Fiber Bragg gratings (FBGs) are frequently used as the reflectors on fiber-laser resonators. These gratings are written in photosensitive fibers that are fused onto the doped-silica fiber that provides the laser gain. Making the fusion joints entails an extra step in fabricating the laser, introduces an additional intracavity loss mechanism and diminishes the laser’s overall robustness. Recently, scientists in Australia wrote an FBG directly into an ytterbium-doped silica fiber, resulting in a fusion-free fiber laser that produces more than 5 W. Figure 1. An ytterbium-doped fiber laser is glowing blue because of a process that is not well understood, but is believed to result from emission from Yb2+ ions instead of Yb3+ ions. External optics — mirrors or bulk gratings — are an alternative way to provide resonator feedback for a fiber laser, and many kilowatts of output have been generated from fiber lasers with external optics. However, mirrors and bulk gratings take up space, require alignment, and, in general, add complexity and expense to the laser. Previous experiments with FBGs written directly into rare-earth-doped silica fibers have produced at most tens of milliwatts. The Australian investigators are associated with the CUDOS (Centre for Ultrahigh-bandwidth Devices for Optical Systems) program at Macquarie University and with the Optical Fibre Technology Centre at Sydney University, both in New South Wales. The researchers wrote the grating in a double-clad, Yb-doped fiber with an 8-μm core and a hexagonal inner cladding whose corner-to-corner spacing was 300 μm. They stripped the outer cladding from the fiber and focused femtosecond pulses from a Ti:sapphire laser into the core with an oil-immersion objective lens. The ∼120-fs pulses contained 220 nJ and were delivered at a 1-kHz rate. The grating that resulted was 15 mm long and had a period of 1.12 μm, corresponding to the third-order diffraction of 1080-nm light. Figure 2. A Ti:sapphire laser wrote the 15-mm-long fiber Bragg grating (FBG)directly into the Yb-doped core of the silica fiber. The grating defined one end of the fiber-laser resonator, and the Fresnel reflection of the output facet defined the other end. Reprinted with permission of Optics Letters. (HR = high reflectivity, HT = high transmission).The grating provided the “back mirror” for the fiber laser, and the Fresnel reflection from the facet at the other end was the output coupler. The researchers end-pumped the fiber with a 980-nm diode and separated residual pump light from the fiber-laser output with a dichroic mirror (Figure 2). With slightly more than 11 W of pump power launched into the 20-m-long fiber, the laser generated a maximum of 5.05 W with a slope efficiency of 46 percent (Figure 3). Figure 3. The output power and lasing bandwidth both increased with pump power. Reprinted with permission of Optics Letters.Because there is no sign of saturation at high pump power, the scientists expect that the laser is readily scalable to higher powers. In a brief experiment, they confirmed that expectation by adding a second pump diode and pumping the fiber from both ends, achieving 10 W of output. But the back mirror of a laser resonator should be a maximum reflector, and the FBG provided less than the ideal 100 percent reflectivity. When the scientists substituted an external 100-percent mirror for the grating, the slope efficiency of the laser increased from 46 to 70 percent. In experiments subsequent to the publication of their paper in Optics Letters, the scientists determined that the grating’s reflectivity was approximately 50 percent. They have improved their fabrication techniques so that their current gratings have reflectivities greater than 90 percent. With these new gratings on the laser, they observe outputs approaching 100 W. The scientists also demonstrated wavelength tuning of the laser as a function of the grating’s temperature and applied strain. They heated the grating as high as 600 °C without damaging it and observed a wavelength shift of 9.4 pm/°C. By applying an external strain, they likewise observed a wavelength shift of 38.7 pm/μm and could tune the output across 5.2 nm without breaking the grating.Optics Letters, June 1, 2007, pp. 1486-1488.