[Technology Report]
Exotic Materials Squeeze More Juice Out Of RF Power Amplifiers
While GaN high-electron mobility transistors target infrastructure equipment, SiGe HBTs race for cellular handset sockets.
Today's designers face an expanding array of transistor choices for use in RF power amplifiers in cellular communication gadgets and third-generation mobile systems. Gallium-nitride (GaN) high-electron mobility transistors (HEMTs) and silicon-germanium (SiGe) heterojunction bipolar transistors (HBTs) are mounting significant challenges to the ubiquitous lateral-diffused (LD) MOSFETs and gallium-arsenide (GaAs) HBTs.
Meanwhile, LDMOS transistors continue to make gains on the infrastructure front, where high voltages prevail, and silicon-carbide (SiC) FETs promise extra juice from smaller dies. Interestingly, even silicon bipolars are edging toward the low-voltage wireless-handset arena.
Seeing the advantages of the new compound semiconductor material in the microwave domain, several proponents of GaN RF transistors have emerged. Their dedicated efforts over the last few years are paying off. Developers have made substantial process improvements, producing high-quality material on larger substrates. Some devices have demonstrated high power, linearity, and density at high frequencies, with enough stability to make them commercially viable for wireless basestation transmitters. In short, GaN transistor technology is no longer a laboratory curiosity, but ready to roll into real production lines.
For example, Nitronex Corp. has combined the RF benefits of GaN with a novel way to fabricate HEMTs on larger, low-cost silicon wafers. Although earlier efforts focused on building these transistors on expensive 2-in. sapphire and silicon-on-insulator substrates, Nitronex has demonstrated the viability of growing GaN on mainstream silicon wafers. Toward that goal, it has developed a unique growth process called Pendeo-epitaxy, which uses the technique of proprietary metal organic chemical vapor deposition (MOCVD). This methodology has solved thermal-expansion and lattice mismatch problems, which are the major barriers to GaN's commercialization.
As a result, Nitronex has built GaN HEMTs on 4-in. silicon substrates (see "Multiple Transistor Types Vie For RF Power-Amplifier Sockets," Electronic Design, April 30, 2001, p. 56). These GaN microwave transistors exhibit 10 to 12 dB of gain with high efficiency and power density. Starting at 8 W and going up to 35 W of peak power in the 2.0-, 2.5-, and 3.0-GHz cellular bands, the devices operate from a 28-V drain voltage with 3 to 4 V on the gate as a pinch-off voltage.
According to the maker, the high current density and electron mobility of the transistor translates into higher output impedance. So it's easier to match these devices to 50 Ω, and they're less lossy.
Nitronex is trying to improve the reliability of metal-to-semiconductor interfaces, bond-wire connections, and packaging over the specified temperature range with respect to aging. The company hopes to take its GaN HEMTs into production by mid-2003.
Although initial HEMTs will be produced on 4-in. substrates, the developer is working toward scaling the growth of GaN by Pendeo-epitaxy on larger substrate wafers. For that, Nitronex is building a 6-in. production line in Research Triangle Park, N.C. The company also plans to create a second source for its units, but details were unavailable when this article was written.
According to Nitronex, the initial release will offer about 60-W continuous-wave (CW) power with a 28-V supply. Its power-added efficiency (PAE) will be nearly twice that offered by LDMOS transistors at the same power level. Also, the linearity will be better than LDMOS. The device will provide 50-Ω input/output matching. Although this HEMT will incorporate discrete passives inside of the package for impedance matching, the company plans to integrate passives on the die.
More Players: Recognizing GaN's benefits, and encouraged by recent improvements in GaN technology for RF applications, other companies have joined the race lately. Among them are Cree Inc., NEC Corp., and RF Micro Devices.
At last year's International Electron Devices Meeting (IEDM), Cree disclosed its GaN HEMT transistor, which hits 108-W CW at 2-GHz output. Its peak drain efficiency is 54%. Unlike Nitronex, which uses a silicon substrate, Cree's GaN microwave transistors were built on a semi-insulating SiC substrate, featuring a very high thermal conductivity (Fig. 1). Consequently, they dissipate very high levels of power in CW operation. According to Cree, with a 24-mm-gate width structure, the device has a power density of 4.5 W/mm.
"The demonstration of power levels over 100 W under CW conditions is a major step forward for this technology," says John Palmour, Cree's director of advanced devices. Still, more research is required to address the issues of reliability and repeatability for GaN RF devices, and to complete understanding of the degradation properties of the material.
As Cree continues to improve the quality of the material and the epitaxial growth, it has developed smaller HEMTs at 3.5 GHz that have the ability to deliver a CW power density of 9.3 W/mm. The pulsed power density reported for these transistors is 12.1 W/mm. Additionally, the developer has extended the reach of its GaN/AlGaN HEMTs to 10 GHz. At this frequency, it has obtained 38-W CW output power from its GaN devices on SiC substrate. The PAE for this transistor structure is 29%.
Unlike Cree, researchers at NEC's Photonics and Wireless Devices Research Labs have developed an SiN passivated AlGaN/GaN heterojunction FET (HJFET) on a thinned sapphire substrate that boasts 113-W pulsed power at 40-V drain bias, despite the lower thermal conductivity of sapphire. The improvement comes from the passivation, which suppresses drain current dispersion and surface traps to boost the power capability of the HJFET. These results, later presented at the IEDM, were attained from a 32-mm wide passivated HJFET with a linear gain of 6.8 dB at 1.95 GHz.
Thank you for a basic intro to all the exotic compounds currently showing up in journal articles and discussions. GaAS, GaN, Ebrium doping are terms I hear more and more often, but the basic question as to Why? these compounds are beneficial brought me no good answer -Till I read your article.
David Herlihy -December 08, 2003
Your Comments:
Enter the text from the image below
Please refresh the page if you have trouble reading this text.
Search Electronic Design
Email Newsletter
Sponsored By:
Electronic Design UPDATE provides readers with late-breaking news, opinions from industry experts, and timely technology stories. It's a unique opportunity to get your product message in front of engineers, engineering managers, and corporate managers while they're reading about critical information online.
Reprints
