Gallium nitride (GaN) offers significant performance benefits for a wide range of analog microchips, from RF-ICs to numerous power control ICs such as high-power HEMTs used in communications, energy, and military applications. GaN is also the material of choice for high-brightness LEDs used in highly energy-efficient solid-state lighting.
However, the best-quality single-crystal GaN is grown via several epitaxial processes requiring expensive single-use silicon carbide (SiC) substrates, which has limited its commercialization into broader markets, including consumer electronics. A recent discovery by IBM T.J. Watson Research Center scientists may change all of that, in a single crystal GaN film growth process called direct van der Waals epitaxy.
The compelling research finding has targeted the use of recyclable SiC substrates with graphene overlayers for the growth of GaN films based on the process flow depicted in figure 1.
Graphene on silicon-carbide produces a reusable growth substrate for GaN.
These GaN films were subsequently lifted off carefully without a detrimental impact to their roughness or crystallinity and transferred to silicon-based device structures, which could be processed further for LEDs or ICs. Nickel was deposited on top of the GaN layer, and a thermal release tape technique was utilized for the GaN layer transfer to Si substrate stacks. The recycled SiC substrates were used to fabricate functioning blue LEDs, as shown in figure 2.
Many companies and research groups have tried growing GaN directly on silicon with and without the use of transition layers to reduce the lattice mismatch and defect density on silicon – but with limited success. Even though the lattice match between GaN and graphene is about 23%, the IBM group generated remarkably high-quality epitaxial films that were structurally stable enough for the layer transfer to low-cost, conventional silicon substrate material stacks. In addition, Raman spectroscopy indicated no remnant graphene remained on the back-side of the GaN after the transfer process, and electron microscopy supported that data.
GaN has the potential to replace GaAs in all analog IC devices due to its superior electronic properties, but comparatively high cost has held it back. Many analog IC companies already produce a small percentage of GaN-based RF ICs primarily used in military electronics, while GaAs has been favored for more price-sensitive mobile devices such as smartphones.
Most of these companies purchase SiC substrates from Cree, which manufacturers the same chips. So companies such as TriQuint Semiconductor and Texas Instruments are seeking to reduce the conflict of interest in their supply chain reliance on Cree, which is also a top LED producer. Reusable SiC substrates using GaN on graphene may be the answer if this research fully materializes into commercialization over the next five years.
IBM has invested millions into graphene research over the last decade and is considered the leader in this field. It is no surprise that one of its research groups made this discovery, recently published September 11 in Nature Communications by Jeehwan Kim et al. (volume 5, page 4,836).
IBM is also interested in developing graphene substrates to replace Si altogether, due to graphene’s superior electronic properties over Si. What’s more, the company is planning to invest $3 billion over the next five years in this technology, which will likely lead to more breakthroughs to boost the prospects of graphene and GaN commercialization in advanced electronic devices.
Brian Coppa has written about electronic materials and devices and worked for Micron and ASM America and now consults on topics including alternative energy and microelectronic applications.
This article first appeared on EE Times’ Planet Analog website.
Related links and articles: