Worldwide LED video display area shipments are expected to increase by 36.7 percent in 2019 as factors including plunging prices make them a more attractive alternative to projectors and flat panel displays, according to IHS Markit.

Global LED video display area shipments are set to reach 1.1 million square meters in 2019, up from 800,000 in 2018. This follows shipment growth of 83 percent in 2018. Revenue this year will amount to $5.1 billion, compared to $4.8 billion last year.

By 2023, area shipments will expand to about 2 million square meters and $7.5 billion in revenue.

“LED video display technology is rapidly replacing front projectors in large venues,” said Tarika Bheda, research and analysis associate director at IHS Markit. “The technology is also aggressively competing with LCD videowalls in indoor applications, such as command-and-control rooms and premium conference rooms. In addition to declining cost, lifespan, size scalability and seamless image display are the main factors driving the increasing adoption of LED video technology.”

Nearly 50 percent of LED video displays shipped this year will be installed in indoor applications, with the 2-to-4.99mm pixel-pitch category dominating the market.

“Pixel-pitch innovations and price erosion are continuing to make LED video-display technology more affordable for indoor applications,” Bheda added.

IHS Markit predicts the 2-to-4.99mm pixel-pitch category will grow to a 52 percent share of the global LED video display market in 2023, up from 40.0 percent in 2019. During the same time period, the share of the smaller-than-or-equal-to 1.99mm pixel-pitch category will expand to 13 percent, up from 9.0 percent.

The smaller-than-or-equal-to 1.99mm pixel pitch category is where most of the innovation is taking place. More brands are now venturing into this space. Sellers of projectors and legacy flat-panel displays are expanding their product lines to include fine-pixel-pitch LED video displays to cater varying customer demands.

LED video display suppliers have increased their manufacturing capacity significantly during the last few years. Competition among brands has intensified, pulling average selling prices (ASPs) lower. However, revenue growth is projected to remain positive as ASP erosion will be balanced by an increase in demand for expensive fine pixel-pitch displays.

In terms of applications, public spaces remain the biggest unit and revenue contributor for the LED video display market, with a share of more than 40 percent in 2018. IHS Markit expects the public-space application to remain on top, while transportation, control room and corporate will grow at the fastest pace during the forecast period between 2019 and 2023.

LED video display manufacturers are moving toward standardisation, so they are devising preconfigured and easy-to-install products to address the needs of certain vertical markets, where standard designs are preferred. These markets include corporate boardrooms, which don't require Regionally, China led the LED video display market with a 34 percent share of shipments in 2018, followed by North America at 22.4 percent and Western Europe at 13 percent. These regions are expected to maintain similar shares in 2019 with North America declining slightly to 19.6 percent due to US-China tariff conflict, which led to lower-than-expected results in the first quarter of 2019.

The Chinese market is thriving because of smart-city initiatives undertaken by local governments. North America leads in terms of volume due to large outdoor sports and public-space projects, which together account for more than 50 percent of total shipments. Western Europe brings significant demand from digital-out-of-home (DOOH) projects. LED video demand in Asia pacific is split between countries with high adoption rates—including India, Australia, Thailand and South Korea—and nations that have strong growth potential.

IHS Markit expects the Asia Pacific region to expand, as established brands develop their sales structures. Shipments in 2019 are expected to rise by 70.0 percent compared to 2018. Japan's LED video display shipments are expected to get a boost in 2019 as the country prepares for 2020 summer Olympics. creative designs.

Alta Devices has announced it is scaling up production of its GaAs solar cells to meet growing demand for small satellites (SmallSats). Tens of thousands of low-earth orbiting (LEO) SmallSats are expected to be launched over the next decade. They will all rely on the sun for power. Alta Devices solar material has the potential to meet the specific needs of this SmallSat boom.

Solar is the single most expensive hardware component in a small satellite. Alta Devices has developed a proprietary process to mass produce its thin film GaAs technology. The production equipment is being scaled up to produce at much higher volumes and lower cost than other space solar manufacturers. Alta's ability to produce at this scale could substantially lower the cost of each SmallSat.

Furthermore, Alta Devices flexible solar cells provide a new level of mechanical and design flexibility for the satellite industry. Alta's technology can be mounted to thin, flexible, low-mass deployable structures allowing creative design approaches to maximising the solar array. These include coiled carbon fibre booms, flat-packed, polymer-based accordion-style arrays, or even inflatable structures. This allows more compact design of the high-power solar arrays required to power high-speed LEO communication satellites.

Traditional photovoltaic assemblies (PVAs) are composed of hundreds or thousands of small solar cells, each protected with a tile of glass, connected to each other through individual metallic welds, which are then carefully grouted to prevent electrical arcing and attack from the space environment. In contrast, Alta's space product will consist of flexible, glass-free units that are 10 to 100 times larger, thus eliminating breakage, lowering cost and having no exposed electrical interconnects. Alta says its space product will enable bonding to customer substrates using a large-area, high-volume vacuum bonding process, as opposed to mounting individual cells or CICs (Coverglass Interconnected Cells).

Overall, Alta's goal is to provide LEO satellite manufacturers with a better balance of solar conversion efficiency, reduced array weight, cell durability, speed of deployment, and cost compared with conventional space solar cells.

According to filings disclosed by the US Federal Communications Commission, over 15,000 LEO SmallSats are planned for deployment in the near future. The applications of these satellites include communications, scientific research, military intelligence, remote sensing, and new technology development.

Over the coming years, Alta Devices plans to produce tens of megawatts of solar. Manufacturing will take place at Alta Devices corporate facility in Sunnyvale, California and at an international manufacturing facility. In April this year, Alta Devices solar was used on a successful mission to power over 60 ThinSats launched from the NASA Wallops Flight Facility in Virginia. Alta Devices is working with top satellite manufacturers for the deployment of its technology on several upcoming launches.

Alta Devices CEO, Jian Ding stated: “Now that we are established in the HALE UAV/HAPS market, we are committed to enabling LEO satellite projects with our unique technology. We believe our solution will meet the needs of this fast-growing market.”

Cree and ON Semiconductor have announced the execution of a multi-year agreement where Cree will produce and supply its Wolfspeed SiC wafers to ON Semiconductor.

The agreement, valued at more than $85 million, provides for the supply of Cree's advanced 150mm SiC (SiC) bare and epitaxial wafers to ON Semiconductor for use in high-growth markets, such as electric vehicle and industrial applications.

“ON Semiconductor continues to be a leader in driving the development of energy efficient innovations and devices,” said ON Semiconductor VP and chief procurement officer Jeffrey Wincel. “Partnering with Cree is essential in maintaining a world-class supply base. This agreement supports our commitment to growing automotive and industrial applications and ensuring the availability of industry-leading SiC that helps engineers solve their unique design challenges.”

“We are committed to leading the global semiconductor market's transition from silicon to SiC solutions and are pleased to support ON Semiconductor as we work to accelerate this market,” said Cree CEO Gregg Lowe. “This is the fourth major long-term agreement for SiC materials that we have announced in the past year and a half. We will continue to drive SiC adoption and availability through ongoing wafer supply agreements, such as this, and our recently announced major capacity expansion.”

LED firms Everlight and Nichia are both claiming victories over a series of courtroom wrangles about patents.

Taiwanese Everlight has issued a press statement announcing successful invalidation trial procedures in a number of countries against a range of YAG patents filed by its Japanese rival Nichia. These procedures include the US Supreme Court's decision that the claims of Nichia's US YAG patents US5,998,925 and US7,531,960 were invalid; and the Full Court of the Federal Court of Australia's decision that the claims of Nichia's Australia YAG patent 720234 were invalid.

In addition, Everlight says that on May 9, 2019, Beijing IP Court decided that Nichia's China YAG patent CN200610095837.4 was invalid; and on April 25, 2019, the Taiwanese Supreme Administrative Court decided that Nichia's Taiwan YAG patent TW383508 was invalid.

Meanwhile, Nichia has issued its own statement. It says that the China National IP Administration (CNIPA) has affirmed a patent invalidation request filed by Nichia against Chinese patent ZL200710161547.X (547 Patent) in relation to flip-chip technology owned by Everlight Electronics (China). Nichia states that after the CNIPA's decision, infringement lawsuits concerning the ‘547 Patent filed by Everlight against Nichia before the Shenzhen Intermediate Court were terminated by Everlight's withdrawal.

In January 2019, Nichia won infringement lawsuits, based on a YAG patent, against Everlight and its distributors in China. Nichia says that the judgment was final, and Nichia received damages of around $460k (RMB 3.2 million) from Everlight and its distributors.

In response to Nichia's press release, Everlight has declared: “We finally decided to withdraw the statements of the lawsuit without prejudice based on the adjustment of the litigation strategy, but the above-mentioned winning facts remain unchanged.”

Plessey has appointed Edward H. Frank to its microLED advisory board, as part of the rapid development of the company's microLED technology.

Keith Strickland, co-CEO and CTO at Plessey, said: “Ed's strong and proven history of developing and bringing revolutionary solutions to market makes him an excellent addition to our advisory board. His expertise in delivering innovative silicon and systems, at companies both large and small, will be invaluable as we continue to provide cutting-edge microLED display solutions to AR applications.”

Frank commented: “MicroLED display has many advantages such as high brightness, ultra-low power consumption, fast response time, very high contrast rate, wide colour gamut, long lifetime, environmental stability, high resolution, the option of flexible backplanes and integration of sensors in the display. Its main challenges are manufacturing complexity and yield issues, all of which Plessey are addressing, I am excited to be working with such an innovative, cutting-edge technology company.”

Previously, Frank was CEO and co-founder of Cloud Parity, a voice-of-the-customer start-up. Earlier, he served as VP of Macintosh Hardware Systems Engineering at Apple, as corporate VP of R&D at Broadcom, a broad-based manufacturer of wireless chips used in many Apple devices and served as a distinguished engineer at Sun Microsystems.Frank currently serves on the boards of Analog Devices, and Marvell.

Frank earned a BSEE and MSEE from Stanford University, and a PhD in computer science from Carnegie Mellon University, where he is presently Vice-chair of its Board of Trustees. He is a member of the National Academy of Engineering in the United States and is a named inventor on over 50 patents.

The US Army has awarded Lockheed Martin three contracts to produce additional Q-53 radars and outfit the GaN-based systems with enhanced capabilities, including extended range and counter unmanned aerial system (CUAS) surveillance.

"We realise the warfighter needs new and improved capabilities. The Q-53 represents a fast path to respond to current and emerging threats," said Rick Herodes, director of the Q-53 program at Lockheed Martin. "The flexibility of the architecture continues to allow the Q-53 to provide capabilities far beyond the original mission and allows for additional upgrades in the future."

The Army awarded Lockheed Martin a contract for a third lot of 15 Full Rate Production systems. Once this contract is delivered the Army will own 189 Q-53 systems. The Lot 3 systems will continue to be produced using GaN transmit-receive modules. This will provide the radar with additional power, reliability and the possibility for enhanced capabilities including extended range, counterfire target acquisition (CTA) and multi-mission, which delivers simultaneous CTA and air surveillance.

The primary mission of the Q-53 is to protect troops in combat by detecting, classifying, tracking and identifying the location of enemy indirect fire in either 90 or 360-degree modes. The Q-53 has protected warfighters around the world since 2010.

Lockheed Martin uses an open GaN foundry model, using relationships with commercial suppliers that use the power of the expansive telecommunications market to provide military-grade GaN modules while taking advantage of commercial cost efficiencies.

Akash Systems, a San Francisco-based GaN-on-diamond startup focused on satellite communications, has announced $14.5 million in Series A funding, including $10 million in new equity funds, and an additional $4.5 million converted from prior convertible notes.

Investors in this round include Khosla Ventures, Founders Fund, ACME Capital, Sriram Krishnan, Correlation Ventures and others. Akash will deploy the new capital toward scaling its transmit/receive radio modules and power amplifier business, moving it closer to profitability.

“The successful close of this funding round marks an important next step in fulfilling Akash's mission statement,” said co-founder, CEO and GaN-on-diamond Inventor Felix Ejeckam. “Our aim is to be the RF communications link for every satellite in space.”

Akash has designed its satellite transmit/receive radio modules to easily integrate with existing ground station and satellite infrastructure for satellite makers in all markets. The company's radios, which are expected to be available for purchase in late 2019, promise impressive performance not yet seen on the market.

“Akash Systems is playing a critical role in meeting the growing and vital need for improved satellite communications infrastructure,” said Delian Asparouhov of Founders Fund, which is focused on assisting entrepreneurs to build impactful new energy and technology companies. “We're proud to be part of Akash's journey as a critical enabler and accelerant in global communications.”

“It's rare to see a young company with such transformative technology already approaching commercialisation,” said Alex Fayette of ACME Capital. “Akash's business lies squarely at the intersection of major trends like global connectivity, high-performance telecommunications and space-based infrastructure. Their foundational technology will underpin other solutions across these industries. We are thrilled to partner with the team at this early stage as they execute against their vision.”

“What Akash has been able to achieve in its first few years is remarkable, as it has already attracted major customers across the ecosystem. The company is quickly shaping trends in global telecommunications. It's one of the most exciting companies we've seen in a while,” said angel investor Sriram Krishnan.

In GaN-on-Diamond, the hottest part of a transistor is brought to within tens of nanometers of synthetic diamond - the most thermally conductive material ever made by humans. The result is a dramatic reduction in the waste heat generated from the power amplifier, and therefore the entire satellite system; in the communications “base stations” of space, the RF power amplifier is typically responsible for the bulk of the power consumed and the heat generated. The cooler state of operation gives RF and satellite system designers a previously unattainable thermal envelope to dramatically improve a satellite's communication bandwidth and energy efficiency while simultaneously shrinking the system's size, weight and operating costs.

Akash is currently manufacturing GaN-on-diamond-based power amplifiers and radio modules for customers who make satellites requiring high frequency and high power efficiency. Its radio products are on track to hit the market in Q4 2019.

Australian semiconductor company BluGlass has announced that the United States Patent and Trademark Office has issued US Patent 10,355,165, Buried Activated p-(Al,In)GaN Layers. This proprietary technology has significant performance potential in a number of applications including high-brightness LEDs, micro-LEDs, laser diodes, solar cells, and other optoelectronic and power electronics devices.

BluGlass' RPCVD-grown buried activated p-GaN layers can help address one of the LED industry's critical challenges: “efficiency droop”. The efficiency of an LED is a measure of how much light output is generated for a given amount of electrical power used to drive the device. Efficiency droop is where the efficiency of an LED reduces as operating power is increased. This challenge results in many of today's high-powered LEDs being operated outside their peak efficiency.

Incremental efficiency improvements continue to be a major objective for LED manufacturers.

One potential resolution of efficiency droop is multi-junction or cascade devices (LEDs or laser diodes as examples). RPCVD's buried activated p-GaN layers are a critical building block to enable the tunnel junction, which in turn allows multiple LEDs to be grown in a continuous vertical stack and interconnected in a single, high efficiency chip, called a cascade LED. As a result, less power is needed to deliver the desired light output, potentially eliminating efficiency droop and significantly increasing device performance.

Cascade LEDs are expected to enable smaller, cheaper and higher performing LEDs - the three key interest areas of the LED industry. To date, functioning tunnel junctions, and therefore cascade LEDs, have been prohibitively difficult to produce due to the challenges in achieving buried activated p-GaN using conventional growth technologies such as MOCVD.

BluGlass is developing and commercialising a unique Australian technology called remote plasma chemical vapour deposition (RPCVD), a revolutionary approach to the manufacture of group III nitrides which are essential components used in millions of electronics devices globally. RPCVD offers better-performing, lower-cost devices and more environmentally sustainable processes for electronics manufacturers producing LEDs for automotive and overhead lighting, microLEDs for wearables and virtual reality display and power electronics for efficient power conversion.

BluGlass' Managing Director, Giles Bourne said “This patent adds an important cornerstone to BluGlass' intellectual property portfolio, protecting our unique RPCVD process, hardware and competitive advantages. RPCVD allows us to develop this elegant new option for resolving the challenges of efficiency droop in high-performance LEDs - something that's been very difficult to resolve using the industry standard MOCVD technology.” “Our growing strategic patent portfolio, comprising 63 internationally granted patents, continues to underpin the commercialisation of our RPCVD technology across a range of market segments with long-term market exclusivity”. BluGlass has received strong interest on the performance potential of RPCVD tunnel junctions and cascade LEDs from the industry. Discussions with a number of groups are continuing.

The Superior Court of California has granted Lumileds a permanent, worldwide injunction against Elec-Tech International (ETI), immediately prohibiting ETI and its affiliates from using, disclosing, licensing, or selling Lumileds' epitaxy technology trade secrets that ETI was found to have misappropriated.

The injunction also prohibits ETI from manufacturing, distributing, importing, exporting, or selling any products that use certain of ETI's own epitaxy growth recipes that the jury concluded were produced using Lumileds' trade secrets.

This injunction is the result of a bench trial (following last August's jury trial), where Lumileds presented experts and significant evidence that clearly demonstrated the appropriateness of a permanent worldwide injunction.

The Court specifically found that Lumileds' trade secrets continue to exist and are valuable. The Court determined that without an injunction, ETI would continue to misappropriate Lumileds' trade secrets and further stated that if ETI continued to use the stolen intellectual property, Lumileds would suffer irreparable harm. Furthermore, the Court has denied ETI's motions for a new trial and to set aside the verdict awarded by the Jury.

“We are very pleased with the Court's determination that a permanent injunction is warranted in this case as it further validates the jury's findings from the 2018 verdict,” said Cheree McAlpine, chief legal officer and SVP of Lumileds. “Protecting our innovations is paramount to our business, and we will continue to vigorously defend our intellectual property.”

After a four-week jury trial beginning in July 2018, a jury of 12 people found that ETI misappropriated and used five Lumileds' trade secrets related to epitaxy technology. These trade secrets include the foundational epitaxy-development technology used by Lumileds to produce its LEDs.

The jury determined that ETI used Lumileds' trade secrets in six of their recipes and the misappropriation and use of Lumileds' trade secrets was a substantial factor in causing ETI to have obtained an unjust benefit of $66 million.

Infineon Technologies utilizes the ANPC topology for its hybrid SiC and IGBT power module EasyPACK 2B in the 1200 V family. Optimising for sweet spot losses of the CoolSiC MOSFET and the TRENCHSTOP IGBT4 chipsets respectively, the module features increased power density and a switching frequency of up to 48 kHz. This is especially well suited for the needs of new generation 1500 V photovoltaic and energy storage applications. The new ANPC topology supports a system efficiency of more than 99 percent. Implementing the hybrid Easy 2B power module in for example, the DC/AC stage of a 1500 V solar string inverter allows for coils to be smaller than with devices with a lower switching frequency. It therefore weighs significantly less than a corresponding inverter with purely silicon components. Additionally, the losses with silicon carbide are smaller than with silicon. For this reason, less heat must be dissipated so that the heat sink can also shrink. Overall, this leads to smaller inverter housings and costs savings at system level. Compared to 5-level topologies, the 3-level design reduces complexity of the inverter design.

The Easy 2B standard package for power modules is characterised by an industry-leading low stray inductance. Additionally, the integrated body diode of the CoolSiC MOSFET chip ensures a low-loss freewheeling function without the need for another SiC diode chip. While the NTC temperature sensor facilitates the monitoring of the device, the PressFIT technology reduces assembly time for mounting the device.

Wafer firm IQE has started initial VCSEL production for a second major customer at the Newport Mega Foundry in Wales and further positive qualifications.

Initial production has now commenced with a second major customer, serving the Android supply chains, and several more are in advanced stages of qualification. Performance data from the Newport Foundry has exceeded previously attained performance levels, and several new qualifications are now in long term reliability testing.

As a result, IQE has signed a contract extension with one of its largest VCSEL customers, extending the current contract until the end of 2021. In addition, two other existing contracts have also been extended with several other new contracts in final negotiations.

VCSELs are a key component in multiple current and future 3D sensor systems, using both structured light and Time of Flight (ToF) technologies. Applications range from 3D Facial Identification (front facing) systems and world facing cameras on mobile handsets to LiDAR and in cabin sensing for autonomous drive vehicles, range-finding and 3D sensing systems for a wide range of industrial and commercial applications.

Production volumes are expected to ramp strongly over the next few years as adoption across multiple mobile platforms and other use cases proliferate. In anticipation of such strong growth, IQE has invested heavily over the last two years in building the worlds's largest VCSEL epi-wafer facility in Newport, UK (IQE's Mega Epi Foundry). IQE is involved in multiple engagements across all of the aforementioned applications and is currently qualifying more than a dozen volume customers at the Newport Mega Foundry.

Mark Furlong, executive VP, Global Business Development, Photonics and InfraRed commented: "The initiation of production for a second major customer at our Newport Mega Epi Foundry, is another important milestone in cementing Newport as the premier location globally for large scale, state of the art, VCSEL production. The availability of the large-scale production capacity which has been established over the last 18 months at Newport, and consequent improvement in production efficiency and performance, is highly valued by our customers as they address mass market adoption across multiple applications."

Vishay Intertechnology is adding to its portfolio of short wavelength ultraviolet (UVC) emitting diodes with a compact new SMD device for sterilisation, sanitation, and purification applications.

Designed to replace mercury UVC lamps, the Vishay Semiconductors VLMU35CL-280-120 delivers an extremely long lifetime in a 3.5 mm by 3.5 mm by 1.2 mm surface-mount package with a quartz window that eliminates the need for an external lens.

The device offers an emission angle of ± 60° and high radiant power to 3.2 mW at 20 mA or 7.5 mW at 50 mA. Built on AlGaN technology, the VLMU35CL-280-12 features maximum forward voltage of 6.5 V, a wavelength range of 265 nm to 285 nm, and forward current up to 70 mA, compared to 30 mA for typical competing parts.

The emitter diode's specifications make it suitable for water and air purification, physical surface sterilisation, medical disinfection, and portable sanitisers. RoHS-compliant, halogen-free, and Vishay Green, the VLMU35CL..-280-120 is compatible with reflow soldering processes and features a Moisture Sensitivity Level of 3 in accordance with J-STD-020.

Power Integrations has announced new members of its InnoSwitch3 families of offline CV/CC flyback switcher ICs. The new ICs feature up to 95 percent efficiency across the full load range and up to 100 W in enclosed adapter implementations without requiring a heatsink. This groundbreaking increase in performance is achieved using an internally developed high-voltage GaN switch technology.

Quasi-resonant InnoSwitch3-CP, InnoSwitch3-EP and InnoSwitch3-Pro ICs combine primary, secondary and feedback circuits in a single surface-mounted package. In the newly released family members, GaN switches replace the traditional silicon high-voltage transistors on the primary side of the IC, reducing conduction losses when current is flowing, and considerably reducing switching losses during operation. This results in substantially less wasted energy and therefore increased efficiency and power delivery from the space-saving InSOP-24D package.

Targeting high-efficiency flyback designs, such as USB-PD and high-current chargers/adapters for mobile devices, set-top boxes, displays, appliances, networking and gaming products, the new ICs provide accurate CV/CC/CP independent of external components, and easily interface to fast-charging protocol ICs. The InnoSwitch3-CP and ‑EP variants are hardware-configurable, while the InnoSwitch3-Pro incorporates a sophisticated digital interface for software control of CV and CC setpoints, exception handing and safety-mode options.

Comments Balu Balakrishnan, president and CEO of Power Integrations: “GaN is a pivotal technology offering significant efficiency and size benefits over silicon. We anticipate a rapid conversion from silicon transistors to GaN in many power applications. InnoSwitch3 has been the clear technology leader in the offline switcher IC market since we launched the silicon variants 18 months ago, and the new GaN-based ICs further extend our lead by advancing both the efficiency and power capability of our flyback products.”

Power Integrations' new InnoSwitch 3 ICs are available now, priced at $4/unit in 10,000-piece quantities. Five new reference designs describing USB-PD chargers from 60 W to 100 W are available on the Power Integrations website, along with an automated design tool, PI Expert, and other technical support documentation.

BY 2024, the market for edge-emitting lasers (EELs) will have more than doubled to $5.1 billion, predicts market research & strategy consulting firm, Yole Développement.

Spurred on as always by the optical communications markets, emerging potential killer applications are set to underpin more growth, driving a 13 percent compound annual growth rate in the next five years.

“Edge-emitting lasers can be used as direct lasers, or coupled with optical fibres or crystals to make diode-pumped solid state lasers,” highlights Martin Vallo, technology and market analyst at Yole. “As a result, the number of applications is impressive, including optical communications, material processing, medical, sensing, optical storage [and more].”

“But as the automotive industry head towards an autonomous era, lidar systems for automotive applications are being developed at a fast pace,” he adds. “So 3D lidar sensing is a key emerging killer application.”

According to Yole, right now, the largest market segment for edge-emitting lasers is optical communications, providing a hefty $1.385 million revenue in 2018 and forecast to swell to $3.397 million by 2025. At the same time, the relatively well-established materials processing and display sectors will provide healthy growth opportunities with markets expected to grow by some $100 million and more, to $533 million and $441 million, respectively. Meanwhile, combined printing and optical storage markets will dwindle from $155 million to $53 million.

“The growth rate of optical communications does depend on large Chinese companies such as Huawei and ZTE that produce many laser-based optical systems for datacoms and telecoms,” points out Vallo.

“So optical communications could compensate for decreases in optical data storage,” he adds. “But emerging applications in automotive lidar systems as well as new applications in the medical and materials processing sectors could also compensate here.”

Technology competition

Right now, different edge-emitting laser technologies, and of course, VCSELs are available for all market sectors.

For example, in optical communications, 1550 nm, InGaAsP/InP Distributed Feedback (DFB) EELs and 1300 nm, InGaAsP/InGaP Fabry-Perot EELs are typically used at transmission distances between 2 km and 40 km and at up to 40 Gbit/s. However,

850 nm InGaAs- and AlGaAs-based VCSELs dominate at 60 m to 300 m transmission distances and lower transmission speeds.

But for the new wave of emerging applications, which technology will be used where?  Perhaps not surprisingly, the pathway is not yet clear. For starters, as Vallo points out, the EEL industry is extremely fragmented.

For example, leading players in materials processing sectors, making, say, laser dicers, are vertically integrated so will manufacture edge-emitting lasers, laser systems and everything in between. Meanwhile, in upcoming sensing applications, numerous device-level challenges mean companies are more specialised and will focus on producing just the edge-emitting laser.

“It is going to be difficult to unify the laser diode device market and we do not expect to see any [market] unification coming soon,” says Vallo. “We see plenty of edge-emitting laser designs for different applications but will carefully watch how he edge-emitting lasers market changes.”

In the meantime, it’s no secret that the VCSEL has a head-start in the world of 3D sensing.  With a small form-factor, narrow wavelength range and being testable at wafer-level, these devices have already infiltrated consumer applications, most notably in facial recognition, in Apple’s iPhone.

However, next-generation, high-end smartphones are set to include rear-facing 3D sensors with ranges up to 5 m to detect and measure objects across a room. And this could favour the higher-power, edge-emitting lasers.

As Vallo puts it: “A huge advantage of the VCSEL is its form-factor and the technology can be easily used for facial recognition in mobile applications, but edge-emitting lasers have potential in rear-3D sensing. It’s really a question of whether the power of a VCSEL is enough here or if [manufacturers] will use the edge-emitting lasers instead,” he adds.

And of course, the future looks bright for both VCSELs and edge-emitting lasers in the sensing systems of future autonomous vehicles. Thanks to high efficiency, narrow bandwidth, VCSEL arrays are vying to replace LEDs for in-cabin sensing, to, say, detect and warn a texting driver.

Meanwhile, both VCSELs and edge-emitting lasers could be used for short- to medium-range lidar for blindspot detection, lane departure and rear-collision warning.

“Edge-emitting lasers are one hundred times brighter than the VCSEL, which will make these lasers more suitable for long-range sensing applications,” says Vallo. “But the VCSEL is a low power device and could save battery power in the future electric car.”

“For each technology, the market share of each application remains questionable and the final integrators will always be considering which is the best option for them,” he adds.

Caption: Rising revenues for edge-emitting lasers. Source: Edge-emitting Lasers – Market and Technology Trends 2019 report, Yole Développement, 2018)

In July this year, Sweden-based SiC power semiconductor developer, Ascatron, revealed it had raised Ä3.5 million in funds, following the sale of shares in Basic Semiconductor, a joint venture business in China.

Now focusing on product development and volume production, the company is currently providing its first devices – 1200 V and 1700 V diodes - to customers for design-in projects, with MOSFETs scheduled for volume production next year.

“Right now we are working on 1200 V devices and higher,” says Ascatron chief executive, Christian Vieider. “We see our competitors having difficulties getting out products with higher voltages than 1700 V so high-voltage devices are our focus.”

“Our next step is to increase capacity, and we will coordinate this in line with customer demand,” he adds. “We have good connections with suppliers of epitaxy equipment and we see this as being feasible within a one-year time period.”

Crucially, the company has spent the past year transferring the process from its in-house clean room at the Electrum Laboratory, just outside Stockholm, to an external foundry. According to Vieider, Ascatron’s design is now cost-compatible with competitors, and the company is currently gearing up for volume production.

“We use this buried grid in our design to shield off electric fields and then fabricate the Schottky diode or MOSFET on top of this,” says Vieider. “This was a new process step for the foundry, but we’ve got this to work and we are now getting the same results at the foundry as in our own clean room, which is a big achievement for us.”

Up until now, Ascatron has been manufacturing some 1000 epiwafers a year on its own equipment, but Vieider is keen to triple production in the short term, and take volumes to at least 1000 a month in the next three to four years. As he points out, the foundry capacity is more than 20,000 wafers a year, although, in his words: “It could take some time for us to get to this level [of production] from where we are now.”

And as a handful of SiC device makers have been making the tricky transition from 100 mm to 150 mm wafers, grappling with yields and defect levels, Vieider anticipates Ascatron production will be centred on the 150 mm size for the foreseeable future. “Six-inch wafers are still not as high quality as four-inch wafers. So in terms of quality, we would be stepping back a couple of years by shifting to eight-inch wafers, even if these were available to buy,” he says.

But could industry issues over SiC wafer supply stymie Ascatron’s plans for volume production? Vieider isn’t worried.

“Everyone is fighting these wafer shortages but we haven’t seen any problems yet,” he says. “For us volume production starts at 1000 wafers a month and from what we understand these volumes are available.”

“Also, if we talk about scaling over a two-year period, our suppliers don’t anticipate this being a problem as they are investing heavily in capacity,” he adds.

So what now for Ascatron? With its fabrication process qualified, the company is working on product qualification, which Vieider reckons will be achieved towards the end of 2020. The company is targeting high-power, high-voltage applications, including industrial and the all-important electric vehicles, and its focus will remain firmly here.

“We have seen interest in 650 V devices but right now we need to focus on our buried-grid technology and there is a clear advantage for this at the higher voltages,” says Vieider. “Also, it’s important for us to be competitive with other silicon carbide vendors by having a design where we can increase the current density and make our chips smaller, so we can get more chips on every wafer.”

And importantly, Vieider is confident that Ascatron is ready to compete with the likes of Infineon and Cree. “From customer feedback, Infineon is the reference, so we have been evaluating our first products against its high-performing devices,” he says. “We have our innovative design which gives us many advantages compared to these bigger companies.”

At the same time, Ascatron can also tailor its diode design to customer requirements. “Until recently, we were considering packaging but we’ve now reduced our product offering to bare dies so that our customers take care of the packaging and offer a complete module,” he says. “In this way Infineon could also be our customer.”

There are some people that always want the latest and the best. They are sure to be ones to splash out on a 5G smartphone, allured by the promise of a faster connection that could enable a 4K movie to be downloaded in a minute or so.

It is likely that these early adopters will be heavily disappointed. If they live in a city that has just rolled out 5G, it is true that they will enjoy better connectivity – and when everyone wants to access the internet at the same time, that’s highly valued. But if they expect to download data at breath-taking speeds, they will probably be in for a shock. According to many experts, the increase in speed will be only about 20 percent.

Why are these expectations going to be so different from reality? Well, it’s because 5G is being deployed in two phases – and it is only the second phase, which doesn’t start rolling out until around 2023 at the very earliest, that is going to deliver a substantial hike in data rates, thanks to a move to wider frequency bands in the millimetre-wave spectrum above 24 GHz. 

Right now, it is only the roll-out of a sub 6 GHz band that is underway. This typically operates at a frequency of 2.6 GHz or more, and as well as increasing coverage for subscribers, it will open up new business opportunities for operators, such as industrial monitoring.

Only by understanding the details of this two-phase deployment, as the analysts at Yole Développement do, is it possible to predict how the emergence of 5G will impact sales of RF compound semiconductor devices. According to experts at Yole, the roll-out of 5G will be good for GaAs and GaN. Production of GaAs-based power amplifiers for mobile phones will increase, as the introduction of the new network will lead to, on average, an additional amplifier in every phone; and sales of GaN HEMTs will rise as carriers deploy 5G infrastructure, which will start forming national networks in the next few years.

In cities in South Korea, the UK, the US, Japan and China, deployment of a 5G, sub 6 GHz network is already underway.

“The technology looks a lot like LTE,” says Antoine Bonnabel, Technology & Market Analyst for the Power & Wireless team at Yole. Thanks to this similarity, Bonnabel argues that it is relatively easy to develop and implement this form of 5G, with existing base stations upgraded to include this new communication technology.

In some dense urban areas, operators are improving their coverage by also introducing a network of small cells. This has happened in South Korea, where Samsung has deployed 53,000 of them, operating at 3.5 GHz. Operators are using these cells to initially provide LTE coverage, but in time they will be upgraded to operate on the 5G network.  

Small cells are also being deployed in large numbers in the US. In 2018, Sprint added 27,000 small cells, spread across nine cities. These cells operate at 2.5 GHz.

The output of these small cells is less than that of conventional base stations. Rather than producing 240 W, it’s typically between 40 W and 80 W. “It’s still high power,’ says Bonnabel. “You can reach up to a few hundred metres.”

GaN versus silicon

For many years, network operators had to choose between silicon LDMOS and GaN to provide the amplification in the base station. For the emerging 5G networks, the base station needs to deliver a high-power at a relatively high efficiency, a requirement that plays into the hands of GaN. According to Bonnabel, for all the 5G base stations built with conventional technology, as frequencies are above 2.6 GHz, GaN is the only viable option.

As carriers have moved from one generation of communication technology to another, they have also, independently, advanced their antennae technologies. One innovation has been the introduction of active antennae, a move that allows beam-forming and massive MIMO. To deliver these technologies, a small front-end is employed at each antenna, rather than a single front-end for all antenna elements. Configuring the base station in this manner slashes the power requirement for each antenna element, opening the door to the use of silicon LDMOS in 5G networks.

However, even though GaN faces competition from silicon LDMOS for some tasks within the 5G network, shipments of this device will still rise fast. According to Bonnabel, sales of GaN devices to the base station market will increase from $300 million this year to $750 million by 2024.

Bonnabel says that today, the only substrate used for the production of these GaN HEMTs is SiC. However, he points out that the opportunity to turn to a silicon substrate has attracted much interest within industry, due to its potential for low cost, with prices approaching those of silicon LDMOS.

Grabbing the lion’s share of the GaN-on-SiC market is the Japanese firm Sumitomo, which ships its GaN amplifiers in the form of discrete devices. “The main reason [for their dominance] is that they are the exclusive supplier of Huawei,” says Bonnabel, who believes that Sumitomo’s success is deserved, because it has the best technology – and it continues to refine it. To cater to the growing market, Sumitomo is increasing its production capacity, a move that will help it to stay ahead of a chasing pack that includes Wolfspeed and Qorvo.  Legislation in several countries, related to security concerns, has dented Huawei’s attempts to win contracts to deploy 5G all over the world.

“China, because of this ban, has invested more in its 5G deployment than it was supposed to,” says  Bonnabel. “The investment that they were supposed to make over the course of three years, they have made it in a year.”

One advantage that Huawei has over its peers in the 5G market is that it has a lot of expertise in making and deploying devices operating at relatively high-frequencies within the sub 6 GHz band. In comparison, the main frequencies in today’s networks in the US and Europe are below 2.6 GHz, a domain well served by Nokia and Ericson.

Ironically, it is only recently that the US has permitted the use of higher frequency bands, such as those at 3.5 GHz and 4.2 GHz. In developing countries, legislators are far more willing to permit the use of higher frequencies, and operators are using them for 4G deployments. “They go for Huawei products, because they are 30 percent cheaper than Ericson,” says Bonnabel.

5G smartphones

Alongside the roll out of 5G networks is the introduction of a handful of compatible smartphones. Aside from a couple of exceptions, 5G capability is limited to the sub 6 GHz network.

So far, 5G smartphone sales have totalled between 5 million and 10 million, according to Cédric Malaquin, a Technology & Market Analyst who specialises in RF devices & technologies and works within the Power & Wireless division at Yole. Malaquin expects that even by the end of this year, sales of 5G phones will not have exceeded 50 million, and it will take until 2025 before they can account for 30 percent of the market. “So, in number, something like 600 million units.” 

The compound semiconductor content in the smartphone is in the form of a GaAs power amplifier. “The power and efficiency is unmatched by any other technology, especially CMOS,” says Malaquin. He expects the dominance of the GaAs power amplifier to continue, as operators are requesting more and more power from the device to the base station. 

Rewind the clock a few years and manufacturers of GaAs power amplifiers, such as the big four from the US – RFMD, Skyworks, TriQuint and Anadigics – would be striving to increase their sales of discrete devices. But that’s no longer the case. The battleground has changed, with competition now surrounding sales of a front-end module, which incorporates a GaAs power amplifier, along with a filter and a chip that switches between the frequency bands.

Rivals in this sector include Skyworks, Murata and Qorvo  –  an entity formed from the merger of RFMD and TriQuint  –  that are drawing on their in-house GaAs power amplifier technology to make front-end modules. Up against them are the likes of Qualcomm and Broadcom, all making their own GaAs power amplifiers by drawing on the services of GaAs foundries, such as WIN Semiconductors of Taiwan.

It is not easy to assess the impact of the introduction of sub 6 GHz 5G networks on the production volumes of GaAs power amplifiers. In some countries, such as Japan, the established 4G networks operate at high frequencies, and it is possible that a design engineer could use one amplifier for both the 4G and 5G signals. But that is not the norm, according to Malaquin, who has calculated that, on average, the introduction of 5G will lead to the addition of one power amplifier per phone. That’s good news for the makers of MOCVD tools, GaAs substrates, metal-organics sources and characterisation equipment, that will all welcome the additional business.

Note that the key component in the front-end module is not the GaAs power amplifier, as its production can be outsourced. “It’s more important to have the right cost and the right performance for the filter,” says Malaquin, who argues that it is the excellence in this technology that is behind the success of Skyworks, Qorvo, Broadcom, Murata and Qualcomm.

Millimetre-wave networks...

The second phase the 5G rollout – the introduction of a millimetre-wave network – will require a revolution, rather than just an evolution, in infrastructure. At frequencies of 24 GHz and above, signal attenuation is far higher, so the smartphone must be far closer to the base-station. Consequently, there needs to be a far denser network, which will be made of small cells. Deploying this only makes sense in areas where there are large crowds, such as sports stadia.    

Trials of millimetre-wave networks are underway. In the US, there has been great interest in this technology, but according to Bonnabel, those working in this area have found it more challenging than expected. “Right now there is a move back from millimetre wave, and larger interest in sub 6 GHz frequencies.”

Bonnabel believes that it will be 2023 before the millimetre-wave market starts to develop, and from then on growth will be modest. “Operators spend a limited amount of money on their annual deployment, so the amount of money for millimetre-wave small cells will be limited, and will not grow,” reasons Bonnabel. He expects a saturation in sales to the small cell market to occur by 2025. “We don’t expect it to be a booming market that will see millions and millions of devices.”

…and smartphones

For those living or working where 5G millimetre-wave trials are taking place, there is the opportunity to experience this revolutionary network by buying one of a very small number of compatible handsets. They include a version of the Samsung SG105G, and a solution from Motorola. It far from easy to cram the additional technology into a phone that enables it to operate in the millimetre-wave band. In fact, Motorola doesn’t even try, preferring instead to have a separate unit for this spectral domain, and strapping it onto one of its Z-Series phones (see the image on the right).

Designing a handset that works well at these frequencies is a tough ask. Antenna gain is low, and current CMOS power amplifiers are inefficient. “It’s a problem for your battery life,” says Malaquin.   

A teardown of the Motorola millimetre-wave unit shows that it employs a silicon-based power amplifier. “The efficiency is quite poor,” says Malaquin, who believes that engineers adopted this approach because it is the quickest way to get to market.

Over the next few years, other technologies could come to the fore. “Above 6 GHz, your HBT today doesn’t do the job,” says Malaquin. “But the pHEMT, it’s a different story – you can use the pHEMT design to do a millimetre-wave power amplifier.” Another candidate is GaN-on-silicon, a technology dogged by concerns over reliability.

This is further evidence that there is much uncertainty surrounding the revolutionary phase of 5G. That’s certainly not the case for the first phase, which is an evolution of 4G LTE that will drive an increase in the production of GaAs HBTs and GaN HEMTs. So, in the short term, 5G is definitely going to be good for the compound semiconductor industry – and looking further ahead, there could be additional sales from the roll out of millimetre-wave networks.

The introduction of 5G will have an impact on the architecture of the RF front-end of mobile devices. Taken from 5G’s Impact on RF Front-End Module and Connectivity for Cell phones report, Yole Développement, 2019.

Demands on data centres are rocketing. That’s no surprise, given the rise of Netflix, Youtube, social media, and far greater use of the cloud.

To cope with this ever-increasing demand, there have been changes taking place within the infrastructure of the data centre. Historically, coaxial cables have routed data between different central processing units and different mother boards. But these cables are bulky, lossy, and highly dispersive – and they fail to accommodate the tremendous increases in both the data rate and the number of channels in the data centre. So, to overcome this bottleneck associated with the total bandwidth capacity in the data centres, coaxial cables have been replaced with optical fibres. They have the upper hand on many fronts, combining a far lower loss with less dispersion, a broader bandwidth and a smaller form factor. The only downside is that by moving from the electrical to the optical domain for the links, additional electrical-to-optical and optical-to-electrical processes have to be inserted on the transmitting and receiving sides of fibre, respectively.

One option for the light source that generates the data for the optical link is the edge-emitting, distributed feedback (DFB) laser. It is a well-established device, deployed in the traditional optical communication network, where the length of the fibre can be up to thousands of kilometres. The DFB laser dominates this application, because of its high optical output power, low chirp, and a narrow spectral width, a set of attributes that make it ideal for transmitting data over vast distances with very low error rates. But the DFB laser is not a good option in datacentres, where many millions of high-speed light sources have to be densely packaged with ICs, to move data at high speeds between different boards and chips. Its Achilles heel is its relatively high power consumption, which can lead to thermal management issues, wasted heat, and a high carbon footprint for the data centre.

For this particular application, a far better candidate is the VCSEL. In data centres, over 95 percent of fibre channels have linking distances less than 50 metres, so fibre loss and dispersion are not major concerns. In these very-short-reach links, what is paramount is to minimise the electrical-to-optical and the optical-to-electrical conversion losses. And when judged in this manner, the VCSEL wins hands down, consuming just one-tenth of the power of the edge-emitter, thanks to the far smaller size of the optical cavity. 

VCSELs started to become the dominant source in data communications more than a decade ago. Over the intervening years, a steady stream of improvements to their design have driven increases in the maximum data rates. Ideally, this trend must continue, as it holds the key to a new-generation of data centres that are more frugal that their predecessors. Fail on this front, and the carbon footprint of the data centre could spiral out of control.

Working at the National Central University, Taiwan, our team has developed a novel form of VCSEL that sets a new benchmark for performance. This device, a prime candidate for serving tomorrow’s datacentres, features a pair of apertures (see Figure 1). One is defined by zinc diffusion, and allows us to manipulate the number of modes of the VCSEL; and the other is created by oxide relief, and provides us with a lever for the device modulation speed.    

A great strength of our device is that it is able to deliver a record-breaking performance at elevated temperatures. For those that build data centres, this is a very valuable attribute. When VCSELs are paired with photodiodes to form transceiver modules, they must be packaged as close as possible to the electronic integrated circuits, which generate a significant amount of heat during high-speed operation. Dense packaging is essential, because it minimises propagation loss and dispersion in the high-speed electrical signals transmitted by the printed circuit board. However, heat generated by the ICs increases the junction temperature of the VCSEL, applying the brakes to the modulation speed. 

To serve in the next generation of optical interconnect channels, VCSELs need to have a high-speed performance that is almost invariant with temperature. A data rate of more than 50 Gbit/sec per lane must be maintained between room-temperature and over 85 °C. 

We have proven that VCSELs based on the GaAs family of materials can meet these requirements. If GaAs-based VCSELs emit at 850 nm, they have an active region that is based on GaAs quantum wells, but if they emit in the spectral range that spans 900 nm to 1000 nm, they have wells made of InGaAs. We work with the latter, because InGaAs wells introduce a larger compressive strain and conduction band offset. These modifications are most welcome, because they minimise the probability of electron leakage under high-temperature operation.

Armed with this form of active region, our novel design of 940 nm VCSEL has realised record modulation bandwidths. Judged in terms of optical-to electrical conversion, for a -3dB loss, we have recorded bandwidths of 40 GHz and 32 GHz under room-temperature and at 85 °C. These milestones follow important successes by other groups, such as: a team led by researchers Chalmers University of Technology that reported a 850 nm VCSEL with a 3 dB electrical-to-optical bandwidth of 30 GHz in 2015; and in the same year, a group at the University of Illinois, Urbana Champaign, that announced a multi-electrode VCSEL with a 3 dB bandwidth up to 37 GHz. Recently, the company VI System demonstrated single-mode VCSEL chips with 3 dB bandwidth up to 30 GHz.

We fabricate our VCSELs by processing epiwafers purchased from Landmark. Grown on an semi-insulating GaAs substrate, our epiwafers feature: an active region with several InGaAs quantum wells; an aluminium-rich AlGaAs oxidation layer that sits just above the active region; and a pair AlGaAs-based distributed Bragg reflectors, which form the mirrors either side of the active region (see Figure 1).

Our first step is to process these epiwafers into mesa structures with diameters of typically 30 µm. After this, we employ wet oxidation and oxide-relief processes to selectively remove the Al_xO_1-x current confined layer. During these two processes, oxidation takes place in the oxide current-confined layer and also at the edge of distributed Bragg reflector layers with high aluminium concentrations. Consequently, both are subsequently etched away (see Figure 2 (a), which shows scanning electron microscopy images of our oxide aperture before and after the oxide relief process).

The oxide-relief process is beneficial. As the refractive index of air is much smaller than that of the Al_xO_1-x layer – it is 1, rather than 1.6 – there is a further reduction in the parasitic capacitance of the oxide-relief device. This leads to an increase in the bandwidth of the VCSEL, so that it can operate at a higher data-rate.

In a typical VCSEL, there is just one aperture, formed by oxidation. But in our devices, we have a second one, created by zinc-diffusion. A downside of the diffusion process is that it supresses the local lasing phenomenon: threshold gain increases in the zinc-diffused region, due to a rise in free-carrier absorption loss; and there is also a reduction in reflectivity. However, there is an upside too, as zinc diffusion enables mode control, a key to improving the quality of data transmission through an optical fibre. Improving mode control trims chromatic dispersion, and in turn increases the transmission distance, which can be valuable in modern data centres that have expanded to several buildings to accommodate the demand for more data, and may contain links of a kilometre or more.

We use a high-quality Si₃N₄ layer to create a diffusion mask that controls the location of the zinc within our device. Inspecting the distributed Bragg reflector after the diffusion process shows that it introduces a great deal of disorder – the individual layers are no longer visible with a scanning electron microscope (see Figure 2 (b)). This disordering is beneficial, as it reduces the contact resistance when the injected current flows through the distributed Bragg reflector.

To evaluate the performance of our VCSELs, we have plotted their current-voltage-light characteristics and their threshold current over temperatures ranging from ambient to 105 °C (see Figure 3). These measurements show that over that temperature range the increase in the threshold current is very small – it is less than 0.2 mA. We attribute this to the large conduction band offset in the AlGaAs/InGaAs active region, which minimizes electron leakage. 

Another attribute of our VCSELs, highlighted in Figure 3, is their low differential resistance. Under a 6 mA bias current, differential resistance is as low as 100 Ω at room temperature, and just 80 Ω at 85 °C. In comparison, conventional 980 nm VCSELs with a similar size of oxide aperture – that is, about 3 mm – have differential resistances of typically 200 Ω at 85 °C. The superiority of our devices stems from the zinc-diffusion process and the topmost current-spreading layers.

We have also measured the bias-dependent electrical-optical frequency responses of our VCSELs, and their corresponding output optical spectra, at both room temperature and 85 °C (see Figure 4). The graphs show that the maximum 3 dB electrical-optical bandwidth can hit 40 GHz at room-temperature, and is still as high as 32 GHz at 85 °C.

Additional measurements have been made to evaluate the capability of our VCSELs to transmit data through a fibre. They include 50 Gbit/s and 60 Gbit/s on-off keying transmission results with eye-patterns and corresponding bathtub curves for the device under room-temperature and 85 °C operation over a 1 m OM5 fibre (see Figures 5 (a) to (d)). These results are very encouraging. At 50 Gbit/s, the bit-error rate is below 1 x 10^-12, all the way from room-temperature up to 85 °C, without any pre-emphasis or equalisation techniques. In other words, over this temperature range, we have produced bit-error-free transmission. What’s more, at room temperature, the maximum error-free data rate can be as high as 60 Gbit/s.

We have also investigated transmission over 100 m, using the same type of fibre (see Figure 6). In this case, standard five taps feed-forward equalisation is used at the receiver end to compensate for the limited bandwidth of around 32 GHz. Results show that we can realise invariant 50 Gbit/s transmission performance, with a bit error rate below 1 x10^-7, from room temperature to 85 °C.

Our results are very encouraging for the future of very-short-reach optical interconnect networks. The capability of these novel VCSELs over a wide temperature range makes them very attractive candidates for data centres. In addition, they could also serve in other harsh environments. They could, for example, deliver data transmission inside autonomous cars and drones, opening up yet more markets for what is arguably the most exciting compound semiconductor device of our age.

UV light emitters serve a broad range of applications. They can be used to provide purification of water and air; deliver disinfection, curing, bioagent analysis and detection; and provide the light source for fluorescence spectroscopy. But all of today’s common UV sources are far from ideal. Whether it is a mercury lamp, an excimer laser, or a diode-pumped solid-state laser that produces UV emission, the source suffers from a large footprint, a low energy efficiency and a limited emission wavelength, and it uses toxic material. What’s more, when viewed from an environmental perspective, all these forms of UV emitter are unsustainable.

Far more promising alternatives are LEDs and laser diodes. These sources, which reach the UV with AlGaN alloys, are far smaller, safe in operation, and able to target a specific wavelength. This expands the number of applications that the UV source can serve. 

However, before UV LEDs and lasers can become considerable commercial successes, a few challenges have to be overcome. High on this list is the limited spectral range – to date, the shortest reported wavelength for an electrically injected AlGaN quantum-well laser is 336 nm. Another drawback of this device is its large threshold current density, which is on the order of at least 10 kA cm^-2. Like the laser, the performance of the LED is vastly inferior to the visible cousins, with an efficiency well below 10 percent in a large part of the UV-C spectrum, which spans 100 nm to 280 nm. The losses come from inefficient light extraction, the presence of a large density of defects, and poor p-type current conduction.

It is widely accepted that the poor performance of UV lasers and LEDs stems from the conventional growth techniques for the AlGaN alloy and the basic material properties. To overcome these bottlenecks, there needs to be new approaches to growing the material and engineering the material properties.

At the University of Michigan we are trying to succeed on these fronts with arrays of AlGaN nanocrystals. This approach has much promise. One of its merits is that it is possible to realise a light extraction efficiency in excess of 90 percent by arranging nanocrystals in a photonic crystal structure, a geometry that is compatible with large-area LEDs. Taking this approach, we have produced devices spanning 210 nm to 360 nm that have efficient p-type doping, one of the issues in conventional devices. What’s more, we have made a low-threshold, electrically pumped nanocrystal laser operating at 369.5 nm, by engineering the optical properties via the photonic crystal structure.

Efficient light extraction

Mid and deep UV LEDs have an extremely low efficiency, due to the dominance of transverse-magnetic-polarized optical emission. In these devices, the electric field of the emitted light is parallel to the c-axis, due to the negative crystal-field splitting energy in aluminium-rich AlGaN. One consequence of this is that the majority of light is emitted out of the active region in the horizontal direction (see Figure 1 (a)), and it cannot be extracted out of the chip due to total internal reflection. The resulting light reabsorption inside the device is highly detrimental, leading not only to an extremely low efficiency, but also causing severe heating and device instability.

To see if we can address these issues, we have undertaken finite-difference-time-domain simulations to investigate whether photonic nanocrystals can improve light extraction. These simulations consider emission at 240 nm, for nanocrystals that have a lateral size of 95 nm and are spaced apart by 160 nm (see Figure 1 (b)). By calculating the relation between the wavelength and the wavevector in these photonic crystals, we have determined that light emission is primarily coupled in the vertical direction (see Figure 1 (d)). Another encouraging aspect of this design is that there is a large overlap between the centre of nanocrystals and the maxima of the electric field intensity profile. This aids the output power of the nanocrystal LED.

The benefits of the nanocrystals are so substantial that they can clear the bottleneck of light extraction, which has been hampering efforts to improve deep UV LEDs for many years. While extraction efficiency for planar structures is below 10 percent, in photonic crystals it can exceed 90 percent.

Note that the photonic crystal structures that we are investigating are not just designs for hero results in a lab. The architecture is robust, offering reasonable tolerance for deviations in dimensions. Even if the 95 nm target is missed and diameters vary from 80 nm to 100 nm, the efficiency of light extraction is still around 85 percent or more, implying that this device has much commercial promise. 

Selective area epitaxy

The majority of III-nitride nanostructures are grown on foreign substrates through a self-organisation process. However, that is no good to us, as it creates nanocrystals with random diameters at random positions. Instead, we employ selective-area epitaxy to achieve site-controlled growth of nanocrystals with a well-defined size and position. This involves using a growth mask to promote the growth of nanocrystals on the template, in favour of growth in the apertures on the mask.

To produce our nanostructures, we take a GaN (or AlN) template formed on sapphire, and deposit a thin titanium mask on top. This is followed by electron-beam lithography and dry etching, to create nanoscale aperture arrays (see Figure 2(a)). Then, through the use of optimised growth conditions, we ensure epitaxy takes place in just the apertures, to create nanocrystals with a precisely controlled size and spacing (see Figure 2(b)). With this approach, we can produce AlGaN nanocrystals with an aluminium-content across the entire compositional range. According to photoluminescence measurements, strong room-temperature emission is possible over the range 210 nm to 360 nm by varying the growth conditions for the AlGaN nanocrystals (see Figure 2 (c)).

We have fabricated LEDs from our nanocrystals. Plotting current as a function of voltage reveals excellent characteristics, with 50 µm × 50 µm devices showing a turn-on voltage of 4.4 V, and a current density of 100 A cm^-2 at 5.5 V. This is a level of performance comparable to, or better than, that of a conventional AlGaN quantum-well LED.

Electroluminescence spectra reveal a peak at 280 nm, coming from emission from the active region (see the inset of Figure 2(e)). There is also a peak at 260 nm, due to electron overflow and resulting emission from the cladding layer. The output power produced by this device, measured at current density of 250 A cm^-2, is 0.9 W cm^-2.

p-type conduction

With conventional planar structures, one of the biggest barriers to realising high-efficiency with mid- and deep-UV LEDs and laser diodes has been the quality of the p-type current conduction. As aluminium content increases in the AlGaN alloy, p-type conduction gets more and more difficult, due to a larger activation energy and reduced hole concentration in the valence band. For example, for the typical p-type dopant, magnesium, the room-temperature activation energy in AlN can be as high as 500-600 meV. Consequently, only a tiny proportion of magnesium dopants are activated and contribute holes in the valence band. Compounding this issue, if the magnesium dopant is cranked up, this promotes the formation of donor-like defects, such as nitrogen vacancies. This self-compensation severely limits the magnesium concentration and ultimately restricts the hole concentration.

With nanocrystals, doping is a very different story. Thanks to efficient lateral strain relaxation, far less energy is required to substitute a magnesium dopant for an aluminium or gallium atom. This significantly enhances magnesium dopant incorporation. What’s more, the growth condition can be tuned to the nitrogen-rich regime, suppressing formation of donor-like nitrogen vacancies. The upshot is a far more effective incorporation of magnesium in the nanocrystal structure, compared with a planar structure.

Evidence for the enhanced incorporation of magnesium in nanocrystals is confirmed by a pronounced photoluminescence peak associated with a magnesium-acceptor-related transition at 235 nm (see Figure 3(a)). Measurements on AlN nanowire transistors reveal that the hole concentration at room-temperature reaches 6×10¹⁷ cm^-3 in magnesium-doped AlN nanocrystals – that is seven orders higher than that in a magnesium-doped AlN epilayer. Our detailed investigations also suggest that the efficient p-type conduction comes from a combination of the significant reduction in activation energy for a portion of magnesium dopant, and hole hopping conduction in the magnesium impurity band. Both of these traits are enabled by the significantly enhanced magnesium-dopant incorporation in Al(Ga)N nanocrystals and suppressed defect formation.

Topological resonators

The performance of the UV laser diodes is held back by the lack of a high-reflectance mirror for optical feedback. Its absence results in a high threshold and severe heating.

One solution is to form a defect-free photonic nanocrystal structure that functions as a topological resonator with a high quality factor. Note that this architecture does not need extra mirrors for optical feedback. Designs with an arbitrary irregular shape can exhibit a high quality-factor resonance at the G point, a feature that is essential for the robustness of device operation.

We have undertaken simulations of irregular photonic crystals with a lattice spacing of 400 nm, and a separation between the nanocrystals of 36 nm (see Figure 4). These calculations suggest a Q-factor of over 11,000. This is a high value that has encouraged us to produce a laser diode with this architecture. The current-voltage characteristics of this device, featuring a topological resonator based on an AlGaN nanowire photonic crystal structure, include a turn-on voltage of only 4 V and a very low reverse bias leakage current (see Figure 5(a)). When operating at room-temperature, as the current increases from below to above threshold, a sharp lasing peak with a linewidth of about 0.2 nm emerges from the broad spontaneous emission (see Figure 5(b)). We have also studied the variation in output with injection current, with results revealing an unambiguous lasing threshold of 10.6 mA (see Figure 5(c)).

Our efforts show that AlGaN nanocrystals have many desirable properties, including negligible defects and dislocations, highly efficient p-type conduction, and efficient light extraction. All these strengths are difficult to realise with conventional AlGaN planar structures. Further work is now needed to develop efficient mid- and deep-UV optoelectronic devices, including LEDs and laser diodes, that are hampered by standard device architectures. 

3D sensing has crept quietly into common use. The advent of the iPhone X facial recognition system has enabled millions of people to interact with 3D sensing systems on a daily basis.

At the heart of these facial recognition systems is a class of laser known as the VCSEL – the vertical cavity, surface-emitting laser. The VCSEL enables structured-light depth-sensing analysis of an image. However, its attributes also make it an attractive option for other forms of 3D sensing, and it is being used in the automotive market to provide time-of-flight sensing, as well as LiDAR, for autonomous driving.

These applications are ensuring that the sensing market has a great future. According to figures from market analysts, global revenue for the 3D sensing market will climb from $2.1 billion in 2017 to $18.5 billion in 2023. That’s a compound aggregate growth rate of 44 percent that will help to spur sales of the VCSEL. By 2023, the market for just VCSEL illumination and associated optics will be worth more than $6 billion.

For the foreseeable future, VCSEL sales will rise across mobile, consumer, industrial and automotive sectors. In each of these sectors, there are different requirements for the illumination sources. However, despite these differences, there are five common considerations: wavelength selection, performance over temperature, reliability, range and cost.

VCSEL virtues

Sales of the VCSEL are rising far faster than its more common cousin, the edge-emitting laser, due to its superior attributes. Merits include its vertical architecture (see Figure 1 (a)), which simplifies production by removing the need for cleaving wafers and coating facets. After processing, all that is needed is to dice the wafer. Another feature of the VCSEL is its relatively short cavity, enabling high data rates. These strengths, plus the circular emission profile that is defined by the aperture, have made the VCSEL a very competitive and successful product for data communications, and are driving its success in 3D sensing. For the latter, higher output powers are required, realised by assembling arrays of apertures on a single die. 

The traditional VCSEL is a ‘top-emitter’ – that is, light is emitted through the top of the semiconductor wafer. But that’s not the only possible configuration. At TriLumina of Albuquerque, NM, our team is pioneering the ‘back-emitter’, an architecture offering greater flexibility in VCSEL array design. To produce our devices, we grow the heterostructure up-side down compared to the top-emitter, and use special layers within the laser structure to realise electrical connections to the anode and cathode (see Figure 1 (b)).

Our back-emitting VCSEL configuration has several notable advantages over its conventional sibling. It requires no wire bonds for electrical connections to the array – instead, the VCSEL die is simply flip-chip bonded to a sub-mount that has all of the required electrical connections defined. Thanks to the absence of wire bonds, back-emitting VCSEL die can be tiled close to one another to form large illumination modules that produce enough power for long-range sensing applications. In addition, eliminating wire bonds trims the parasitic impedance of the package, and in turn speeds rise and fall times, improving the performance for pulsed depth-sensing applications. 

Another merit of back-side emission is that it allows micro-lenses to be monolithically etched directly into the emitting surface. Armed with this feature, beam shaping is possible: either reducing the divergence of each emitter, or diffusing light output. What’s more, lenses can be offset from the emitter aperture to steer the beam off-axis. This enables fully solid-state beam steering, by combining beam pointing with individual addressability of zones of the VCSEL apertures. Armed with these features, the overall system cost for 3D sensing can be reduced by eliminating optical elements or beam steering components.

What wavelength?

For engineers that design 3D sensing systems, one of their biggest decisions is the selection of the  operating wavelength. Common lasing wavelengths include 850 nm, 905 nm, 940 nm and 1550 nm. The wavelength that an engineer selects will influence: eye safety; visibility, as sources that meet eye safety requirements can still be distracting; interference from ambient light; and the availability of sensors for detection. Note that all these considerations can impact other aspects of system design.

To meet eye safety requirements, engineers must comply with the maximum permissible exposure level. This varies with wavelength and the duration of the pulse (see Figure 2). Independent of wavelength, shorter pulses have a lower maximum permissible exposure (this is a result of the way the plot, shown in Figure 2, is constructed, as the same amount of energy is being delivered in a short pulse and a long pulse). For 3D sensing systems, pulse widths can be as short as 10 ns – and even less in some cases. This increases the burden on eye safety considerations. Moving to longer wavelengths can ease this, as sources operating at 1550 nm allow for the greatest pulse power for eye safety limits.

VCSEL arrays typically operate at 850 nm to 980 nm. This might suggest, based on Figure 2, that eye safety would limit output powers compared to sources at longer wavelengths. However, it’s not that simple. That’s because eye safety considerations account for the difference between an array of lower-power sources and a single, high-brightness emitter, such as an edge-emitting laser. Far from the source, VCSEL arrays appear to be a single, circular output beam, and at these distances they behave like point sources. But as one gets closer to a VCSEL array, they become an extended source and eye safety considerations change, now being determined by the output of individual emitters. That’s welcome, as it can slash the overall power factors by up to four orders of magnitude, and make 940 nm VCSEL array sources nearly equivalent to point sources at 1550 nm in regard to eye safety.

Another consideration, which affects the performance and cost of the system, is the level of background illumination at or near the source wavelength. When background illumination is high, it must be suppressed with additional filtering. To gauge the level of filtering required, one can consider the solar radiation at air mass 1.5. Data provided by NREL (see Figure 3) reveals the level of background at the wavelengths 850 nm, 940 nm, and 1550 nm. At 850 nm, radiation is relatively high, and additional filtering is required for systems operating at this wavelength. Both 940 nm, which is not coincidentally where the iPhone X VCSELs operate, and 1550 nm, have reasonably low background radiation.

One final wavelength-related consideration relates to the availability of suitable sensors. There are three different commonly available detector types used for 3D sensing systems: silicon avalanche photodiodes; multi-spectral CMOS imagers; and compound semiconductor detectors and arrays, like InGaAs avalanche photodiodes. The first two types are compatible with 850 nm and 940 nm sources, while the latter is needed for those operating at 1550 nm (see Figure 4). When cost is a concern, silicon sensors are preferred. There are a wide variety of silicon sensors with imagers supporting many different approaches to 3D sensing.

Coping with the heat

All of the key markets for 3D sensing systems – consumer, industrial and automotive – require the optical source to maintain performance over an extended temperature range. The toughest requirements are found in the automotive sector, where operation may need to extend from -40 °C to +125 °C. This challenge is particularly tough because, to prevent system costs from escalating, the illumination source must ideally run without temperature control. Unfortunately, the characteristics of the laser change with temperature, including its wavelength and its output power.

With VCSELs, changes with temperature are far smaller than those of edge-emitters, eliminating the need for temperature control for appropriately designed 3D sensing systems. Shifts in wavelength are lower by a factor of five, down from typically 0.4 nm/K to 0.07 nm/K. The superiority of wavelength stability stems from the use of far shorter cavities, which result in the resonator modes within the VCSEL being widely separated in wavelength. When the VCSEL heats up, the resonator wavelength and the gain profile shift to a longer wavelength. However,  only one resonant mode overlaps the gain profile, and this determines the wavelength. In contrast, in an edge-emitter, resonator modes are narrowly separated in wavelength, so when the gain profile shifts, a new lasing mode emerges at the gain peak, leading to a new lasing wavelength.

The shift in the output power of semiconductor lasers with temperature tends to be associated with the resonator mode and gain profile shifts, and changes in non-radiative recombination rates. The extent of the variation in power is strongly dependent on the design, but in a typical edge-emitter it can plummet to near zero at 85 °C. In contrast, our VCSEL arrays operate at up to 125 °C at close to 70 percent of the rated power at 25 °C (see Figure 5), meeting the most demanding automotive requirements without the need for temperature control systems.

Ensuring reliability

Reliability is a key concern, especially in automotive applications where user safety is critical. Our VCSELs excel in tests to judge this characteristic, because the back-emitting architecture is inherently robust to the typical stresses used for assessing reliability. The n- and p- contacts sit on the same side of the VCSEL array, and are both entirely clad in metal or silicon nitride. Thanks to this design, there are no ingress paths for moisture into the active area of the device. 

That’s not the case for the top-emitting VCSEL, which needs to have an opening in the p- contact to allow light to exit the device. It is only fair to point out that these top-emitting VCSELs have been shown to be highly-reliable – but we have no doubt that our back-emitting configuration provides even greater confidence for long-term operation. We find no evidence of degradation when driving our arrays under accelerated conditions for over 9,000 hours. We are now undertaking full evaluation of our 8 W illumination module to the automotive quality standard, AEC-Q102, and expect to complete qualification before this article is published.

Expanding the range

Depth-sensing systems use various approaches to determine distance. For short-range systems, either a structured light approach is adopted, such as in the iPhone X, or one based on stereoscopic vision. For medium to long-range systems, including LiDAR, the more common approaches are based time-of-flight. They can be either direct, using a flash; or indirect (see Figure 6).

We have designed and delivered back-emitting VCSEL array modules to customers covering every range, from consumer applications requiring depth sensing over a few metres to front-facing automotive LiDAR operating at 250 m or more (see Figure 7 for examples of our products, and Table 1 for key metrics for the illumination modules). 

Our lasers are going into short-range systems, providing a key ingredient in world-facing cameras for mobile applications. The unique advantage of illumination modules formed with our VCSELs is that they can be soldered directly to a printed circuit board, like any other electrical component. This makes them a low-cost option for replacing LEDs with VCSEL sources. Modules run at low duty cycles – typically a 1 percent to 3 percent average duty cycle – and they produce an output power of about 4 W.

Offering a little more power are our illumination modules that target automotive safety applications, such as occupant monitoring. Examples include two 1 mm x 2 mm die that are connected in series, mounted onto a ceramic sub-mount, and produce an output of around 8 W at a 3 percent duty cycle.

For the 50 m to 100 m range, our illumination modules are being deployed in multi-mode time-of-flight cameras. These are used in indirect time-of-flight systems, featuring multi-spectral CMOS imagers that attain high-resolution depth maps, along with black-and-white and infrared images. The cameras are positioned at the four lamp locations on an automobile and enable the driver to benefit from increased awareness and safety. For this application, we produce 100 W modules, operating at a 2 percent duty cycle, formed from ten 2 mm x 2 mm VCSEL arrays on a ceramic sub-mount.

We ship 600 W illumination modules for the longest ranges. Assembled from twenty 1 mm x 2 mm die, and housed on a ceramic sub-mount, they are suitable for direct time-of-flight and flash front-facing LiDAR systems. To realise a narrow divergence – it is approximately 12 degrees, in terms of its full-width at half maximum – micro-lenses are etched into the chip. Using pulses as short as 10 ns or so, delivered at a 100 kHz repetition rate, LiDAR realises the specified range. Note that it is the low duty cycle of just 0.1 percent that allows the module to hit its high peak power.

They key to the competitive performance of all our modules is our unique back-emitting VCSEL array design. Often, to develop a new module, all we need to do is to produce a new ceramic sub-mount design.  By tiling our VCSEL array die in appropriate series or parallel networks, or sometimes a combination of the two, we can fulfil the desired electrical and optical specifications. 

Looking ahead

VCSEL technology is continuing to make its way into new applications. The advent of the 3D sensing system has made the VCSEL array a very strong contender for infrared illumination. Armed with our back-emitting laser with integrated microlenses, we are able to offer a compelling product that helps to drive down the total system cost for advanced 3D sensing systems. The back-emitting configuration paves the way for higher levels of integration that will see VCSEL technology used in innovative designs at the forefront of 3D sensing.