We report on the highly efficient, resonantly diode-pumped Er:YAG-core, double-clad, all-crystalline eye-safe waveguide laser. A 500 × 500 μm Er3+(1%):YAG single-crystalline core with an ultra low numerical aperture (NA) of ~0.02 was surrounded by a 700 × 700 μm undoped single-crystalline YAG cladding. The entire Er:YAG/YAG core/clad structure was over-clad by transparent magnesium aluminum spinel (MgAl2O4) ceramic. The waveguide was continuously (CW) clad-pumped by a spectrally-narrowed, fiber-coupled InGaAsP/InP laser diode module at ~1532 nm. We achieved 25.4 W of output power at 1645 nm with a beam quality of M2 ~2.6. The achieved 56.6% slope efficiency with respect to the absorbed pump was derived by factoring out scattering loss of the pump light in the inner cladding. With a wavelength-selective cavity, the waveguide laser delivered ~8 W of output power at 1616.6 nm. To the best of our knowledge, it is the first reported laser experiment with a crystalline Er3+:YAG-core and a truly double-clad crystalline waveguide structure.
© 2012 OSA
Solid state lasers with a waveguiding gain element were introduced soon after the first laser was launched . In his original implementation E. Snitzer utilized a round cross-section solid clad/core glass structure (double-clad fiber, with the second cladding simply being the ambient air). Originally, this structure was aimed predominantly at achieving good beam quality based on a low numerical aperture (NA) of the core. This round step-index waveguide architecture has become the mainstream design of all glass-based fiber lasers and amplifiers (e.g., ). In their “true double-clad” implementation (the second cladding is a low-index polymer coating) the large cladding NA and the relatively large cladding diameter (versus core) greatly facilitate pump power scaling and pump confinement within the fiber, while the small waveguide core diameter and its small NA serve the purpose of forming a nearly diffraction-limited output beam.
The idea of substituting low thermal conductivity glass in fibers with a high thermal conductivity single crystal for further power scaling has led to fiber architectures in the form of relatively thin and long crystalline rods manufactured by either laser heated pedestal growth (LHPG)  or micro-pulling-down  techniques. These air-clad, rod-like fibers have a very high NA, so they are utilized to facilitate the collection and confinement of a laser diode pump emission along the entire rod-fiber length [4, 5]. But in this case, the laser’s good beam quality cannot be achieved without a proper external cavity, which compromises the original idea of forming an output beam with waveguiding alone. In addition, a spatial matching of pump and laser modes, formed by the cavity design, is never complete. For that reason the rod-like fibers will always have to remain short, thus, forfeiting the benefits of the fiber’s large surface area to volume ratio-for efficient thermal management.
Recently, successful efforts in true double-clad (DC) crystalline core fibers were reported, where the DC structure was fabricated using the co-drawing laser-heated pedestal growth (CDLHPG) technique (e.g., [6, 7]). The first cladding was made of glass, which can provide a good laser beam quality through core waveguiding. However, glass introduces a much higher thermal resistance between the crystalline core and the heat sink - a major hindrance for power scaling.
To this day, the idea of fully crystalline double-clad waveguide devices for bulk solid-state lasers has only been proven feasible when researchers turned away from conventional round cross-section clad/core structures to rectangular shapes – a planar waveguide or a channel waveguide. Among the best achievements was a Yb:YAG or Nd:YAG planar waveguide, directly clad-pumped by two laser diode bars . This laser demonstrated an optical-to-optical slope efficiency of 50% and was single-mode in the guided direction (~8 μm) and highly multimode in the unguided direction (~500 μm). In another demonstration, a Yb:YAG channel waveguide with the nearly square core-clad structure was directly pumped into the core by a single laser diode bar . It yielded a nearly diffraction-limited output in both directions and a slope efficiency of 43% with respect to the absorbed pump. In both experiments, transverse pumping was used. The end pumping of the planar waveguide = structure was successfully implemented in . In the [8–10] the gain elements were fabricated using adhesive-free bonding (AFB) of the thin, Nd- or Yb-doped YAG cores and the undoped YAG cladding to form a planar, waveguiding structure. Based on some experimental demonstrations pursuing significant power scaling (e.g., [8–11]), AFB is a promising technique for fabrication of true double-clad and even triple-clad, fully crystalline, bulk structures (Depending on the source, the AFB process is also referred to in literature as the “diffusion bonding” , “thermal bonding”  or “direct bonding” ). Available data on utilizing AFB-fabricated double-clad waveguides for power scaling is still scarce, requiring further investigation of these structures for better understanding of their potential.
Here we report our first laser results obtained with a fully crystalline, double-clad, channel-waveguide structure, fabricated by AFB, which we consider to be a first step on the path toward a highly scalable, single-mode, waveguide laser. We demonstrated a resonantly clad-pumped, waveguide laser with an Er3+:YAG core. It operated at 1645 nm with slope efficiency of over 56% by the absorbed pump when directly pumped at ~1532 nm by a low-brightness, fiber coupled, InGaAsP/InP laser diode module. To the best of our knowledge, this is the first reported experiment with a Er3+-doped, true double-clad, crystalline, channel waveguide structure.
2. Experimental setup
The investigated 30 mm long, double-clad waveguide gain element was fabricated using the AFB process. The waveguide has a square active core made of a single-crystalline Er3+(1%):YAG with a 500 × 500 μm cross-section, see Fig. 1 . The core was clad by an undoped YAG forming the first cladding with a 700 × 700 μm cross-section. The refractive index of the core is higher than the refractive index of an undoped YAG in the first cladding by ~1.04 × 10−4. This difference provides the required waveguiding. Taking the refractive index of Er:YAG at 1532 nm as 1.8073 , the core has an ultra-low NA of ~0.02. The outer cladding with the 1 × 8 mm cross-section was fabricated from transparent spinel ceramic (MgAl2O4) with a refractive index of 1.695 at 1532 nm. Thus, the NA of the first cladding for the pump emission (1532 nm) was ~0.627. The large NA allowed us to use pump beams with an angular divergence of up to ~78 degrees for pump delivery into the cladding. The rendering of the entire double-clad structure with its dimensions is depicted in Fig. 1(b). Both ends of the Er:YAG waveguide were anti-reflection (AR) coated for the spectral range of 1500 - 1680 nm. The waveguide was clamped between two water-cooled, copper plates and conductively cooled from the top and the bottom. The temperature of the cooling water was maintained at 18 C.
A simplified optical layout of our experimental set-up is shown in Fig. 2 . The pump source was a spectrally narrowed (~2 nm full width half maximum, FWHM), fiber coupled, water cooled, InGaAsP/InP laser diode module (FCLDM). The fiber core had a 1 mm diameter and a 0.22 NA. The wavelength of this diode laser could be temperature-tuned by varying the coolant temperature of the stack. A combination of a polarizer and a half-wave plate inside the laser diode module was used as a variable pump power attenuator.
The pump radiation was collimated by a lens L1 with a focal length of F1 = 40 mm. The collimated pump beam was focused through a flat dichroic mirror M1 (T > 98% at 1520 - 1540 nm, R > 99.55% at 1590 - 1650 nm) into the inner cladding of the Er:YAG waveguide by an aspheric lens L2. Three different aspheric lenses with focal lengths of F2 = 30 mm, 25.4 mm and 20 mm were used in the L2 position, which resulted in a pump NA of 0.425, 0.5 and 0.635, respectively. The best performance was obtained with a F2 = 20 mm. In this case, the pump beam diameter on the waveguide face was ~650 μm (at e−2 level). The laser cavity was formed by a flat dichroic mirror, M1, and a concave output coupler (OC), M2. Radii of curvature (RoC) of the OC varied between 50 and 100 mm and their reflectivities varied between 85% and 68%. The cavity length was set to 35 mm for the OC with a RoC of 50 mm and to 110 mm for the OC with the RoC of 100 mm.
An absorption spectrum of Er:YAG at room temperature (see Fig. 3 ), associated with resonant 4I15/2→4I13/2 transitions of the Er3+ ion, has several maxima between 1524 and 1540 nm which can be used for pumping. The pumping wavelength of the FCLDM was adjusted to the vicinity of the strongest 1532 nm absorption peak, see Fig. 3. The pump wavelength was monitored by an optical spectrum analyzer (Yokogawa, model AQ6370C) with the spectral resolution set to 0.1 nm.
3. Experimental results: Er:YAG double-clad waveguide laser with nonselective cavity
The optimized CW performance of the resonantly pumped, double-clad Er:YAG waveguide laser is presented in Fig. 4 . The 35 mm long, plano-concave laser cavity used an OC with the RoC of 50 mm and ~85% reflectivity. The best optical-to-optical slope efficiency with respect to the incident pump power was 31.4%. This efficiency was lower than we expected based on our preliminary modeling, therefore, we carefully analyzed the power balance of this laser. Despite the considerable reduction in the effective pump absorption of the double-clad, Er:YAG-core waveguide (per clad-to-core ratio), the pump saturation effects can be important for resonantly pumped Er:YAG . The absorption of Er-doped crystals is usually saturated in the gain medium without lasing and experiences partial or full “recovery” to its unsaturated level with growing intensity of the laser emission in the cavity. To evaluate the absorbed pump power correctly, we measured the incident and the transmitted pump power as well as the laser output power simultaneously. It should be noted, that in the case of the clad-pumped waveguide laser, the calculated absorbed pump power, determined as the difference between the incident and the transmitted pump power, can be severely affected by the scattering losses in the inner cladding. In order to estimate the possible contribution of this effect we separately measured the passive transmission of the waveguide. A widely tunable, narrow-bandwidth (1 MHz) semiconductor laser source (Santec, model TLS-210) was tuned to the 1524 nm (or 1540 nm) wavelength, where there is practically no Er3+ absorption (see Fig. 3). Its output beam was shaped to reproduce the NA and dimensions of an actual pump beam at the entrance face of the waveguide. We also assumed that passive losses in the Er:YAG core at pumping (1532 nm) and lasing (1645 nm) wavelengths are the same.
The passive transmission of the waveguide measured in this manner was found to be 78%. After taking into account the residual reflection of the AR coatings at these wavelengths (~1.0% on each waveguide end, according to the vendor’s specification), we arrived at the passive loss figure of α = 0.07 cm−1 (which translates to a ~30 dB/m). This steep loss figure can explain the relatively low waveguide laser efficiency versus the incident pump power. In this situation the laser output versus the absorbed pump power is a better representation of laser efficiency and scaling potential.
Figure 5 indicates the Er:YAG channel waveguide laser output power at 1645 nm versus the absorbed pump power at 1532 nm, calculated using the procedure described above and taking into account the measured pump scattering loss. The best slope efficiency of 56.6% with respect to the absorbed pump power was achieved with 85% reflectivity of the OC. A lower slope efficiency of 50.3% was achieved with 72% OC reflectivity. The maximum achieved CW output power at 1645 nm was 25.4 W. The measured fraction of the absorbed pump power during lasing varied from ~52% to ~63% depending on the pump power density. As seen in Fig. 5, the dependence is perfectly linear, so power scaling of the laser at this point is strictly pump limited.
With the expected multimode nature of the Er:YAG core in mind, we measured the output beam divergence with a IR camera (Spiricon, model LW230). The camera was placed in the focal plane of the concave reflector with a 500 mm radius of curvature (f = 250 mm). We found that at the maximum output power of the waveguide laser, the beam divergence was ~13.4 mrad in both directions. The cavity length was L = 35 mm and the output corresponded to the multimode regime with M2 ~2.6. This beam quality is comparable to that achieved in the guided direction with a 35 μm-core, single-mode planar waveguide . Compared to a rod-like fiber laser , which demonstrated beam quality of M2 ~15, our laser has 2.5 times higher brightness, despite having 10-times lower output power.
With a longer laser cavity of 110 mm and the output coupler with the RoC of 100 mm, by slightly misaligning the OC, it was possible to achieve nearly TEM00 operation, but at a considerably lower (3 - 4 times) output power. A measured angular divergence with misaligned cavity was ~6 mrad. Typical intensity distributions in the focal plane of the reflector (far field) for the highly multimode (left) and nearly-TEM00-mode beams are shown in Fig. 6 .
4. Experimental results: Er:YAG double-clad waveguide laser with selective cavity
The pump power density, which was achieved in the active Er:YAG core by cladding-pumping with a FCLDM, is very high, in excess of 20 kW/cm2. In fact, it is higher than the pump power density reported in most of the end-pumped Er:YAG lasers, even those pumped by Er-fiber lasers (e.g., ). This feature is very important for Er:YAG lasers designed to exploit the 1617-nm laser transition of Er3+, for the following reasons. Both 1645-nm and 1617-nm laser lines correspond to different inter-Stark transitions between the energy levels of 4I13/2 and 4I15/2 manifolds of Er3+ in YAG. The terminal level of the 1617 nm laser transition is much closer to the ground state, and its Boltzmann population at room temperature is ~0.036, versus ~0.021 for the 1645 nm one. As a result, the 1617 nm lasing has higher ground-state absorption (GSA) loss. This difference in GSA loss is not fully compensated by the higher emission cross-section of the 1617 nm transition, especially at low pump densities. Therefore, in order to “force” the Er:YAG waveguide laser to operate at 1617 nm, it is necessary to use a wavelength-selective cavity . Alternatively, one can achieve relatively larger inversion density which will help to suppress the 1645 nm lasing . In our experiment we used a 35 mm long cavity formed by a flat nonselective HR mirror, and a concave, wavelength-selective output coupler (ROC = 62% at 1617 nm, ROC = 21% at 1645 nm) with the RoC of 50 cm. With this cavity, the laser operates only at ~1617 nm (the peak of laser emission was measured at 1616.6 nm, Δλ ~0.6 nm at e−2 level). Figure 7 shows the output power of the Er:YAG waveguide laser operating at 1616 nm versus the incident pump power.
As can be seen from Fig. 7, the laser output grew linearly with the pump power, reaching a maximum of 8.05 W (slope efficiency ~13%). Due to a limited choice of available concave output couplers with the required wavelength selectivity, this laser has not been optimized. The accurate measurements of the absorbed pump power in these experiments were also not feasible because the wavelength-selective OC had a very low transmission at the pump wavelength (less than 0.01%), and the transmitted pump power could not be determined correctly.
In conclusion, we demonstrated, what is believed to be, the first resonantly diode-pumped, Er:YAG-core, double-clad, channel waveguide laser. The waveguide was designed with the single-crystalline Er3+(1%):YAG core, undoped single-crystalline YAG cladding and overclad with transparent magnesium aluminum spinel (MgAl2O4) ceramic. The resonant clad-pumping was performed by a spectrally narrowed (~2 nm FWHM), fiber coupled InGaAsP/InP laser diode module at ~1532 nm. The highly efficient waveguide laser delivered a pump power limited ~25.4 W of the multimode CW output power at 1645 nm. Despite the multimode gain element core used in this first experiment, the measured beam quality of M2 ~2.6 was achieved with the slope efficiency of 56.6% with respect to the absorbed pump power. This efficiency factors out heavy scattering loss (~0.07 cm−1) of the pump light in the inner cladding owing to marginal second cladding material quality. The same waveguide laser, but with an un-optimized wavelength-selective cavity, delivered ~8 W of CW output power at 1616.6 nm. These results, obtained with the fully crystalline, truly double-clad, channel waveguide, are our first step on the path toward a highly power scalable, nearly diffraction-limited waveguide laser.
References and links
1. E. Snitzer, “Optical maser action of Nd3+ in a barium crown glass,” Phys. Rev. Lett. 7(12), 444–446 (1961). [CrossRef]
2. J. W. Dawson, M. J. Messerly, R. J. Beach, M. Y. Shverdin, E. A. Stappaerts, A. K. Sridharan, P. H. Pax, J. E. Heebner, C. W. Siders, and C. P. J. Barty, “Analysis of the scalability of diffraction-limited fiber lasers and amplifiers to high average power,” Opt. Express 16(17), 13240–13266 (2008). [CrossRef] [PubMed]
3. R. S. Feigelson, W. L. Kway, and R. K. Route, “Single-Crystal Fibers by the Laser-Heated Pedestal Growth Method,” Opt. Eng. 24, 1102–1107 (1985).
4. D. Sangla, I. Martial, N. Aubry, J. Didierjean, D. Perrodin, F. Balembois, K. Lebbou, A. Brenier, P. Georges, O. Tillement, and J.-M. Fourmigué, “High power laser operation with crystal fibers,” Appl. Phys. B 97(2), 263–273 (2009). [CrossRef]
5. C. Bibeau, R. J. Beach, S. C. Mitchell, M. A. Emanuel, J. Skidmore, C. A. Ebbers, S. B. Sutton, and K. S. Jancaitis, “High-Average-Power 1-μm Performance and Frequency Conversion of a Diode-End-Pumped Yb:YAG Laser,” IEEE J. Quantum Electron. 34(10), 2010–2019 (1998). [CrossRef]
6. C.-Y. Lo, K.-Y. Huang, J.-C. Chen, S.-Y. Tu, and S.-L. Huang, “Glass-clad Cr4+:YAG crystal fiber for the generation of superwideband amplified spontaneous emission,” Opt. Lett. 29(5), 439–441 (2004). [CrossRef] [PubMed]
7. C.-C. Lai, C.-P. Ke, S.-K. Liu, D.-Y. Jheng, D.-J. Wang, M.-Y. Chen, Y.-S. Li, P. S. Yeh, and S.-L. Huang, “Efficient and low-threshold Cr4+:YAG double-clad crystal fiber laser,” Opt. Lett. 36(6), 784–786 (2011). [CrossRef] [PubMed]
8. R. J. Beach, S. C. Mitchell, H. E. Meissner, O. R. Meissner, W. F. Krupke, J. M. McMahon, W. J. Bennett, and D. P. Shepherd, “Continuous-wave and passive Q-switched cladding-pumped planar waveguide lasers,” Opt. Lett. 26(12), 881–883 (2001). [CrossRef] [PubMed]
11. X. Mu, H. Meissner, H.-C. Lee, and M. Dubinskii, “True Crystalline Fibers: Double-Clad LMA Design Concept of Tm:YAG-Core Fiber and Mode Simulation,” Proc. SPIE 8237, 82373M (2012). [CrossRef]
13. Y. Sato and T. Taira, “Saturation factors of pump absorption in solid-state lasers,” IEEE J. Quantum Electron. 40(3), 270–280 (2004). [CrossRef]
14. J. Mackenzie, “An efficient high-power 946 nm Nd:YAG planar waveguide laser,” Appl. Phys. B 97(2), 297–306 (2009). [CrossRef]
15. X. Délen, S. Piehler, J. Didierjean, N. Aubry, A. Voss, M. A. Ahmed, T. Graf, F. Balembois, and P. Georges, “250 W single-crystal fiber Yb:YAG laser,” Opt. Lett. 37(14), 2898–2900 (2012). [CrossRef] [PubMed]
16. K. Spariosu, V. Leyva, R. A. Reeder, and M. J. Klotz, “Efficient Er:YAG Laser Operating at 1645 and 1617 nm,” IEEE J. Quantum Electron. 42(2), 182–186 (2006). [CrossRef]
17. S. D. Setzler, M. P. Francis, Y. E. Young, J. R. Konves, and E. P. Chicklis, “Resonantly pumped eyesafe erbium,” IEEE J. Sel. Top. Quantum Electron. 11(3), 645–657 (2005). [CrossRef]