Technical Insight
Output power rockets for green II-VI lasers
A novel material system produces powerful, continuous-wave laser emission at 536 nm, an ideal wavelength for the green source in pico-projectors
A Japanese team has smashed its own record for the continuous-wave (CW) output power of green lasers built from a beryllium-based family of II-VI alloys.
The latest green laser produced by researchers from Hitachi and the National Institute of Advanced Industrial Science and Technology (AIST) combines a new device structure with improved material quality to deliver a 70 mW output. This is a tremendous hike in power compared to team’s 1.5 mW, 545 nm laser fabricated in conjunction with Sony and reported in summer 2010.
Green lasers made by Hitachi and AIST can produce up to 70 mW at 536 nm.
Beryllium-based lasers are an attractive alternative to those based on InGaN, which are targeting the market for the green source for pico-projectors, but emit at shorter wavelengths that would compromise the colour range produced by a red-green-blue laser display, according to AIST’s Ryouichi Akimoto.
He explains that when engineers try to push their nitride lasers to 530 nm and beyond, threshold current density rockets, driving down efficiency.
This, plus difficulties in growing high-quality InGaN with the high indium content needed to push emission to longer wavelengths, have pegged back the wavelength of the first commercial green nitride lasers, which are made by Nichia and have an emission between 510-520 nm.
Nichia’s lasers are made on c-plane, polar GaN substrates. Switching to a semi-polar platform, an approach pioneered by Sumitomo, can extend the reach of GaN lasers to just beyond 530 nm. However, this adds to manufacturing costs.
“I have no exact number for the price of semi-polar GaN substrates, but I presume that they are one order of magnitude higher than those of GaAs substrates, [which we used for making our beryllium-based lasers],” says Akimoto. “This will make a significant difference in the fabrication costs of II-VI and InGaN lasers.”
Back in the early 1990s, prior to the first big successes of GaN, II-VI materials were the most promising devices for delivering blue and green laser emission. However, lifetime was restricted to 500 hours or so, due to the ease of propagation of defects in these alloys, which have a high degree of ionic bonding.
Lasers produced by Akimoto and his co-workers are markedly different from the ZnSe-based devices of the 1990s – they also contain beryllium, which increases ‘covalency’ and promises to lead to longer laser lifetimes.
This lifetime boost is yet to be quantified. “According to Sony's review paper on blue-green lasers reported in 2000, the crystal defect density in the epitaxial layers of a laser is the critical factor that determines the lifetime of II-VI lasers,” says Akimoto. “Based on that report, we presume that the crystal defect density in our lasers is still high, hence the lifetime is short.”
The team’s 536 nm, gain-guided laser was fabricated by wet etching the p-type contact to form a mesa structure. After Ti/Pt/Au and AuGe/Ni/Au electrodes were added by evaporation, cleaving formed 800 µm-long cavities featuring 7 nm-thick BeZnCdSe quantum wells.
Laser chips with a 2 µm-wide mesa and front and rear facet reflectivities of 90 percent produced a threshold current and voltage of 55 mA and 9.8 V under CW operation at 25 °C. Driven at 118 mA, light output power hit 50 mW, rising to 70 mW when front-fact reflectivity was reduced to 70 percent.
Akimoto believes that it will be possible to reduce the operating voltage of these lasers. “We presume that p-contact and p-cladding layers are not optimized yet, hence electrical resistivity is high. We expect that the threshold voltage will go down to about 5 V after optimization.”
To make their lasers suitable for the pico-projector market, the team are trying to build 50 mW devices with 2000-hour lifetimes, a 5 V threshold voltage and a threshold current density of 1 kA cm-2.
“The improvement of the lifetime is the most important challenge,” admits Akimoto. “We expect that the keys to achieving this goal are a decrease in crystal defect density in p-layers and a full use of the ‘berylium effect’, which will strengthen chemical bonds in a BeZnCdSe active layer.”
S. Fujisaki et al. Appl. Phys. Express 5 062101 (2012)
