Panasonic eyes non-polar GaN substrates
Researchers from Panasonic s Matsushita Electric Works (MEW) have successfully grown non-polar m-plane GaN on low-cost sapphire substrates for the first time.
They subsequently used this MOCVD-grown material as a seed for sodium-flux liquid-phase epitaxy (LPE) to produce higher-quality crystals, saying that the method is superior to current techniques for using non-polar material.
The sapphire substrate work was published in Applied Physics Letters at the beginning of March, but had already contributed to LPE research published at the same time last year. That paper described making m-plane GaN crystals with less than 1x107 dislocations per cm2.
According to the MEW team, growing boules by LPE should result in more economical m-plane substrates than current approaches, which cut vertical slices from a c-axis oriented crystal grown by hydride vapor-phase epitaxy (HVPE).
“With that method you need to grow a GaN crystal at least two inches thick in order to get a two inch m-plane wafer,” said Robert Armitage, one of the researchers involved in the work.
“It is difficult and time consuming to grow such a thick GaN crystal, which is why the m-plane substrates sold by Mitsubishi and Kyma have sizes only in the range of a few tens of millimeters.”
Armitage says that making 2-inch m-plane GaN wafers by LPE would be more straightforward and faster than HVPE methods.
Seeding where others have failed
Teaming with Japanese research institute RIKEN, Armitage first grew an AlN nucleation layer on an m-sapphire substrate. 2-3 micron thick m-plane GaN crystals were then deposited on the nucleation layer by MOCVD.
Previous attempts at growth on sapphire had failed and Armitage narrowed the cause down to substrate annealing prior to depositing the nucleation layer. The MEW method avoids annealing in ammonia at moderate temperatures, or in air at high temperatures.
“Such layers might be applied in devices which are relatively more tolerant of defects, like transistors, optical detectors and modulators,” Armitage said. “I am not optimistic that the quality of the thin layers can be improved enough for efficient LEDs.”
In the 2007 Japanese Journal of Applied Physics paper Panasonic researchers also teamed with the group of Takatomo Sasaki at the University of Osaka for the LPE growth.
Panasonic extended its existing expertise in sodium flux epitaxy (see related stories) to grow bulk m-plane GaN crystals. In this case researchers from the company and its collaborators heated gallium and sodium metal to 850°C in a furnace pressurized to 50 atm with nitrogen.
Varying the sodium:gallium ratio changed the morphology of the growing GaN and the researchers found that m-plane crystals appeared when this ratio stood at around 80:20.
The paper describes crystals of non-polar GaN that measure 10 mm along each horizontal direction and 0.8 mm in depth, using the MOCVD grown material as a seed.
LEDs made from non-polar, m-plane, GaN have already delivered very high quantum efficiencies and could address wavelengths inaccessible to conventional c-plane GaN.
The emergence of Panasonic as a player in the non-polar GaN arena is another surprise from the conglomerate after it announced its late entry to the LED market last year. Despite this, Armitage claims to have been researching m-plane GaN since February 2006 and says that the results in the sapphire substrate paper also date from 2006.
