We demonstrate, for the first time, passively Q-switched laser operation of Yb:Y3Ga5O12 garnet crystal. An output power of 4.53 W at 1031 nm was generated at a pulse repetition rate of 55.6 kHz, with a Cr4+:YAG crystal acting as saturable absorber whose initial transmission was 97.5%, the corresponding pulse energy, duration and peak power were respectively 81.5 μJ, 28.5 ns and 2.86 kW. Laser pulses at a lower repetition rate of 18.2 kHz were also achieved at 1025 nm while the initial transmission of the Cr4+:YAG crystal was 85.0%, with an output power measured to be 2.56 W, the resulting pulse energy, duration, and peak power being 140.8 μJ, 5.9 ns, and 23.9 kW, respectively.
© 2013 OSA
Due to its excellent chemical, mechanical, and thermal properties, yttrium aluminum garnet, Y3Al5O12 (YAG), has become the most common crystal serving as host medium for trivalent rare earth active ions such as Nd, Yb, Tm, Ho, Er, etc., to make various solid-state laser materials, in particular for applications in high-power or high-energy laser systems. However, it has been found that in certain circumstances, YAG suffers from some disadvantages; e.g., the thermal conductivity of Yb:YAG tends to drop largely with the increase in Yb doping level . To avoid such drawbacks while exploiting other desirable properties possessed by Yb:YAG, several isomorphic Yb garnets have been developed, including Yb:Lu3Al5O12 (Yb:LuAG) , Yb:Gd3Ga5O12 (Yb:GGG) [1, 3, 4], Yb:Y3Ga5O12 (Yb:YGG) [5, 6], and Yb:Lu3Ga5O12 (Yb:LuGG) . Among these Yb:LuAG has been studied most extensively, from high-power thin-disk lasers producing 5 kW of output to microchip lasers operating in continuous-wave (cw) or passively Q-switched modes, and also mode-locked laser operation generating ultrashort pulses [8–11]. In the class of Yb doped gallium garnets, Yb:GGG is the first one developed, demonstrating comparable laser performance to that of Yb:YAG . Another member in this garnet class, Yb:YGG, was developed later in 2009 , its growth, thermal and spectroscopic properties have been studied . In passively mode-locked operation, an output power of 570 mW was produced with pulse duration of 245 fs from a diode pumped Yb:YGG laser ; while cw laser operation was demonstrated with a folded resonator, generating 2.65 W of output power with a slope efficiency amounting to as high as 84.5% . Despite these promising results, however, the research into its basic laser performance is still limited, with some important laser properties, such as passive or active Q-switching, and wavelength tenability, remaining unknown.
In this paper we report on the passive Q-switching laser performance of Yb:YGG crystal, demonstrated with Cr4+:YAG crystals utilized as saturable absorber. Employing several crystal samples of different lengths, the cw laser performance has also been evaluated.
2. Description of experiment
As in the previous work [5, 6, 12], the Yb:YGG crystal studied here was also grown by the optical floating zone technique. The Yb ion concentration in crystal was 7.35 at. %, which was lower than the doping level of the previously used Yb:YGG (9.8 at. %). Three crystal samples, cut along the  crystallographic direction, with a square aperture of 3 mm × 3 mm and different lengths of 4, 5, and 6 mm, were prepared. The laser resonator was formed by a plane reflector and a concave mirror serving as the output coupler. The plane reflector was coated for high reflectance (>99.9%) at 1030−1200 nm and high transmittance (>98%) at 820−990 nm. As the output coupler a number of different concave mirrors, with radius-of-curvature (R2) of 25 or 50 mm, and output couplings (T) in a range of 0.5−30%, were used. Several Cr4+:YAG crystals were utilized as saturable absorbers for passive Q-switching action, their initial transmissions (T0) were 97.5%, 94.4%, 90.0%, and 85.0%, respectively. These Cr4+:YAG plates were coated for antireflection (R <0.2%) at 1.06 μm on both faces. The uncoated Yb:YGG sample was held in a water-cooled copper block, and was placed close to the plane reflector inside the resonator. The pump source was a high-power fiber-coupled diode laser, with fiber core diameter of 200 μm and NA of 0.22. Depending on the power level, the wavelength of pumping could be varied between 970 and 974 nm, matching with the strong absorption peak at 970 nm . The pump beam was transferred into the laser crystal through a re-imaging optics of 1:1 ratio, with a beam spot radius of approximately 100 μm. To measure the laser emission spectra, a spectrometer (AvaSpec-3648, Avantes B.V., resolution of 0.3 nm) was employed. A digital oscilloscope (Infiniium DSO80304B, Agilent Co. Ltd., bandwidth of 3GHz, sampling rate of 40 GSa/s) was used to monitor the laser pulses and measure the pulse parameters.
3. Results and discussion
First the cw laser performance of the Yb:YGG garnet was studied in a plano-concave resonator composed of couplers with R2 of 25 mm, the physical cavity length being 24 mm. With each crystal sample laser oscillation was achieved at room temperature with output couplings ranging from T = 0.5% to T = 30%. For the 4 mm thick crystal, the optimum coupling was found to be T = 1%; whereas for the 5 or 6 mm thick crystal it increased to T = 3%. This variation in optimum output coupling resulted mainly from the increasing resonant absorption losses of the longer crystals.
Figure 1 shows the output power as a function of absorbed pump power (Pabs) for the three crystal samples, measured under conditions of optimum output coupling. The absorbed pump power was determined from the residual pump power that was measured behind the laser crystal while the output coupler was removed from the cavity. Due to the strong divergence of the pump beam, it was not feasible to measure Pabs under lasing conditions. Indeed, the amount of Pabs determined in this way might deviate from its real value when the laser was on, hence overestimating the laser efficiency to some extent. One can see obviously from Fig. 1 that the cw laser performance achieved with the 4 mm thick crystal was much better than obtained with the other two longer samples. Several reasons were responsible for this result, including the specific resonator configuration, the pumping conditions, the Yb ion concentration of the crystal, and the thermal effects occurring inside the crystal during the laser action. A maximum output power of 6.75 W was generated at Pabs = 9.93 W, which was the highest available pump power, resulting in an optical-to-optical efficiency of 68%. The slope efficiency determined for Pabs > 4.5 W amounted to as high as 85%. In comparison with the previous results reported for the Yb:YGG crystal, the output power and optical-to-optical efficiency achieved in the current experiment were much higher; while the slope efficiency proved to be comparable . Compared to the Yb:GGG and Yb:LuGG lasers that were also longitudinally pumped by similar diode lasers emitting around 971 nm, producing cw output powers of ~4−5 W with optical-to-optical efficiencies of 51−52% and slope efficiencies of 63−64% [1, 13], the laser performance of Yb:YGG demonstrated here proves to be much more promising. With the longer crystals, the highest output power produced was measured to be 6.20 and 5.37 W, for the 5 and 6 mm thick crystals, respectively. One can also note from Fig. 1 that in excess of Pabs ≈10 W, the output power produced with the two longer crystals exhibited a tendency of rolling off, arising from the increasingly strengthening thermal effects. It should be pointed out that the optimum crystal thickness of 4 mm was determined for the specific resonator configuration and pumping conditions utilized in the current experiment. As mentioned above, with these conditions changed, the optimum crystal length may vary accordingly.
In the cw laser operation achieved with a Yb:YGG crystal of given thickness, the oscillation wavelengths varied only slightly with pump power for a certain output coupling. Shown in Fig. 2 are the laser emission spectra, measured at an intermediate pump level (a) and the highest absorbed pump power (b) for each crystal. In the case of the 4 mm thick crystal, the laser oscillated in a wavelength range of 1065−1074 nm at Pabs = 5.82 W; at the highest pump power of Pabs = 9.93 W, the emission range changed to 1067−1076 nm. Similar situation occurred also in the cases of the two longer crystals. The oscillation wavelengths measured at the highest pump power covered a range of 1044−1049 nm (5-mm crystal) and 1045−1050 nm (6-mm crystal).
Figure 3 illustrates the variation of the laser emission spectrum with the output coupling, measured for the 4 mm thick crystal in a wide range from T = 0.5% to T = 30%. With the output coupling increased, the laser oscillation was forced to occur at shorter wavelengths, as higher gain was required to balance the greater losses. One sees that in the case of T = 3%, the laser oscillated at 1042−1047 nm, which was very close to the laser emission range measured for the 5 mm or 6 mm thick crystal (Fig. 2). Similarly, the emission spectrum measured for the 5 mm or 6 mm thick crystal in the case of T = 1% also covered a range of roughly 1065−1075 nm. As a result, the different emission spectrum for the 4 mm thick crystal compared to those for the longer crystals, as shown in Fig. 2, was attributed to the different output couplings utilized.
Theoretically, the oscillation of a quasi-three-level laser operating in free-running mode will occur at wavelengths (λ) where the gain cross section, σg(λ) = βσem(λ) − (1−β)σabs(λ), reaches its maximum. Here σabs and σem are absorption and stimulated emission cross sections, respectively, whereas β is the fraction of population excited to the upper manifold. For a practical laser of which the output coupling remains extremely low, e.g., T ≤ 1%, the magnitude of β is very small, usually less than 0.1. Depicted in Fig. 4 is a group of σg(λ) versus λ curves for different parameter β ranging from 0.07 to 0.045, calculated by use of the absorption and emission spectra for the Yb:YGG crystal reported in . One can see that as the parameter β decreases from 0.07 to 0.055, the maximum of σg(λ) will shift from 1036 to 1043 nm. Considering the fact that the magnitude of β for a practical laser under certain operational conditions is determined mainly by the output coupling of this laser, one can understand qualitatively the dependence of oscillation wavelengths upon the output coupling illustrated in Fig. 3 for the Yb:YGG laser. It is also seen from Fig. 4 that with β reduced further, the maximum in the σg(λ) curve will become less pronounced; in fact, for β = 0.05, the peak value of σg(λ) at 1045 nm has already become roughly equal to the values for the wavelength range of about 1064−1073 nm; and eventually for β = 0.045, the highest gain will be reached in this longer wavelength range. This explains the behavior of oscillation wavelength shifting from 1042−1047 nm to 1065−1074 nm, observed at Pabs = 5.82 W in the cw laser operation achieved with the 4 mm thick crystal, when the output coupling was changed to T = 1% from T = 3%, as shown in Figs. 2 and 3. As is known, under steady-state oscillation conditions, the gain and hence the magnitude of β will be clamped at some certain value as long as the overall losses of the resonator remain unchanged, regardless of the variation in pump power level. Therefore, it should not be difficult to understand that the laser action obtained with the 4 mm thick crystal in the case of T = 1%, occurring at 1067−1076 nm which proved far from the peak wavelength of the stimulated emission cross section (1024 nm ), produced the highest output power.
To study the passive Q-switching laser performance of Yb:YGG garnet, only the 4 mm thick crystal was employed. The plano-concave resonator was formed by the plane reflector and a concave mirror of R2 = 50 mm, with output coupling limited to T ≥ 20%, in order to prevent any damage to the internal elements. The physical cavity length was 35 mm for the laser using a Cr4+:YAG of T0 = 97.5% or 94.4%, which was optimized experimentally to generate highest output power; it was shortened to 28 mm when the saturable absorber utilized was the one of T0 = 90.0% or 85.0%, in order to obtain laser pulses of shortest duration by reducing the cavity photon lifetime. The average output power, produced under several different operational conditions, is plotted in Fig. 5 against the absorbed pump power. In the case of T = 20%, T0 = 97.5%, the Q-switched laser oscillation reached threshold at Pabs = 2.05 W. Above this threshold the output power increased approximately linearly with Pabs, reaching 4.53 W at Pabs = 8.95 W, leading to an optical-to-optical efficiency of 50.6%, while the slope efficiency was determined to be 65%. With the output coupling increased to T = 30%, the threshold, as expected, became higher; it was measured to be Pabs = 2.50, 2.53, and 2.56 W, for T0 = 94.4%, 90.0%, and 85.0%, respectively. In connection with the closeness in threshold, one sees from Fig. 5 that the output characteristics of the Yb:YGG laser essentially differ only slightly for the three different saturable absorbers employed, over a wide operational range from the threshold up to Pabs ≈7.0 W, for which a common slope efficiency of 60% was measured approximately. In excess of this pumping level, while the output power produced in the case of T0 = 94.4% would continue to increase, reaching 3.81 W at the highest pump power of 8.95 W; the Q-switched laser operation that was generated by the saturable absorber of T0 = 90.0%, became less efficient, producing a maximum output power of 3.15 W at Pabs = 8.95 W. In the case of T0 = 85.0%, an average output power of 2.56 W was measured at Pabs = 7.14 W, resulting in an optical-to-optical efficiency of 35.9%. Above this power level damage was likely to occur on the crystal surface, due to the high peak power circulating inside the laser cavity. Similarly, the highest pump power absorbed in the crystal was limited to 8.95 W in the cases of T0 = 97.5%, 94.4%, and 90.0%, in order to prevent any damage to the internal elements.
Upon increasing pump power, the pulse repetition frequency (PRF) of the Yb:YGG laser was found to rise, which is a typical behavior for passively Q-switched lasers. Depending on the operational conditions, the variation ranges for PRF were measured to be 9.5−55.6 kHz (T = 20%, T0 = 97.5%); 5.6−50.0 kHz (T = 30%, T0 = 94.4%); 2.0−28.6 kHz (T = 30%, T0 = 90.0%); and 0.6−18.2 kHz (T = 30%, T0 = 85.0%).
The energy contained in a single laser pulse can be estimated from the average output power and the corresponding PRF. As usual, the laser pulse energy was found to remain more of less unchanged when the laser was operated far above threshold. From the results shown in Fig. 5, one can determine the pulse energy obtained under different operational conditions: it was 81.5 μJ (T = 20%, T0 = 97.5%); 76.2 μJ (T = 30%, T0 = 94.4%); 110.1 μJ (T = 30%, T0 = 90.0%); and 140.8 μJ (T = 30%, T0 = 85.0%), at the highest output power.
The pulse duration was found to be independent of pump power, it depended only on the T and T0 used. Illustrated in Fig. 6 are typical profiles of the laser pulses generated under different conditions of T and T0, measured at an intermediate power level of Pabs = 6.02 W. The pulse duration amounted to 28.5 ns for T = 20%, T0 = 97.5%; it was reduced to 5.9 ns in the case of T = 30%, T0 = 85.0%. In the other two cases, it was measured to be 14.0 (T = 30%, T0 = 94.4%) and 7.0 ns (T = 30%, T0 = 90.0%). From the pulse energy (Ep) and duration (tp), one can estimate the corresponding peak power by Pp = Ep/tp, giving 2.86 (T = 20%, T0 = 97.5%); 5.44 (T = 30%, T0 = 94.4%); 15.7 (T = 30%, T0 = 90.0%); and 23.9 kW (T = 30%, T0 = 85.0%). The intensity fluctuations from pulse to pulse were estimated to be less than 10%, while the time jitters were lower than 20%.
In Q-switched operation far above threshold, the Yb:YGG laser varied very slightly in oscillation wavelength, which was determined actually by only the output coupling utilized. Shown in Fig. 7 are the laser emission spectra measured at Pabs = 6.02 W, under different conditions of T and T0. The Q-switched laser oscillation occurred at 1031 nm under conditions of T = 20%, T0 = 97.5%; with the output coupling increased to T = 30%, it would shift to 1025 nm, regardless of the amount of T0 used. It is instructive to note from Fig. 7 that for each case, the emission spectrum consisted of only a single wavelength, with a linewidth less than 0.5 nm that was limited by the resolution of the spectrometer. This stands in sharp contrast to the emission spectra of cw operation illustrated in Fig. 3. During the passive Q-switching process, the absorption saturation in the Cr4+:YAG crystal played a crucial role in the discrimination of oscillating axial modes.
Table 1 summarizes the primary parameters characterizing the passive Q-switching laser performance of Yb:YGG garnet, in which the additional symbols are defined as follows: Pavr, maximum average output power; ηopt, optical-to-optical efficiency; ηs, slope efficiency; and λ, oscillation wavelength. By a comparison with a Yb:LuGG laser formed with an identical cavity, which was also passively Q-switched under similar conditions (T = 20%, T0 = 97.6%) , one sees that while most parameters were very close, the output power and slope efficiency proved to be considerably higher for the Yb:YGG laser; more importantly, the Q-switching laser performance of the Yb:YGG achieved with a Cr4+:YAG of T0 = 85.0%, which was likely not to work with Yb:LuGG, turns out to be much more promising.
With the help of a theoretical model developed for passively Q-switched Yb lasers, which was used successfully in analyzing a Yb:GGG/Cr4+:YAG laser , we have calculated the parameters of laser pulses generated with the two lower initial transmission Cr4+:YAG crystals (T0 = 90.0%, 85.0%) in the 28 mm long resonator with a coupler of T = 30%. Using the same symbols as in  to denote the parameters needed in the calculation, we list the numbers for these parameters as follows: σgsa = 4.3 × 10−18 cm2, σesa = 8.2 × 10−19 cm2 , γσ = 2.5 × 10−20 cm2 (γ ≈1) , tr = 0.20 ns, na0 = 4.0 × 1018 cm−3, L = 0.01, ωL = 70 μm, and ωp = 100 μm. The calculated pulse energy and duration, 117.9 μJ, 5.8 ns for T0 = 90.0%, and 165.8 μJ, 4.1 ns for T0 = 85.0%, are in fair agreement with the experimental results listed in Table 1.
In summary, the passive Q-switching laser performance of the Yb:Y3Ga5O12 garnet crystal has been studied employing different Cr4+:YAG saturable absorbers. An average output power of 4.53 W at 1031 nm was generated at a pulse repetition rate of 55.6 kHz, with optical-to-optical and slope efficiencies being 50.6% and 65%, respectively. In the pulsed operation achieved with a Cr4+:YAG of T0 = 85.0%, the output power could reach 2.56 W at 1025 nm with a slope efficiency of 60%, the resulting pulse energy, duration, and peak power were 140.8 μJ, 5.9 ns, and 23.9 kW, respectively. In cw laser operation the output power reached 6.75 W, with an optical-to-optical efficiency of 68%, while the slope efficiency amounting to 85%.
This work was supported by the National Natural Science Foundation of China (Grants 60978023, 51272131, and 51025210).
References and links
1. S. Chénais, F. Druon, F. Balembois, P. Georges, A. Brenier, and G. Boulon, “Diode-pumped Yb:GGG laser: comparison with Yb:YAG,” Opt. Mater. 22(2), 99–106 (2003). [CrossRef]
2. A. Brenier, Y. Guyot, H. Canibano, G. Boulon, A. Ródenas, D. Jaque, A. Eganyan, and A. G. Petrosyan, “Growth, spectroscopic, and laser properties of Yb3+-doped Lu3Al5O12 garnet crystal,” J. Opt. Soc. Am. B 23(4), 676–683 (2006). [CrossRef]
3. B. Jiang, Z. Zhao, X. Xu, P. Song, X. Wang, J. Xu, and P. Deng, “Spectral properties and charge transfer luminescence of Yb3+:Gd3Ga5O12 (Yb:GGG) crystal,” J. Cryst. Growth 277(1-4), 186–191 (2005). [CrossRef]
4. Y. Guyot, H. Canibano, C. Goutaudier, A. Novoselov, A. Yoshikawa, T. Fukuda, and G. Boulon, “Yb3+-doped Gd3Ga5O12 garnet single crystals grown by the micro-pulling down technique for laser application. Part 2: Concentration quenching analysis and laser optimization,” Opt. Mater. 28(1-2), 1–8 (2006). [CrossRef]
5. Y. Zhang, Z. Wei, B. Zhou, C. Xu, Y. Zou, D. Li, Z. Zhang, H. Zhang, J. Wang, H. Yu, K. Wu, B. Yao, and J. Wang, “Diode-pumped passively mode-locked Yb:Y3Ga5O12 laser,” Opt. Lett. 34(21), 3316–3318 (2009). [CrossRef] [PubMed]
6. H. Yu, K. Wu, B. Yao, H. Zhang, Z. Wang, J. Wang, Y. Zhang, Z. Wei, Z. Zhang, X. Zhang, and M. Jiang, “Growth and characteristics of Yb-doped Y3Ga5O12 laser crystal,” IEEE J. Quantum Electron. 46(12), 1689–1695 (2010). [CrossRef]
7. K. Wu, L. Hao, H. Zhang, H. Yu, Y. Wang, J. Wang, X. Tian, Z. Zhou, J. Liu, and R. I. Boughton, “Lu3Ga5O12: exploration of new laser host material for the ytterbium ion,” J. Opt. Soc. Am. B 29(9), 2320–2328 (2012). [CrossRef]
8. K. Beil, S. T. Fredrich-Thornton, F. Tellkamp, R. Peters, C. Kränkel, K. Petermann, and G. Huber, “Thermal and laser properties of Yb:LuAG for kW thin disk lasers,” Opt. Express 18(20), 20712–20722 (2010). [CrossRef] [PubMed]
9. J. Dong, K. Ueda, and A. A. Kaminskii, “Laser-diode pumped efficient Yb:LuAG microchip lasers oscillating at 1030 and 1047 nm,” Laser Phys. Lett. 7(10), 726–733 (2010). [CrossRef]
12. Y. Zhang, Z. Wei, Q. Wang, D. Li, Z. Zhang, H. Yu, H. Zhang, J. Wang, and L. Lv, “Diode-pumped efficient continuous-wave Yb:Y3Ga5O12 laser at 1035 nm,” Opt. Lett. 36(4), 472–474 (2011). [CrossRef] [PubMed]
14. X. Zhang, A. Brenier, Q. Wang, Z. Wang, J. Chang, P. Li, S. Zhang, S. Ding, and S. Li, “Passive Q-switching characteristics of Yb3+:Gd3Ga5O12 crystal,” Opt. Express 13(19), 7708–7719 (2005). [CrossRef] [PubMed]