Lasers in the 2- to 10-µm region are important for a number of spectroscopic, remote sensing and military applications. Raman fiber lasers offer one approach to such lasers, but the relatively low Raman gain of conventional silica fibers is an impediment. Recently, scientists from SFA Inc. in Largo, Md., and the Naval Research Laboratory in Washington have shown that the Raman gain of a small-core AsSe fiber has a Raman coefficient 340 times as great as that of silica. Such a high Raman gain opens the possibility that the fiber could be instrumental in the development of robust mid-IR lasers.The fibers investigated by the group were produced by the double-crucible technique, rather than drawn from a preform. The core diameter was 7 µm, and the numerical aperture was 0.45 at 1.56 µm. Because of this large aperture, the fiber was not single-mode. The fiber loss was measured at 0.7 dB/m at 1.56 µm, but process improvements are expected to improve this number, according to Peter A. Thielen, one of the group members. Unlike silica fiber, the AsSe fiber has a similarly low loss over most of the region between 2 and 8 µm. Because Raman gain scales inversely with wavelength, the fiber could be used for a Raman laser over most of this spectral region.Figure 1. The Raman gain coefficient of the AsSe fiber was measured by pumping it with 1502-nm pulses from the optical parametric oscillator. Seed photons at 1560 nm were supplied by the CW laser diode.In the experimental arrangement for measuring the Raman coefficient, an optical parametric oscillator launched 8-ns, 1502-nm, 10.8-W peak-power pump pulses into the fiber (Figure 1). A CW laser diode provided seed photons at 1560 nm. The Raman shift of approximately 58 nm in chalcogenide glass is less than the approximately 100-nm shift in silica because the heavier atoms in chalcogenide glass have lower vibrational frequencies. A monochromator with an effective resolution of 3 nm separated the pump from the amplified signal, and an InGaAs detector measured the pump power at the end of the fiber.Figure 2. Experimental results showed that the Raman gain coefficient of the fiber was 340 times that of a conventional silica-glass fiber.In the experimental results (Figure 2), the trace on the left represents the pump pulse from the optical parametric oscillator, and the three traces on the right show the signal for the cases when only the seed photons from the CW diode are present, only the pump photons are present, and both pump and seed are present. The seed-only trace is amplified by a factor of 10 to make it visible on this scale. The peak at 1560 nm in the pump-only case is the result of spontaneous Raman scattering of pump photons. The intensity of the 1560-nm peak when both seed and pump photons are present indicates a Raman gain coefficient of 2.3 X 10-11 m/W, a value 340 times greater than the coefficient of silica.Although the scientists measured the Raman gain with a pulsed pump laser, they recognize that a CW Raman laser is preferable in most applications. They say they would expect no problems substituting a CW source for the pulsed parametric oscillator in their experiment and note that they have pumped as much as 0.5 W of mid-IR radiation into the fiber without incident.