Photonic crystal fiber enables designers to avoid several of the problems inherent with conventional fiber. In particular, the chief limitation on the power achievable with fiber lasers involves nonlinear effects whose onset scales with the product of the fiber length and the power density in the fiber core. This is especially severe in pulsed lasers, where the peak power in the fiber can be much higher than the average power in a continuous-wave laser. Because photonic crystal fiber lasers can support larger modes — and, hence, reduced power density in the core — they offer a way to avoid these nonlinearities. Recently, investigators at Friedrich Schiller Universität in Jena, Germany, and at Femlight in Talence, France, demonstrated a laser whose unusually short length and low intrafiber power density make it an excellent candidate for the generation of high-peak-power pulses. Already, the scientists have generated 320 W of continuous-wave, single-mode power from a 0.5-m length of fiber. They believe that the resulting ratio, 550 W/m, is the highest power per unit length reported for any type of fiber laser.The experimental laser comprised a 58-cm-long, ytterbium-doped “rod type” fiber. The “rod type” designation results from the fiber’s 1.5-mm outer diameter, which keeps it stiff enough to avoid the bending losses unavoidable in large-mode-area fibers. Whereas many photonic crystal fibers are fabricated with a single missing hole for the core, in this case there were 19 missing holes, resulting in a fundamental-mode diameter of ~50 μm, or a mode-field area of 2000 μm2 (Figure 1). The precise control of the cladding’s refractive index, enabled by the precision with which the airhole geometry was fabricated, allowed for the large mode-field area.Figure 1. An optical microscope image of the fiber shows the air cladding surrounding the pump-guiding inner cladding (a). A scanning electron microscope image of the core and inner cladding shows the microstructure of the inner cladding and the core (b). The core is formed by the absence of 19 airholes from the structure. Images ©OSA. The fiber’s core measured 60 μm in diameter and was composed of a mixture of ytterbium- and fluorine-doped glass such that its index was slightly less than that of the silica in the surrounding inner cladding. (The effective index of the inner cladding, due to the airholes, was less than that of the core, so the fundamental mode was guided in the core by total internal reflection.) The 175-um-diameter inner cladding was bounded by the thin silica bridges that formed the outer cladding. The bridges were ~400 nm wide and ~10 μm long, resulting in a high numerical aperture (∼0.6 at the 975-nm pump wavelength) for the inner cladding. The investigators calculated the mode pattern expected from the fiber using a finite-difference algorithm to solve the scalar Helmholtz equation. They found the result in good agreement with the measured pattern (Figure 2). Figure 2. The measured mode pattern of the laser (a) closely resembles the calculated pattern (b). Because the onset of deleterious nonlinear effects is proportional to the product of a fiber laser’s length and the inverse of its mode-field diameter, it is instructive to compare those two numbers for the experimental fiber and a conventional large-mode-index fiber. The latter typically is 5 m long with a mode-field diameter of 25 μm, while the former was 0.5 m long with a 50-μm mode-field diameter. Quick arithmetic yields the result that a conventional large-mode-area fiber is 40 times more susceptible to nonlinear effects than the experimental photonic crystal fiber. A similar calculation comparing the experimental fiber with a standard step-index fiber shows that the latter is 2000 times more susceptible to nonlinearities than the photonic crystal fiber. In their initial experiments, the researchers generated as much as 320 W of output from a launched pump power of 425 W. The 78 percent slope efficiency is significantly higher than that of conventional large-mode-area fibers because this laser avoids the bending losses that are necessary to force conventional lasers to guide only a single mode.Optics Express, April 3, 2006, pp. 2715-2720.