Chromium-doped lithium strontium aluminum fluoride (Cr3+:LiSrAlF6) lasers have many promising characteristics, including a long spontaneous lifetime and a wide spectral emission range. But the crystal’s poor thermal conductivity has limited the lasers’ commercial appeal. During the past year, researchers at Centro de Lasers e Aplicações in São Paulo, Brazil, have investigated tailoring these lasers’ pump light with optical filters and have reported what they believe are previously unachieved power levels — as high as 30 W of average power at 851 nm — from Cr:LiSAF.Figure 1. Cr:LiSAF absorbs pump light in three broad absorption bands, but the quantum efficiency of the shorter-wavelength bands can be very low. To decrease thermal loading on the laser rod that results from these low quantum efficiencies, the researchers filtered the pump light as indicated to avoid absorption at the shorter wavelengths. Images reprinted with permission from OSA.Previous experiments conducted with Cr:LiSAF have encountered catastrophic results that included cracked laser rods when pushing toward repetition rates greater than 10 or 12 Hz or pulse energies in excess of a few joules. The laser has three broad absorption bands, centered at 290, 450 and 650 nm (Figure 1). The enormous quantum defect associated with pumping to these bands and lasing at ~850 nm leads to large amounts of waste heat deposited in the crystal, so the researchers in Brazil investigated filtering the pump light to eliminate the shorter-wavelength pump bands, using the spectral filters indicated in Figure 1.Figure 2. The optical filters between the flashlamps and the rod divided the pump cavity into three compartments. Cooling water circulated first around the laser rod and then around the pump lamps.In a close-coupled, or nonelliptical, pump cavity, the investigators placed spectral filters between the laser rod and the flashlamps (Figure 2). In their initial experiments, they allowed only pump light at the longest-wavelength pump band to reach the laser rod (filter set 1 in Figure 1). They timed the pump-lamp pulse duration to correspond to the ~67-μs upper-state lifetime of Cr:LiSAF, eliminating much of the heating that results from spontaneous emission. As a result, they pushed the repetition frequency of their laser as high as 30 Hz, observing average powers of 20 W and peak powers of 10 kW in ~65-μs pulses. But the strong spectral filtering of the pump lamp diminished the laser’s overall electrical efficiency.To improve the electrical efficiency, they investigated less-severe filtering of the pump light with various filters (filter sets 2 and 3 in Figure 1). Interestingly, they noticed that the intracavity circulating-power loss depended strongly on pump-lamp filtering. By measuring the laser threshold, they saw the resonator losses increase from ∼5 percent with filter set 1 to ~12 percent with filter set 2 to ~16 percent with filter set 3. They attributed this phenomenon to excited-state absorption: A laser photon interacting with a chromium ion excited to one of the higher pump bands could be sacrificed by boosting that ion to an even higher level. Such an event would be impossible if there were no ions in the higher levels — i.e., with filter set 1 in place.Figure 3. At higher pump-pulse energies (given in the inset), the laser-pulse energy decreased with increasing repetition rate. At the highest pump energy (100 J per pulse), the laser produced 2-J pulses at a 15-Hz rate, producing what the researchers believe is the highest average power — 30 W — obtained from Cr:LiSAF. These data were taken with filter set 2 (from Figure 1) in place; the roll-off of pulse energy with repetition rate was more severe with filter set 3.Despite the existence of this increased intracavity loss, the researchers obtained 2-J pulses at a 15-Hz repetition rate from their laser with filter set 2 in place. They believe that this is the highest average power extracted from a Cr:LiSAF laser rod. But the benefits of reducing the pump-light filtering further — by installing filter set 3 — were dubious. Although the laser’s electrical efficiency increased slightly, the pulse energy decreased rapidly at higher repetition rates, and the highest average power obtainable with filter set 3 was 16 W.Although the decrease in pulse energy with increasing repetition rate was less with filter set 2 than with filter set 3, it was nonetheless significant (Figure 3). To understand and perhaps reduce the effect, the researchers sought its cause. One possibility was thermal quenching of the laser transition; that is, heating the laser crystal enough so that the population inversion was diminished by a nonradiative, phonon-based, depopulation channel from the higher laser level. But when they examined the effective lifetime of the upper laser level as a function of pump-pulse energy, they found that the results were independent of repetition rate (Figure 4).Figure 4. The effective lifetime of the upper laser level decreased as expected with increasing pump-pulse energy because the chromium ions cycled more quickly through their four-level laser system. But the lifetime was independent of repetition frequency, indicating that thermal effects were not significant.They concluded that the roll-off in Figure 3 was caused not by a uniform thermal rise across the laser rod, which would quickly depopulate the upper laser level, but rather by a spatial thermal gradient, which introduced thermal lensing with a resulting deleterious effect on the intracavity mode geometry.Applied Optics, May 10, 2006, pp. 3356-3360; Optics Letters, Jan. 1, 2007, pp. 50-52.