IQE Prepares To Thrive In A Buoyant Market
Richard Stevenson reports.
For IQE there is good reason to look forward to the future. It is serving a compound semiconductor industry that is expected to enjoy double-digit growth in the coming years, and some of its products are operating in sectors that could grow even faster.
Both of these trends were discussed on 15 December at a Technology Update meeting attended by IQE at College Hill, London. Philippe Roussel from Yole Développement kicked-off the presentations with an overview of the state of the compound semiconductor industry, and this was followed by presentations from several players in the supply chain, including UK solar start-up Quantasol and the US-based manufacturer of GaAs-based wireless products, Anadigics.
Roussel had some good news for everyone involved in the compound semiconductor community: revenues will grow in all the major sectors, and total global sales for compound semiconductor chips will increase from $12-13 billion in 2008 to just over $20 billion in 2014.
According to him, the LED market is outpacing overall growth, and by 2014 it will be worth $10 billion, almost twice the value that it is today. The main driver behind this high compound annual growth rate of 15 percent is the increasing deployment of LEDs in automobile headlamps, forms of solid-state lighting and types of displays. Roussel says that these applications require ultra-high-brightness LEDs with an output of at least 1W, and he predicts revenue from this particular class of LED to rocket from $280 million in 2009 to $3.1 billion by 2014. These highperformance emitters will command a very hefty price tag: in 2014 their market share, in terms of volume, will be less than 2 percent, but they will account for more than 30 percent of total LED revenue.
Up until recently, IQE has had minimal involvement in the LED sector, but it could now start to tap into this market thanks to its acquisition of UK start-up NanoGaN. The University of Bath spin-out has a novel nanotechnology for making GaN substrates, and the promise of low-cost production processes may allow this platform to be used for high-power LED manufacture.
Today IQE generates a substantial proportion of its revenue from the sale of epiwafers for wireless products, a market that Roussel tips for growth, thanks to the roll out of new wireless applications. He did not evaluate this growth in terms of the dollar, but he said that the volumes of GaAs semi-insulating substrates that are used exclusively for manufacturing these types of products will enjoy a compound annual growth rate of 15 percent.
Roussel sees an opportunity for GaN and SiC devices in the power electronics market. One of the attractions of turning to these wider bandgap materials is their ability to handle far higher current densities: silicon devices must be operated at less than 1 A mm-2, while those made from GaN can be driven at 3-5 A mm-2. In addition, SiC and GaN devices can operate ate far higher temperatures than their silicon brethren.
The French analyst told the attendees that the attractive attributes of SiC and GaN devices have caught the eye of all the leading silicon power electronics manufacturers. Market leader Infineon has pioneered the commercialization of SiC Schottky diodes, and its closest rival, Fairchild, dabbled with SiC before exiting this part of the business in 2009. The next biggest silicon power electronics manufacturer, International Rectifier, is developing GaN products, and ST Microelectronics and Toshiba are pursuing both types of material technologies.
Roussel expects the market for GaN and SiC devices to be worth just $27 million in 2010, but will rocket to over $800 million by 2019. He believes that the main driver behind this tremendous growth is the uptake of wide bandgap diodes and transistors in hybrid electric vehicles. Switching from devices made from silicon to those based on SiC or GaN increases the power electronics bill, but the far higher operating temperature of the wide bandgap chips allows the removal of a second water cooling system that leads to significant cost and weight savings.
The wide bandgap power electronics market offers a great opportunity for epiwafer providers, according to Roussel. He said that many of the companies operating in this arena tend to already work with external silicon foundries, and they will not want to invest in their own growth facilities. IQE is yet to offer any products in this area, but there is no reason why it could not move into this sector over the coming years.
A quantum leap for PV
The economic crisis took its toll on the solar market in 2009, and according to market analyst Display Search, global demand experienced a year-on-year fall of 13 percent to 5.2 GW. However, a rebound is expected, with demand predicted to hit 26.5 GW by 2013.
Today systems based on concentrating photovoltaics (CPVs) are a negligible fraction of the current solar market, but deployment of this technology is expected to grow strongly throughout this decade. Quantasol, a spinout of Imperial College London, hopes to profit from this growth. “Our company is now ready for production and commercial manufacture," claimed Chris Shannon, who has been the CEO of the company since fall 2009.
If CPV technology is to enjoy significant commercial success, then its power generation costs must fall. Shannon claimed that there are three ways to do this: increase cell efficiency; reduce the cost of the cell; or cut system costs.
Quantasol has chosen the first of these options, and has developed a novel solar cell technology that features quantum well layers in the cells. The cells in a triple-junction device are connected in series, and this means that the current densities generated by every cell must be matched, which limits overall efficiency. In a conventional device the middle cell produces the least current, and the other cells are modified to reduce their current density.
“What most people do to optimize this is to make the top cell thin, but this is a compromise," explained Shannon, who went on to reveal that it is possible to increase the density in the middle cell from 12.5 A cm-2 to 14 A cm-2 by adding quantum wells.
He believes that this increase in current density will make a significant impact on the overall solar cell conversion efficiencies, and backed this claim up with data from recent trials that revealed an increase from 34.9 percent to 39.6 percent, thanks to the addition of multiple quantum wells.
This efficiency is still below that produced by the leading developers of traditional multi-junction cells, such as Spectrolab and Fraunhofer Institute for Solar Energy Systems (ISE), which both reported values of more than 41 percent in 2009. And the target for Quantasol is going up, because this record tends to increase at about one percent per year. However, Shannon claimed that commercial device efficiencies trail the record-breaking values by about 3 percent, and he believes that customers will be attracted Quantasol’s launch of devices with 41.5 percent efficiency in 2010.
Shannon estimates that the potential revenue for Quantasol’s cells will rise over the next few years, and will be over $300 million by 2012. By then the efficiency of its cells is expected to be over 43 percent, rising to more than 46 percent by 2015. This gain will result from the introduction of quantum wells to the top cell.
If these efficiencies can be reached, then the benefits are not limited to just lowering the cost of the CPV systems – they will also increase the addressable market, thanks to an increase in the proportion of the world where the cost of this technology is viable. Shannon explained that today it only makes sense to consider CPV systems in areas where direct normal irradiance (DNI) is greater than 6, such as Spain, Australia and California. However, as efficiencies improve, this technology can compete financially in many other parts of the world, including all of continental US, southern Europe, and India.
Opportunities for VCSELs
A Japanese optoelectronics industry expert, Takeshi Nakamura, provided an insight into the opportunities in the VCSEL market. He explained that datacoms is the biggest market for VCSELs today, generating sales of about 20 million units.
The light source for optical mice is the only other significant market for VCSELs, and total shipments for this application are 10-20 million a year. Despite the small size of this particular sector, competition is fierce with about 10 manufacturers competing for sales, including Finisar, Avago Technologies and JDSU.
Although the current market for VCSELs is relatively small, Nakamura believes that volumes will go through the roof over the next few years, thanks to the emergence of this class of laser in consumer applications. He tips this device for deployment in PCs and mobile phones, where it will be used in conjunction with optical fiber to fulfill the demand for high speed routing of data. Intel is leading the introduction of this technology, and last year it unveiled its development of “Light Peak", which is claimed to combine fewer, smaller connectors with longer, thinner cables to deliver higher bandwidth and multiple input/output protocols on a single cable.
Light Peak interconnects could replace copper links between the PC and the screen, facilitate video downloading from the network, transfer data from a gaming machine to its display, and improve data transfer within a PC. Intel’s Light Peak technology would backward compatible to USB 1.0-3.0, and be capable of data rates in the Gbit/s range. The cost-per-Gigabit is relatively low – about one-tenth of 8G Ethernet.
Nakamura pointed out that cell phones could also benefit from VCSEL-based optical links. He detailed the advantages of an optical approach, which include the removal of electromagnetic interference issues, the size of the technology (the thin-film waveguides can have a diameter of 0.1 mm, and pass through 3 mm diameter hinges), and the opportunity to simplify design layout.
One of IQE’s key growth sectors over the last few years has been the outsourcing of epiwafers to the wireless market, and Mario Rivas, the new CEO of Anadigics, the US manufacturer of GaAs-based MMICs, gave the keynote wireless presentation.
Rivas is a firm believer in the outsourcing model. He pointed out that the silicon industry has adopted this approach over the last few years, and the only large inhouse manufacturers today are Intel, IBM and TSMC.
Anadigics does not just outsource epiwafers. It also works with the Taiwanese firm WIN Semiconductors, a company that Rivas expects to become the TSMC of the GaAs foundry business. This relationship with WIN has helped Anadigics to increase its share of a GaAs market that should be worth $5 billion in 2011.
Just over four-fifths of this is wireless products, an area that accounts for 70 percent of Anadigics’ sales. Cable TV is the other significant sector for the company, and Rivas describes the products for this market as “very profitable".
He explained that the short-term goals for Anadigics include operating at “cash-neutral", and generating quarterly revenues of $80-100 million. According to him, the key behind the company’s recent success is its agility, and this should help it to enjoy further success in emerging applications based on 4G/LTE, mobile WiMax and femtocells technologies. If it executes on these fronts, then Anadigics will enjoy the spoils. And if IQE can garner powerful relationships with growing fabless companies, then it too shall reap rewards in years to come.
Declaration: the author of this article, Richard Stevenson, holds a small number of shares in IQE