Diode emits blue-purple pulses at room temperature

Diode emits blue-purple pulses at room temperature

Paul Mortensen

Researchers at Toshiba Materials and Devices Laboratories (Kawasaki, Japan) have developed an indium gallium nitride-based blue-purple-emitting (417 nm) diode laser that emits a pulsed beam at room temperature; they are reportedly working "around the clock" to achieve continuous wave (CW) operation. None theless, Toshiba researchers are lagging behind those at Nichia Chemical Industries Ltd. who have achieved at least 35 hours of CW operation from a gallium nitride (GaN) blue-emitting device (see Laser Focus World, April 1996, p. 18). Nichia is expected to be the first company in Japan to develo¥a practical blue-light-emitting semiconductor laser, perhaps within the next two years.

Some Japanese electronics companies, however, are assuming that Nichia`s primary application for its expertise will be development of the huge market for blue-light-emitting diodes (LEDs) rather than the smaller market for semiconductor lasers for mass storage. Toshiba, though, is targeting its blue-emitting diode laser at future digital versatile disk (DVD) technologies. High-definition motion pictures, for example, will require 420-nm laser diodes to read dual-sided DVDs with a capacity of 15 Gbyte per side--current DVD technology uses a red-emitting laser diode (650 nm) for a 4.7-Gbyte disk.

It is about two years since Toshiba shifted from working with the II-VI compound zinc selenide (ZnSe), which offered the possibility of a low-dislocation material, to the III-V compound GaN. The company be lieves that III-V compound material is more suitable for producing a stable, short-wavelength laser beam than other materials. Indeed, Sony, Matsushita, and NTT may share that opinion and are shifting at least part of their re search focus away from ZnSe. Another advantage, notes Toshiba, is that metal-organic chemical vapor deposition (MOCVD), which is used for the deposition of GaN, has none of the disadvantages of molecular beam epitaxy (MBE) used with ZnSe, such as low mass-production rate and the need for a high vacuum.

Toshiba researchers credit two developments for progress with their blue-purple-emitting laser, which operates at a threshold current of 5 A and voltage of 20 V. First was success in fabricating very thin layers of GaN crystals on a multiple-quantum well (MQW) structure using MOCVD. They were able to grow a GaN buffer layer on a sapphire C-face substrate followed by a 2-µm-thick undoped GaN layer, a 4-µm Si-GaN contact layer, a 0.1-µm undoped GaN optical guiding layer, and 25 periods of a InGaN/InGaN MQW structure consisting of 20-Å-thick InGaN well layers and 40-Å-thick InGaN barrier layers (see figure). These layers were grown continuously at temperatures ranging from 550°C to 1100°C with MOCVD.

The second breakthrough came in a new technology for cleaving the GaN crystal grown on the C-face sapphire substrate and assuring a smooth surface. Cleaving GaN crystals on a sapphire substrate usually produces a rough surface unsuitable for laser emission, so dry etching has been used instead. Cleaving is less expensive than dry etching, however, and requires fewer manufacturing processes. The Toshiba researchers predict that the new cleavage technology will be more practical when blue-purple diode lasers are commercialized.

The researchers are under pressure to deliver CW operation, which they hope will be achieved early in 1997. Meanwhile, Nichia`s 35 hours of CW blue-purple emission was achieved at 80°C. For optical disk applications, though, Nichia researchers will need to stabilize at least 10,000 hours of operation and, for telecommunications, 1 million hours. Fortunately for Nichia, since the development cycle typically follows a logarithmic scale, achieving 10,000 hours from a start of 35 hours is much less of a struggle than achieving the first few hours of CW operation--the problem now faced by Toshiba researchers.

Once the diode becomes a practical reality other problems will have to be solved. The most serious is that the sensitivity of current optical detectors (Si-PIN) drops precipitously in the region of the operating wavelength (420 nm). Other issues include such conventional design problems as reducing noise due to reflection of light back into the laser from the disk surface, reducing the beam-aspect ratio, and reducing the price. Here, the use of sapphire substrate instead of GaAs will not result in a major price increase--Toshiba notes that for a $500 DVD player, the optical head comprises 5%-10% of the total system cost and the laser diode, just 1%-2%.