At the University of California, Los Angeles, scientists have reported the demonstration of what they believe to be the first silicon laser, a feat that exploits the Raman effect. Ozdal Boyraz and Bahram Jalali have described pulsed Raman laser emission at 1675 nm with a 25-MHz repetition rate, using a silicon waveguide as the gain medium. The laser has a clear threshold at 9 W of peak pump pulse power and a slope efficiency of 8.5 percent.Figure 1. The silicon laser exploits the Raman effect. The key components are the silicon-on-insulator rib waveguide and the fiber that forms a loop around the chip.The approach is fundamentally different from those being investigated by other groups in pursuit of the silicon laser (see "Seeking a Silicon Laser," Photonics Spectra, February 2003, page 62). Jalali, a professor of electrical engineering at the university's Henry Samueli School of Engineering and Applied Science, explained that it uses the natural atomic vibrations of silicon to create or amplify light. In contrast, the other methods rely on electronic transitions, which demand a material with a direct electronic band structure."Silicon does not have this, and so many groups have tried adding impurities or developing complicated device structures," he said. "With our approach, we do not need this."Figure 2. The next-generation silicon Raman laser, under development, integrates a microdisk resonator and an optical waveguide. Laser action occurs when the pump signal interacts with the atomic vibrations of silicon as it circulates around the disk. The laser light is produced at the desired wavelength (different from the pump wavelength) and exits the device through the other end of the waveguide.The Raman effect has been exploited in fiber lasers for many years, but several kilometers of fiber are required to make a useful device because glass fiber has an amorphous structure with a random atomic arrangement, weakening the Raman effect. Crystalline silicon, however, has a well-ordered atomic arrangement, so the Raman effect is up to a million times stronger in a confined silicon waveguide, making it ideal for creating a laser only millimeters in size. Jalali is the first to admit that the new laser has its limitations. It is an optically pumped, pulsed device, whereas the scientific community is striving toward an electrically pumped, CW emitter. Pulsed operation currently is necessary to avoid the accumulation of free carriers that are generated by two-photon absorption. He nevertheless believes that CW operation will be possible by using a diode to remove the free carriers that compete with the Raman effect.Figure 3. An electron micrograph of a prototype of the next-generation silicon laser reveals the microdisk overlapping the optical waveguide.For fundamental reasons, electrical pumping seems not to be achievable, however. "This experiment is very elegant and a nice piece of science," said Kevin Homewood of the University of Surrey in Guildford, UK, whose group also is trying to develop a silicon laser. "But this is not the device people need for the key applications in microelectronics or silicon photonics."The Raman effect, Homewood explained, is a purely optical-to-optical transition. "The original photon is scattered by a phonon producing another photon emerging, in this case, ~60 meV lower in energy," he said. "This approach, therefore, gives you a device that basically down-converts a pump laser by this amount of energy. There is certainly no route from here, using this approach, to developing an electrically pumped device."Nevertheless, the silicon Raman laser can be switched or the free-carrier concentrations can be electrically modulated using integrated diodes, and Jalali believes that the device has many potential applications. He said it is not intended to compete with laser diodes but rather to do what they cannot."Potential applications include a tunable laser source for biochemical detection, free-space optical communication and defense applications," he said. "A key attribute of the new technology is that it can produce mid-infrared radiation without any cooling. This is a drastic improvement over current technology, where antimonide-based material plus cryogenic conditions are required to achieve lasing."