Fujitsu Laser is Highlight of LEOS Rump Session

A small group of researchers at Fujitsu's Research and Develpment Center in Atsugi, Japan, has surprised the GaN community by achieving an InGaN/GaN diode laser to compete with international leaders such as Nichia and Cree. The group, headed by A. Kuramata, consists of only 5 hard-working researchers. One member of this group, K. Domen, showed the data demonstrating laser operation at the rump session of the LEOS Summer Topical Workshop held 14 August 1997 in Montreal, Canada.

The Fujitsu laser was grown on Shin Nippon Steel SiC substrates by low-pressure metallorganic vapor-phase epitaxy (LP-MOVPE). This is the first time that a low-pressure process has been reported to produce a nitride-based laser. The Fujitsu success came only one month after CREE reported the first nitride laser fabricated using SiC substrates. The Fujitsu laser operated in pulsed mode, with a 300ns pulse width and a 0.03% duty cycle. The threshold current was 800 mA, and the output power was 20 mW. The lasing wavelength was 414 nm. The MQW laser structure had a 3.5 µ stripe width, and used a backside contact (through the substrate) and the far-field emission pattern was that characteristic of a high order transverse mode. The laser was run for more than 5 hours under pulsed conditions.

"It was very beautiful" said Dr. Domen of the Fujitsu laser. "It made me feel very cool in the hot Japanese summer".

The SiC versus sapphire choice figured prominently in the discussion which followed. "I have used SiC long ago" said Nichia's Shuji Nakamura. "The threshold current is very large. I would use SiC if the price was the same as sapphire." "My group would have to spend $1000/day to use SiC" said Prof. S. DenBaars of UC Santa Barbara. In response to a question of whether the noise performance and temperature stability of GaN lasers would be good enough for applications, Dr. Domen suggested that the stability of lasers should be good on SiC. Dr. Nakamura responded that temperature stability of lasers was fine on sapphire, and that he had measured the noise; it was not a problem.

The other issue that generated controversy was whether the mechanism responsible for lasing in nitride quantum wells was related to quantum dots, composition fluctuations, and the like. "How could a dot-laser have such a high gain?" posed one panel member. Y. S. Park of ONR cited this as a matter of central importance, adding that he had 4 proposals of this type already on his desk. "Make that 5!" shouted one conference participant. "He sent his by UPS!" cracked another.

Several questioners tried to coax secrets out of the panel, for example by asking what the most important aspect of their reactors was. The panelists politely demurred.

The pace of laser commercialization is likely to accelerate towards the end of the year. Dr. Nakamura promised to present a new laser structure with much better lifetime at the International Conference on Nitride Semiconductors in Tokushima, Japan this October. The current Nichia structure has run as long as 300 hours CW. Dr. Nakamura was asked what the cause of failure is. His answer was that "after seeing our new structure, you can see what causes our short lifetime".

To give a historical perspective to the evening, P. Zory recounted the rump session held at the same meeting 4 years ago. A vote was taken at that meeting to see which material system the participants thought would prevail in the marketplace for visible light emitters, nitrides or II-VI's. They were prescient in voting for nitrides for wavelengths shorter than 450 nm, because they had just heard of the new results from Japan. But they were confident that II-VI's would prevail for wavelengths longer than 450 nm. The vote was not repeated at this meeting.

Eric Hellman