Open Access
Add to BibSonomyAbstract
First operation of the 4F9/2 → 6H13/2 laser transition in dysprosium doped yttrium aluminum garnet is reported. Efficient room temperature operation at 583nm was obtained using 447nm GaN diode lasers pumps. Gaussian single-mode operation was demonstrated with a non-optimized slope efficiency of 12%. Millisecond pulsed operation generated 150mW with power limited by the pump diodes’ brightness.
©2012 Optical Society of America
Introduction
The emergence of commercial gallium nitride laser diodes opens possibilities for many new optical devices in the visible. With emission from the violet to the green, these laser diodes offer new capabilities for a wide range of applications [1]. One such application is their use as pump sources for solid-state lasers. This approach has been successfully applied to visible lasers based on praseodymium [2]. Here we report operation of a visible dysprosium laser based on direct laser diode pumping. Laser action on this yellow transition has previously been reported in a Dy:ZBLAN fiber pumped by an Argon ion laser [3], and in cryogenic flashlamp pumped Dy:KY(WO4)2 [4]. Recently a Dysprosium doped fluoro-aluminate fiber laser, pumped with a 399nm diode, was demonstrated with operation in the yellow [5]. Gain has also been reported for the yellow transition in Dy:LiNbO3 crystals pumped with a blue OPO [6].
The four-level dysprosium laser scheme investigated here is shown in Fig. 1 . Blue pump light is absorbed by the Dy3+ ground state via the transition 6H15/2 → 4G11/2. Multi-phonon quenching of the upper level rapidly transfers energy to the metastable 4F9/2 level. At low concentrations this level has a fluorescence lifetime of 0.9 ms in Dy:YAG [7,8]. While many laser transitions are possible from the 4F9/2 level into any of the 6F and 6H levels shown in Fig. 1, most of them would experience significant ground state absorption [9]. In this paper, we are concerned only with the yellow emission from 4F9/2 → 6H13/2 transition.
Spectroscopic experiments
Optical pumping with GaN laser diodes is similar in many respects to near-IR pumping with GaAs based diodes, but with a few important differences. In particular, the shorter wavelength GaN lasers leads to potentially higher intrinsic brightness and smaller pump spots. This can facilitate mode-matching in end-pumped laser systems.
Spectral control of these lasers can be a bit more challenging than for GaAs-based devices. Current devices have a useful operating temperature range of only −10C to 50C. With temperature tuning coefficients of around 0.05nm/°C, the total temperature tuning is limited to about ± 1 nm. Diode selection for pumping narrow absorption lines is therefore more critical than with GaAs diodes.
For these experiments we used a pair of single-stripe 1W laser diodes. Their wavelength was selected to approximately match the strong 447 nm Dy:YAG pump band. Fine tuning of the pump wavelength was accomplished with a simple external grating in the Littrow configuration. The grating-locked laser diodes produced a linewidth of 0.17 nm FWHM. Anamorphic prisms were used to circularize the beam. A half-wave plate and a polarizer were used to combine the diode beams. A slightly elliptical pump spot of 70 μm x 60 μm was measured at the focus of a 15 cm spherical lens. The position and size of the pump spot were adjusted to approximately mode match the laser cavity. To reduce the risk of premature diode failure, operation was limited to millisecond pulses and currents below 1.6 A. This generated peak powers of up to 2.3 W. The diodes required about 100 µs to stabilize wavelength and power. Although the diodes are rated for cw operation, low repetition rates were used to simplify temperature stabilization which was accomplished with a small thermo-electric control unit.
Dysprosium-doped YAG crystals were grown for these laser experiments. Doping concentrations of 1, 2, and 3% atomic relative to yttrium were targeted. The 3%Dy:YAG growth was select for these experiments as a trade between pump absorption and concentration quenching. Right angle 3x20 mm laser rods were fabricated with anti-reflection coatings for both the pump and the yellow emission line at 582nm.
The room temperature absorption of 3%Dy:YAG was measured with 0.1nm resolution using a Cary 5G spectrophotometer and is shown in Fig. 2 . The sharp blue absorption peak at 447 nm is shown in the inset. Tuning the grating-locked pump diodes to the peak at 447 nm yielded 70% single-pass absorption in the 2.0 cm laser rods. Assuming a 3% atomic Dy3+ concentration, these two measurements generate effective absorption cross sections at 447.4 nm of 1.6E-21cm2 and 1.5E-21cm2, respectively. These are somewhat higher than the value of 1.1E-21cm2 recently reported for ceramic YAG [10].
Fig. 2 Blue absorption of 3% Dy:YAG at room temperature. Inset shows diode pump band on expanded scale. Sample was uncoated and 3.39 mm thick.
Room temperature emission spectra were recorded from the 3% Dy:YAG rod pumped at 447nm. The spectra were collected using an Ando AQ6315A spectrometer from 450 nm to 1500 nm at 0.1 nm resolution. Strong emission bands were observed from the 4F9/2 → 6H13/2 (yellow) and 4F9/2 → 6H15/2 (blue-green) transitions. Weak emission bands were also observed in the near-IR at 670 nm, 760 nm, 830 nm, 845 nm, and 1016 nm. The yellow emission band is shown in Fig. 3 . The strongest of the emission lines is at 582.7 nm, corresponding to the observed laser line. Our measurement shows this line is significantly sharper, 0.20 nm FWHM, than that reported in ceramic YAG [10]. We have also observed similar sharp yellow emission lines in samples of Dysprosium doped YLiF4 and Y2O3.
Fig. 3 Room temperature yellow emission cross sections of Dy:YAG calculated from the spectral intensity.
The emission shown in Fig. 3 was converted to cross section using the Füchtbauer-Ladenburg relation and the 2.0ms radiative lifetime calculated by Lupei et al. [10]. However, our value at 582.7nm of 3.0E-21cm2 is significantly lower than their reported value of 1.5E-20 cm2. Threshold analysis in the next section is more consistent with the lower value.
Fluorescent decay of the 4F9/2 level was found to be independent of pump intensities. Low doping concentrations produced fluorescent decays of 0.9 ms, comparable to those previously reported [7,8,10]. No dependence of the lifetime on pump wavelength was observed between350-450 nm. Increased dysprosium doping leads to the non-exponential decays shown in Fig. 4 . Although the storage lifetime is clearly reduced, we found that the steady-state yellow fluorescence rose almost linearly with pump intensity in the 3% Dy:YAG. These observations are consistent with the weak cross relaxation mechanisms shown in Fig. 1 and previously report by Klimczak [7].
Fig. 4 Room temperature fluorescence decay of Dy:YAG. Emission measured at 580nm after blue excitation. Lifetimes are exponential fits to first 500µs.
The lower laser level, 6H13/2, is rapidly quenched via electron-phonon coupling. Prior measurements and energy gap calculations reported a lifetime in YAG of 10 μs [11]. Our attempts to measure its 3 μm emission revealed the lifetime to be less than 50 μs. From the measured lifetimes we estimate the lower level density should remain less than 1% of the upper laser level density; therefore, it is reasonable to treat this as a four-level laser system.
Laser experiments
A simple end-pumped laser resonator was set up to test the Dy:YAG. A pair concave of mirrors, placed 7 cm apart, defined the resonator. The mirrors had 5 cm radii of curvature and were positioned symmetrically around the 2 cm laser rod. The predicted TEM00 radius in the rod varied from 66 µm to 72 µm. The elliptical diode pump beam described above fell withinthe TEM00 mode at the center of the rod, but overfilled the mode at the ends of the rod. Using the end-pumping model developed by Laporta and Brussard [12], we calculated a geometric overlap efficiency, ηo, of 68% for the TEM00 mode volume. Single-pass pump absorption in the rod was found to be 70% of the incident power. The back mirror had a reflectivity of 99.9% at 583nm. Both mirrors were less 5% reflective at the pump wavelength. Laser performance was measured with several output couplers as shown in Fig. 5 . Best efficiency was obtained with about 2% output coupling. Laser powers up to 150mW were generated with 1.9W of incident pump power.
The threshold for the yellow transition was higher than expected for a low-loss four-level laser, so a modified Findlay-Clay analysis was conducted to determine gain and loss [13]. For this experiment, the output coupler was replaced with a high reflector and a low-loss fused silica plate was inserted into the cavity. Small rotations of the plate away from Brewster’s angle allowed a controlled introduction of loss to the resonator. This laser cavity was pumped to threshold as described above for the full range of constant diode powers. The time required to reach laser threshold, τt, was recorded along with the fused silica plate angle and transmitted pump energy. The absorbed pump power, Pabs, was computed from the transmitted energy and pump pulse duration. The threshold excited state densities averaged over the laser mode volume, Nthres, were computed from the pump data according to:
Here Veff and η0 are the effective mode volume and pump geometric overlap efficiency calculated from the resonator geometry and measurement of the diode pump beam profile according to the model of Laporta and Brussard [12]. The pump photon energy is hvp and the storage lifetime of the upper laser level is τs. To calculate the storage time, we took advantage of the fact that the threshold excited state density depends only on the cavity losses; thus, threshold measurements at different pump powers but fixed cavity losses should produce the same threshold excitation density. This assumption allowed us to find a best fit storage lifetime for each level of cavity loss. Storage lifetimes were found to vary from 0.68ms to 0.50ms as the cavity losses were increased. The plot of threshold densities as a function of cavity loss is shown in Fig. 6 . Fitting this data to the standard expression for laser threshold:yields the laser emission cross section, σe, and bulk optical loss coefficient, δ. Here Lc is the cavity length and R is the combined round-trip reflectivity of the Brewster plate, end mirrors, and AR coatings. The emission cross section obtained from the threshold analysis, 4.1E-21 cm2 is comparable to the peak value in Fig. 3. It is also comparable to the emission cross section reported in Dy:LiNbO3 [6]. This emission cross section implies a laser saturation intensity of 140kW/cm2, which is slightly higher than the intracavity intensity for the conditions shown in Fig. 6. Resonators with increased gain-to-loss ratios are therefore indicated for improved optical efficiency.This analysis also reveals the laser rod to have relatively low bulk loss. The calculated bulk loss of 0.002 cm−1 is typical for experimental growths of single-crystal YAG. Bulk loss measurements were consistent over the course of these experiments, therefore no evidence of photo-induced absorption at the pump or laser wavelength was observed. We also found no evidence of significant excited state absorption at 447nm for intensities up to 14kW/cm2.
Summary
A new yellow laser transition was demonstrated in Dy:YAG at room temperature. This four-level system can be efficiently pumped with GaN laser diodes operating at 447nm. An optical efficiency of 8% was obtained in a non-optimal end-pumped resonator. Millisecond pulsed operation generated up to 150mW in a TEM00 beam. Output power was limited by the brightness of the diodes. Laser operation should be possible on several sharp transitions in the range from 576nm to 591nm. The emission cross section of the strongest transition at 582.7nm was calculated to be 4.1E-21cm2 from Findlay-Clay analysis and 3.0 E-21cm2 from the Füchtbauer-Ladenburg relation. The 4F9/2 fluorescence lifetime varied between 0.67ms to 0.37ms as the Dy3+ doping was increased from 1% to 3% atomic. This concentration quenching appears to be the result of a previously reported cross relaxation mechanism. No evidence of energy transfer upconversion, excited state pump absorption, or photo-darkening was observed. Future increases in GaN laser diode brightness should improve both the power and efficiency of this new laser material.
Acknowledgments
This work was supported by the Office of Naval Research.
References and links
1. S. Nakamura, S. J. Pearton, and G. Fasol, The Blue Laser Diode: The Complete Story (Springer-Verlag, 2000).
2. A. Richter, N. Pavel, E. Heumann, G. Huber, D. Parisi, A. Toncelli, M. Tonelli, A. Diening, and W. Seelert, “Continuous-wave ultraviolet generation at 320 nm by intracavity frequency doubling of red-emitting Praseodymium lasers,” Opt. Express 14(8), 3282–3287 (2006). [CrossRef] [PubMed]
3. J. Limpert, H. Zellmer, P. Riedel, G. Maze, and A. Tunnermann, “Laser oscillation in yellow and blue spectral range in Dy3+:ZBLAN,” Electron. Lett. 36(16), 1386–1387 (2000). [CrossRef]
4. A. A. Kaminskii, U. Hommerich, D. Temple, J. T. Seo, K. Ueda, S. Bagayev, and A. Pavlyulk, “Visible laser action of Dy3+ ions in monoclinic KY(WO4)2 and KGd(WO4)2 crystals under Xe-flashlamp pumping,” Jpn. J. Appl. Phys. 39(Part 2, No. 3A/B), 208–211 (2000). [CrossRef]
5. Y. Fujimoto, O. Ishii, and M. Yamazaki, “Yellow laser oscillation in Dy3-doped waterproof fluoro-aluminate glass fibre pumped by 398.8 nm GaN laser diodes,” Electron. Lett. 46(8), 586 (2010). [CrossRef]
6. P. Haro-González, L. L. Martín, I. R. Martín, G. Grazyna Dominiak-Dzik, and W. Ryba-Romanowski, “Pump and probe measurements of optical amplification at 584 nm in dysprosium doped lithium niobate crystal,” Opt. Mater. 33(2), 196–199 (2010). [CrossRef]
7. M. Klimczak, M. Malinowski, J. Sarnecki, and R. Piramidowicz, “Luminescence properties in the visible of Dy:YAG/YAG planar waveguides,” J. Lumin. 129(12), 1869–1873 (2009). [CrossRef]
8. S. R. Bowman, S. P. O'Connor, S. Biswal, N. J. Condon, and A. Rosenberg, “Diode pumped dysprosium laser materials,” Proc. SPIE 7325, 31–38 (2009).
9. P. Grünberg, S. Hufner, E. Orlich, and J. Schmitt, “Crystal field in Dysprosium garnets,” Phys. Rev. 184(2), 285–293 (1969). [CrossRef]
10. A. Lupei, V. Lupei, C. Gheorghe, A. Ikesue, and M. Enculescu, “Spectroscopic characteristics of Dy3+ doped Y3Al5O12 transparent ceramics,” J. Appl. Phys. 110(8), 083120 (2011). [CrossRef]
11. A. A. Kaminskii, Crystalline Lasers: Physical Processes and Operating Schemes (CRC Press, 1996), Chap. 4.
12. P. Laporta and M. Brussard, “Design criteria for mode size optimization in diode-pumped solid-state lasers,” IEEE J. Quantum Electron. 27(10), 2319–2326 (1991). [CrossRef]
13. D. Findlay and R. A. Clay, “The measurement of internal losses in 4-level lasers,” Phys. Lett. 20(3), 277–278 (1966). [CrossRef]
