Many “quasi three level” neodymium lasers, in which the lasing transition terminates on an upper sublevel of the ground state, have been operated in the past 40 years. But researchers at Centre Scientifique d’Orsay in France have demonstrated what they believe is the first true three-level neodymium laser and have frequency-doubled it to generate deep-blue light at 440 nm. There are multiple applications for such rugged and efficient deep-blue lasers in fields such as displays, medicine and spectroscopy. A quasi-three-level laser terminates on a level that has significant thermal population, making it more difficult to obtain a population inversion (Figure 1). A true three-level laser terminates on the lowest sublevel of the ground level and generates the shortest possible wavelength for a given transition. Realizing a true three-level laser is difficult because a significantly greater thermal population must be overcome to create a population inversion.Figure 1. This energy-level diagram for Nd:GdVO4 shows the emission from quasi-three-level lasers at 912 and 893 nm, and the emission from a true three-level laser at 879 nm. The normal ~1.06-μm transition in neodymium terminates on the 4 /11/2 level, which is not shown here.(Nonetheless, the first laser, Theodore H. Maiman’s Cr:ruby, was a true three-level. Legend has it that one of the originators of the theoretical explanation of lasers ridiculed Maiman’s initial attempts because he felt it would be impossibly difficult to create a population inversion in a three-level system.)The researchers selected a neodymium-vanadate laser crystal because they calculated that the transparency intensity — that is, the 808-nm pump intensity required to equalize the populations of the upper and lower laser levels — was significantly less for the vanadates than for other neodymium hosts. They experimented with both Nd:GdVO4 and Nd:YVO4 in a folded resonator (Figure 2).Figure 2. An intracavity etalon defeated lasing on the quasi-three-level transitions and forced the laser to oscillate on the true three-level transition at 879 nm. For the frequency-doubling experiments, the researchers removed mirror M3 and added the components inside the dotted line.Initial experiments with Nd:GdVO4 yielded 800 mW of 879-nm power from ~16 W of 808-nm pump power and a threshold of ~10.5 W (Figure 3). With Nd:YVO4, however, the 1.06-μm transition lased at low pump powers, despite the high transmission of all the resonator mirrors at that wavelength. When the pump power reached 15.6 W, the 879-nm transition finally reached threshold, extinguished the 1.06-μm lasing and generated up to 330 mW. The researchers concluded that parasitic lasing at 1.06 μm — which did not occur in Nd:GdVO4 because of the 1.06-μm transition’s lower emission cross section in that material — was responsible for the high threshold in Nd:YVO4. They conducted the rest of their investigation with Nd:GdVO4.Figure 3. Parasitic lasing at 1.06 μm apparently reduced the performance of the Nd:YVO4 laser at 879 nm.To frequency-double the laser, they first considered straightforward intracavity doubling of the continuous wave laser but decided that that was impractical because the maximum intracavity fundamental power was only ~5 W. So they Q-switched the laser with an acousto-optic modulator to obtain a high peak power for efficient doubling (Figure 2). They used a 10-mm-long BiBO crystal oriented inside the resonator for type I phase matching, and generated up to 178 mW of second-harmonic average power at 439.5 nm, which was divided between two beams exiting the resonator.The researchers believe that several straightforward steps could significantly enhance their results. The intracavity etalon introduced losses that could be avoided by using highly selective mirrors to quash laser oscillation on the competing transitions. And better nonlinear crystals, such as KNbO3, would result in more efficient second-harmonic conversion. Moreover, the standard technique of reflecting the second harmonic for another pass through the crystal would not only combine the two output beams into a single beam, but also increase the second-harmonic conversion efficiency.Optics Letters, Sept. 15, 2006, pp. 2731-2733.