Semiconductors symbolize the state of progress for this industry—arguably, moreso than any other technology or segment. When silicon was the established semiconductor technology and gallium arsenide (GaAs) was just a novelty, foundries were closely held (and proudly displayed during visits) by companies associated with defense applications. These included big names like Raytheon Co. and TRW, along with large test equipment manufacturers with sufficient resources to fund a semiconductor foundry, such as Hewlett-Packard Co. (now Agilent Technologies).
But the industry has changed drastically over the intervening three decades. Open foundries now abound, with services and semiconductor processes available to any customer with sufficient funds. And what processes: These are much more than the silicon bipolar and early GaAs metal-epitaxial-semiconductor field-effect-transistor (MESFET) semiconductor processes that first started replacing vacuum tubes in defense systems. The number of different high-frequency semiconductor processes available today is almost staggering. And newer processes—such as those based on silicon carbide (SiC) and gallium nitride (GaN) substrates—offer the promise of the tube-like power levels often mentioned by the keepers of those earlier captive silicon and GaAs foundries.
From the outset, GaAs substrates offered higher electron mobility than silicon materials; even early developers of discrete and integrated semiconductor devices knew that GaAs could support circuits such as amplifiers and oscillators well into the microwave frequency range. While GaAs technology and its supporting foundries no doubt would have succeeded in this industry on the merits of the performance it brought to a large number of commercial communications applications, GaAs had more than a little push from the defense community—most notably, the Defense Advanced Research Projects Agency (DARPA). DARPA is sometimes accused of recklessly funding technologies that may appear to benefit modern warfare, but the organization’s generous support of GaAs technology is the most likely reason that GaAs devices and monolithic microwave integrated circuits (MMICs) became affordable enough for commercial and consumer applications.
DARPA’s Microwave/Millimeter-Wave Monolithic Integrated Circuits (MIMIC) program invested about one-half billion dollars of US taxpayers’ money over an extended period, beginning in 1986, to nominally develop reliable devices for military applications [such as radar and electronic-warfare (EW) receivers]. Of course, the enormous investment not only funded the semiconductor efforts of major defense contractors, but also made it possible for commercial foundries such as TriQuint Semiconductor to grow and develop many commercial versions of GaAs MMIC products. It is safe to say DARPA’s MIMIC program made the United States the world leader in GaAs MMIC technology, a position that it holds to this day.
DARPA may not always find the most efficient ways to spend money, but there is no doubt that that organization has played a significant role in fostering new curiosities (such as GaAs semiconductors) into proven technologies. DARPA and its Microsystems Technology Office (MTO) is now interested in the potential of gallium nitride (GaN) semiconductor technology for defense electronics applications. In some ways, GaAs MMICs met many of the high-frequency needs of the military, but ran out of power. While GaAs devices are usable at millimeter-wave frequencies, they are limited in output power compared to GaN. Not only do military technologists believe that GaN can replace vacuum-tube electronics for high-power applications at microwave frequencies, but they also feel that GaN discrete and integrated devices may be usable into the terahertz frequency range (100 GHz through 10 THz).
DARPA hopes to develop next-generation GaN devices through its Nitride Electronic NeXt-Generation Technology (NEXT) program, which is headed by Program Manager Dr. Dev Palmer. The program is seeking GaN transistors capable of terahertz speeds, but also with the large voltage swings that bring high-output-power capabilities. The program is also working on GaN-based enhancement/depletion (E/D) mode logic circuits to create next-generation terahertz-speed analog-to-digital converters (ADCs) and digital-to-analog converters (DACs). DARPA showed what it could do for GaAs; should anyone doubt that it will make GaN the next big thing in microwave/millimeter-wave active circuits?