1-mJ Q-switched diode-pumped Nd : BaY 2 F 8 laser

We report what is to our knowledge the first high repetition rate Q-switched Nd:BaY2F8 (Nd:BaYF) laser pumped with a multiwatt fibercoupled diode array tuned at 806 nm. As much as 2.42 W of average power and up to 1.05 mJ of pulse energy were obtained with 6.1 W of absorbed pump power, with excellent beam quality (M<1.2) and linear polarization. © 2004 Optical Society of America OCIS codes: (160.3380) laser materials; (140.3480) lasers, diode-pumped; (140.3540) lasers, Qswitched References and links 1. W.A. Clarkson, P.J. Hardman, D.C. Hanna, “High-power diode-bar end-pumped Nd:YLF laser at 1.053 μm,” Opt. Lett. 23, 1363-1365 (1998). 2. J. Harrison, P.F. Moulton, G.A. Scott, “13-W, M<1.2 Nd:YLF laser pumped by a pair of 20-W diode bars,” Postdeadline Paper CPD-20, CLEO 1995 3. A.A. Kaminskii, “New room-temperature diode-laser-pumped efficient quasi-cw and cw single-mode laser based on monoclinic BaY2F8:Nd 3+ crystal,” Phys. Stat. Sol. (a) 137, 61-63 (1993). 4. N.P. Barnes, K.E. Murray, A. Cassanho, K.M. Dinndorf, H.P. Jenssen, “Flashlamp pumped Nd:BaY2F8,” in OSA Proceedings on Advanced Solid-State Lasers, A.A. Pinto and T.Y. Fan, eds., Vol. 15 (Optical Society of America, Washington, D.C., 1993), 24-27. 5. A. Agnesi, A. Guandalini, G. Reali, E. Sani, A. Toncelli, M. Tonelli, “Spectroscopic analysis and diode pumped laser results of Nd: BaY2F8,” IEEE J. of Quantum Electron. 39, 971-978 (2003). 6. A. Agnesi, A. Guandalini, A. Lucca, E. Sani, A. Toncelli, M. Tonelli, “Medium-power diode-pumped Nd: BaY2F8,” Opt. Express 11, 1149-1155 (2003). 7. Y.-F. Chen, “Design criteria for concentration optimization in scaling diode end-pumped lasers to high powers: influence of thermal fracture,” IEEE J. of Quantum Electron. 35, 234-239 (1999). 8. W. Koechner, Solid-State Laser Engineering (Springer, Berlin, 1996).


Introduction
Neodymium fluoride laser crystals such as Nd:LiYF 4 (Nd:YLF) and Nd:BaYF are attractive for medium-high power diode-pumped Q-switched lasers at 1-µm wavelength with high pulse energy, because of their relatively long upper-state lifetime τ≈500 µs, allowing higher energy storage with respect to more widely used commercial laser crystals based on oxide materials, i.e.Nd:YAG (τ≈230 µs) and Nd:YVO 4 (τ≈100 µs).In addition, fluoride crystals benefit of more favorable thermo-optical properties, which can be conveniently exploited in multiwatt diode-pumped lasers for the generation of linearly polarized beams with excellent spatial quality and significantly reduced thermal lensing.
Nd:YLF has been extensively investigated for generation of high-energy pulses at kHz repetition rate in diode-pumped lasers, and excellent results were achieved both with endpumping [1] and side-pumping [2] configurations.
Nd:BaYF has long been known as a laser material with spectroscopic characteristics similar to Nd:YLF [3,4] (with the exception of a fluorescence bandwidth nearly two times larger), but only recently it was considered as an interesting alternative to Nd:YLF in diodepumped solid-state lasers.In particular, a detailed spectroscopic and structural investigation, along with low-power diode-pumping operation at 1.05 µm and at 1.32 µm was reported in Ref. [5].A 2.4-W efficient cw Nd:BaYF laser pumped by a multiwatt fiber-coupled diode array was also reported [6], with an accurate characterization of the thermal lens and its associated diffractive losses, which resulted significantly lower than those measured in a Nd:YVO 4 laser used for comparison.
Here we summarize the results of a high-energy, high repetition rate Q-switched Nd:BaYF laser based on a low-doping crystal sample (0.6% at).As much as 1.05-mJ pulse energy was obtained in Q-switching at 0.9 kHz from a 2.4-W cw diode-pumped Nd:BaYF laser.Short Nd:BaYF samples yielded an internal slope efficiency dP out /dP abs as high as 66% in cw operation, confirming the excellent optical quality of the grown crystals.At this doping level, the maximum incident power leading to thermal fracture was ≈20 W.

Experiments and modeling
The crystal was grown in a homemade Czochralski furnace with resistive heating and automatic diameter control system.Particular care was taken in order to avoid contamination in the sample: the starting materials were 5N fluoride powders (BaY 2 F 8 , NdF 3 , and BaF 2 ), and the growth process took place in high purity Ar atmosphere.During the growth process the temperature was around 976°C, the pull rate was 0.5-mm/hour, and the rotation rate was 5 revolutions per minute.The seed was a piece of undoped crystal oriented along the a crystallographic axis.The doping level in the melt was 0.6 Nd 3+ at %.The boule was pencilshaped with a maximum diameter of 15 mm and a length of about 63 mm.It was oriented and a 3.8x1.5x11mm 3 sample was Brewster-cut for beam propagation perpendicular both to the b and c axes and electric field polarization parallel to the c-axis.This particular shape was chosen to allow also the investigation of mode-locking performance in a separate experiment, where even small Fresnel reflections are undesired.Shorter samples were cut for operation at normal incidence.
A 25-W fiber-coupled diode array was used as the pump source.The output fiber core had a diameter of 400 µm and a numerical aperture of 0.22.To assess the thermal fracture power, initially we tested a 2.6-mm long a-cut sample of Nd:BaYF used at normal incidence, with uncoated end faces, polished flat and parallel.The crystal, mounted in a water-cooled copper holder, using an indium foil for optimum thermal contact, was put in a 45-mm long plane-concave Fabry-Perot cavity.The input mirror was a 100-mm radius-of-curvature (RoC) concave mirror, HR at 1049 nm and HT at 806 nm, whereas the output coupler (OC) was flat.
The pump beam diameter was ≈500 µm, and the fiber-coupled pump was tuned at ≈806 nm.
Figure 1 shows some test results.It is worth noting the high internal optical-to-optical slope efficiency of 66% owing to the effective overlap between the pump beam and the resonant mode, which is allowed by the use of a short crystal.Of course, the overall opticalto-optical efficiency was much lower (≈10%), because of the small absolute absorption ≈17%.Nevertheless, this result proves the high quality of the grown Nd:BaYF crystals.The critical incident pump power for thermal fracture was ≈20.4 W, more than three times the level allowed by the 1.8%-doped crystal tested in an earlier experiment [6].These results are in reasonable agreement, since the thermal fracture parameter scales with the effective absorption coefficient, which is approximately proportional to the doping level [7].
For the Q-switching experiments we built a Z-folded resonator, using the Brewster-cut 11-mm long Nd:BaYF crystal which absorbed ≈47% of the pump power.The resonator (C) 2004 OSA schematic is shown in Fig. 2. Its length (≈66 cm) was chosen to yield sufficiently long pulses also at 1-kHz repetition rate, thus avoiding excessive peak intensity, which could produce surface damage of the optical components.The folding angles were kept <7° to minimize mode-mismatching with the pump within the laser crystal, however the output beam was horizontally elongated owing to the asymmetric thermal lens, with an aspect ratio of ≈1:1.4.To reduce the risk of thermal damage of the Nd:BaYF crystal we limited the level of the incident pump power to ≈13 W.
Using an OC with R=84%, which optimized the Q-switching operation, the cw output power was 2.42 W at 6.1 W of absorbed pump power, corresponding to an internal slope efficiency of ≈56%, thus the performance was slightly better than for a relatively long 1.8%doped Nd:BaYF crystal [6].The laser was Q-switched with a fused-silica acousto-optic Brewster-angled modulator (NEOS, model N36027-5-.8-BR).Figure 3 summarizes the best performance.Pulse energy as high as 1.05 mJ with pulse duration ≈70 ns was obtained at the lowest frequency accessible with the electronic driver (0.9 kHz), whereas the average power approached closely the cw value when the frequency was set above 25 kHz.
The beam quality at the maximum output power was M 2 ≈1.1 in both directions, indicating nearly diffraction-limited operation.Investigation of the pump-power dependent mode size on the OC showed that the uncoated, Brewster-cut Nd:BaYF slab exhibited a negative thermal lens with focal length of ≈-500 mm in the sagittal (i.e., vertical) plane with 6.1 W of absorbed pump power.In our earlier study we measured instead a positive thermal focal lens, possibly because the positive contribution of the bulging end of the laser crystal that was used as the cavity pump mirror overbalanced the negative refractive term dn/dT.
Passive Q-switching with two Cr 4+ :YAG crystals with initial transmission T 0 =91% and T 0 =85% was also investigated with this resonator setup (the saturable absorber was inserted near the OC).At an absorbed pump power of 5.2 W we compared the pulse energy obtained with passive and active Q-switching, at the same repetition frequency: 220 µJ at 6.3 kHz and 368 µJ at 2.9 kHz were obtained with T 0 =91% and T 0 =85% respectively, yielding ≈77% of the pulse energy available through active Q-switching.Owing to the filtering provided by the nonlinear absorber, the beam quality was even better, M 2 <1.1, and the pulse width slightly shorter.In order to investigate the possible presence of parasitic effects (i.e., upconversion etc.), which are related to the gain in excess stored in the laser crystal before the Q-switching, we compared the experimental data at different pump power with the numerical results for the output pulse energy E yielded by the standard model of repetitive Q-switching at the frequency f [8]: where g t is the gain at threshold, i.e. g t = L − ln(R) , L is the intracavity passive loss, g i and g f are the gain before and after the Q-switching, respectively, g ∞ is the small-signal gain at steady state at a given pump level, and E sat is the saturation energy given by the product of the saturation fluence (including level degeneracy) and the effective mode area.For a particular pump level, Eqs. ( 2) and (3) can be solved numerically to find the initial and final gain, g i and g f .The parameter E sat can be conveniently determined by a best-fit procedure.
Figure 4 shows the comparison between the experiment and the model.The value for the upper level lifetime τ=490 µs was measured with spectroscopic techniques [5] for the 0.6%doped Nd:BaYF.Considering the relative pump intensity (g ∞ /g t ) experimentally measured for each curve, we find a reasonable agreement of the experimental data with the model predictions, which neglect parasitic effects.Furthermore, the data reported in Fig. 4 show that the output pulse energy grows almost linearly with the cw output power up to the maximum pump power exploited, and the low-frequency Q-switching output energy is approximately that expected given the cw performance and the fluorescence lifetime.Therefore, pumpinduced nonlinear losses are not significant in this laser and should not prevent at least a moderate energy upscaling.

Conclusions
Nd:BaYF has been proved an interesting material for efficient diode-pumped operation, also at relatively high pump power.Its thermal fracture limit is confirmed to be ≈3 times smaller than that of Nd:YLF [6].This drawback can be partially overcome by improving the thermal management of the laser head (for example using a side-pumped slab design) and also by reducing the doping level, which also helps to minimize parasitic effects such as upconversion.Indeed, 0.6%-doped crystals of high optical quality have shown efficient operation (up to 66% internal slope efficiency) and could be pumped by as much as 20 W of incident power before the thermal fracture occurred.In particular, we have shown that Nd:BaYF is an attractive material for efficient multiwatt-pumped kHz Q-switched diodepumped lasers.Pulse energy as high as 1.05 mJ at 1049 nm was obtained in nearly TEM 00 mode, with linear polarization.A numerical model of the high-frequency Q-switching operation suggests that at the pump level required for this performance the gain-dependent parasitic effects such as upconversion do not represent a major limitation to further energy upscaling.

Fig. 3 .
Fig. 3. Output pulse energy and duration (fwhm) as a function of the repetition frequency (optimum OC with R=84%).

Fig. 4 .
Fig. 4. Output pulse energy at different pump levels (the corresponding cw output power is indicated in the inset).Continuous lines represent the numerical results.