GaAs is Green & Energy Efficient

 Dr. Mike Heimlich, Microwave & RF Marketing, AWR

Complementary metal oxide semiconductor (CMOS) technology is truly amazing. The degree to which technological progress and international commerce have benefited from IC miniaturization cannot be underestimated, and device integration has led to astounding new capabilities that technologists of fifty years ago could never have dreamed of.

Today, nearly every aspect of analog and mixed-signal design has been realized in silicon at frequencies or applications that no one would have thought possible just a few years back. But will the futures' yet-to-be-defined applications result in the end of the line for compound semiconductors once and for all?

GaAs microwave monolithic integrated circuits (MMICs) exploded on to the scene in the mid- to late- 1980s. Many of the processing challenges in making bulk, crystalline silicon had been solved. The semi-insulating properties of GaAs, so valuable to microwave applications, could be reliably and somewhat repeatedly manufactured for three- and/or four-inch wafers. MMICs had come of age. Fast forward to the new millennium… Despite the benefits of integration and low-cost, high-volume production, silicon has not taken over. So what happened? What are the lessons of the recent past that provide a basis for prognosticating the compound semiconductor future in light of silicon’s never-ending progress? Two words: batteries and modules.

For portable devices, battery life is paramount and modules inexpensively combine multiple technologies leading to more of it. The first generation or two of cell phones taught the industry many things, not the least of which was that people want more “talk-time.” Compound semiconductors fit the bill better than silicon when it comes to energy efficiency. This is not so much a billboard for being “green” as it is a reason for forestalling the recharging of energy-hungry handsets. While the higher threshold voltages of many compound semiconductor technologies would seem to be a detractor to efficiency, since power is about both voltage and current, when operation gets down below 5V then AlGaAs and InP really come through.  The balance between voltage and current when determining power and efficiency is a complex interplay of effective device operation, voltage breakdown, and current loss. Efficiencies of compound semiconductor-based designs come much closer to their topologically ideal values than their silicon counterparts mostly because of the physics behind the operation of the active devices.

The next frontiers in wireless are a bit hazy. Now that W- and V-band CMOS is available, what role will GaAs play? The technology winners and losers won’t be determined by single-chip integration capability. To be sure, solutions will need to be well-integrated and small. However, if the wireless growth of the last decade is an indicator, solutions will most definitely need to be energy efficienct. 

CMOS designers have published results of integrated radios at 60GHz and discrete solutions above that so it looks like silicon is poised to continue its dominance. However compound semiconductors (GaAs) will also see use in new applications and higher frequencies. Wherever battery life reigns supreme, circuits will need to be implemented in technologies yielding the highest power-added efficiency (PAE), the greatest linearity, and the lowest noise figure. The players, in terms of the chemistry and devices, may change, but there is every reason to expect that FET and bipolar devices will be found in handset modules using InP, AlGaAs, or related substrates. If these technologies are to be successful, geometries will need to be reduced. But more importantly, the module technologies will need to ride the technology curve to higher performance as well, boding well for GaAs –as its green and energy efficient!