• 繁體中文
  • 简体中文
Home> News


InP Substrates for Solar Cell

Published on 16th January 2013
By Future Timeline.net

Multi-junction solar cell to break efficiency barrier?

U.S. Naval Research Laboratory scientists, in collaboration with Imperial College London and MicroLink Devices, have proposed a novel triple-junction solar cell with the potential to break the 50 percent conversion efficiency barrier, which is the current goal in multi-junction photovoltaic development.

"This research has produced a novel, realistically achievable, lattice-matched, multi-junction solar cell design with the potential to break the 50 percent power conversion efficiency mark under concentrated illumination," said Robert Walters, Ph.D., NRL research physicist. "At present, the world record triple-junction solar cell efficiency is 44 percent under concentration and it is generally accepted that a major technology breakthrough will be required for the efficiency of these cells to increase much further."

In multi-junction (MJ) solar cells, each junction is 'tuned' to different wavelength bands in the solar spectrum to increase efficiency. High bandgap semiconductor material is used to absorb the short wavelength radiation with longer wavelength parts transmitted to subsequent semiconductors. In theory, an infinite-junction cell could obtain a maximum power conversion percentage of nearly 87 percent. The challenge is to develop a semiconductor material system that can attain a wide range of bandgaps and be grown with high crystalline quality.

By exploring novel semiconductor materials and applying band structure engineering, via strain-balanced quantum wells, the NRL research team has produced a design for a MJ solar cell that can achieve direct band gaps from 0.7 to 1.8 electron volts (eV) with materials that are all lattice-matched to an indium phosphide (InP) substrate.

"Having all lattice-matched materials with this wide range of band gaps is the key to breaking the current world record" adds Walters. "It is well known that materials lattice-matched to InP can achieve band gaps of about 1.4 eV and below, but no ternary alloy semiconductors exist with a higher direct band-gap."

The primary innovation enabling this new path to high efficiency is the identification of InAlAsSb quaternary alloys as a high band gap material layer that can be grown lattice-matched to InP. Drawing from their experience with Sb-based compounds for detector and laser applications, NRL scientists modeled the band structure of InAlAsSb and showed that this material could potentially achieve a direct band-gap as high as 1.8eV. With this result, and using a model that includes both radiative and non-radiative recombination, the NRL scientists created a solar cell design that is a potential route to over 50 percent power conversion efficiency under concentrated solar illumination.

Recently awarded a U.S. Department of Energy (DoE), Advanced Research Projects Agency-Energy (ARPA-E) project, NRL scientists, working with MicroLink and Rochester Institute of Technology, N.Y., will execute a three year materials and device development program to realise this new solar cell technology.

Through a highly competitive, peer-reviewed proposal process, ARPA-E seeks out transformational, breakthrough technologies that show fundamental technical promise but are too early for private-sector investment. These projects have the potential to produce game-changing breakthroughs in energy technology, form the foundation for entirely new industries, and to have large commercial impacts.

InP, GaAs, Si, Sapphire, SiC, GaN Single Crystal Substrates Growth Rate vs. Price

By Northeast University, Japan

US DoD Launches Competition For Integrated Photonics Manufacturing Institute

Published on 3rd October 2014
By Compound Semiconductor

$100 million in federal investment matched by $100 million or more in private investment to the winning consortia

The US Department of Defense is launching a competition to award more than $100 million in federal investment matched by $100 million or more in private investment to the winning consortia to build a new Institute for Manufacturing Innovation (IMI) focused on Integrated Photonics, President Obama announced today.

This Institute will focus on developing an end-to-end photonics ‘ecosystem’ in the US, including domestic foundry access, integrated design tools, automated packaging, assembly and test, and workforce development.

Photonics, the use of light for applications as diverse as lasers and telecommunications, powers the Internet as we know it today. Integrated photonics manufacturing, the next generation of this of technology, involves integrating electronic and photonics technologies into single systems rather than the assortments of discrete units that exist now. Realising these opportunities requires development of high-volume mass-manufacturing, assembly, and packaging technologies and processes that are reliable and cost-effective.

Compound III-V semiconductors such as GaA, InP, and InGaAs are considered key technologies within photonics. They are used for making light emitting structures including solid state lasers and LEDs and components such as optical amplifiers

Beyond the Internet and telecommunications, integrated photonics is expected to be important in medicine – from the development of 'needleless' technologies for monitoring diabetics’ blood sugar levels to tiny cameras smaller than pills that can travel within arteries. It is also hoped that integrated photonic technology will make human genomes sequencing cheaper bringing it well below $1,000 as compared to $5,000 today.  In defence applications potential uses of integrated photonics range from improving battlefield imaging to advances in radar. 

The Integrated Photonics Manufacturing Institute  - with over $200 million in public and private resources - is expected to comprise the largest Federal investment to date, reflecting the complexity of this technology, its importance to national security, and its revolutionary potential. 

Record-Breaking DARPA InP Amplifier Runs At 1012GHz

Published on 4th November 2014
By Compound Semiconductor

Terahertz electronics program could pave way for new sub-millimeter wave applications

DARPA has made a series of strategic investments in terahertz electronics through its HiFIVE, SWIFT and TFAST programs. The objective of the Terahertz (THz) Electronics program is to develop device and integration technologies necessary to realize compact, high-performance electronic circuits that operate at center frequencies exceeding 1.0 THz.

"Terahertz circuits promise to open up new areas of research and unforeseen applications in the sub-millimeter-wave spectrum, in addition to bringing unprecedented performance to circuits operating at more conventional frequencies," said Dev Palmer, DARPA program manager.

"This breakthrough could lead to revolutionary technologies such as high-resolution security imaging systems, improved collision-avoidance radar, communications networks with many times the capacity of current systems and spectrometers that could detect potentially dangerous chemicals and explosives with much greater sensitivity."

Developed by Northrop Grumman Corporation, the InP Terahertz Monolithic Integrated Circuit (TMIC) exhibits a measured power gain of 9dB at 1.0 terahertz and 10dB at 1.03 terahertz. "Gains of six decibels or more start to move this research from the laboratory bench to practical applications-nine decibels of gain is unheard of at terahertz frequencies" said Palmer. "This opens up new possibilities for building terahertz radio circuits."

Current electronics using solid-state technologies have largely been unable to access the sub-millimeter band of the electromagnetic spectrum due to insufficient transistor performance. To address the 'terahertz gap'  engineers have traditionally used frequency conversion-converting alternating current at one frequency to alternating current at another frequency-to multiply circuit operating frequencies up from millimeter-wave frequencies. This approach, however, restricts the output power of electrical devices and adversely affects signal-to-noise ratio. Frequency conversion also increases device size, weight and power supply requirements.

DARPA has made a series of strategic investments in terahertz electronics through its HiFIVE, SWIFT and TFAST programs. The objective of the Terahertz (THz) Electronics program is to develop device and integration technologies necessary to realize compact, high-performance electronic circuits that operate at center frequencies exceeding 1.0 THz.

Intel goes to immersion 32nm with InP wafer

Published on 29 October 2008

by Electronics news desk

Intel will take the wraps off its new 32nm process technology for microprocessors at the International
Electron Device Meeting (IEDM) in December.

Intel claimed it is onto a second-generation high-k/metal gate technology and will, for the first time,
adopt 193nm immersion lithography to scale down the gate length. The company claimed its process
has the highest drive currents reported to date for a 32nm technology. The NMOS saturated drive current is 1.55mA/μm while the corresponding PMOS value is 1.21mA/μm.

Intel researchers used the process to build the largest fully functional SRAM array yet reported: a 291Mb SRAM array test chip with a cell size of 0.171μm2 and an array density of 4.2 Mb/mm2. The test chip operated at 3.8GHz at 1.1 V.

Also at the conference, researchers from HRL Laboratories will describe a possible shortcut to put the high speed of indium phosphide-based transistors into CMOS chips. They built entire wafers of high-performance 250nm double-heterostructure bipolar transistors able to switch at up to 300GHz on IBM’s 130nm RF-CMOS technology.

A partially fabricated wafer is bonded to a full-thickness, but smaller, InP epitaxial wafer. The InP wafer first is temporarily bonded to a handle wafer which allows the InP growth substrate and etch-stop layers to be removed. An aluminum heat-spreader layer is deposited as a blanket film, then the InP DHBT layers are permanently bonded to the IBM CMOS wafer’s top surface. In a first, the CMOS transistors showed no sign of degradation, while the InP transistors showed only minor performance impacts.

A further late paper is work from Tohoku University where magnetic tunnel junctions (MJTs) rather than SRAM cells are used to store data in a high-density 3D processor architecture. The researchers used the MTJs to construct a spin-transfer torque memory. Then, they used the spin-memories to drive reconfigurable 3D logic blocks fabricated with a standard 0.14μm CMOS process.


All comments

You need to be registered with the IET to leave a comment. Please log in or register as a new user.

IPRM InP & ISCS Compound Semiconductors Supply Chain For GaAs, GaP, GaN, 
InP, Sic, and Sapphire

Acme Corporation, AdTech, Agility, Alpha Crystal, Alphion, Ammono Sp, Anadigic, AOI, Avago, Avanex, AWSC, AXT, Bookham, Bridgestone, ComSeCore, Cotomat, Covega, Cree, Crysband, Crystal Applied Technology, Crystal On, Crystal Photonics, Crystal Q, Crystalwise, Cyoptics, Denselight, Denso, Dow Corning, Dowa, DTI/ Ding Ten, Elma Malachit JSC, Epistar, Epiworks, Eudyna, Exiton, Finisar, Freiberger, Fujitsu, Furukawa, GCS, GE, GigaCom, Goldeneye Inc., Hitachi(OpNext), Hitachi Cable, Hittite Microwave, HRL, II/VI, IIjin Display, Infinera, Inlustra Technologies LLC, InPACT, IntelliEPI, Intexys Photonics, IQE, JDSU, Juropol, Kopin, Kyma, Kyocera, Lumileds, LumiLOG, M/A-Com, Magnachip, Mimix Broadband, Mitsubishi Chemical, Mitsubishi Electric, MJ Corporation, Modulight, Monocrystal PLC, Multiplex,Namiki, NanoGaN Ltd., N-Crystals, NEC, Neomax, NeosemiTech, Nikko, Materials / Acrotec, Nippon Steel, Norstel, Northrop Grumman, NTT, OKI, OMMIC, Ostendo, Oxford Inst. / TDI, PAM Xiamen, Panasonic, Peregrine, Phostec, Picogiga, Renesas, RMFD, Rohm, Rubicon, Saint-Gobain Crystals, Samsung-Corning, Santur, Sapphicon, Sapphire Technology, SEH, Sensor Electronic Technology: SET, Shinkosha, Showa Denko, SiCrystal, Silian, Sino American SAS, Shyworks, Sony, Soraa, Inc., sumika, Sumitomo Electric SEI, Sumitomo Metal Mining, Svedice, TankeBlue, Tera Xtal Technology, Thales, TOPCO Scientific, TopGaN, TriQuint, UCSB, UMS, Unipress, Vitesse, VPEC, Wafer Tech. / IQE, Wafer Works, Win Semi, Xiamen Powerway, Xindium.

Swagelok develops new process to harden stainless steel

Apr 09, 2009

Steel users such as Swagelok haven't given up on value-added projects that reduce the cost and improve the quality of the metal products they buy and use in making valves and fluid and gas system components. Case in point: The self-developed Swagelok SAT12 process, a surface-hardening procedure that boosts basic stainless steel's surface hardness and wear and corrosion resistance to mimic the mechanical properties of titanium, Hastelloy and other specialty alloys.
Solon, Ohio-based Swagelok produces and sells more than $1.3 billion annually in high-quality products used in the instrumentation, pharmaceutical, oil and gas, power, petrochemical, alternative fuels and semiconductor industries.
Within Swagelok, there is a multi-functional Metallurgical Research and Development team which performs applied research for various product-manufacturing groups. In recent years, the team was responsible for the development of the SAT12 surface-treatment method that reduces throughput time and improves the mechanical properties of precision parts made into complex shapes and structures from austenitic stainless steel.
Previously, the company had bought expensive alloys to attain such key performance characteristics as wear resistance, surface hardness, fatigue resistance and corrosion resistance. Now, using the SAT12 process, also known as low-temperature colossal super-saturation (LTCSS) service, the company can use basic austenitic stainless steel as the substrate for eventual use as the source material for high-quality components. In the process, carbon is diffused interstitially into the lattice structure of the stainless steel. The result is stainless steel exhibiting tool steel-like hardness, nickel alloy-like corrosion resistance, high-levels of wear resistance, austenitic ductility, no distortion or dimensional change, no carbide formation and resistance to galling.
Most market applications use 300 Series grades of austenitic stainless steel but further field research has found that the SAT12 treatment of 316 Series of stainless steel can replace more expensive, exotic materials.
Senior research fellow Sunniva Collins has been charged with added research and evaluation that will lead to commercialization of the new technology. Swagelok has worked with Case Western Reserve University in Cleveland to help analyze the process, which also is a low-temperature case carburization process for tube fittings. This project is being co-funded with a $5.5 million grant from Ohio's Third Frontier Project, a state-funded program to turn high-tech research into home-market manufacturing.
Swagelok developed the process itself to heat-treat austenitic stainless steel rather than have it done by a stainless steel producer because surface science and metallurgy are core competencies of Swagelok. A corporate statement says that "the patented SAT12 treatment process still remains unique in how it achieves significantly increased tool steel-level surface hardness and retained corrosion resistance in the austenitic target materials." Additionally, the SAT12 process was developed to treat the back ferrule, a narrow circular ring of metal used for fastening, joining or reinforcement in the Swagelok tube fitting, a core product for company.
A corporate spokesman says that "the SAT12 process has its sweet spot where its application makes the greatest economic and technological sense." That's because complete processing can include cleaning and handling operations in addition to the low temperature carburizing treatment. So, end-product pricing is dependent on things like the condition of parts received, batch size, part configuration and mass. "Customers have found it cost effective because upon treatment, some corrosion-resistance properties of relatively low priced stainless steels have demonstrated the performance of more expensive, exotic alloys," he explains.
The majority of applications desire improved surface wear or scratch resistance while maintaining or improving corrosion resistance. So, currently, the SAT12 process is being used on Swagelok products used by the military and commercial marine, medical device, biopharmaceutical production and food production equipment.

Sources from: http://www.purchasing.com/article/CA6649917.html?industryid=48389

InP Related R&D in Japan, Taiwan and Asia

InP Related R&D in Japan, Taiwan and Asia


Copyright © 1995-2009 by the Asian Technology Information Program (ATIP).

This material may not be published, modified, or otherwise redistributed in whole or in part, in any form,

without prior approval by ATIP, which reserves all rights.

ABSTRACT: The 16th International Conference on InP and Related Materials (IPRM04), held in Kagoshima (May 31 to June 4, 2004), provided an opportunity to assess recent developments in III-V compound semiconductor materials and devices in Japan and other Pacific rim countries.

This report is a review of the status of InP related R&D in Asia based on presentations given during the IPRM04, with particular emphasis on trends in the development of quantum dot lasers, large area InP substrates, and new materials research on thallium compound semiconductors for infrared devices.

KEYWORDS: Advanced Materials, Conferences, Nanotechnology, Physics, Regional Overview

of Science and Technology, Photonics/Optoelectronics, Semiconductors,





DATE: June 30, 2004







3. IPRM04


4.1 Korea

4.2 Singapore

4.3 Taiwan

ATIP Document ID: 040630AR ATIP04.028: InP Related R&D in Japan and Pacific Rim


4.4 China

4.5 Japan





Over the last thirty five years, Japan has made significant contributions to the development of III-V (read as “three-five”) compound semiconductor materials and devices. Examples of major contributions by Japanese engineers and scientists include: 1) the development of GaAs semiconductor lasers by Dr. Izuo Hayashi, 2) the development of optical fiber communications, in particular long wavelength semiconductor lasers and photonic integrated circuits by the Tokyo Institute of Technology group, 3) the invention and commercialization of HEMT (high electron mobility transistor) by Fujitsu in the early 1980s, and most recently, 4) the fabrication of GaN blue LEDs by Nakamura and Akasaki. Although eclipsed by the R&D on gallium nitride of late, Japan’s indium phosphide community is still active and waiting for resumption of investment in InP based optoelectronics, following the excesses the IT fever and subsequent downturn in 2000. (For a companion report on InN research, see ATIP04.026.)


The InP and Related Materials conference was established sixteen years ago, following an

initiative led by Japanese researchers. This event is held annually, moving between the US, Japan, and Europe, and is the most important meeting for researchers in the field. The Kagoshima conference was sponsored by the Japan Society of Applied Physics, IEEE Lasers and Electro-Optics Society, IEEE Electron Devices Society, and several industrial sponsors.


This year’s conference was chaired by Dr. Yuichi Matsushima of the National Institute of Information and Communications Technology (NICT) in Tokyo. The event gathered 350 participants from 16 countries and included 3 plenary talks, 25 technical sessions, and 199 presentations. The proceedings covered the six areas of characterization and bulk materials, epitaxy, processing and materials integration, electron devices, optoelectronics and nanostructures, and novel materials.


The present report reviews the presentations and technical discussions of IPRM04 and assesses trends such as outsourcing, the role of Asian academia and industry, and new devices in the next decade.


ATIP offers a full range of information services, including reports, assessments, briefings, visits,

sample procurements, workshops, cultural/business sensitivity training, and liaison activities, all

performed by our on-the-ground multilingual experts.

Email: info@atip.org Website: http://www.atip.org


Japan Office:


MBE 221, Akasaka Twin Tower

2-17-22 Akasaka, Minato-ku,

Tokyo 107-0052 JAPAN

Tel: + 81 3 5411-6670


U.S. Office (HQ):


PO Box 4510

Albuquerque, NM



Tel: +1 (505) 842-9020

Fax: +1 (505) 766-5166


China Office:


QingYun Modern Plaza, #2029

No. 43, W. Northern 3rd Ring


Haidian District

Beijing 100086 China

Tel: +86 (10) 6213-6752

Fax: +86 (10) 6213-6732


Complete ATIP reports on Asian Science and Technology are provided to subscribers and

collaborating organizations by direct distribution, or via electronic access. These contain text and often, charts, graphs, and pictures. Reports for unrestricted distribution often contain summarized or abstracted information. Sponsors can also obtain specific follow up information, including copies of proceedings, selected papers, exhibition particulars, updates, translations, query searches, etc.


[The remaining sections of this report are available to ATIP subscribers]

ATIP04.028: InP Related R&D in Japan, Taiwan and Asia

Copyright © 1995-2009 by the Asian Technology Information Program (ATIP).

This material may not be published, modified, or otherwise redistributed in whole or in part, in any form,

without prior approval by ATIP, which reserves all rights.


l            InP and related III-V compound semiconductors are used for fabricating HEMT, HBT (heterostructure bipolar transistor), as well as heterostructure lasers and other devices used in communications (fiber optics, wireless, direct broadcast satellite); computers (information, DVD, speed), and other consumer products (data storage and health care).

l           In spite of the severe downturn in the InP industry in 2000, with many bankruptcies and takeovers, according to a 2002 report published by Strategies Unlimited, the world market for InP was predicted to exceed $500 million by 2007.

l           The InP and Related Materials Conference (IPRM) was established sixteen years ago and is held annually; it is the most important meeting for engineers working in the field. A significant increase in participants from Pacific Rim countries was noted at IPRM04, which was held in Kagoshima, Japan, reflecting notable technological developments in Asia over the last five years: growth (MBE, MOCVD, LEC), fabrication (EBL), and packaging of state of the art optoelectronic devices.

l           Engineers in Japan and neighboring countries have succeeded in fabricating operational quantum dot lasers manufacture of large area InP substrates. Commercial lasers and systems can be expected within the next 3 to 5 years from the Asian industry.

l           The launch of Eudyna Devices Inc. (April 2004;capitalized at 19.5 billion yen, approx US$190 million), a 50-50 joint venture between Fujitsu Quantum Devices Limited and Sumitomo Electric, emphasizes the growing competition in the III-V and InP semiconductor market as well as the expectation of an upturn in the next 3 to 5 years.



InP-based products and devices are used by industries involved in fiber optics, direct broadcast satellite, DVD, and data storage. The exponential increase in demand for realtime global networking will necessitate the procurement of optoelectronic components manufactured, using InP-based semiconductors. The present report offers valuable insights into newcomers in Asia who will be part of the network of manufacturers competing with their Western counterparts for a share of the next generation of the global telecommunications market.



The following table provides ATIP’s evaluation and assessment of current and evolving country-level capabilities in a variety of technologies related to InP. It is based on the conference presentations as well as interactions with people directly working in the noted fields. Capabilities are rated from zero to four “+” with added comments as appropriate.





Topic [InP substrate manufacturing]

                    2004             2007 (estimate)

China                   +           +++

     (strategically important for the development of China's optoelectronics industry)

Japan                   +++          ++++

     (solid technological base; just waiting for the markets to improve)

Taiwan                zero            +

      (mainly due to outsourcing from Japan)

Singapore              zero           zero

Korea                  zero           +

     (Korean conglomerates want this technology; it is only a matter of time)


Topic [Quantum dot laser products]

                    2004           2007 (estimate)

China                   zero         zero


Japan                   zero          ++

      (domestic development)

Taiwan                  zero          ++

      (mainly due to outsourcing from Japan)

Singapore               zero         zero

Korea                   zero         +

     (as a result of intense national efforts)



Topic [mass production of HEMT, HBT devices]

                      2004           2007 (estimate)

China                zero             +

     (due to new venture businesses)

Japan                +++        ++++

     (strong base; waiting for market upturn)

Taiwan                 +            ++

      (mainly due to outsourcing from Japan)

Singapore            zero            +

       (but only with government support)

Korea                  +            ++

     (Samsung et al; challenge Japanese dominance)


Topic [mass production of lasers and PICs]

                    2004            2007 (estimate)

China                  zero           +

     (very large telecommunication market will necessitate domestic industry )

Japan                  +++          ++++

     (strong base; waiting for market upturn; Chinese telecom. market will be triggered)

Taiwan                 +             ++

      (mainly due to out sourcing from Japan)

Singapore              zero           +

       (but only with government support)

Korea                  ++            +++

      (again Samsung et al; challenge Japanese dominance)


ATIP comments on table:

Outsourcing from Japan to Taiwan: ATIP expects companies from other countries (mainly Japan) to outsource their work to Taiwan once the opto-electronics market improves again and the supply cannot meet demand.


Readers may be surprised that China is rated above Korea and Taiwan in one specific area, the manufacture of InP substrates, both today (2004) and in 2007. This was based on the papers presented at the conference (no papers from Korean or Taiwanese engineers about InP substrate manufacture but ~3 from Chinese engineers) as well as conversations with other experts at the conference.


It is relatively easy to start the manufacture of HEMT/HBT and laser structures: buy the substrates, buy the deposition technology such as MOCVD or MBE, read papers about the basic device structures, and off you go. The manufacture of III-V substrates, in particular InP, takes a great deal of time and effort with much more in-house, craftsman-type know how in order to produce a useful product. It takes time to develop InP substrate manufacturing technology, and there are no simple 'plug-in and use' solutions.


The Chinese have a tradition of working with crystal pulling techniques such as LEC for the fabrication of III-V substrates and as the conference papers show, there have been dramatic improvements in their ability to manufacture InP substrates over the last 5 years.


On the other hand, there were no papers from Korean or Taiwanese institutes/industry on

this topic and ATIP has not heard either of any concerted efforts from these countries to

develop such technology.


ATIP believes China will build on its recent success on InP substrates. ATIP also expects this country to have a noticeable industry within the next 3-5 years for its domestic use, hence the +++ for China in 2007.




InP is extensively used for 1300–1550nm wavelength optical communication devices with 75% of the market being InP based lasers, 20% optical receivers, and the remainder electronic devices such as HEMT and HBT. Expanded use of optical fiber networks for WDM systems has led to greater demand for long wavelength InP laser diodes and photodiodes. In addition, compound semiconductor devices are increasingly used for cell phones, DVD systems, and visible light LEDs. The InP market is greatly affected by the investment in optical fiber network equipment by local authorities and governments. In the medium term, the optical communications market is expected to expand in China and Asia. According to a 2002 report published by Strategies Unlimited, the world market for InP was predicted to exceed $500 million by 2007.


Driven by demands for high performance automobile radar and high speed optical and

microwave/millimeter wave communication systems, there has been a growing interest  for ultra high frequency devices and peripheral technology operating at 100Gbit/sec. The competing technologies are InP HBT, InP HEMT, GaAs p-HEMT, and SiGe BiCMOS. The successful technology will be the one offering the highest bit rate density at the lowest cost, that is, the highest integration with the lowest power consumption. Existing WDM (wavelength division multiplexing) systems use 10Gbit/s devices and will be replaced by 40Gbit/s networks as Internet traffic continuous to increase. Silicon has shown tremendous resilience and Si integrated circuits capable of 40 Gbit/s have been produced; however, their shortcoming is that they consume more power than their 10Gbit/s predecessors and do not represent a solution to the problem of increasing speed and reducing power consumption.


InP based compound semiconductors and ICs have been studied for more than 30 years and many experts agree that they will be the key technology for the niche areas in optoelectronics where silicon cannot be used. The characteristic features of InP based devices offer the following advantages over silicon and GaAs technology: 1) better thermal characteristics, 2) higher breakdown voltage, 3) higher frequency response, and 4) lower threshold voltages.


The bursting of the IT bubble led to a severe down turn in the InP industry with many bankruptcies and takeovers. In Japan, many university projects shifted their emphasis to other areas such as nitrides and industry has reduced its efforts on speculative projects. Basic research has become a luxury that even the major Japanese companies are not funding unless it is part of national projects led by JST, NEDO, etc., where the government funds most of the equipment and human resources.


The April 1, 2004 announcement of the launch of Eudyna Devices Inc, a 50-50 joint venture (capitalized at 19.5 billion yen; approx US$190 million) between Fujitsu Quantum Devices Limited and the electronic devices division of Sumitomo Electric, reflects not only the growing competition in the III-V and in particular InP semiconductor market, but also the expectation that there will be an upturn in the near future.


The last few years of reduced activity in the InP sector has enabled researchers in Korea,

Taiwan, and Singapore to develop technology to compete with industries in the US, EU and Japan. One of the results of this activity has been that engineers in Korea and Singapore are now able to fabricate state of the art InP based HEMT and HBT devices. Further, China has made tremendous advances in the growth of large area InP substrates (Hebei Semiconductor Research Institute) that will affect market trends given the advantages in manpower costs in China compared with Japan and US/EU.


The InP industry could be seen as having matured and experienced its first cycle of growth followed by contraction that is usually associated with the silicon industry. The future will require greater emphasis on outsourcing and globalization.



3. IPRM04

IPRM04 was chaired by Dr. Yuichi Matsushima (NICT) and held at the Kagoshima Bunka Hall. Kagoshima city is called the Napoli of Japan because it is located on the coast facing

Sakurajima, a 1200m active volcano.


The proceedings covered the six areas of characterization and bulk materials (17), epitaxy (27), processing and materials integration (18), electron devices (47), optoelectronics (51), as well as nanostructures and novel materials (39). The numerals in brackets indicate the number of papers in the session.

Excluding on-site registration, there were 18 participants from Korea, 11 from Taiwan, 7 from Singapore, and 5 from China. This increase in the number of participants from Asia reflects the major advances in the field in these countries over the last three years.




4.1 Korea

The majority of reports from Korea originated from institutes funded as part of the 21 Century Frontier Programs. The main funding agencies are the Ministry of Science and Technology as well as the Ministry of Information and Telecommunications.


InP HEMT devices have tremendous potential for use in space and military systems due to

their extremely high cut off frequencies and transcondunctances.


The Seo group from Seoul National University (as part of the National program for Tera hertz level integrated circuits funded by the Ministry of Science and Technology) described the fabrication and electrical characteristics of a W-band MMICs (mixer, oscillator, distributed amplifier) using 60nm gate length InAlAs/InGaAs HEMT grown by solid source MBE, which showed Vth=-0.65V, VB=-4.1V, extrinsic transconductance = 1.15 S/mm, ft=250 GHz and fmax= 263 GHz. A 1.5 x0.7mm2 ultra broad band distributed amplifier showed a small signal gain of 6.6dB over 0.4~110 GHz. This group has developed electron beam lithography and MBE growth technology for producing these state of the art devices. The technology to fabricate such devices was not readily available in Korea five years ago.


Impact ionization in the narrow band gap of InGaAs (< 0.75eV ) strained channel layers in metamorphic InAlAs/InGaAs HEMT structures leads to premature channel breakdown and excessive microwave noise, thus necessitating technology for the passivation of the sides of InGaAs gate recesses. This group successfully used RPCVD for depositing silicon nitride layers in order to passivate InGaAs metamorphic HEMT.


There were three other related papers on HEMT devices. This group is working with WAVICS Co Ltd, Seoul, on HEMT-MMIC related research.


The Electronics and Telecommunications Research Institute (ETRI)


The use of GaAs substrates is a low cost alternative to InGaAs/InAlAs HEMT devices that are usually grown lattice matched to InP substrates. The problem of the mismatch between GaAs substrate and InAlAs/InGaAs epilayers is overcome by growing a metamorphic structure. The growth and subsequent fabrication of metamorphic HEMT structures is a technologically demanding undertaking that requires a deep understanding of band gap engineering, crystal growth, and processing procedures. The Electronics and Telecommunications Research Institute (ETRI) group showed that they have the ability to fabricate such devices and described the microwave performance of 150nm T-gate metamorphic InAlAs/InGaAs M-HEMT on GaAs substrates with ft and fmax of 150GHz and 240 GHz, respectively. The small signal gain of 1.7mmx 2mm chip, 3 stage M-HEMT MMIC amplifier was reported to be 13.5 dB at a maximum output of 7dBm.


Distributed Bragg reflector (DBR) lasers are attractive as light sources for monolithic photonic integrated circuits such as the master oscillator power amplifier. One of the important requirements for implementing high performance DBR lasers is to reduce the absorption loss in the DBR region. Further, tunable laser diodes are expected to play an important role in WDM networks. High output power of the order of 10mW is desired but this is considered difficult to achieve without sacrificing the wide wavelength tenability. This group reported on the fabrication of multisection tunable single mode ampled-grating DBR lasers monolithically integrated with semiconductor optical amplifiers. Fiber coupled output power exceeding 10dBm at CW operation was reported.


The delta function like density of states resulting from the incorporation of quantum dots (QDs) in the active layers of semiconductor lasers are expected to lead to improvements in their performance, including lower threshold current density, larger differential gain, lower chirp, and reduced temperature dependence compared to quantum well and bulk lasers. This group reported on the optimization of the growth of InGaAs QDs (dimensions of between 32nm x 4.5nm to 66nmx6.6nm; density of 8x10 10 cm-2) inside InGaAsP barriers by MOCVD on (001) InP substrates. The emission wavelength of the lasers was between 1350nm to 1650nm and controlled by varying the composition of the InGaAs dots. The threshold current density was measured to be 3.3kA/cm2 at 200K. This research was funded by the National R&D project for Nano Science and Technology.


Samsung Electronics Co.


This company is internationally known for the manufacture of electronics components such as flat panel displays. As presentations made at this IPRM04 conference showed, in recent years, Samsung has also been investing in InP based research on InGaAsP lasers.


The Network Research Team of Samsung’s Telecommunications R&D center reported on the fabrication of InGaAsP quantum well spectrum sliced amplified spontaneous emission (ASE) injected wideband gain lasers, covering the C-band wavelength channels. The WBG laser diodes were used as 155 Mb/s wavelength division multiplexed-passive optical network (WBG-PON) with a feeder fiber of 25 km and AWGs with 100 GHz channel spacing.


Members of the same center also described their work on the selective growth of AlGaInAs on silicon dioxide masks for fabricating ridge waveguides for spot size converter integrated laser diodes for low cost laser modules by eliminating expensive components such as the lens and thermo-electric cooler.


A group from Samsung’s Photonic Solutions Laboratory reported on the fabrication of

10Gb/s InGaAs/InP avalanche photodiodes fabricated using the non local mode. Control of the multiplication layer thickness is important in order to achieve high gain and large bandwidth. The InGaAs/InP APD were designed based on the non local mode and grown by MOCVD. The dark current was less than 1 nA and gain bandwidth larger than 80GHz.

4.2 Singapore

Nanyang Technological University

The incorporation of nitrogen (N) into InGaAs QDs is expected to reduce the bandgap (1% N will cause ~200meV shift which is larger than the 154meV shift needed to tune the wavelength from 1300 to 1550nm) and also compensate the compressive strain due to the indium for fabricating long wavelength lasers. An InGaNAs laser was recently

demonstrated with a low threshold current density of 350 A/cm2 at 270K and lifetime of over 1000 hours. Self organized GaInNAs QDs were grown by gas source MBE to a density of ~10 10 cm-2. Oxide stripe edge-emitting, laser diodes with GaInNAs QD were raised showing RT CW operation at 1200nm. The output power and threshold current were 16mW and 2kA/cm2, respectively. The observation of room temperature laser oscillation is notable and will add momentum to this field of study.


Four other papers on the use of polyimides for InP/InGaAs HBT passivation, drain transients in InAlAs/InGaAs HEMT, and the use of photoluminescence for studying structural stability of metamorphic InP/InGaAs HBT and HEMT structure after rapid thermal annealing were presented.


4.3 Taiwan

The group from National Sun Yat-Sen University used the Kramers-Kronig model for a theoretical study of the refractive index of InP quaternary compounds at photon energies near and above the band gap.


A group from the same university in collaboration with ITRI (Industrial Technology Research Institute), Taiwan’s largest public laboratory, reported on the reverse bias dependence of the photocurrent and differential absorption of InGaAsN/GaAs (N=2%) single (SQW) and double quantum well structures (DQW). These so called, ‘dilute nitrides’ are considered alternative materials for InGaAsP/InP photonic devices operating in the 1300 to 1600nm wavelength range. The absorption spectra increased in amplitude and showed a red shift above 1200nm wavelength as the bias decreased from 0 to -6 volts. Further, the electroluminescence of the DQW structure was found to be 2.6 times larger than SQW structures. These results make these structures particularly promising for chipfree traveling wave EA modules.


The National Central University

Researchers from NCU reported on improvements in the on-resistance and power performance of InGaP/GaAs DHBT by the introduction of a 350nm thick GaAs layer into the collector region. The improved performance was due to the higher electron mobility of the GaAs layer. The group used MOCVD to grow the devices on 4-inch GaAs substrates and non-self aligned technology for fabrication. Growth on large area substrates indicates the maturity of the technology in this Taiwan.

The Chan group, in work sponsored by the MOE (Ministry of Education) program for promoting academic excellence of Universities, reported on the MBE growth of InAlAs/InGaAs metamorphic HEMT on GaAs substrates. The 250nm to 650nm gate length structures were fabricated using electron beam lithography and found to have extremely large electron velocities of 2.8x107 cm/s at indium content of 60% in the InGaAs channel. These devices are promising for low noise applications in the millimeter range.

In work funded by the National Science Council, the HH Lin group used a VG V80H solid source MBE system used for growing three stacks of InAs self organized QDs on GaAs substrates for InAs/InGaAs QD lasers of 1214nm wavelength and threshold current of 124 A/cm2. The internal loss in quantum efficiency of the QD lasers was 2.5 cm-1 and 74%, respectively. The present report reflects the strong interest and expectations in the development of QD lasers in Pacific Rim countries.


4.4 China

Contributions from China were focused on the development of indigenous industries for growth of large area InP substrates. The in-situ injection of phosphorous into the melt during LEC growth has been shown to yield InP wafers with low defect density. However, the LEC process is difficult due to the high phosphorous pressure at the melting point (27.5 atm and 1335K). N. Sun of Hebei Semiconductor Research Institute described the growth of 3 and 4 inch Sn doped InP wafers by a method combining phosphorous injection followed by high pressure liquid encapsulated Czochralski (HP-LEC) method using only a single chamber. The figures of merit of a typical 160 mm diameter (100) InP wafers were:

    - Three -inch: mobility= 2260 cm2/Vs; etch pit density=0.3x 104 cm-2; carrier conc.= 0.7x 1018 cm-3

    - Four-inch: mobility = 1600 cm2/Vs; etch pit density=3.7x104 cm-2;carrier conc.=

2.8x 1018 cm-3

The use of a ‘large shoulder angel’ was demonstrated to enable the realization of twin-free



Doped InP substrates are the foundations of optical devices such as LEDs, and photodetectors. The capability to grow such wafers in China will be important for the development of the Chinese optoelectronic industry.


4.5 Japan



Dr. Murata of NTT, Atsugi, grew InAlAs/InGaAs/InP HEMT structures on 3” InP substrates by MOCVD and fabricated devices exhibiting ft=186 GHz, fmax=320 GHz, gm=1.2 S/mm with a Vth variation of 13 mV over a 3inch wafer. The NTT group is also working on reaching the goals of its HBT roadmap, including HBTs for full rate 100 Gbit/s circuitry such as 3rd generation InP D-HBTs. Recent results of D-HBTs grown on 3-inch InP substrates, with a 0.6μm emitter, 150nm collector and 30 nm pseudomorphic base, showed the devices with ft= 340 GHz, fmax= 492GHz,Jc= 6 mA/μm2, and ECL gate delay=3.48 pS. Prototype 100 GHz PRBS pulse generators were shown to exhibit ‘true’ signals with an RMS jitter of ~700 fs. Further progress in this field will depend on overcoming shortcomings due to electrical connections by use of OEIC technology.


Materials and Optical In-Situ Monitoring

Carbon is used as a low diffusivity dopant for fabricating high performance optoelectronics devices. The Mitsubishi Chemical group presented its results on carbon doped InAlGaAs/InP 1mm cavity length laser diodes for 1.3 μm wavelengths. The use of InAlGaAs/InP offers a large conduction band offset with carbon doping enabling sharper doping profiles due to the low diffusion coefficient of carbon in InAlAs. The InP/InAlAs/InP (S-doped (100) substrate) structures were grown by MOVPE, with the InAlAs layer grownat 550 C and CBr4 used to dope it ~1x10 18 cm-3. The carbon doped devices showed better performance than Zn doped structures, with no phase separation and a maximum power of 150 mW.

The importance of in-situ monitoring of material parameters for increasing the yield of production epitaxial wafers was emphasized by several speakers. Dr. Watatani (Mitsubishi Electric Corp) described the use of reflectance spectroscopy for in-situ monitoring of AlGaInP during MOVPE growth. Their method employs Fabry-Perot oscillations and the film growth rate and Al content can be determined independently of the structure. This method is expected to be applicable for monitoring other materials as well.


The session on ‘new materials for MBE’ yielded conflicting opinions about the incorporation of thallium (Tl) in semiconductors for infrared devices covering the 1.7-10μm range. The addition of Tl has the added benefit that the band gap energy of such semiconducting compounds is predicted to be independent of temperature.

Dr. Asahi’s group from Osaka found a reduction in the temperature coefficient of the refractive index and energy gap of TlInGaAs/InP (gas source MBE) with increasing Tl content. The temperature dependence of TlInGaAs/InP laserd diode structures (Tl=6%; 70 μm wide; 300 μm long; emission wavelength =1665 nm) in the LED mode was 0.06nm/K between 200-330K. Pulsed laser action was also observed between 77-310K, where the temperature coefficient of the LO peak was 0.06nm/K, which is smaller than the

0.1nm/K for InGaAsP laser diodes. These results contradict the reports by the ECL (Ecole Centrale de Lyon) from France, which found that is was possible to incorporate a maximum of 4% of Tl into an In0.46Ga0.53As matrix at relatively growth temperatures of 200-230°C. The optimum conditions for growth of Tl semiconductors still requires clarification.


Dr. Kawase (Sumitomo Electric Industries) reported on the growth of 3” and 4”, semiinsulating Fe doped InP substrates grown by liquid encapsulated growth (LEC), vapor controlled Czochralski (VCZ), and the recently developed vertical boat (VB) method. The VB method was demonstrated to produce superior results, where 4” Fe doped substrates had a dislocation density of 2,500 cm-2 and a resistivity uniformity variation of 3.9% from a nominal value of 2x107 ohm-cm. The superiority of the VB method is attributed to the ability to grow single crystals under a low temperature gradient of 5-20 °C/cm because InP decomposition is suppressed by total encapsulation.


Nikko Materials grew low dislocation 3 and 4 inch Fe doped InP substrates by phosphorous vapor controlled LEC. Dislocations densities of the 3 and 4 inch substrates were ~500 cm- 2 and ~5000 cm-2, respectively. These substrates will be particularly useful for fabricating photo diodes with low dark currents.



Quantum Dots

In an invited talk, Dr. Tsukamoto of Tokyo University, described his work on the use of STM for in-situ monitoring of InAs quantum dot growth on GaAs(001) substrates during MBE growth.


Quantum dots are promising for 1.3μm to 1.55μm range laser applications as predicted by Arakawa and Sakaki, in 1982. The use of self-organizing growth has been extensively investigated by many groups worldwide. The problem of in-situ monitoring during MBE growth is still unresolved since RHEED, although being useful for monitoring 2D growth, cannot be used to monitor and control the size of QDs that occurs as 3D growth. The use of docked, but separate MBE and STM chambers, where the deposition QDs is followed by rapid quenching (50°C/s) and samples are transferred (~2s) to the adjacent STM chamber, offers snap shots, and not real time images. Real time imaging during growth is preferable but there are still problems related to the contamination of STM instrumentation and probes located inside an MBE chamber. Dr. Tsukamoto has ingeniously managed to locate the STM within an MBE chamber and recorded the evolution of InAs QDs on GaAs over a wide range of growth conditions. His group has found no evidence for a highly mobile layer of In or InAs. On average, the InAs QDs were 1.7nm high and 5nm wide with a density of ~6x10 11 cm-2.


Dr. Yamamoto (National Inst. Of Information and Comm., Koganei) observed continuous wave laser emission from InGaSb QD VCSEL structures grown on GaAs substrates. The density of InGaSb QDs in the active region was increased by irradiating the GaAs surface with silicon atoms (~4x1011 cm-2) under an Sb flux of ~4x10-7 Torr. A record, 4.4x109 cm-2 InGaSb QDs were produced by use of Si irradiation, which is a factor of 100 greater than without Si irradiation. The InGaAs QD VCSEL structure showed cw operation at room temperature, at an emission wavelength of ~1.34μm, and threshold current density of 310 A/cm2.


In the session dedicated to electroabsorption modulators, Dr. Fukano (NTT Photonics, Atsugi) demonstrated the development of a 40 Gbit/s InGaAlAs/InAlAs modulator, with an RF extinction ratio of 10dB, operating at Vpp of 1.1V, a voltage which is less than half that  of other such devices terminated at 50 ohm.



Short Summaries of Other Notable Contributions


Dr. Yamada, Sumitomo Electric Industries, Ltd: PL, PR, and FTIR measurements showed the origin of the weak luminescence from GaInNAs grown by MOVPE to be due to nonradiative centers arising from the spatial inhomogeneity of nitrogen and the existence of N-H bonds.


Dr. M. Yoshimoto, Kyoto Institute of Technology: epitaxial layers of GaNAsBi were grown on GaAs (001) by MBE. GaBi molar fraction of up to 4.0% and GaN molar fraction up to 8.0% were achieved.


Dr. Uchiyama, Hitachi CRL: Reported that fluorine penetration into the channel layer and stacking of fluorine atoms in the Si-planar doped layer was responsible for RIE damage in InGaAs/InAlAs pseudomorphic HEMT. A n ultra-thin InSb barrier layer above the Si-planar doped layer was found to suppress degradation.


Dr. Fukuyama, NTT Photonics Laboratories: Development of a parallel feedback amplifier using InP/InGaAs double heterojunction bipolar transistor IC technology with a gain of 14 dB and 91GHz bandwidth, at 100-Gbit/s operation. It is said to be the highest bandwidth attained to date for base band parallel feedback amplifiers.

Dr. Shirai, Mitsubishi Electric Corp: Development of high performance10Gbps, direct modulation 1.3μm AlGaInAs-MQW ridge waveguide DFB LDs with output power of 5mW, operating at 120ºC.




The key drivers in the III-V and InP industry are communications (fiber optics, wireless, direct broadcast satellite), computers (information, DVD, speed), and consumers (data storage, and health care). The decline in sales of III-V related technology in the year 2000 or so resulted in a major down scaling of the InP industry. However, the vast majority of industrialists were optimistic about a brighter future as excess capacity is consumed with the prospects for increases in demand within the next few years. In particular, return on investment will depend on the implementation of new strategies, which include using large wafers sizes, (12 x 4” OMVPE; 1x12” by MBE; 12” silicon substrates) and outsourcing the production of wafers and device fabrication to reduce development and production costs.


The InP industry has shown signs of maturity following the ‘boom-bust’ cycles associated with the silicon technology. Future industrial strategies will have to allow for more cyclic movements, and it is expected that there will be an increasing need to establish new infrastructure such as chip foundries and outsourcing as used in the Si industry.

In-situ monitoring will become indispensable for increasing the yield of structures produced by both MOCVD and MBE. The use of optical methods such as RAS will be key technologies for attaining good thickness, temperature and growth rate uniformity during the growth of optoelectronic device structures.

Quantum dot lasers, grown by both MBE and MOCVD, have been produced by engineers in Japan and neighboring countries. The next 3-5 years will see reports of commercial

applications of such devices from Asian industries.

The IPRM05 will be held in Glasgow, Scotland, May 8-12, 2005.







Yuichi Matsushima, IPRM04, Conference Chairman

National Institute of Information and Communications Technology

Information & Network Systems Department

4-2-1 Nukui-Kitamachi, Koganei, Tokyo 184-8795 Japan

Fax: +81-42-327-6672

E-mail: matsushima@nict.go.jp



Seoul National University



The Electronics and Telecommunications Research Institute (ETRI)



Samsung Electronics Co.



Nanyang Technological University



National Sun Yat-Sen University



The National Central University




The research for this report was funded in part by the Office of Naval Research (ONR) grant #



InP Device History

  • 1960s

  Demonstration of LEDs and initial laser designs

  Substrate and epi development

• 1970s

  ♦ Development of 1310 and 1550 nm FP and DFB

  ♦ Epitaxial growth and fabrication advancements


  ♦ Initial manufacturing of InP lasers, detectors, LEDs for comm.

  ♦ Advanced designs to improve quality and reliability


  ♦ Development of InGaAIP/InGaN designs for visible lasers/LEDs

  ♦ Tunable InP designs manufactured


  ♦ Manufacturing 1310, 1550 nm comm. Sources/detectors

  ♦ High volume manufacturing blue/red visible sources

  ♦ Array manufacturing InP tunable sources


• 1960s and 1970s

  ♦ Nearly all research in central labs(internal use only)

  ♦ AT&T Bell Labs, IBM, HP

• 1980s

  ♦ Competitive research in central labs(internal use only)

  ♦ Bell Labs, Bellcore, Alcatel, Nortel Siemens

• 1990s

  ♦ Emergence of new vendors to InP industry

  ♦ Wafers: Sumitomo, AXT, Crystalcomm

  ♦ Epitaxial growth: IQE(EPI), EMCORE

  ♦ Devices: Mitsubishi, Hitachi

Movement towards non-centralized research

  ♦ Exploration of independent InP organizations


  ♦ Acquisitions, Spin-outs and Independence..

  ♦ Nortel→Bookham; Lucent→Agere→Triquint→Cyoptics

  ♦ Siemens→Infineon→Finisar;

  ♦ Alcatel/Corning→Avanex


  ♦ Fujitsu/Sumitomo→Eudyna

InP Industry Dynamics

Original Host Company

New Entity/New Owner




Optoelectronics components for telecom and datacom (design, fab, manufacture)



Optoelectronics datacom division (design, fab, manufacture)



Optoelectronics components for telecom (design, fab, manufacture)



Optoelectronics components for telecom (design, fab, manufacture)



Optoelectronics components for telecom (design, fab, manufacture)



Optoelectronics components for telecom (design, fab, manufacture)


JDSU(acquired SDL in whole)

Laser diodes(design, fab and manufacture)


JDSU(acquired ETek in whole)

Passive telecom optoelectronic components(design and manufacture)



Optoelectronic components for datacom and telecom


Agilent→SPG(new entity)

Fiber optics division for telecom and datacom

Major Industry Players Merits


Epitaxial Growth


Pachages Moduless

Indium based Fab

Indium absed aT&T

Fab Capaciyt Availabe

















































Consumer& Communications