Technical Insight
GaN advances into defence electronics
ABI Research analyst, Lance Wilson, predicts GaN will drive pulsed RF power device market growth. Compound Semiconductor finds out more.
With the markets for pulsed RF power semiconductor devices set to top $250 million by 2018, the future for GaN in defence electronics applications looks very healthy.
As Lance Wilson, ABI Research analyst and author of a recent study 'Pulsed RF Power Semiconductors' puts it: "Despite everything in the news about defence markets being cut, we are going to see a higher percentage of spending going towards defence electronics."
Wilson has identified what he calls a 'quantum shift' in defence-related procurement away from capital-intensive equipment such as fighter aircraft and towards relatively cheap, electronics-weighty applications such as radar and electronics warfare.
"This does not mean the defence electronics industry gets a free ride - the next three to five years will be tough - but once medium term budget battles are over, I think defence electronics is going to come out as the big winner," he says.
Right now, the pulsed RF power device market is crowded to say the least. Myriad manufacturers from Cree, Freescale, M/A-Com to Microsemi, RFMD, TriQuint, and more, are developing devices based on pure or various blends of silicon, SiC, GaAs and GaN.
Silicon-based devices dominate the sub-1GHz radar market, both GaN and silicon development is rife for 1-3GHz markets while in radar applications at 3GHz and more, market share is all going to GaN. But still, product differentiation isn't easy and consolidation is to be expected.
"Certainly over the next five years we will see consolidation within these frequency divisions," says Wilson. "This isn't based on technical issues but rather on economic and market forces."
The analyst won't name names, but asserts companies with a track record of working with government and defence organisations will have an advantage over new entrants. "Northrop Grumman and Raytheon, for example, also make their own GaN and use it in radar systems," he adds. "I don't want to say this gives them an advantage but it gives them a manufacturing efficiency that other companies might not have."
GaN drives growth
Crucially, however, Wilson believes GaN is going to drive most of the market growth between now and 2018. Beyond 4GHz frequencies, GaN, with its high voltage, high power and high frequency performance, is the only truly viable option.
As Wilson explains, silicon devices can't operate at these frequencies while GaAs cannot produce the peak power levels demanded by most pulsed applications here. As a result, all the above 4GHz frequency applications that were served by microwave tubes are now open to GaN, and so solid state device manufacturers are now turning to GaN-on-silicon and GaN-on-SiC devices.
"Most pulsed silicon device suppliers are going into GaN as they recognise a large portion of the business will drift over to the material and they don't want to lose market share," says Wilson. "And for those companies that don't appear to have a GaN program, they do, they all do. Some companies are less vocal than others but all the principle silicon manufacturers have robust GaN development."
Still, this doesn't sound the death knell for manufacturers of vacuum tubes for radar applications. As Wilson asserts, these manufacturers 'have not been sitting still and watching their business disappear' and have been developing smaller and smaller micro travelling wave tubes (TWTs), for high power and wideband RF transmission, that are similar in size to solid state amplifiers.
"The wide bandwidths of these TWT amplifiers are very difficult to replicate in solid state, so not all of the tube market will go over to GaN," he says. "And at very higher power - hundreds of kilowatts or megawatts of power - TWTs will never be replaced by solid state amplifiers."
But when it comes to GaN, does Wilson see a winning technology emerging? Not yet.
While many manufacturers are focusing on GaN-on-SiC devices, US-based Nitronex, for example, is busy churning out GaN-on-silicon RF power devices based on its proprietary SIGANTIC manufacturing process. Here, buffer layers, including an AlN layer, are deposited on the silicon wafer, which in Wilson's words is 'not an easy process'.
"Most people are opting for the epitaxial GaN device on a SiC substrate as there is much better thermal matching," he says. "From a practical standpoint, GaN on silicon substrates are the way things are going, but which one is better from a technology standpoint? The jury is out."
