International Defense Security & Technology Your trusted Source for News, Research and Analysis
Related Articles
Radar and communication systems are ubiquitous in both military and commercial applications. Whether they are being used to navigate, control air traffic, track weather patterns, carry out search-and-rescue missions, map terrain, or countless other functions, radar technologies are constantly advancing.
As radio-frequency (RF) systems, radar capabilities hinge on the ability to sense and communicate across long distances while maintaining signal strength. Powerful RF signal capabilities extend mission-critical communications and situational awareness, but the microelectronic technologies that strengthen RF output – specifically, high power density transistors – must overcome thermal limitations to operate reliably and at significantly higher capacity.
The performance of these systems depends on the signal-to-noise (S/N) ratio achievable at the receiver, which is proportional to the RF output power of the transmitter. In the Department of Defense (DoD) platforms, where the RF aperture is often limited in size, the only practical approach for improving system performance (e.g., increasing radar or communication system range) therefore is by increasing the RF output power of the transmitter power amplifier (PA). The latter is directly proportional to the output power density of the PA transistor (i.e., transistor output power divided by transistor periphery width).
The operating output power densities achieved in today’s DoD RF transmitters are thermally
limited to values substantially below theoretical electronic limits. Wide bandgap (WBG)
transistors, such as gallium nitride (GaN), were developed specifically to improve output power
density in PAs. Indeed, GaN provides a 5X improvement in RF power output (Pout) compared to
the previous generation transistor technology, gallium arsenide (GaAs). But while it is known that
a further order-of-magnitude increase in Pout is possible in GaN, this cannot be realized in
sustained operation today due to excessive waste heat in the transistor channel layer.
DARPA launched Technologies for Heat Removal in Electronics at the Device Scale (THREADS) program in Nov 2022 with aim to overcome the thermal limits inherent to internal circuitry operations in general, and to critical power-amplifying functions specifically. Today, RF systems operate well below the limits of electronic capacity simply because the transistors, the basic building blocks of RF amplifiers, get too hot. With new materials and approaches to diffusing the heat that degrades performance and mission life, THREADS targets thermal management challenges at the transistor level.
The waste heat is generated because DC-to-RF conversion efficiency in GaN transistors is less than unity (e.g., ~60% at X-band), causes elevated channel temperatures and results in rapid transistor performance and lifetime degradation (device lifetime is cut in half for every 10 oC rise in channel temperature). The maximum channel temperature for safe
operation in GaN is 225 oC.
The Microsystems Technology Office at DARPA seeks innovative proposals to develop
technologies that will overcome the thermal limitations preventing transistors from operating
reliably at RF output power density close to their fundamental electronic limit. Proposed research
should investigate innovative approaches that enable revolutionary advances in science, devices,
or systems.
While GaN transistors have been shown to operate in pulsed mode at high power (Pout = 40 W/mm), operating these devices in PAs under real-world waveforms (long pulse-width, ~30% duty cycle) would result in unacceptably high channel temperature (>450 °C,7 equating to five orders of magnitude reduction in transistor lifetime). Achieving the transistor output power near the GaN fundamental electronic limit while maintaining a channel temperature below the nominal maximum temperature (225 oC) requires a significant reduction in thermal resistances of the transistor (to improve heat removal from the channel) while preserving the superior electronic properties of WBG semiconductors.
Through its investment in prior programs like Dynamic Range-enhanced Electronics and Materials (DREaM), DARPA has successfully increased transistor power density, leading to even greater concerns over thermal dissipation. Programs like Thermal Ground Plane (TGP) developed novel packaging approaches (e.g., phase change heat spreaders) to improve thermal management, but these techniques did not address thermal resistance within the transistor.
On the other hand, the Near Junction Thermal Transport (NJTT) and Intrachip/Interchip Enhanced Cooling (ICECool) programs provided device-level cooling by incorporating high thermal conductivity substrates (e.g., GaN on diamond) as well as microfluidic cooled backplanes, to moderate channel temperature at high power density. But while promising, NJTT enabled just a 3X increase in power density and did not improve the thermal resistance of the GaN transistor epilayers stack or at the interface to the diamond substrate.
In contrast, THREADS will focus on achieving high power density through a reduction in transistor thermal resistance, both within and outside the intrinsic device In this example, the intrinsic device consists of the epilayer stack and individual gate finger, whereas the extrinsic (i.e., outside the intrinsic) device consists of a multi-finger transistor and includes the gate, drain and source fingers, pads, and buses. The extrinsic device includes heat-spreading layers or structures next to the intrinsic device. The combined thermal resistance of the device is modeled as the parallel combination of the intrinsic (within) and the extrinsic (outside) thermal resistances.
Program Description
The THREADS program seeks to develop technologies that overcome transistor thermal
limitations and realize robust high-power density devices that operate near their fundamental
electronic limit of radio frequency (RF) output power. Specifically, the THREADS program will
demonstrate:
High efficiency, X-band (8-12 GHz) transistors and PA test vehicles whose output stage
transistors have an output power density of 81 W/mm;
8X reduction in transistor thermal resistance; and,
Reliable operation with a predicted mean-time-to-failure (MTTF) of 106 hours at 225 oC
channel temperature (comparable to today’s production GaN operated at ~5 W/mm output
power density).
Significant advances have recently been made in WBG and ultra-wide bandgap (UWBG)
semiconductor materials, thermal interface engineering, and advanced three-dimensional heat
spreading. The THREADS program seeks to apply these recent insights to realistic submicron
transistor geometries to reduce transistor thermal resistance and enable operation at high power
density while maintaining a maximum channel temperature of 225 oC. Through a combination of
material thermal resistance improvements, novel transistor topologies and heat spreading
layer(s)/structures, the THREADS program will demonstrate a net 8X reduction of transistor
thermal resistance.
TC 1: Reducing thermal resistance within the device while maintaining good channel current transport properties.
THREADS seeks to reduce interfacial and thin film thermal resistance within the intrinsic device
(epitaxial layer stack). Approaches may include but are not limited to:
Novel nucleation and buffer layer growth processes to reduce defect density at substrate epilayer (e.g., GaN-SiC) interfaces and enable the use of thin buffer layers;
Phonon bridges (e.g., controlled defect incorporation at heterointerfaces;18 nano-structuring
techniques at heterogeneous interfaces;19 ballistic thermal injection;18 strain-enhanced
thermal boundary conductance19);
Phonon engineering through the use of specific isotopes (such as nitrogen-15 vs.
nitrogen-14) during epitaxial growth;
Graded channel GaN HEMTs that uniformly spread the electrons in the channel to reduce
scattering, lower electron temperature, and enhance saturation velocity;
Digital AlN/GaN alloys to increase channel bandgaps while reducing alloy scattering,
channel/buffer thermal resistance and interface scattering and lowering thermal boundary
resistance;
Alternate high thermal conductivity substrates (e.g. diamond, AlN) in combination with
approaches that reduce interfacial thermal resistance;
Alternate high thermal conductivity buffer layers (e.g. AlN); and,
Homoepitaxial growth (e.g. AlN/AlN).
TC 2: Moving heat away from high-power transistors more efficiently without degrading RF performance.
THREADS seeks to develop approaches to spread waste heat and reduce transistor thermal
resistance to maintain channel temperature of 225 °C. Approaches may include but are not limited to:
Topside and/or embedded 2D and 3D cooling structures with high thermal conductivity (e.g.
diamond, AlN, c-BN) and low thermal boundary resistance that do not degrade RF
performance; and,
Novel gate layouts and multi-finger transistor topologies, such as a segmented gate or ring
HEMT, uniform, nonuniform and honeycomb geometries, combined with 3D thermal
conduction geometries and heterogeneous material integration to spread heat efficiently
reducing hot spot peak temperatures.
DARPA awards
Raytheon won a $14.9 million THREADS contract on 29 Sept. 2023, and Northrop Grumman won a $14.2 million THREADS contract on 13 Sept. 2023.
In the THREADS program, Raytheon and Northrop Grumman will focus on achieving high power density by reducing transistor thermal resistance in two ways: reducing thermal resistance within the device while maintaining good channel current transport properties; and moving heat away from high-power transistors more efficiently without degrading RF performance.
US-based Raytheon, a business of aerospace & defense company RTX, has been awarded a four-year, $15m contract from the US Defense Advanced Research Projects Agency (DARPA) to increase the electronic capability of radio frequency sensors with high-power-density gallium nitride (GaN) transistors. The improved transistors will have 16 times higher output power than traditional GaN with no increase in operating temperature.
Significant advancements in transistor design: Raytheon, a contractor for the THREADS program, has successfully developed a gallium nitride (GaN) transistor with a demonstrated power density of 75 W/mm. This represents a significant improvement over the program’s initial target of 50 W/mm.
“Our engineers have unlocked a new way to produce gallium nitride, where thermal management is no longer a limiting factor,” says Colin Whelan, president of Advanced Technology at Raytheon. “These new system architectures will result in sensors with enhanced range.”
Raytheon is partnering with the Naval Research Laboratory, Stanford University and Diamond Foundry to grow diamond, the world’s best thermal conductor, for integration with military-grade GaN transistors and circuits. Cornell University, Michigan State University, the University of Maryland and Penn State University are also providing technology and performance analysis.
Breakthroughs in heat dissipation: Researchers at MIT, another THREADS contractor, have demonstrated a novel thermal management technique using diamond substrates, achieving a 20% reduction in transistor thermal resistance.
BAE Systems and Qorvo Join Efforts to Cool GaN Components for Military Applications
- BAE Systems: Secured a $12.4 million contract in November 2023 to explore microfluidic cooling solutions for GaN electronics. This technology uses tiny channels to circulate fluids, directly removing heat from the source.
- Qorvo: Won a $12.7 million contract in November 2023 to investigate advanced phase-change materials for GaN cooling. These materials undergo reversible phase changes when absorbing or releasing heat, offering efficient thermal management.
BAE Systems is exploring microfluidic cooling, employing tiny channels to directly remove heat. Qorvo is investigating advanced phase-change materials that absorb and release heat efficiently. Effective cooling translates to enhanced performance, smaller systems, and increased reliability for crucial military equipment.
Early prototypes under development: Several contractors are currently developing early prototypes of high-power RF amplifiers based on the new transistor designs and thermal management technologies.
Phase 2 completion: Phase 2 of the program, focused on transistor design and fabrication, is nearing completion.
Phase 3 transition: Transition to Phase 3, focusing on integration and prototyping, is expected to begin in early 2024.
Current funding level: The program has received approximately $60 million in funding for Phase 2.
“Wide bandgap transistors, such as gallium nitride (GaN), were developed specifically to improve output density in power amplifiers – and GaN does provide a greater than 5x improvement compared to previous-generation transistor technology. We also know that a further order-of-magnitude increase in power output is possible in GaN, but it can’t be realized in sustained operation today due to excessive waste heat,” said Thomas Kazior, the DARPA program manager for THREADS. “If we can relax the heat problem, we can crank up the amplifier and increase the range of radar. If the program is successful, we’re looking at increasing the range of radar by a factor of 2x to 3x.”
