Market report: VCSELs edge out the edge emitters

17 Jun 2002

In less than five years VCSELs have become the darling of telecoms networkers. If promising R&D can be translated into products, VCSELs could push out the edge emitter from datacoms and become the key component in displays, says Roy Szweda.

From Opto & Laser Europe July/August 2001

The Aixtron reactor

Two years ago, the VCSEL was a new type of laser that was only just being commercialized for the communications market. Today the VCSEL market is worth USD 500 million worldwide - according to a report about to be published by Reed Electronics Research, US - with new European start-ups with innovative designs trying to get in on the act.

Today's VCSEL-component market is dominated by a handful of companies - Agilent, Honeywell, Infineon and Mitel. These command more than 80% of the VCSEL market-place. However, recent years have seen numerous start-ups in Europe and the US that are now contesting the remaining 20% of the market.

At present, datacoms modules - produced in their millions every year - constitute 95% of the VCSEL market. However, the total available market for the VCSEL could double. This is because researchers have plans to broaden its applications to include lighting products, optical interconnects, displays, sensors and printing, to name but a few. By 2005, datacoms will command only 85% of the total VCSEL market.

Europe is home to one of the leading commercial VCSEL epiwafer firms, IQE of Cardiff, Wales, and component device leader Infineon of Germany. Geoff Duggan, VCSEL marketing manager at IQE, said: "VCSELs are implicitly simpler and therefore cheaper to produce than the well established edge-emitter lasers due to on-chip testing of the devices. However, manufacturing tolerances on VCSEL growth are much tighter, so require control of the thickness of layers to within 1%."

These qualities are being further developed by an increasing number of start-ups, including Avalon Photonics - a spin-off from the Swiss Centre for Electronics and Microtechnology (CSEM), Zurich, Switzerland; FireComms - spun out of the National Materials Research Centre (NMRC), Cork, Ireland; and ULM Photonics - spun out of the University of Ulm, Germany.

Avalon Photonics, formed at the end of last year, is a subsidiary of the CSEM. CEO Karl-Heinz Gulden said: "Owing to the opportunity existing within the telecoms markets, we have reduced our activity on singlemode VCSELs for sensing applications. CSEM has transferred all of its VCSEL activities to Avalon Photonics."

In March, a new start-up - FireComms - was born out of eight years of research carried out at the NMRC. The company will build light sources and modules that enable plastic optical fibre to be used in short, high-bandwidth, low-cost links with a huge data capacity. Its new products will include visible VCSELs.

Thomas Moriarty, CEO of FireComms, said: "We see red VCSELs being used in miniaturized bar-code readers as well as high-speed plastic-optical-fibre links. A particularly attractive idea is to put bar-code readers into mobile phones, which would then allow people to make purchases through the Internet by simply scanning a bar-code on a magazine article, for example. The key attribute of the red VCSEL would be its low operating current."

There are, however, some technical difficulties associated with producing red/visible VCSELs, the degree of which increases with decreasing wavelength. The bar-code application uses 670 nm VCSELs, which are easier to make than the 650 nm VCSELs that are used in plastic optical fibre. The biggest challenge for 650 nm devices is to get the VCSEL to operate at the elevated temperatures required of the application.

Ulm University's VCSEL array

ULM Photonics, a joint venture of the University of Ulm and Schott Communications Technology, Germany, is to commercialize oxide-confined VCSEL products, including single devices, linear arrays and large-scale two-dimensional arrays.

Max Kicherer at the University of Ulm told OLEabout what trends exist in VCSELs: "We see that 2.5 Gbit/s operation will not really challenge the VCSEL. There are a number of groups, including ourselves, that have reported 10 and even 12.5 Gbit/s operation - of course not bias-free - from an 850 nm VCSEL over about 100 m of standard 50 µm multimode fibre."

The University of Ulm has achieved many notable firsts in VCSEL development. For example, in 1992 it announced the first continuous-wave VCSEL in Europe that was proton-implanted and emitted at 980 nm. More recently, the university has reported a 4 ¥ 8 array for flip-chip CMOS integration and a 2.5 Gbit/s bias-free data transmission via 2.5 m of plastic optical fibre that has a core diameter of 125 µm.

Today, the interest in VCSELs is driven by telecoms and in particular the wavelengths coupled to fibre-optic attenuation windows. But interest is focusing on the 1300 to 1500 nm wavelengths. These VCSELs extend the potential range of fibre-optic communications and are cheaper, more reliable and have greater power than edge-emitters. This is being achieved via "dilute nitrides", whereby nitrogen is incorporated into quaternary alloys. Leading the commercialization of this technology, Infineon Technologies has achieved a breakthrough in 1300 nm VCSELs - previously, these were available only as edge emitters.

Commercially available next year, the new 1300 nm VCSELs can be modulated at up to 10 GHz, providing the output power required for fibre-optic transmission systems operating at OC-192 data rates. With the 1300 nm technology, tightly spaced laser arrays can optimize port densities.

In contrast to other players that favour MOVPE approaches, Infineon builds its new VCSELs using molecular beam epitaxy. The laser is grown on a GaAs substrate and has an active region that consists of multiple quantum wells of InGaAsN. Mastery of this novel compound semiconductor was crucial for the new devices.

However, several researchers worldwide have also had encouraging results for blue and near-ultraviolet VCSEL emission from AlGaN/GaN structures. The first blue GaN VCSELs were demonstrated only in the past 12 months, with lasing action observed at 399 nm under optical excitation. Extension of these device concepts to the near-ultraviolet is relatively unexplored, so far.

Blue GaN VCSELs have great potential in a market sector unrelated to telecoms - commercial lighting. The cost of manufacturing VCSELs enables them to compete with LEDs in established markets.

Jung Han

Late last year, the first near-ultraviolet (380 nm) solid-state microcavity laser was demonstrated at the US Department of Energy's Sandia National Laboratories and Brown University. Although the laboratory prototype uses optical pumping, electrical pumping - as required for commercial devices - is under development.

Sandia's Jung Han commented: "No-one before now has produced the technology to create a compact laser source for ultraviolet excitation. It is important for security work because many molecular bonds of interest do not respond to longer wavelengths of light. Using an AlGaN device, we chose to add indium, which brought the VCSEL efficiency to a tolerable starting point of 20%, even though it pushed the emitted wavelength into the near-ultraviolet range."

While blue LEDs are already used commercially to create white light with the aid of special phosphor-coated packages, the light is considered to be "cold".

Novalux of Sunnyvale, California, US, claims to have the first high-power surface-emitting lasers for long-haul optical networks - a 100-fold increase in power over other surface-emitting lasers, making it the most powerful laser of its kind. The Novalux range of extended-cavity surface-emitting lasers (NECSELs) comprises two 980 nm pumps for pumping erbium-doped fibre amplifiers in optical networks. The first is available in 200 and 360 mW, and the second is a multimode 750 mW pump laser for the emerging dual-clad fibre market.

"It is the nature of the NECSEL as a surface-emitting laser that makes it possible to produce high volumes of pump lasers and test them efficiently in situ," said Malcolm Thompson, CEO of Novalux.

The future of the VCSEL will depend on the development of mass-market applications, such as the lighting sector. At present these are highly speculative, but should they succeed, then the VCSEL market has the potential to reach the billion-dollar-per-annum class within the next 10 years.

VCSELs have already challenged the well-established edge-emitter diode lasers and, thanks to their greater speed, lower beam divergence and advantageous economics, also threaten LEDs in some applications. Such advantages have gained the attention of manufacturers of printing equipment, displays and novel new sensors.