Intracavity frequency doubling of diode-pumped solid-state lasers has been for many years the favored approach to creating solid-state lasers with visible outputs. Dozens of commercial products are on the market, with applications such as medical diagnosis and treatment, underwater surveillance, full-color solid-state displays and the ubiquitous green laser pointers. But an intracavity-doubled laser can be subject to the so-called “green problem,” an undesirable and chaotic fluctuation in output power that results from sum-frequency generation among the laser’s longitudinal modes. Recently, scientists at Universität Hamburg and at Coherent Lambda Physik GmbH in Lübeck, both in Germany, developed an up-conversion laser that avoids the green problem. The laser takes advantage of the spectral overlap between ground-state absorption and excited-state absorption in an Er-doped LiLuF4 crystal (Figure 1). Two 974-nm photons create a population inversion in the Er ions, which then can lase at 552 nm. Figure 1. The 974-nm photons can excite an Er ion in LiLuF4 from the ground state ( 4/15/2 ) to the excited state ( 4 I11/2 ), and from the excited state to a higher excited state ( 4 F7/2 ). From there, a spontaneous decay causes a population inversion between 4 S3/2 and the ground state. A small splitting of the ground state creates a quasi-three-level laser system.The researchers grew their own Er:LiLuF4 crystals using the Czochralski technique, keeping the Er doping level low (~1 percent atomic) to avoid lifetime quenching of the excited levels. The drawback of low doping was relatively weak pump absorption, so they designed a configuration to pass the pump radiation through the 1.6-mm-long crystal four times (Figure 2). The coating on the right side of the crystal (designated “P” in the figure) was highly transmissive at the 552-nm laser wavelength and highly reflective at the 974-nm pump wavelength. The coating on the left side of the crystal (designated “L”) had the opposite characteristics. Figure 2. The experimental arrangement allowed a quadruple pass of the pump radiation through the weakly absorbing crystal. In the photograph, the laser crystal is in the center, the quadruple-pass pump cavity is at left, and the laser resonator — defined by the coating on the crystal and the output coupler — is at right.Thus, the quadruple-pass pump cavity and the laser resonator could be aligned independently, and the arrangement provided a good overlap in the crystal between the pump radiation and the laser mode. An optically pumped semiconductor laser provided the pump power. The scientists could tune this laser smoothly and continuously over the pump wavelengths of interest — 962 to 976 nm — by adjusting an intra-cavity birefringent filter. The laser produced up to 9 W with a relatively constant 1-nm bandwidth across the tuning range. They obtained the best results with a pump wavelength of 974 nm and a 4 percent output coupler on the up-conversion laser (Figure 3). There was a rollover in the output power at pump levels above ~1 W, so they chopped the pump power with a 50 percent duty cycle, enabling the (peak power) output to continue linearly up to 760 mW. They concluded that the rollover was due to thermal population of the terminal laser level of the quasi-three-level system. Figure 3. Thermal population of the lower laser level limited the output of the up-conversion laser. Circles show the output of a continuous-wave laser; squares show the (peak) output when the pump power was chopped with a 50 percent duty cycle.The investigators believe that the 760-mW output is the highest reported for a room-temperature crystalline up-conversion laser, and they expect that it should be possible to extrapolate the technique to realize multiwatt visible lasers. Whether such lasers eventually might compete with well-established frequency-doubled lasers in an already crowded marketplace is an open question. Applied Physics Letters, Feb. 6, 2006, 061108.