Cecchi, S., Isa, F., et al. (2012). (2013). Tensile-strained germanium microdisks. Appl. Phys. Lett. 102, 221112.
Direct-gap gain and optical absorption in germanium correlated to the density of doi:10.1364/OE.23.006722
photoexcited carriers, doping, and strain. Phys. Rev. Lett. 109, 057402. doi:10.1103/ Ghrib, A., Kurdi El, M., Prost, M., Sauvage, S., Checoury, X., Beaudoin, G., et al. (2015).
PhysRevLett.109.057402 All-around SiN stressor for high and homogeneous tensile strain in germanium
Chang, G.-E., and Cheng, H. H. (2013). Optical gain of germanium infrared microdisk cavities. Adv. Opt. Mater. 3, 353–358. doi:10.1002/adom.201400369
lasers on different crystal orientations. J. Phys. D Appl. Phys. 46, 065103. Golikova, O. A., Moizhes, B. Y., and Stilbans, L. S. (1962). Hole mobility of germa-
doi:10.1088/0022-3727/46/6/065103 nium as a function of concentration and temperature. Sov. Phys. Solid State 3,
Chen, R., Gupta, S., Huang, Y.-C., Huo, Y., Rudy, C. W., Sanchez, E., et al. (2013a). 2259–2265.
Demonstration of a Ge/GeSn/Ge quantum-well microdisk resonator on silicon: Gray, A. X., Papp, C., Ueda, S., Balke, B., Yamashita, Y., Plucinski, L., et al. (2011).
enabling high-quality Ge(Sn) materials for micro- and nanophotonics. Nano Lett. Probing bulk electronic structure with hard X-ray angle-resolved photoemission.
14, 37–43. doi:10.1021/nl402815v Nat. Mater. 10, 759–764. doi:10.1038/nmat3089
Chen, R., Huang, Y.-C., Gupta, S., Lin, A. C., Sanchez, E., Kim, Y., et al. (2013b). Greil, J., Lugstein, A., Zeiner, C., Strasser, G., and Bertagnolli, E. (2012). Tuning the
Material characterization of high Sn-content, compressively-strained GeSn epitaxial electro-optical properties of germanium nanowires by tensile strain. Nano Lett.
films after rapid thermal processing. J. Cryst. Growth 365, 29–34. doi:10.1016/j. 12, 6230–6234. doi:10.1021/nl303288g
jcrysgro.2012.12.014 Gunn, J. B. (1963). Microwave oscillations of current in III–V semiconductors. Solid
Chen, R., Lin, H., Huo, Y., Hitzman, C., Kamins, T. I., and Harris, J. S. (2011a). State Commun. 1, 88–91. doi:10.1016/0038-1098(63)90041-3
Increased photoluminescence of strain-reduced, high-Sn composition Gupta, S., Chen, R., Vincent, B., Lin, D., Magyari-Kope, B., Caymax, M., et al. (2013a).
Ge1−xSnx alloys grown by molecular beam epitaxy. Appl. Phys. Lett. 99, 181125. (Invited) GeSn channel n and p MOSFETs. ECS Trans. 50, 937–941. doi:10.114
doi:10.1063/1.3658632 9/05009.0937ecst

Gupta, S., Magyari-Köpe, B., Nishi, Y., and Saraswat, K. C. (2013b). Achieving direct Nam, D., Sukhdeo, D., Cheng, S.-L., Roy, A., Chih-Yao Huang, K., Brongersma, M.,
band gap in germanium through integration of Sn alloying and external strain. et al. (2012). Electroluminescence from strained germanium membranes and
J. Appl. Phys. 113, 073707. doi:10.1063/1.4792649 implications for an efficient Si-compatible laser. Appl. Phys. Lett. 100, 131112.
Hartmann, J. M., Sanchez, L., Van Den Daele, W., Abbadie, A., Baud, L., Truche, R., doi:10.1063/1.3699224
et al. (2010). Fabrication, structural and electrical properties of compressively Nam, D., Sukhdeo, D., Roy, A., Balram, K., Cheng, S.-L., Huang, K. C.-Y., et al. (2011).
strained Ge-on-insulator substrates. Semicond. Sci. Technol. 25, 075010. Strained germanium thin film membrane on silicon substrate for optoelectronics.
doi:10.1088/0268-1242/25/7/075010 Opt. Express 19, 25866–25872. doi:10.1364/OE.19.025866
Harwit, A., Pukite, P. R., Angilello, J., and Iyer, S. S. (1990). Properties of diamond Nam, D., Sukhdeo, D. S., Kang, J.-H., Petykiewicz, J., Lee, J. H., Jung, W. S., et al.
structure SnGe films grown by molecular beam epitaxy. Thin Solid Films 184, (2013). Strain-induced pseudoheterostructure nanowires confining carriers at
395–401. doi:10.1016/0040-6090(90)90437-I room temperature with nanoscale-tunable band profiles. Nano Lett. 13, 3118–3123.
He, G., and Atwater, H. A. (1997). Interband transitions in SnxGe1−xAlloys. Phys. doi:10.1021/nl401042n
Rev. Lett. 79, 1937–1940. doi:10.1103/PhysRevLett.79.1937 Nedeljkovic, M., Khokhar, A. Z., Hu, Y., Chen, X., Penades, J. S., Stankovic, S., et al.
Heck, M. J. R., and Bowers, J. E. (2014). Energy efficient and energy proportional optical (2013). Silicon photonic devices and platforms for the mid-infrared. Opt. Mater.
interconnects for multi-core processors: driving the need for on-chip sources. Express 3, 1205. doi:10.1364/OME.3.001205
IEEE J. Sel. Top. Quantum Electron. 20, 332–343. doi:10.1109/JSTQE.2013.2293271 Newman, R., and Tyler, W. (1957). Effect of impurities on free-hole infrared
Huldt, L. (1971). Band-to-band auger recombination in indirect gap semiconductors. absorption in p-type germanium. Phys. Rev. 105, 885–886. doi:10.1103/
Phys. Status Solidi A 8, 173–187. doi:10.1002/pssa.2210080118 PhysRev.105.885
Huldt, L. (1974). Auger recombination in germanium. Phys. Status Solidi A 24, 221–229. Oda, K., Okumura, T., Tani, K., Saito, S.-I., and Ido, T. (2013). Improvement of photo-
doi:10.1088/0957-4484/22/43/435401 luminescence from Ge layer with patterned Si3N4 stressors. Thin Solid Films 557,
Huo, Y., Lin, H., Chen, R., Makarova, M., Rong, Y., Li, M., et al. (2011). Strong enhance- 355–362. doi:10.1016/j.tsf.2013.08.117
ment of direct transition photoluminescence with highly tensile-strained Ge grown Oehme, M., Buca, D., Kostecki, K., Wirths, S., Holländer, B., Kasper, E., et al. (2013).
by molecular beam epitaxy. Appl. Phys. Lett. 98, 011111. doi:10.1063/1.3534785 Epitaxial growth of highly compressively strained GeSn alloys up to 12.5% Sn. J.
Jacoboni, C., Nava, F., Canali, C., and Ottaviani, G. (1981). Electron drift velocity and Cryst. Growth 384, 71–76. doi:10.1016/j.jcrysgro.2013.09.018
diffusivity in germanium. Phys. Rev. B 24, 1014–1026. doi:10.1103/PhysRevB.24.1014 Passaro, V., Tullio, C., Troia, B., Notte, M., Giannoccaro, G., and Leonardis, F. (2012).
Jain, J. R., Hryciw, A., Baer, T. M., Miller, D. A. B., Brongersma, M. L., and Howe, R. T. Recent advances in integrated photonic sensors. Sensors 12, 15558–15598.
(2012). A micromachining-based technology for enhancing germanium light emis- doi:10.3390/s121115558
sion via tensile strain. Nat. Photonics 6, 398–405. doi:10.1038/nphoton.2012.111 Peschka, D., Thomas, M., Glitzky, A., Nurnberg, R., Gartner, K., Virgilio, M., et al.
Kao, K.-H., Verhulst, A. S., Rooyackers, R., Douhard, B., Delmotte, J., Bender, H., et al. (2015). Modeling of edge-emitting lasers based on tensile strained germanium
(2014). Compressively strained SiGe band-to-band tunneling model calibration microstrips. IEEE Photon. J. 7, 1–15. doi:10.1109/JPHOT.2015.2427093
based on p-i-n diodes and prospect of strained SiGe tunneling field-effect tran- Pukite, P. R., Harwit, A., and Iyer, S. S. (1989). Molecular beam epitaxy of meta-
sistors. J. Appl. Phys. 116, 214506. doi:10.1063/1.4903288 stable, diamond structure SnxGe1-x alloys. Appl. Phys. Lett. 54, 2142–2144.
Kasper, E., Kittler, M., Oehme, M., and Arguirov, T. (2013). Germanium tin: silicon doi:10.1063/1.101152
photonics toward the mid-infrared [Invited]. Photon. Res. 1, 69. doi:10.1364/ Reboud, V., Widiez, J., Hartmann, J.-M., Dias, G. O., Fowler, D., Chelnokov, A., et al.
PRJ.1.000069 (2015). Structural and optical properties of 200 mm germanium-on- insulator
Koerner, R., Oehme, M., Gollhofer, M., Schmid, M., Kostecki, K., Bechler, S., et al. (GeOI) substrates for silicon photonics applications. Proc. SPIE 9367, Silicon
(2015). Electrically pumped lasing from Ge Fabry-Perot resonators on Si. Opt. Photonics X 2015. doi:10.1117/12.2079393
Express 23, 14815–14822. doi:10.1364/OE.23.014815 Saito, S., Gardes, F. Y., Al-Attili, A. Z., Tani, K., Oda, K., Suwa, Y., et al. (2014). Group
Liu, J., Sun, X., Camacho-Aguilera, R., Kimerling, L. C., and Michel, J. (2010). Ge-on-Si IV light sources to enable the convergence of photonics and electronics. Front.
laser operating at room temperature. Opt. Lett. 35, 679. doi:10.1364/OL.35.000679 Mater. 1:15. doi:10.3389/fmats.2014.00015
Liu, J., Sun, X., Pan, D., Wang, X., Kimerling, L. C., Koch, T. L., et al. (2007). Tensile- Sánchez-Pérez, J. R., Boztug, C., Chen, F., Sudradjat, F. F., Paskiewicz, D. M., Jacobson,
strained, n-type Ge as a gain medium for monolithic laser integration on Si. Opt. R. B., et al. (2011). Direct-bandgap light-emitting germanium in tensilely strained
Express 15, 11272. doi:10.1364/OE.15.011272 nanomembranes. Proc. Natl. Acad. Sci. U.S.A. 108, 18893–18898. doi:10.1073/
Lochmann, W. (1978). Phonon-assisted auger recombination in indirect gap semicon- pnas.1107968108
ductors. Phys. Status Solidi A 45, 423–432. doi:10.1002/pssa.2210450208 Satta, A., Nicholas, G., Simoen, E., Houssa, M., Dimoulas, A., De Jaeger, B., et al.
Low, K. L., Yang, Y., Han, G., Fan, W., and Yeo, Y.-C. (2012). Electronic band structure (2006). Impact of germanium surface passivation on the leakage current of shallow
and effective mass parameters of Ge1-xSnx alloys. J. Appl. Phys. 112, 103715. planar p-n junctions. Mater. Sci. Semicond. Process. 9, 716–720. doi:10.1016/j.
doi:10.1063/1.4767381 mssp.2006.08.037
Marin, E., Bonzon, C., Geiger, R., Zabel, T., Sigg, H., and Faist, J. (2015). Proceedings Sau, J., and Cohen, M. (2007). Possibility of increased mobility in Ge-Sn alloy system.
CLEO-Europe 2015. (in print) Phys. Rev. B 75, 045208. doi:10.1103/PhysRevB.75.045208
Metzger, W. K., Wanlass, M. W., Ellingson, R. J., Ahrenkiel, R. K., and Carapella, J. J. Schetzina, J. F., and McKelvey, J. P. (1969). Strain dependence of the minority
(2001). Auger recombination in low-band-gap n-type InGaAs. Appl. Phys. Lett. carrier mobility in p-type germanium. Phys. Rev. 181, 1191. doi:10.1103/
79, 3272–3274. doi:10.1063/1.1418032 PhysRev.181.1191
Michel, J., Liu, J., and Kimerling, L. C. (2010). High-performance Ge-on-Si photode- Schubert, E. F. (2006). Light-Emitting Diodes, 2nd Edn. Cambridge University
tectors. Nat. Photonics 4, 527–534. doi:10.1038/nphoton.2010.157 Press.
Miller, D. A. B. (2009). Device requirements for optical interconnects to silicon chips. Senaratne, C. L., Gallagher, J. D., Jiang, L., Aoki, T., Smith, D. J., Menendez, J., et al.
Proc. IEEE 97, 1166–1185. doi:10.1109/JPROC.2009.2014298 (2014). Ge1-ySny (y = 0.01-0.10) alloys on Ge-buffered Si: synthesis, microstructure,
Miller, D. A. B. (2010). Optical interconnects to electronic chips. Appl. Opt. 49, F59. and optical properties. J. Appl. Phys. 116, 133509. doi:10.1063/1.4896788
doi:10.1364/AO.49.000F59 Shockley, W., and Read, W. (1952). Statistics of the recombinations of holes and
Minamisawa, R. A., Süess, M. J., Spolenak, R., Faist, J., David, C., Gobrecht, J., et al. electrons. Phys. Rev. 87, 835–842. doi:10.1103/PhysRev.87.835
(2012). Top-down fabricated silicon nanowires under tensile elastic strain up to Soref, R. (2010). Mid-infrared photonics in silicon and germanium. Nat. Photonics 4,
4.5%. Nat Commun 3, 1096. doi:10.1038/ncomms2102 495–497. doi:10.1038/nphoton.2010.171
Nakatsuka, O., Tsutsui, N., Shimura, Y., Takeuchi, S., Sakai, A., and Zaima, S. (2010). Strocov, V. N., Wang, X., Shi, M., Kobayashi, M., Krempasky, J., Hess, C., et al. (2014).
Mobility behavior of Ge1-xSnx layers grown on silicon-on-insulator substrates. Soft-X-ray ARPES facility at the ADRESS beamline of the SLS: concepts, technical
Jpn. J. Appl. Phys. 49, 04DA10. doi:10.1143/JJAP.49.04DA10 realisation and scientific applications. J. Synchrotron. Radiat. 21, 32–44. doi:10.1107/
Nam, D., Kang, J.-H., Brongersma, M. L., and Saraswat, K. C. (2014). Observation S1600577513019085
of improved minority carrier lifetimes in high-quality Ge-on-insulator Süess, M. J., Geiger, R., Minamisawa, R. A., Schiefler, G. L., Frigerio, J., Chrastina, D.,
using time-resolved photoluminescence. Opt. Lett. 39, 6205. doi:10.1364/ et al. (2013). Analysis of enhanced light emission from highly strained germanium
OL.39.006205 microbridges. Nat. Photonics 7, 466–472. doi:10.1038/nphoton.2013.67

Sukhdeo, D. S., Nam, D., Kang, J.-H., Brongersma, M. L., and Saraswat, K. C. (2014). Wirths, S., Buca, D., Mussler, G., Tiedemann, A. T., Holländer, B., Bernardy, P., et al.
Direct bandgap germanium-on-silicon inferred from 5.7% 〈100〉 uniaxial tensile (2013a). Reduced pressure CVD growth of Ge and Ge1-xSnx alloys. ECS J. Solid
strain [Invited]. Photon. Res. 2, A8. doi:10.1364/PRJ.2.0000A8 State Sci. Technol. 2, N99–N102. doi:10.1149/2.006305jss
Sun, G., Soref, R. A., and Cheng, H. H. (2010). Design of a Si-based lattice-matched Wirths, S., Ikonić, Z., Tiedemann, A. T., Holländer, B., Stoica, T., Mussler, G., et al.
room-temperature GeSn/GeSiSn multi-quantum-well mid-infrared laser diode. (2013b). Tensely strained GeSn alloys as optical gain media. Appl. Phys. Lett. 103,
Opt. Express 18, 19957–19965. doi:10.1364/OE.18.019957 192110. doi:10.1063/1.4829360
Tahini, H., Chroneos, A., Grimes, R. W., Schwingenschlögl, U., and Dimoulas, A. Wirths, S., Geiger, R., den Driesch von, N., Mussler, G., Stoica, T., Mantl, S., et al.
(2012). Strain-induced changes to the electronic structure of germanium. J. Phys. (2015). Lasing in direct-bandgap GeSn alloy grown on Si. Nat. Photonics 9, 88–92.
Condens. Matter 24, 195802. doi:10.1088/0953-8984/24/19/195802 doi:10.1038/nphoton.2014.321
Takeuchi, S., Sakai, A., Yamamoto, K., Nakatsuka, O., Ogawa, M., and Zaima, S. Wirths, S., Stange, D., Pampillón, M.-A., Tiedemann, A. T., Mussler, G., Fox, A., et al.
(2006). Growth and structure evaluation of strain-relaxed Ge 1− xSn xbuffer (2014). High-k gate stacks on low bandgap tensile strained Ge and GeSn alloys for
layers grown on various types of substrates. Semicond. Sci. Technol. 22, S231–S235. field-effect transistors. ACS Appl. Mater. Interfaces 7, 62–67. doi:10.1021/am5075248
doi:10.1088/0268-1242/22/1/S54 Xiaochen, S., Jifeng, L., Kimerling, L. C., and Michel, J. (2010). Toward a germanium
Velha, P., Dumas, D. C., Gallacher, K., Millar, R., Myronov, M., Leadley, D. R., et al. laser for integrated silicon photonics. IEEE J. Sel. Top. Quantum Electron. 16,
(2013). Strained Germanium Nanostructures on Silicon Emitting at >2.2 µm 124–131. doi:10.1109/JSTQE.2009.2027445
Wavelength. IEEE, 142–143. Xu, C., Beeler, R. T., Jiang, L., Grzybowski, G., Chizmeshya, A. V. G., Menéndez, J., et al.
Vincent, B., Gencarelli, F., Bender, H., Merckling, C., Douhard, B., Petersen, D. H., (2013). New strategies for Ge-on-Si materials and devices using non-conventional
et al. (2011). Undoped and in-situ B doped GeSn epitaxial growth on Ge by hydride chemistries: the tetragermane case. Semicond. Sci. Technol. 28, 105001.
atmospheric pressure-chemical vapor deposition. Appl. Phys. Lett. 99, 152103. doi:10.1088/0268-1242/28/10/105001
doi:10.1063/1.3645620 Zhu, Y.-H., Xu, Q., Fan, W.-J., and Wang, J.-W. (2010). Theoretical gain of strained
Virgilio, M., Manganelli, C. L., Grosso, G., Schroeder, T., and Capellini, G. (2013). GeSn0.02/Ge1−x−y′SixSny′ quantum well laser. J. Appl. Phys. 107, 073108.
Photoluminescence, recombination rate, and gain spectra in optically excited n-type doi:10.1063/1.3329424
and tensile strained germanium layers. J. Appl. Phys. 114, 243102. doi:10.1063/1.4849855
Wang, L., Su, S., Wang, W., Gong, X., Yang, Y., Guo, P., et al. (2013). Strained germa- Conflict of Interest Statement: The authors declare that the research was conducted
nium-tin (GeSn) p-channel metal-oxide-semiconductor field-effect-transistors in the absence of any commercial or financial relationships that could be construed
(p-MOSFETs) with ammonium sulfide passivation. Solid State Electron. 83, 66–70. as a potential conflict of interest.
doi:10.1016/j.sse.2013.01.031
Wegscheider, W., Eberl, K., Menczigar, U., and Abstreiter, G. (1990). Single-crystal Sn/ Copyright © 2015 Geiger, Zabel and Sigg. This is an open-access article distributed under
Ge superlattices on Ge substrates: growth and structural properties. Appl. Phys. the terms of the Creative Commons Attribution License (CC BY). The use, distribution
Lett. 57, 875. doi:10.1063/1.104264 or reproduction in other forums is permitted, provided the original author(s) or licensor
Wen, H., and Bellotti, E. (2015). Rigorous theory of the radiative and gain characteristics are credited and that the original publication in this journal is cited, in accordance with
of silicon and germanium lasing media. Phys. Rev. B 91, 035307. doi:10.1103/ accepted academic practice. No use, distribution or reproduction is permitted which does
PhysRevB.91.035307 not comply with these terms.

MATERIALS
doi: 10.3389/fmats.2015.00030
Direct growth of Ge1−x Snx films on Si using a cold-wall

ultra-high vacuum chemical-vapor-deposition system
Aboozar Mosleh 1,2 *, Murtadha A. Alher 2,3 , Larry C. Cousar 1,4 , Wei Du 2 , Seyed Amir Ghetmiri 1,2 ,Thach Pham 2 ,
Joshua M. Grant 5 , Greg Sun 6 , Richard A. Soref 6 , Baohua Li 4 , Hameed A. Naseem 2 and Shui-Qing Yu 2
1
Microelectronics-Photonics Graduate Program (µEP), University of Arkansas, Fayetteville, AR, USA
2
Department of Electrical Engineering, University of Arkansas, Fayetteville, AR, USA
3
Mechanical Engineering Department, University of Karbala, Karbala, Iraq
4
Arktonics, LLC, Fayetteville, AR, USA
5
Engineering-Physics Department, Southern Arkansas University, Magnolia, AR, USA
6
Department of Engineering, University of Massachusetts Boston, Boston, MA, USA
Edited by: Germanium–tin alloys were grown directly on Si substrate at low temperatures using a cold-
wall ultra-high vacuum chemical-vapor-deposition system. Epitaxial growth was achieved
Reviewed by:
by adopting commercial gas precursors of germane and stannic chloride without any carrier
Fabio Iacona, National Research
Council, Italy gases. The X-ray diffraction analysis showed the incorporation of Sn and that the Ge1−x Snx
Christophe Labbé, Ecole Nationale films are fully epitaxial and strain relaxed. Tin incorporation in the Ge matrix was found to
Supérieure d’Ingénieurs de Caen, vary from 1 to 7%. The scanning electron microscopy images and energy-dispersive X-ray
France
spectra maps show uniform Sn incorporation and continuous film growth. Investigation
*Correspondence:
of deposition parameters shows that at high flow rates of stannic chloride the films were
Aboozar Mosleh, Engineering
Research Center (ENRC), 700 etched due to the production of HCl. The photoluminescence study shows the reduction
Research Center Boulevard, of band-gap from 0.8 to 0.55 eV as a result of Sn incorporation.
Fayetteville, AR 72701, USA
Keywords: chemical-vapor-deposition, Si photonics, Ge alloys, photoluminescence, Ge–Sn
e-mail: amosleh@gmail.com
INTRODUCTION both gas source and solid source MBE have been used by different
The discovery and development of Ge1−x Snx epitaxy technology groups to grow Ge1−x Snx films (Gurdal et al., 1998; Takeuchi et al.,
has enabled silicon photonics to be explored in a different scope of 2007; Chen et al., 2011; Werner et al., 2011; Stefanov et al., 2012;
a material platform. The ability of band-gap engineering by vary- Bhargava et al., 2013; Oehme et al., 2013; Wang et al., 2013).
ing Sn mole fraction, along with its compatibility to the comple- The other parallel approach of Ge1−x Snx growth is CVD. The
mentary metal–oxide–semiconductor (CMOS) process, has paved early results of CVD growth by Kouvetakis and Chizmeshya (2007)
the way for highly competitive Si-based near and mid-infrared at Arizona State University (ASU) showed the ability to grow
optoelectronic devices. Recent reports on the fabrication and Ge1−x Snx film directly on Si using a hot-wall ultra-high vacuum
characterization of high performance Ge1−x Snx devices such as CVD (UHV-CVD) system with deuterated Stannane (SnD4 ) as
modulators (Kouvetakis et al., 2005), photodetectors (Conley et al., the Sn precursor along with different chemistries of germanium.
2014a,b), and light emitting diodes (LEDs) (Du et al., 2014a) show Due to the high cost and instability of SnD4 , other precursors
great potential for Ge1−x Snx being adopted by industry in the such as tetramethyl tin [Sn(CH3 )4 ] and stannic chloride (SnCl4 )
near future. Cutting-edge reports on Ge1−x Snx , achieving a direct have been explored to grow Ge1−x Snx alloys. Vincent et al. (2011)
band-gap group IV alloy (Du et al., 2014b; Ghetmiri et al., 2014a; (from IMEC using atmospheric pressure CVD) and Kim et al.
Li et al., 2014; Wirths et al., 2014), is a turning point for the tech- (Chen et al., 2013) [from Applied Materials/Stanford University
nology to be pursued for the demonstration of an efficient group using reduced pressure-CVD (RP-CVD)] have reported successful
IV laser. In addition, due to the tunable lattice constant and forma- growth of Ge1−x Snx by using SnCl4 and a high cost Ge pre-
tion of Lomer dislocations, Ge1−x Snx has been shown to work as cursor digermane (Ge2 H6 ) and carrier gases on a Ge-buffered
a universal compliant buffer layer to grow high quality lattice mis- Si substrate. Using the same SnCl4 and Ge2 H6 precursors and
matched materials, like III–Vs, on Si (Beeler et al., 2011a; Mosleh carrier gases, Mantl et al. (Wirths et al., 2013) (from PGI9-IT)
et al., 2014). demonstrated direct growth of Ge1−x Snx on Si using showerhead
A variety of challenges exist for the growth of Ge1−x Snx alloys technology in an RP-CVD chamber. In the recent report, Tolle
on Si such as large lattice mismatch between Ge1−x Snx and Si et al. (Margetis et al., 2014; Mosleh et al., 2014a) (ASM company)
(more than 4.2%), low solid solubility of Sn in Ge (less than 0.5%), have achieved Ge1−x Snx growth using an industry prevail RP-
and instability of diamond lattice Sn (α-Sn) above 13°C. There- CVD reactor in collaboration with University of Arkansas (UA).
fore, growth can only possibly be done under non-equilibrium Low-cost Germane (GeH4 ) and SnCl4 with carrier gasses of N2 /H2
conditions. Different growth methods have been demonstrated were used to grow Ge1−x Snx . A Ge buffer was deposited between
for Ge1−x Snx growth in which molecular beam epitaxy (MBE) the Si substrate and the Ge1−x Snx layer in order to compensate
and chemical-vapor-deposition (CVD) have obtained device qual- the lattice mismatch between the layers. Table 1 lists the different
ity material and high Sn incorporation. For the MBE method, research groups that have grown Ge1−x Snx using CVD. Different

Mosleh et al. Direct Ge–Sn growth on Si using UHV-CVD
Table 1 | A summary of reports on Ge1−x Snx growth using CVD methods by different research groups.
Growth team Deposition system Deposition gas precursors Carrier gas Buffer layer
Ge Cost Sn Cost
ASU (Kouvetakis and Chizmeshya, 2007) UHV-CVD Different chemistries High SnD4 High Yes No
IMEC (Vincent et al., 2011) AP-CVD Ge2 H6 High SnCl4 Low Yes Ge
Applied materials (Chen et al., 2013) RP-CVD Ge2 H6 High SnCl4 Low Yes Ge
PGI9-IT (Wirths et al., 2013) RP-CVD Ge2 H6 High SnCl4 Low Yes No
ASM/UA (Margetis et al., 2014; Mosleh et al., 2014a) RP-CVD GeH4 Low SnCl4 Low Yes Ge
UA (this work) UHV-CVD GeH4 Low SnCl4 Low No No
growth methods and the cost effectiveness of the gas precursors ratio and deposition pressure, a growth rate of 20–3.3 nm/min
are compared. was achieved.
In this paper, we report direct growth of strain-relaxed
Ge1−x Snx films on Si substrates with Sn mole fractions up to 7% CHARACTERIZATION METHOD
using a cold-wall UHV-CVD system. Stannic chloride and ger- Analysis of Sn mole fraction, lattice constant, growth quality,
mane were chosen as the precursors which are low-cost and com- and strain in the Ge1−x Snx films were conducted using a high-
mercially available. The growth of Ge1−x Snx films was achieved resolution X-ray diffractometer. High-resolution TEM (TITAN)
without using any carrier gases and buffer layers. In order to with an accelerating voltage of 300 kV was used to investigate
investigate the material quality, the X-ray diffraction (XRD), crystal orientation and defects in the grown epi-layers as well as
high-resolution transmission electron microscopy (TEM), energy- determining the thicknesses of the samples. Surface morphology
dispersive X-ray spectroscopy (EDX), Raman spectroscopy, and of the samples was investigated by a scanning electron micro-
photoluminescence (PL) measurements have been conducted. scope equipped with EDX. Room temperature PL measurements
were carried out using a 690-nm excitation laser. The PL signal
EXPERIMENT was collected by a grating-based spectrometer equipped with a
GROWTH METHOD thermoelectric-cooled PbS detector (cut-off at 3 µm) for spectral
A cold-wall UHV-CVD system was adopted to grow Ge1−x Snx analysis.
films (see Figure 1 for machine schematic). The system com-
poses a load-lock chamber with a base pressure of 10−6 Pa and RESULTS AND DISCUSSION
a process chamber whose base pressure reaches 10−8 Pa using the MATERIAL CHARACTERIZATION
turbo-molecular and cryogenic pumps, respectively. Due to low- The 2θ-ω XRD scan was performed from the symmetric (004)
temperature growth of the films, removal of oxygen and water plane to obtain the out-of-plane lattice constant of the Ge1−x Snx
vapor is critical which was achieved by using a cryogenic pump. films. Figure 2A shows the peak at 69° corresponding to a satisfac-
The turbo-molecular pumps are backed by mechanical pumps. tion of the Bragg condition by Si (001) substrate, and the peaks at
The heating stage consisted of a pyrolytic graphite heater with a lower angles of 66–65° due to larger lattice size of the Ge1−x Snx lay-
thermocouple placed at the same distance away from the heater ers. The difference in the position of Ge1−x Snx peaks is due to the
as the wafer. The sample holder rotates up to 80 rpm for uniform difference in the Sn mole fractions of Ge1−x Snx layers. Different
film growth. The gas flow is through a side entry port, controlled compositions were achieved from 1 to 7% with desirable crystal
by mass flow controllers (MFCs). Stannic chloride is a volatile liq- quality. The Ge1−x Snx peaks are broadened for two reasons: (1)
uid with vapor pressure of 2.4 kPa at one atmospheric pressure. thin film thickness of the layers and (2) presence of mosaicity in the
Therefore, the evaporation could produce enough pressure to be Ge–Sn crystal and formation of defects as a result of strain relax-
passed through the MFC. ation. The full width at half maximum (FWHM) of the Ge1−x Snx
Germanium–tin films were grown on 400 (001) p-type Si sub- peaks are between 0.28 for 1% Sn film and 0.36 for 7% Sn film. The
strates with 5–10 Ω cm resistivity. Prior to loading the samples, change in FWHM depends on various factors such as film thick-
they were cleaned in a two-step process: (1) Piranha etch solution ness, relaxation, and quality and there is no trend showing that the
[H2 SO4 :H2 O2 (1:1)], (2) oxide strip HF dipping [H2 O:HF (10:1) FWHM of the peaks change as the Sn composition increases.
using 48% pure HF] followed by nitrogen blow drying. The final In order to calculate the total lattice constant and the strain
oxide strip step was not followed by a water rinse as it reduces the in the film, an asymmetric reciprocal space mapping (RSM)
life-time of hydrogen passivation and exposes the surface to ambi- from (−2, −2, 4) plane was performed. The RSM scans provide
ent oxygen (Mosleh et al., 2013, 2014b). The experiments were measurement of the in-plane (a k ) and out-of-plane (a ⊥ ) lattice
carried out at reduced pressures of 13, 40, 65, 95, 130, 200, and constant of Ge1−x Snx alloys. The total lattice constant a0GeSn was
260 Pa and at temperatures as low as 300°C. Germane (GeH4 ) and calculated by taking into account the elastic constants of Ge1−x Snx
stannic chloride (SnCl4 ) were used as the precursors for Ge1−x Snx (Beeler et al., 2011b). Knowing the total lattice constant, the Sn
growth. The gas flow ratio (GeH4 /SnCl4 ) was set to 5, 3.3, 2.5, mole fractions is calculated through Vegard’s law with the bow-
and 1.6. Depending on the growth parameters such as gas flow ing factor of b = 0.0166 Å (Moontragoon et al., 2012). Figure 2B
FIGURE 1 | Cold-wall UHV-CVD system with a substrate rotation. Samples are transferred through a load-lock chamber equipped with a turbo-molecular
pump. The growth chamber is equipped with a turbo-molecular pump and a cryogenic pump. Side entry of the gases is controlled by mass flow controllers.
shows the RSM of 6% Sn sample. The x-axis shows Q z in recip- at the Ge1−x Snx /Si interface. In addition, Figure 2B shows that
rocal lattice unit (rlu) which is related to the out-of-plane lattice strain relaxation occurred by formation of misfit dislocations at
constant (L) and the y-axis shows Q x which is related to the in- the interface. The TEM image shows that the grown film was fully
plane lattice constant (H or K). Direction of the spread in the epitaxial. Film thickness of the samples is listed in Table 2.
Ge0.94 Sn0.06 peak does not show a compositional gradient in the The SEM scan/EDX spectra of the samples show surface mor-
sample because it is related to the relaxation of the lattice on Si phology of the sample as well as Sn incorporation in the Ge matrix.
substrate. Large lattice mismatch between Sn and Ge is the main The EDX spectra in Figure 2D show the presence of Ge, Si, and Sn
reason for a large spread in the omega direction. The relaxation in the Ge0.94 Sn0.06 film. Due to the high count collection of sec-
line in Figure 2B shows that the films which are grown above are ondary electrons from the substrate, the ratio of Sn and Ge cannot
tensile strained and the films grown underneath are compressively exactly reveal the percentage of Sn in Ge. The presence of carbon
strained. The Ge0.94 Sn0.06 peak is observed to be on the relaxation and oxygen in the EDX spectra is mainly due to the contamina-
line and the relaxation is measured to be 97%. tion and oxidation of the film after exposure to ambient air. The
Calculation of total strain in other samples shows that all the EDX maps for Ge (Figure 2E) and Sn (Figure 2F) display uniform
films are more than 95% relaxed. Table 2 shows the lattice con- incorporation of Sn. The SEM image shows continuous growth
stants of the Ge1−x Snx alloys, their Sn mole fraction, and strain of Ge1−x Snx without observation of locally crystalline patches.
relaxation percentage. Ge1−x Snx films were almost fully relaxed No segregation and precipitation of Sn was observed on the films
mainly due to large lattice mismatch between Si (5.431 Å) and which indicates robust and stable growth of the films.
Ge1−x Snx (above 5.658 Å) and small critical thickness (Mosleh
et al., 2014a). The other reason for relaxation of Ge (and sim- GROWTH MECHANISM
ilarly Ge1−x Snx ) films on Si is the thermal mismatch between Growth of Ge1−x Snx on a Si substrate requires considering the
these two materials. High temperature growth (above 500°C) and reaction of byproducts and reduction of activation energy by
rapid cool down has been the main method for achieving tensile introducing carrier gases. Stannic chloride has a tendency to etch
strained Ge on Si (Conley et al., 2014a). The Ge1−x Snx samples Ge due to the presence of chlorine in the chemistry of the mole-
were grown at 300°C for 30 min and we have not achieved ten- cule. The byproduct of GeH4 + SnCl4 reaction is HCl which is an
sile strained films; however, the thermal mismatch between Si etchant gas for germanium and silicon (Bogumilowicz et al., 2005).
and Ge1−x Snx has helped relaxing the compressive strain. The Following reactions show different mechanisms of film deposition
strain has been mainly relieved through formation of misfit dis- as well as HCl production in the chamber:
locations including Lomer misfit dislocation. The cross-sectional
TEM image in Figure 2C shows formation of such dislocations GeH4 → Ge + 2H2 (1)

FIGURE 2 | (A) Symmetric (004) 2θ-ω scan of Ge1− x Snx films which are Ge0.94 Sn0.06 on the relaxation line shows that the film is strain relaxed.
grown on a Si substrate. The peak at 69° shows the Si substrate peak and (C) Transmission electron microscopy images of Ge0.94 Sn0.06 film shows
the peaks between 66° and 65° belong to Ge1− x Snx films. (B) Reciprocal epitaxial growth of Ge–Sn on a Si substrate. Arrows show misfit
space map from asymmetrical plane (−2, −2, 4) for Ge0.94 Sn0.06 grown on a dislocations formed at the Ge1− x Snx /Si interface. (D) The EDX spectrum of
Si substrate. The x -coordinate shows out-of-plane lattice constant and the Si/Ge0.94 Sn0.06 film shows the presence of Si (substrate), Ge and Sn (film),
y -coordinate shows in-plane lattice constant in units of reciprocal lattice O (native oxide), and C (carbon contamination from the ambient). (E) The
unit. The relaxation line shows that the films grown above are tensile EDX surface maps of Ge and (F) Sn taken from scanning electron
strained and below are compressively strained. Presence of the micrographs for Ge0.94 Sn0.06 film shows uniform growth of Ge1− x Snx alloy.
Table 2 | Tin mole fraction calculation, lattice constant, and relaxation so that growth is the dominant process in the chamber. The Ge/Sn
percentage of the grown samples. film will be etched by HCl through the following reactions:
Sample Sn (%) a kΠ (nm) a ⊥ (nm) a (nm) Relaxation Thickness

4HCl + Ge → GeH4 + 2Cl2 (4)
no. (%) (nm)
4HCl + Sn → SnCl4 +2H2 (5)
1 1.2 5.666 5.671 5.668 98 615
2 2.1 5.673 5.679 5.676 98 423 Domination of etching over growth is the main mechanism
3 2.9 5.678 5.687 5.682 97 295 that prevents direct growth of Ge1−x Snx on Si.
4 4.2 5.689 5.695 5.692 98 207 By controlling the flow through MFCs, we have grown
5 5.8 5.699 5.712 5.706 97 108 Ge1−x Snx films on Si at different pressures with a fixed flow ratio
6 7.0 5.715 5.719 5.717 99 532 of GeH4 /SnCl4 = 1.6. Growth was observed at 13 Pa of deposition
pressure and continued until the deposition pressure increased to
130 Pa. Figure 3A shows the thickness of Ge1−x Snx films versus
2H2 + SnCl4 → Sn + 4HCl (2) deposition pressure of the chamber as well as Sn incorporation
GeH4 + SnCl4 → Ge + Sn + 4HCl (3) percentage. Incorporation of Sn in the Ge lattice is increased by
raising the pressure due to the higher residence time of the precur-
Higher temperature of the substrate results in higher density sors in the chamber. The residence time of the gases has increased
of depositing ad-atoms (Ge and Sn); however, it will result in pro- from 2 s at 13 Pa to 19 s at 130 Pa. Meanwhile, HCl etched more
duction of HCl at a higher rate. In addition, higher flow rate of of the Ge1−x Snx films after deposition at higher pressures. This
SnCl4 increases the production rate of HCl as well. Controlling the trend has continued to 130 Pa and no growth has been observed
temperature and flow rate of the gases could control the process at 200 and 265 Pa. The increase in Sn composition from 1 to 6%
FIGURE 3 | (A) Variation of Sn incorporation percentage versus deposition as a result of increase in the deposition pressure. (B) Tin incorporation and
pressure. Films were etched away for deposition pressures higher than film thickness of the samples grown at 65 Pa growth pressure versus
130 Pa. The secondary axis on the right shows the reduction of film thickness GeH4 /SnCl4 flow ratios.
FIGURE 4 | (A) Raman spectra of the Ge1− x Snx film grown on a Si Sn mole fraction. The solid symbols are experimental data and the
substrate. The shift in the Ge–Ge peak is due to the incorporation of Sn in curves are theoretical predictions for relaxed films. The Ge–Ge peak is
Ge lattice. The shoulder on the left side of the Ge–Ge peak is due to the expected to shift 0.8310 cm−1 for every 1% Sn incorporation in relaxed
Ge–Sn peak at 285 cm−1 . The Ge–Sn peak is shown at lower films. The expected shift (0.8311 cm−1 ) for Ge–Sn peak is very close to
wavenumber of 250–260 cm−1 . (B) Ge–Ge and Ge–Sn peak shifts versus that of Ge–Ge.
has been accompanied with reduction in the thickness from 615 to energy of the growth (Wirths et al., 2013). Although reduction
108 nm. Films that were expected to have higher than 6% Sn con- of activation energy enables easier breakdown of the molecules
tent were totally etched off. Therefore, in order to grow higher Sn on the surface and enhances the growth quality and growth rate,
content films, growth mechanism under fixed pressure and chang- it would prepare the conditions for easier etch due to the pres-
ing the SnCl4 flow was studied. Higher film thickness and higher ence of an etchant agent. Therefore, the presence of carrier gases
Sn incorporation was achieved as a result of domination of growth pushes the competition between growth and etching toward etch-
over etching. Figure 3B shows Sn incorporation in Ge1−x Snx films ing, resulting in film etching at even lower flow rates of carrier
versus SnCl4 flow rate at 95 Pa deposition pressure. The secondary gases when the flow rate of SnCl4 is of the same order of GeH4 .
axis of Figure 3B shows film thicknesses of the samples. Due to the
dominance of etching for higher SnCl4 flow rate, the films were OPTICAL CHARACTERIZATION
mostly etched and the film thickness was less than 100 nm. Raman spectroscopy
Introduction of carrier gases has different effects on the growth The Ge1−x Snx films were further investigated by Raman spec-
of Ge1−x Snx films. Hydrogen changes the balance in the reaction to troscopy in order to analyze the crystal structure. Room tempera-
produce more HCl. Consequently, the GeH4 /SnCl4 ratio at which ture Raman spectra of the grown samples as well as a Ge reference
the Ge1−x Snx films were depositing will not result in growth when sample are plotted in Figure 4A. The Ge–Ge longitudinal optical
hydrogen is introduced in the chamber. In addition, introduction (LO) peak was observed at 300 cm−1 for the Ge reference sample
of nitrogen and argon as carrier gases will reduce the activation while the Ge–Ge peak in the Ge1−x Snx films was shifted to lower

FIGURE 5 | (A) Photoluminescence spectra of the Ge1− x Snx films bowed Vegard’s law interpolation for the direct (solid line) and
with 2, 4, 6, and 7% Sn mole fraction showing a red-shift in the indirect band-gap (dash line) of Ge1− x Snx alloy is plotted for different
band-gap of the films. Incorporation of Sn has shifted both direct Sn compositions and is overlaid with experimental data (solid
band-gap and indirect band-gap toward lower energies. (B) The symbols).
wavenumbers due to the change in bonding energy of Ge–Ge by Ghetmiri et al. (2014b). In Ge0.94 Sn0.06 and Ge0.93 Sn0.07 samples,
incorporation of Sn atoms. The intensity of the Ge–Ge LO peak the energies difference between direct and indirect transitions are
at 300 cm−1 is normalized for all the samples for comparison of very small, therefore the PL emissions from these indirect and
the peak positions. In addition to the main Ge–Ge peak, Raman direct transitions cannot be identified. A temperature-dependent
spectra of Ge1−x Snx films show other peaks that are induced as study is needed to differentiate the direct and indirect peak posi-
a result of Sn incorporation. The Ge–Sn LO peaks for different tions which will be reported in the future. The PL peaks from
Sn mole fractions were observed at 250–260 cm−1 in the films. A the samples with 2, 4, 6, and 7% Sn compositions are shown in
second peak of Ge–Sn is observed at 285 cm−1 , which can be seen Figure 5B as solid symbols. The solid and the dashed lines show
as a shoulder of Ge–Ge main peak. the direct and indirect band-gap energies based on bowed Veg-
The peak positions are obtained by Lorentzian fitting to find ard’s law for the relaxed Ge1−x Snx alloy (Ghetmiri et al., 2014b),
the exact position for further analysis. The shift in the Ge–Ge respectively. Since the Ge1−x Snx films are almost strain-free, as
LO peak depends on both strain and Sn composition of the confirmed by XRD measurements, the experimental results closely
films. Theoretical calculations for ∆ω are different for strain- follow the predicted values from Vegard’s law.
relaxed films and strained films for different Sn (x) content
[∆ωGe−Ge (x) = bx cm−1 ]. The Ge–Ge peak is expected to shift CONCLUSION
by a factor of b = −30.30 for a strained alloy while this factor Direct growth of Ge1−x Snx layers on Si substrates was achieved
varies to b = −83.10 for a strain-relaxed film (Cheng et al., 2013). using a cold-wall UHV-CVD system. The films were grown by
Figure 4B shows the experimental data obtained for Ge–Ge and employing low-cost commercial available GeH4 and SnCl4 precur-
Ge–Sn Raman shift from the sample compared with the theoret- sors without using any carrier gases and buffer layers. Characteri-
ical calculations. The peak shifts match well with the theoretical zations of the samples with XRD showed successful incorporation
calculations for strain-relaxed films. of Sn up to 7%. The TEM images show fully epitaxial growth of the
samples without any precipitation of Sn from the Ge lattice. The
Photoluminescence Raman results verified the Sn incorporation and PL measurements
Germanium has an indirect band-gap in the L valley with the showed reduction of the band-gap to 0.55 eV for 7% Sn sample.
energy of 0.644 eV and a direct band-gap at the γ point with 0.8 eV The low-cost and CMOS compatible growth method and the per-
energy at room temperature. Incorporation of Sn in Ge lattice low- formance of the samples indicate a promising future for Ge1−x Snx
ers the conduction band edge at the γ-point at a faster rate than applications in Si photonics. Moreover, the samples were grown
that at the L-point. PL measurements on Ge1−x Snx samples allow strain-relaxed enabling this material to be a universal compliant
determination of the band-gap edge for various Sn compositions. buffer layer which can be used in hybrid integration.
Figure 5 depicts room temperature PL intensity spectra for
as-grown Ge1−x Snx films with 2, 4, 6, and 7% Sn mole frac- ACKNOWLEDGMENTS
tions. As indicated in Figure 5A, increase of the Sn mole fraction The work at the UA was supported by NSF (EPS-1003970),
results in a band-gap reduction. Both direct and indirect PL peaks the Arkansas Bioscience Institute, the Arktonics, LLC (Air Force
exhibit red-shift with Sn compositions increase from 2 to 7%. A SBIR, FA9550-14-C-0044, Dr. Gernot Pomrenke, Program Man-
Gaussian fitting function was employed to extract the PL peak ager), and DARPA (W911NF-13-1-0196, Dr. Dev Palmer, Program
positions of both direct and indirect transitions as described in Manager). Drs. RS and GS acknowledge support from AFOSR
(FA9550-14-1-0196, Dr. Gernot Pomrenke, Program Manager). using a commercial CVD system. ECS Trans. 64, 1830–1830. doi:10.1149/06406.
JG acknowledges the support of NSF REU Program under Grant 0711ecst
Moontragoon, P., Soref, R., and Ikonic, Z. (2012). The direct and indirect bandgaps
number EEC-1359306.
of unstrained Si x Ge1-xy Sny and their photonic device applications. J. Appl. Phys.
112, 073106–073106–8. doi:10.1063/1.4757414
REFERENCES Mosleh, A., Benamara, M., Ghetmiri, S. A., Conley, B. R., Alher, M. A., Du, W., et al.
Beeler, R. T., Grzybowski, G. J., Roucka, R., Jiang, L., Mathews, J., Smith, D. (2014). Investigation on the formation and propagation of defects in GeSn thin
J., et al. (2011a). Synthesis and materials properties of Sn/P-doped Ge on Si films. ECS Trans. 64, 1845–1845. doi:10.1149/06406.0895ecst
(100): photoluminescence and prototype devices. Chem. Mater. 23, 4480–4486. Mosleh, A., Ghetmiri, S. A., Conley, B. R., Abu-Safe, H., Waqar, Z., Benamara,
doi:10.1021/cm201648x M., et al. (2013). “Nucleation-step study of silicon homoepitaxy for low-
Beeler, R., Roucka, R., Chizmeshya, A., Kouvetakis, J., and Menéndez, J. (2011b). temperature fabrication of Si solar cells,” in Photovoltaic Specialists Conference
Nonlinear structure-composition relationships in the Ge1-y Sny Si (100) (y< (PVSC), IEEE 39th . Tampa, FL.
0.15) system. Phys. Rev. B 84, 035204. doi:10.1103/PhysRevB.84.035204 Mosleh, A., Ghetmiri, S. A., Conley, B. R., Hawkridge, M., Benamara, M., Nazzal, A.,
Bhargava, N., Coppinger, M., Gupta, J. P., Wielunski, L., and Kolodzey, J. (2013). Lat- et al. (2014a). Material characterization of Ge1-x Snx alloys grown by a commer-
tice constant and substitutional composition of GeSn alloys grown by molecular cial CVD system for optoelectronic device applications. J. Electron. Mater. 43,
beam epitaxy. Appl. Phys. Lett. 103, 041908. doi:10.1063/1.4816660 938–946. doi:10.1007/s11664-014-3089-2
Bogumilowicz, Y., Hartmann, J. M., Truche, R., Campidelli, Y., Rolland, G., and Bil- Mosleh, A., Ghetmiri, S. A., Conley, B. R., Abu-Safe, H. H., Benamara, M., Waqar,
lon, T. (2005). Chemical vapour etching of Si, SiGe and Ge with HCl; applications Z., et al. (2014b). Investigation of growth mechanism and role of H2 in very low
to the formation of thin relaxed SiGe buffers and to the revelation of threading temperature Si epitaxy. ECS Trans. 64, 967–975. doi:10.1149/06406.0967ecst
dislocations. Semicond. Sci. Technol. 20, 127. doi:10.1088/0268-1242/20/2/004 Oehme, M., Buca, D., Kostecki, K., Wirths, S., Holländer, B., Kasper, E., et al. (2013).
Chen, R., Huang, Y., Gupta, S., Lin, A. C., Sanchez, E., Kim, Y., et al. (2013). Epitaxial growth of highly compressively strained GeSn alloys up to 12.5% Sn. J.
Material characterization of high Sn-content, compressively-strained GeSn Cryst. Growth 384, 71–76. doi:10.1016/j.jcrysgro.2013.09.018
epitaxial films after rapid thermal processing. J. Cryst. Growth 365, 29–34. Stefanov, S., Conde, J., Benedetti, A., Serra, C., Werner, J., Oehme, M., et al. (2012).
doi:10.1016/j.jcrysgro.2012.12.014 Laser synthesis of germanium tin alloys on virtual germanium. Appl. Phys. Lett.
Chen, R., Lin, H., Huo, Y., Hitzman, C., Kamins, T. I., and Harris, J. S. (2011). 100, 104101. doi:10.1063/1.3692175
Increased photoluminescence of strain-reduced, high-Sn composition Ge1-x Snx Takeuchi, S., Sakai, A., Yamamoto, K., Nakatsuka, O., Ogawa, M., and Zaima, S.
alloys grown by molecular beam epitaxy. Appl. Phys. Lett. 99, 181125. doi:10. (2007). Growth and structure evaluation of strain-relaxed Ge1-x Sn buffer lay-
1063/1.3658632 ers grown on various types of substrates. Semicond. Sci. Technol. 22, S231.
Cheng, R., Wang, W., Gong, X., Sun, L., Guo, P., Hu, H., et al. (2013). Relaxed and doi:10.1088/0268-1242/22/1/S54
strained patterned germanium-tin structures: a Raman scattering study. ECS J. Vincent, B., Gencarelli, F., Bender, H., Merckling, C., Douhard, B., Petersen, D.
Solid State Sci. Technol. 2, 138–145. doi:10.1149/2.013304jss H., et al. (2011). Undoped and in-situ B doped GeSn epitaxial growth on
Conley, B. R., Mosleh, A., Ghetmiri, S. A., Du, W., Soref, R. A., Sun, G., et al. (2014a). Ge by atmospheric pressure-chemical vapor deposition. Appl. Phys. Lett. 99,
Temperature dependent spectral response and detectivity of GeSn photoconduc- 152103–152103–3. doi:10.1063/1.3645620
tors on silicon for short wave infrared detection. Opt. Express 22, 15639–15652. Wang, L., Su, S., Wang, W., Gong, X., Yang, Y., Guo, P., et al. (2013). Strained
doi:10.1364/OE.22.015639 germanium–tin (GeSn) p-channel metal-oxide-semiconductor field-effect-
Conley, B. R., Margetis, J., Du, W., Tran, H., Mosleh, A., Ghetmiri, S. A., et al. (2014b). transistors (p-MOSFETs) with ammonium sulfide passivation. Solid State Elec-
Si based GeSn photoconductors with a 1.63 A/W peak responsivity and a 2.4 µm tron. 83, 66–70. doi:10.1016/j.sse.2013.01.031
long-wavelength cutoff. App. Phys. Lett. 105, 221117. doi:10.1063/1.4903540 Werner, J., Oehme, M., Schmid, M., Kaschel, M., Schirmer,A., Kasper, E., et al. (2011).
Du, W., Zhou, Y., Ghetmiri, S. A., Mosleh, A., Conley, B. R., Nazzal, A., et al. (2014a). Germanium-tin pin photodetectors integrated on silicon grown by molecular
Room-temperature electroluminescence from Ge/Ge1-x Snx /Ge diodes on Si sub- beam epitaxy. Appl. Phys. Lett. 98, 061108–061108–3. doi:10.1063/1.3555439
strates. Appl. Phys. Lett. 104, 241110. doi:10.1063/1.4884380 Wirths, S., Buca, D., Mussler, G., Tiedemann, A., Holländer, B., Bernardy, P., et al.
Du, W., Ghetmiri, S. A., Conley, B. R., Mosleh, A., Nazzal, A., Soref, R. A., (2013). Reduced pressure CVD growth of Ge and Ge1-x Snx alloys. ECS J. Solid
et al. (2014b). Competition of optical transitions between direct and indirect State Sci. Technol. 2, N99–N102. doi:10.1149/2.006305jss
bandgaps in Ge1-x Snx . Appl. Phys. Lett. 105, 051104. doi:10.1063/1.4892302 Wirths, S., Geiger, R., Scherrer, P., Ikonic, Z., Tiedemann, A. T., Mussler, G., et al.
Ghetmiri, S. A., Du, W., Margetis, J., Mosleh, A., Cousar, L., Conley, B. R., et al. (2014). “Epitaxy and photoluminescence studies of high quality GeSn het-
(2014a). Direct-bandgap GeSn grown on Silicon with 2230 nm photolumines- erostructures with Sn concentrations up to 13 at.%,” in 11th International Con-
cence. Appl. Phys. Lett. 105, 151109. doi:10.1063/1.4898597 ference on Group IV Photonics 27–29 August. Paris.
Ghetmiri, S. A., Du, W., Conley, B. R., Mosleh, A., Naseem, H. A., Yu, S., et al.
(2014b). Shortwave-infrared photoluminescence from Ge1-x Snx thin films on
silicon. J. Vac. Sci. Technol. B 32, 060601. doi:10.1116/1.4897917 Conflict of Interest Statement: The authors declare that the research was conducted
Gurdal, O., Desjardins, P., Carlsson, J., Taylor, N., Radamson, H., Sundgren, J., in the absence of any commercial or financial relationships that could be construed
et al. (1998). Low-temperature growth and critical epitaxial thicknesses of fully as a potential conflict of interest.
strained metastable Ge1-x Snx (x.0.26) alloys on Ge (001) 2 × 1. J. Appl. Phys.
83, 162–170. doi:10.1063/1.366690 Received: 28 January 2015; accepted: 23 March 2015; published online: 27 April 2015.
Kouvetakis, J., and Chizmeshya, A. V. G. (2007). New classes of Si-based photonic Citation: Mosleh A, Alher MA, Cousar LC, Du W, Ghetmiri SA, Pham T, Grant JM,
materials and device architectures via designer molecular routes. J. Mater. Chem. Sun G, Soref RA, Li B, Naseem HA and Yu S-Q (2015) Direct growth of Ge1−x Snx films
17, 1649–1655. doi:10.1039/b618416b on Si using a cold-wall ultra-high vacuum chemical-vapor-deposition system. Front.
Kouvetakis, J., Menendez, J., and Soref, R.A. Strain-Engineered Direct-Gap Mater. 2:30. doi: 10.3389/fmats.2015.00030
Ge/SnxGe1-x Heterodiode and Multi-Quantum-Well Photodetectors, Laser, Emit- This article was submitted to Optics and Photonics, a section of the journal Frontiers
ters and Modulators Grown on SnySizGe1-y-z Buffered Silicon. United States in Materials.
patent US 6897471 B1 (2005). Copyright © 2015 Mosleh, Alher, Cousar, Du, Ghetmiri, Pham, Grant , Sun, Soref,
Li, H., Brouillet, J., Salas, A., Chaffin, I., Wang, X., and Liu, J. (2014). Low Li, Naseem and Yu. This is an open-access article distributed under the terms of the
temperature geometrically confined growth of pseudo single crystalline GeSn Creative Commons Attribution License (CC BY). The use, distribution or reproduction
on amorphous layers for advanced optoelectronics. ECS Trans. 64, 819–827. in other forums is permitted, provided the original author(s) or licensor are credited
doi:10.1149/06406.0819ecst and that the original publication in this journal is cited, in accordance with accepted
Margetis, J., Ghetmiri, S. A., Du, W., Conley, B. R., Mosleh, A., Soref, R., et al. (2014). academic practice. No use, distribution or reproduction is permitted which does not
Growth and characterization of epitaxial Ge1-X Snx alloys and heterostructures comply with these terms.

MATERIALS
published: 23 February 2015
doi: 10.3389/fmats.2015.00008
Room-temperature near-infrared electroluminescence

from boron-diffused silicon pn-junction diodes
Si Li , Yuhan Gao, Ruixin Fan, Dongsheng Li and Deren Yang*
State Key Lab of Silicon Materials, Department of Materials Science and Engineering, Zhejiang University, Hangzhou, China
Edited by: Silicon pn-junction diodes with different doping concentrations were prepared by boron
Dan-Xia Xu, National Research
diffusion into Czochralski n-type silicon substrate. Their room-temperature near-infrared
Council Canada, Canada
electroluminescence (EL) was measured. In the EL spectra of the heavily boron doped
Reviewed by:
Jifeng Liu, Dartmouth College, USA diode, a luminescence peak at ~1.6 µm (0.78 eV) was observed besides the band-to-band
Tatiana S. Perova, The University of line (~1.1 eV) under the condition of high current injection, while in that of the lightly boron
Dublin, Ireland doped diode only the band-to-band line was observed. The intensity of peak at 0.78 eV
*Correspondence: increases exponentially with current injection with no observable saturation at room tem-
Deren Yang, State Key Lab of Silicon
perature. Furthermore, no dislocations were found in the cross-sectional transmission
Materials, Department of Materials
Science and Engineering, Zhejiang electron microscopy image, and no dislocation-related luminescence was observed in
University, Zheda Road 38, Hangzhou the low-temperature photoluminescence spectra. We deduce that the 0.78 eV emission
310027, China originates from the irradiative recombination in the strain region of diodes caused by the
e-mail: mseyang@zju.edu.cn
diffusion of large number of the boron atoms into a silicon crystal lattice.
Keywords: boron diffusion, silicon pn-junction diode, near-infrared electroluminescence
INTRODUCTION ~1.6 µm (0.78 eV) EL emission related to dislocation-related cen-

With the development of integrated circuits (ICs), the disad- ter D1 at room temperature from dislocation-rich silicon diodes.
vantage of traditional metal interconnection structure, such as Sun et al. (2004) reported two luminescence bands around 1.05
interlayer interference, energy dissipation, and signal delay, has and 0.95 eV related to doping spikes in boron-implanted silicon pn
become a bottleneck restricting the development of ultra-large- diodes. These emissions show great application potential in silicon
scale integration circuits (USLIs). Optical interconnection, which optoelectronics. But the mechanism is still in dispute.
uses photons to transform information, will be an ultimate solu- In this paper, we fabricated silicon pn-junction diodes with
tion for future progress in USLIs. Because silicon is an indirect- different boron doping concentrations. Their room-temperature
band-gap semiconductor and fundamentally unable to emit light EL was measured and their cross-sectional transmission electron
efficiently, achieving efficient silicon-based light sources compati- microscopy (TEM) images were studied. The result shows that
ble with current IC manufacturing technology has become the key the heavily boron doped silicon pn-junction diode without dis-
issue of silicon optoelectronics. Many routes to fabricate efficient location loops can emit strong 0.78 eV luminescence under the
silicon light emitters have been proposed: porous silicon (Can- condition of high current injection besides the band-to-band
ham, 1990; Qin et al., 1996; Bisi et al., 2000; Zhao et al., 2005a,b), emission. It is considered that the 0.78 eV emission originates
Si nanoprecipitates in SiO2 (Pavesi et al., 2000; Wang et al., 2007), from the irradiative recombination in the strain regions caused by
erbium-doped Si (Ennen et al., 1983; Zheng et al., 1994; Polman the diffusion of large number of boron atoms into silicon crystal
et al., 1995), Si/SiO2 superlattice structures (Lu et al., 1995), and lattice.
silicon pn-junction diodes (Sveinbjörnsson, 1996; Martin et al.,
2001; Ng et al., 2001; Sun et al., 2004; Lourenco et al., 2005). Among MATERIALS AND METHODS
these ways, silicon pn-junction diodes have attracted much atten- Two kinds of boron diffusion sources were prepared by dissolv-
tion. The most standout advantage of this kind of light-emitter ing B2 O3 into SiO2 latex with B3+ concentration of 0.203 mol/L
is that the fabrication process is totally compatible with USLI (marked as A) and 0.569 mol/L (marked as B), respectively. Boron
technology. Both ion-implantation (Sveinbjörnsson, 1996; Mar- sources were spin onto the surface of (100) oriented n-type
tin et al., 2001; Ng et al., 2001; Sun et al., 2004; Lourenco et al., Czochralski-grown Si substrates (2 ~ 10 Ω cm, 500 µm in thick-
2005; Sobolev, 2010) and thermal diffusion (Kveder et al., 2004; ness) after the substrate wafer was cut into 15 mm × 15 mm slices
Hoang et al., 2006, 2007) have been used to manufacture silicon and carefully cleaned by standard RCA process. Rapid thermal
pn diodes. The past few years has seen great advances in the devel- treating method was used to form shallow pn junction by boron
opment of silicon pn-junction diodes. Electroluminescence (EL) diffusion at 1100°C for 5 min in the flowing high-purity N2 atmos-
efficiency of 0.1–1% has been achieved (Martin et al., 2001; Ng phere. After a pn junction was formed, an indium tin oxide (ITO)
et al., 2001). In addition to the band-to-band emission around electrode with a thickness of 100 nm was deposited on the p-layer
1.1 µm, other near-infrared emissions have been found in boron- side by magnetron sputtering, and an Al electrode with a thickness
implanted and boron-diffused silicon pn diodes (Sveinbjörnsson, of 100 nm was evaporated on the n-layer side. Thus, a pn-junction
1996; Sun et al., 2004). Sveinbjörnsson (1996) reported strong diode was prepared.
Frontiers in Materials | Optics and Photonics February 2015 | Volume 2 | Article 8 | 106
Li et al. Electroluminescence from silicon pn junction
The carrier concentration and the depth of pn junctions were EL spectra of Sample A, only band-to-band emission is observed
studied by an SSM350 instrument of spreading resistance pro- as the current increases. In contrast, Sample B emits 0.78 eV EL
file (SRP). The microstructure of pn junctions was measured by a besides the band-to-band emission under the condition of high
transmission electron microscope (TEM, JEOL 2010). Photolumi- current injection (>705 mA) and its intensity increases greatly
nescence (PL) and EL signals were recorded using an Edinburgh with the current. The band-to-band emission of both pn diodes
FLS920P Spectrometer with a nitrogen-cooled near-infrared pho- demonstrates a small red shift with the increasing current; this is
tomultiplier tube. The low-temperature PL measurements were related to the device heating in response to the current injection.
performed over the range of 20 ~ 300 K by using a helium flow Sveinbjörnsson (1996) and Xiang et al. (2012a,b) have reported
cryostat. strong emission of 0.78 eV EL at room temperature from sil-
icon pn diodes containing dislocations. They have also found
RESULTS AND DISCUSSION
Figure 1 is the SRP results of pn junctions made from the two
boron sources (A and B). It is clear from the figure that shallow pn
junctions were formed. It can be seen that Sample B made from the
B boron source has the higher carrier concentration than Sample A
made from the A boron source. The surface carrier concentration
of Sample B reaches 4.6 × 1017 cm−3 , while that of Sample A is
5 × 1016 cm−3 . It is necessary to notify that the SPR measures the
activated dopant density only, which does not take into account
the possible dopant clustering at the surface, so there can be a large
amount of boron doping, which is inactive. The depth of Sample
B pn junction is about 250 nm, a little deeper than that of Sample
A, which is about 200 nm.
Figure 2 shows the I –V curves of the two pn diodes. As shown
in the figure, both the pn diodes perform good rectifying proper-
ties. The forward current increases quickly with the voltage while
the reverse current stays low. Under the same forward bias, the cur-
rent of Sample B is greater than that of Sample A. This is because
Sample B has the higher carrier concentration and less resistance.
FIGURE 2 | I –V curves of the two pn diodes at room temperature.
In addition, the turn-on voltage of Sample B is about 0.7 V, bigger
than that of Sample A, which is about 0.6 V. The reason is that the
pn-junction barrier of Sample B with higher doping concentra-
tion is larger, so that the forward voltage needed to overcome the
barrier is larger. A
The room-temperature EL spectra of Sample B and Sample
A are different, as shown in Figure 3. In the room-temperature
FIGURE 1 | Spreading resistance profile spectra of pn junctions made FIGURE 3 | Room-temperature EL spectra at different electrical
from Sample A and Sample B fabricated with two boron sources. currents of two pn diodes: (A) for sample B and (B) for sample A.
www.frontiersin.org February 2015 | Volume 2 | Article 8 | 107

FIGURE 4 | (A) Photoluminescence spectra registered at different temperatures under laser excitation of 808 nm (500 mW) and (B) cross-sectional TEM image
of sample B.
lattice distortion is caused by large number of boron atoms dif-

fusing into silicon lattice. Under the condition of high current
injection, these regions can trap carriers and form effective irra-
diative recombination centers, which are related to the 0.78 eV
luminescence.
Figure 5 shows the dependence of peak intensity of 0.78 eV
on the input power. When the input power is low, there is no
0.78 eV EL emission. When the input power reaches ~15 W, Sam-
ple B starts to emit 0.78 eV luminescence and its intensity increases
almost exponentially with the input power with no observable
saturation. This means that if the turn-on power of the o.78 eV
emission could be sufficiently decreased, a highly efficient light
source would be achieved.
CONCLUSION
In this paper, two silicon pn diodes with different boron dop-
ing concentrations were fabricated by boron diffusion. We studied
their room-temperature near-infrared EL. The results show that
FIGURE 5 | The dependence of peak intensity of EL band at 0.78 eV on in the EL spectra of the heavily boron doped diode, a lumi-
the input power of pn-junction diodes. nescence peak at ~1.6 µm (0.78 eV) was observed besides the
band-to-band line (~1.1 eV) under the condition of high current
injection, while in that of the lightly boron doped diode, only
the band-to-band line was observed. In addition, no dislocations
dislocation-related bands in the low-temperature PL spectra. were found in the cross-sectional TEM image and no dislocation-
They tended to regard the peak at 0.78 eV in PL spectrum at related luminescence was observed in the low-temperature PL
room temperature as a red-shifted luminescence band D1. How- spectra. The 0.78 eV emission is proved to have no direct connec-
ever, for Sample B, no other luminescence band is found in tion with dislocations or dislocation-related luminescence bands.
the low-temperature PL spectrum (the pump power intensity is In fact, lots of lattice damage regions can be seen near the sur-
0.22 W/cm2 ) except the band-to-band emission and no disloca- face of the highly doped diode in the TEM image. We deduce that
tions are observed in the cross-sectional TEM image as shown the 0.78 eV emission may originate from the irradiative recom-
in Figure 4. It can be seen from Figure 4B that Sample B is bination in these regions. What is more, the intensity of 0.78 eV
free from dislocations. Our work suggests that the 0.78 eV EL emission increases exponentially with the input power without
at room temperature has no direct connection with dislocations observable saturation, which may be used as an efficient light
or dislocation-related luminescence bands. In fact, lots of lattice source in future.
damage regions can be seen near the surface of Sample B in the
TEM image. As mentioned before, although the measured dopant ACKNOWLEDGMENTS
concentration by SPR is relatively low, a significant amount of This work is supported by the National Basic Research Program
inactive boron doping may exist, so we think that the observed of China (973 Program) (No. 2013CB632102).
Frontiers in Materials | Optics and Photonics February 2015 | Volume 2 | Article 8 | 108
REFERENCES boron-implanted silicon pn diodes. Phys. Rev. B 70, 1–11. doi:10.1103/PhysRevB.

Bisi, O., Ossicini, S., and Pavesi, L. (2000). Surf. Sci. Rep. 38, 1–126. doi:10.1016/ 70.155316
S0167-5729(99)00012-6 Sveinbjörnsson, E. Ö, and Weber, J. (1996). Room temperature electroluminescence
Canham, L. T. (1990). Silicon quantum wire array fabrication by electrochem- from dislocation-rich silicon. Appl. Phys. Lett. 69, 2686–2688. doi:10.1063/1.
ical and chemical dissolution of wafers. Appl. Phys. Lett. 57, 1046–1048. 117678
doi:10.1063/1.103561 Wang, M., Li, D., Yuan, Z., Yang, D., and Que, D. (2007). Photoluminescence of
Ennen, H., Schneider, J., Pomerenke, G., and Axmann, A. (1983). 1.54-µm Lumi- Si-rich silicon nitride: defect-related states and silicon nanoclusters. Appl. Phys.
nescence of erbium implanted III–V? semiconductors and silicon. Appl. Phys. Lett. 90, 1–3. doi:10.1063/1.2717014
Lett. 43, 943–945. doi:10.1063/1.94190 Xiang, L., Li, D., Jin, L., and Yang, D. (2012a). Dislocation-related electrolumines-
Hoang, T., Holleman, J., LeMinh, P., Schmitz, J., Mchedlidze, T., Arguirov, T., et al. cence of silicon after electron irradiation. Solid State Commun. 152, 1956–1959.
(2007). Influence of dislocation loops on the near-infrared light emission from doi:10.1016/j.ssc.2012.08.011
silicon diodes. IEEE Trans. Electron Devices 54, 1860–1866. doi:10.1109/TED. Xiang, L., Li, D., Jin, L., Pivac, B., and Yang, D. (2012b). The origin of 0.78eV
2007.901072 line of the dislocation related luminescence in silicon. J. Appl. Phys. 112, 1–4.
Hoang, T., LeMinh, P., Holleman, J., and Schmitz, J. (2006). Strong efficiency doi:10.1063/1.4754825
improvement of SOI-LEDs through carrier confinement. IEEE Electron Device Zhao, Y., Yang, D., Li, D., and Jiang, M. (2005a). Annealing and amorphous silicon
Lett. 27, 383–385. doi:10.1109/LED.2007.895415 passivation of porous silicon with the blue light emission. Appl. Surf. Sci. 252,
Kveder, V., Badylevich, M., Steinman, E., Lzotov, A., Seibt, M., and Schroter, W. 1065–1069. doi:10.1016/j.apsusc.2005.01.176
(2004). Room-temperature silicon light-emitting diodes based on dislocation Zhao, Y., Yang, D., Li, D., and Jiang, M. (2005b). Photoluminescence of oxidized
luminescence. Appl. Phys. Lett. 84, 2106–2108. doi:10.1063/1.1689402 porous silicon under UV-light illumination. Mater. Sci. Eng. B 116, 95–98.
Lourenco, M. A., Milosavljevic, M., Galata, S., Siddiqui, M. S. A., Shao, G., Gwilliam, doi:10.1016/j.mseb.2004.08.014
R. M., et al. (2005). Silicon-based light-emitting devices. Vacuum 78, 551–556. Zheng, B., Michel, J., Ren, F. G., Kimerling, L. C., Jacobson, D. C., and Poate, J.
doi:10.1016/j.vacuum.2005.01.085 M. (1994). Room-temperature sharp line electroluminescence at 1.54 µm from
Lu, Z. H., Lockwood, D. J., and Baribeau, J. M. (1995). Quantum confinement an erbium-doped silicon light-emitting diode. Appl. Phys. Lett. 64, 2842–2844.
and light emission in SiO2 /Si superlattices. Nature 378, 258–260. doi:10.1038/ doi:10.1063/1.111977
378258a0
Martin, A. G., Zhao, J., Wang, A., Peter, J. R., and Michal, G. (2001). Efficient silicon
light-emitting diodes. Nature 412, 805–808. doi:10.1038/35090539 Conflict of Interest Statement: The authors declare that the research was conducted
Ng, W. L., Lourenco, M. A., Gwilliam, R. M., Ledain, S., Shao, G., and Hoomwood, in the absence of any commercial or financial relationships that could be construed
K. P. (2001). An efficient room-temperature silicon-based light-emitting diode. as a potential conflict of interest.
Nature 410, 192–194. doi:10.1038/35069092
Pavesi, L., Dal, N. L., Mazzoleni, C., Franzo, G., and Priolo, F. (2000). Optical gain Received: 17 August 2014; accepted: 20 January 2015; published online: 23 February
in silicon nanocrystals. Nature 408, 440–444. doi:10.1038/35044012 2015.
Polman, A., Van, D., Hoven, G. N., Custer, J. S., Shin, J. H., and Sema, R. (1995). Citation: Li S, Gao Y, Fan R, Li D and Yang D (2015) Room-temperature near-infrared
Erbium in crystal silicon: optical activation, excitation, and concentration limits. electroluminescence from boron-diffused silicon pn-junction diodes. Front. Mater. 2:8.
J. Appl. Phys. 77, 1256–1262. doi:10.1063/1.358927 doi: 10.3389/fmats.2015.00008
Qin, G. G., Song, H. Z., Zhang, B. R., Lin, J., Duan, J. Q., and Yao, G. Q. (1996). This article was submitted to Optics and Photonics, a section of the journal Frontiers
Experimental evidence for luminescence from silicon oxide layers in oxidized in Materials.
porous silicon. Phys. Rev. B 54, 2548–2555. doi:10.1103/PhysRevB.54.2548 Copyright © 2015 Li, Gao, Fan, Li and Yang . This is an open-access article distributed
Sobolev, N. A. (2010). Defect engineering in implantation technology of silicon under the terms of the Creative Commons Attribution License (CC BY). The use, dis-
light-emitting structures with dislocation-related luminescence. Semiconductors tribution or reproduction in other forums is permitted, provided the original author(s)
44, 1–23. doi:10.1134/S106378261001001X or licensor are credited and that the original publication in this journal is cited, in
Sun, J. M., Dekorsy, T., Skorupa, W., Schmdt, B., Mucklich, A., and Helm, accordance with accepted academic practice. No use, distribution or reproduction is
M. (2004). Below-band-gap electroluminescence related to doping spikes in permitted which does not comply with these terms.
www.frontiersin.org February 2015 | Volume 2 | Article 8 | 109

ADVANTAGES OF PUBLISHING IN FRONTIERS
FAST PUBLICATION COLLABORATIVE RESEARCH NETWORK

Average 90 days PEER-REVIEW Our network
from submission Designed to be rigorous – increases readership
to publication yet also collaborative, fair and for your article
constructive
OPEN ACCESS TRANSPARENT GLOBAL SPREAD

Articles are free to read, Editors and reviewers Six million monthly
for greatest visibility acknowledged by name page views worldwide
on published articles
COPYRIGHT TO AUTHORS IMPACT METRICS SUPPORT

No limit to Advanced metrics By our Swiss-based
article distribution track your editorial team
and re-use article’s impact
EPFL Innovation Park · Building I · 1015 Lausanne · Switzerland

T +41 21 510 17 00 · info@frontiersin.org · frontiersin.org

Photonic

Uploaded by

Copyright:

Available Formats

Photonic

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Photonic

Uploaded by

Copyright:

Available Formats

What is the main topic of discussion in the document?

What is the main topic of discussion in the document?

What are some of the advantages of publishing in Frontiers according to the document?

What are some of the advantages of publishing in Frontiers according to the document?

PHOTONIC INTEGRATION AND

Silicon photonics technology, which has

Cover image: Large-scale photonic integration on silicon wafer.

05 Editorial: Photonic integration and photonics–electronics convergence

Editorial: Photonic integration and

National Institute of Advanced Industrial Science and Technology, Tsukuba, Japan

Frontiers in Materials | www.frontiersin.org 5 October 2015 | Volume 2 | Article 65

REFERENCES vacuum chemical-vapor-deposition system. Front. Mater. 2:30. doi:10.3389/

Frontiers in Materials | www.frontiersin.org 6 October 2015 | Volume 2 | Article 65

Silicon photonic integration in

Acacia Communications, Hazlet, NJ, USA

Silicon photonics is the guiding of light in a planar arrangement of silicon-based materials

Frontiers in Physics | www.frontiersin.org 7 August 2015 | Volume 3 | Article 37

Frontiers in Physics | www.frontiersin.org 8 August 2015 | Volume 3 | Article 37

eV (1.1 µm). The silicon is a pure crystal grown in a boule and

FIGURE 3 | Some of the possible PIC integration elements.

Frontiers in Physics | www.frontiersin.org 9 August 2015 | Volume 3 | Article 37

FIGURE 7 | Surface-emitting 1-D grating couplers in silicon (A) and

Frontiers in Physics | www.frontiersin.org 10 August 2015 | Volume 3 | Article 37

FIGURE 9 | Various polarization splitters.

Frontiers in Physics | www.frontiersin.org 11 August 2015 | Volume 3 | Article 37

Frontiers in Physics | www.frontiersin.org 12 August 2015 | Volume 3 | Article 37

FIGURE 12 | (A) configuration of an optical isolator that uses a tandem

Frontiers in Physics | www.frontiersin.org 13 August 2015 | Volume 3 | Article 37

FIGURE 14 | Silicon photonics 8-PSM transceiver. Courtesy of Luxtera.

Frontiers in Physics | www.frontiersin.org 14 August 2015 | Volume 3 | Article 37

Frontiers in Physics | www.frontiersin.org 15 August 2015 | Volume 3 | Article 37

FIGURE 21 | First reported vector modulator. It was in GaAs.

FIGURE 20 | PIC for coupling to the multimodes of a ring-core fiber by

Frontiers in Physics | www.frontiersin.org 16 August 2015 | Volume 3 | Article 37

FIGURE 22 | Early reported silicon photonic vector modulators.

FIGURE 23 | 2-bit optical DAC in silicon photonics.

Frontiers in Physics | www.frontiersin.org 17 August 2015 | Volume 3 | Article 37

FIGURE 25 | Single-chip silicon photonic coherent transceiver.

Frontiers in Physics | www.frontiersin.org 18 August 2015 | Volume 3 | Article 37

Frontiers in Physics | www.frontiersin.org 19 August 2015 | Volume 3 | Article 37

Figure 26D shown for comparison is the performance of discrete

References 6. Mekis A, Gloeckner S, Masini G, Narasimha A, Pinguet T. A grating-coupler-

Frontiers in Physics | www.frontiersin.org 20 August 2015 | Volume 3 | Article 37

Frontiers in Physics | www.frontiersin.org 21 August 2015 | Volume 3 | Article 37

Frontiers in Physics | www.frontiersin.org 22 August 2015 | Volume 3 | Article 37

Small sensitivity to temperature variations of Si-photonic

Edited by: We demonstrated a small sensitivity to temperature variations of delay-line Mach–Zehnder

INTRODUCTION materials, in this work, we minimized the temperature sensitivity

www.frontiersin.org March 2015 | Volume 2 | Article 26 | 23

FIGURE 3 | Microscope image of fabricated Si–SiN-waveguide DL MZI.

Frontiers in Materials | Optics and Photonics March 2015 | Volume 2 | Article 26 | 24

Table 1 | Measured FSRs and dλ/dT.

Sample Arm 1 Arm 2 FSR (GHz) dλ/dT (pm/K)

Si–SiN DL MZI Si SiN 40.5 −2.8

SiN waveguide, and 24 times smaller than that of the conven-

www.frontiersin.org March 2015 | Volume 2 | Article 26 | 25

variation of +4 K is -1.4 GHz, which could cause only about

Frontiers in Materials | Optics and Photonics March 2015 | Volume 2 | Article 26 | 26

www.frontiersin.org March 2015 | Volume 2 | Article 26 | 27