Pure gallium nitride wafers likely to change blue-violet laser market
The short-wavelength blue-violet laser diode, with its potential for increased data density, higher operating speed and smaller system size, is expected to become the main light source in such applications as data storage, communications, measurement and medical inspection.
Sony Corp. will unveil the first blue-violet laser-based product, the Blu-Ray Disc recorder, on April 10. With a suggested retail price tag of about $3,800, however, the product will be out of the reach of most buyers in the consumer market. But the potential demand for short-wave lasers is expected to grow significantly.
Although shortwave laser diodes were originally dubbed "blue" lasers, they are really blue-violet ones that emit light at 405 nanometers. Gallium nitride (GaN) is used for the lasers since they oscillate most efficiently around this wavelength. (Blue is defined as light in the range of 455 to 485 nm, according to a Japanese industrial standard).
In the still-small blue-violet laser market, Nichia Corp. has been the sole supplier. But the company's strength may be undermined by the emergence of GaN wafers. Once GaN wafers become available in volume, most of the technical obstacles currently hindering the production of blue-violet lasers in volume will be solved, said industry sources.
GaN wafers may solve patent issues as well. Nichia is fabricating lasers by growing a GaN layer epitaxially on a sapphire substrate and owns patents covering a wide range of related technologies. GaN wafers have the potential to help other laser developers avoid infringing on Nichia's patents, since those companies may be able to design GaN lasers in different structures.
GaN wafers
After researchers succeeded in getting GaN to emit light in the 1980s, GaN devices have been grown on a sapphire substrate. But sapphire has not been an ideal substrate for GaN device fabrication because of the big gap — over 16 percent — between the lattice constant for GaN and sapphire.
That difference in the lattice constants causes a lot of dislocations, or defects, in the GaN crystal layer that is grown. Those defects have a negative impact on the laser's quality and make it difficult to develop lasers that last long. The difference also makes it difficult to cleave the crystal and obtain a laser reflective surface.
Sapphire substrates have other problems, like low thermal conductivity, which may cause the laser diode grown on the substrate to heat up. And because sapphire is electrically nonconductive, the laser diodes grown on the sapphire substrate have to have both electrodes on the surface of the GaN layer. That constraint makes the size of the chip larger than a device that can have an electrode on the bottom as well as on the top.
GaN devices that are expitaxially grown on GaN substrates, on the other hand, do not suffer from those problems. But the GaN crystal has only been available in millimeter-sized pieces, which have been produced in high-temperature, high-pressure conditions. As GaN does not exist in a liquid state, the pulling method that is widely used to produce single-crystal ingots, which are then sliced into wafers, cannot be applied to separate single-crystal GaN. Thus, sapphire and silicon carbide are mainly used as substrates for GaN device fabrication.
Wafer developers have been trying to find a solution to the problems associated with the epitaxial growth process. Three companies, Sumitomo Electric Industries Ltd. (SEI) (Osaka, Japan), Hitachi Cable Ltd. (Tokyo) and Crystal Photonics Inc. (Sanford, Fla.) have developed freestanding GaN wafers. Sumitomo and CPI have already started offering samples to potential laser suppliers and Hitachi Cable will start sampling soon.
These wafer developers grew a GaN crystal layer on a substrate — usually sapphire — that was several hundred microns thick, and removed the substrate from the GaN layer, leaving the layer as a freestanding wafer. But this process causes the same problem — many dislocations in the crystal — as GaN devices grown on a sapphire substrate have. Each company had a different approach to lowering the number of dislocations in the crystals.
Sumitomo Electric announced in February 2000 that it had developed the world's first freestanding single-crystal GaN wafer, which measured 2 inches in diameter. It grew a GaN layer on a sapphire substrate and reduced the number of dislocations using the company's proprietary Deep (Dislocation elimination by epitaxial growth with inverse-pyramidal pits) method.
In the Deep method, many microscopic pits are generated in the GaN crystal layer while the crystal layer grows. Each side of a pit is matched with a facet of the GaN crystal. Along with the epitaxial growth of the crystal layer, dislocations move and gather to the center of the pits. The pits, which are several 100 microns in diameter, remain on the surface of the crystal layer, leaving the rest of the crystal with fewer dislocations.
Sumitomo Electric engineers said that in the first generation, it was impossible to control the position of the pits. In the second generation, announced last June, they added a control technology to arrange the pits in a more orderly way. As the position of the pits is controlled, the areas that have fewer dislocations appear in a more orderly fashion. Users can grow laser diodes efficiently on those areas, each of which is about 500 microns in diameter.
The dislocation density per square centimeter of these usable areas is between 10,000 and 100,000, which is about 1/100,000 of the density of conventional GaN epitaxial layers on sapphire substrates, according to the company.
SEI started shipping samples with a thickness of 0.25 to 0.6 mm last June and is presently preparing volume production. The company plans to begin operation sometime in April with a capacity of 300 wafers a month.
"Multiple potential users are evaluating our wafers and are giving us feedback about the result of the evaluation. Lower dislocation is mandatory because it has a direct effect on the quality of lasers, on which we have confidence. The next several months will be the most important period for us because major users will select which wafer to use for their laser production," said a spokesman of Sumitomo Electric.
Hitachi Cable, a Hitachi group company, was the second company in Japan to announce the development of a freestanding 2-inch GaN wafer. The company announced the accomplishment in February, and will begin sampling this spring.
Hitachi Cable expitaxially grows a single GaN crystal wafer on a sapphire substrate and the sapphire substrate is removed after the crystal layer has been fully grown. Hitachi Cable said that it becomes difficult to remove the sapphire base completely without damaging the GaN wafer when the wafer diameter becomes as large as 2 inches. Hitachi Cable's solution is to use an approach they call the Void Assistance Separation (VAS) method.
In the VAS method, a nitride titanium film is inserted between the sapphire substrate and the GaN growth layer. The film is actually meshwork consisting of 20- to 30-micron stripes of nitride titanium. During epitaxial growth, large numbers of microscopic voids are formed in the openings of the nitride titanium meshwork. After the GaN crystal layer has been fully grown, it can be removed from the substrate without being damaged at the void layer. This method will allow the company to make wafers larger than 2 inches in diameter in the future, according to a Hitachi Cable spokesman.
Competitors Sanyo Electric Co. Ltd. and Sharp Corp. also have aggressive plans for blue-violet lasers in place.
Sanyo characterizes its blue-violet laser business as one of the new mainstays of the company, along with several other areas such as organic light-emitting diodes and a charge-coupled device module for mobile gears.
Sanyo developed a blue-violet laser with 35-mW output early this year using GaN substrates. The laser has the highest output power thus far announced. Sanyo is the first company to claim the use of a GaN substrate to fabricate a laser diode, though the company has not disclosed the name of the supplier. As the diode does not have an insulating sapphire substrate, it can have one electrode on the top and anther on the bottom, which reduces the chip size to about half that of other devices, according to Sanyo.
Sanyo plans to start operating a blue-violet laser production line at a Tottori Sanyo fab sometime this spring and will ramp volume production this autumn. The first sample will be available in May for about $1,680 each.
Sanyo said it plans to take its aggressively named Blueimpulse blue-violet laser family to about $127 million in sales, including blue LEDs, by 2006.
Sanyo's laser diode does not infringe on Nichia's patents because its structure is different from Nichia's sapphire-based lasers, said Yukinori Kuwano, president of Sanyo.
Sharp Corp., another major player in the red-laser market, is also positioning a blue-violet laser as the next key device but has been keeping silent about its laser diode, whose development it apparently completed at the end of last year. Last May, Sharp opened a new fab in Mihara, Hiroshima Prefecture, for compound semiconductor production. The fab will be used as the production base of blue-violet lasers and will begin volume production by the end of this year.
Toyoda Gosei Co. Ltd., which is fiercely competing with Nichia in the blue-LED market, also completed development of a 410-nm laser in April 2001, as the result of a research project funded by Japan Science and Technology Corp.. The company has continued the development of shortwave laser diodes in cooperation with an undisclosed system manufacturer. Last autumn Toyoda Gosei developed a 405-nm laser with an output power of 30 mW using a sapphire substrate, and it continues to evaluate the device.
Rohm Ltd. is working with Cree Inc. for silicon carbide-based lasers and with Pioneer Corp. for sapphire-based lasers.
With practical applications in sight, blue-violet lasers have already entered into a volume-production phase, and several companies will come out with their own products this year.
Marcos Pinto D`derlee
C.I 17862728
EES
Explore the seven wonders of the world Learn more!
No hay comentarios:
Publicar un comentario