Will we recover from the III-V decline?
Apr 12, 2012
Yes, says Yole Développement; the compound semiconductor III-V market should see a $6.1 billion opportunity by 2020
MOCVD and MBE are the main epitaxial methods used to manufacture III-V based devices .
According to Yole Développement, the combined revenue opportunity for MBE and MOCVD is estimated to be around $6.1 billion for the 2012-2020 period.
LED is by far the single largest application for MOCVD. In 2010 and 2011, the MOCVD market experienced the largest investment cycle in its history driven by a combination of the demand for LED backlit LCD TVs and subsidies by the Chinese central and local governments. Expected demands for the general lighting market was also a big factor.
This has put the market into a significant overcapacity situation that could take 12-18 months to recover from. The next investment cycle, driven by lighting applications and expected to start in 2013, will be more limited than the previous cycle due to improvements in equipment throughput and yields.
Following this cycle, further Cost of Ownership (COO) improvements offered by the next generation of MOCVD reactors should justify the replacement of 2-generation-old reactors installed during the 2010-2011 boom and drive a last small equipment cycle in the second half of the decade. By then, GaN power will also represent a substantial upside for reactor manufacturers.
Overall MBE use is heavily driven by R&D systems, capturing over 50% of the total market.
In the future, Yole expects the MBE market will be essentially driven by the continuous growth in the cell phone and wireless applications that are making heavy use of GaAs based RF components. Emerging applications like smart grid and the trend toward increasing connectivity and “intelligence” incorporated in many consumer products will provide further opportunities.
However, alternative technologies such as silicon CMOS, LDMOS, SoS, and HR SOI, represent a potential threat and could capture shares of the GaAs RF market and reduce the opportunity for MBE.
What's more, MOCVD is making progress in HEMT manufacturing. HCPV, however, could provide a small potential upside for MBE makers.
The MOCVD and MBE equipment markets are duopolies but many emerging players could change the landscape.
Aixtron and Veeco are leading in MOCVD and together represented 96% of the market in 2011. MOCVD production systems are complex. Design and optimisation require expertise in multiple fields including flow dynamics, thermodynamics, chemistry, mechanical and electrical engineering.
Applied Materials is also working with imec to create the ideal MOCVD reactor for 8-inch GaN-on-Silicon for use in LEDs and power devices.
Technological barriers to entry are fairly high. More than fifteen emerging players have been identified but so far have been struggling to capture any sizeable share of the market. But the pressure is mounting and established MOCVD makers will need to maintain that technology gap to keep emerging competitors at bay.
The main problem is total cost of ownership. Established MOCVD makers all have technology roadmaps to enable COO reduction by three or four times within the next 5 years through a combination of improved yields, throughputs and precursor utilisation efficiency.
For MBE, Yole says Riber and Veeco are the only two players offering large capacity / large throughput MBE production tools for volume manufacturing. The market research firm expects both firms will maintain their dominant positions. However, there are about ten other MBE manufacturers offering R&D or pilot production systems that also have a strong presence on the general MBE market, such as DCA, SVT and Eiko.
The Metal Organic precursor market will also be driven by LED applications. But MOCVD reactor technology improvements, including yield, consumption and wafer size, will lower the amount of TMGa and TMIn needed per cm˛ of epiwafer.
The 2010 metal organic shortage ended in mid 2011 thanks to aggressive capacity expansion by leading suppliers. Further capacity expansion plans from established and emerging suppliers could come online within the next 3 years.
If executed as announced, Yole expects significant oversupply starting from 2012 that could continue through 2016 and beyond. This situation could put pressure on prices. Further MO synthesis technology improvements could provide opportunity for cost reductions. However, the usually volatile prices of raw indium and gallium also have a significant impact on cost.
Yole's report provides Metal Organic precursor price, volume & revenue trends and data on over 150 companies.