Energy transfer upconversion induced thermal effects are mainly responsible for the low efficiency of laser diode pumped Er:YAG lasers. The current work adopts Er:YAG rods with 0.25% Er3+ doping concentration, instead of the commonly used rods with 0.5% Er3+ doping concentration. Results show that the thermal effect is greatly alleviated. A continuous-wave output of 10.2 W is obtained using 31 W incident pump power. Optical–optical efficiency is approximately 33%. Slope efficiency, with respect to the absorbed pump power, is as high as 83%, which is close to the quantum efficiency. In a Q-switched operation, 7 mJ pulses with a pulse width of ~65 ns are obtained at 100 Hz PRF.
©2011 Optical Society of America
Multi-watt lasers emitting in the eye-safe band have important applications in many aspects, such as range finding, spectroscopy, Doppler wind lidar, and so on [1,2]. One of the most promising eye-safe lasers, 1645 nm hybrid fiber-bulk erbium lasers, has been proven to be effective in achieving high output power and good beam quality [3–7]. The best results were reported to be a 60 W output with a slope efficiency of 80% for a continuous-wave (CW) operation  and 30 mJ pulses with a pulse width less than 20 ns at 20 Hz PRF for a Q-switched operation . This two-step method, however, adds complexity, weight, and volume. In addition, the optical–optical efficiency of the two-step system is limited by the efficiency of the first step: 1532 nm fiber lasers pumped by 976 nm laser diodes (LDs). Consequently, only 28% optical–optical efficiency is achieved for the two-step system [5,7].
Due to its simple structure and potential for relatively high optical–optical efficiency, room temperature LD-directly pumped 1645 nm Er:YAG lasers have been developed intensively in recent years [4,8–10]. High efficiency is prevented by two serious problems. First, there is conflict between the broad bandwidth of LD (~10 nm ) and the narrow absorption bandwidth of Er:YAG crystals (~1 nm for the highest peak, around 1532 nm). Second, the energy transfer upconversion (ETU) induces a strong thermal effect.
Considering that Er:YAG has a relatively broader absorption bandwidth at around 1470 nm, compared with that at around 1532 nm, LDs lasing 1470 nm lasers are often used, although this method has lower quantum efficiency [4,8–10]. CW operations with 0.5% Er3+ doped Er:YAG slab have obtained 6.1 W CW output with approximately 33 W incident pump power, resulting in an optical–optical efficiency of 19% . One of the few studies  on Q-switched operations pumped in this manner reported 38 mJ pulses with pulse widths of 60 ns, but the authors admitted that the pumping system used was highly inefficient.
Volumetric Bragg grating was used to obtain narrow-band 1470 and 1532 nm pump sources with a loss of less than 10%. By adopting the former, 80 W CW output was achieved with a 0.5% Er3+ doped Er:YAG slab [11,12]. The maximum optical–optical efficiency for both pump sources was around 25%. Narrow-band 1532 nm LDs (less than 1 nm) have recently become commercially available, and they have been used as pump sources for Er:YAG lasers using 0.5% Er3+ doped conventional rod crystals  and novel fiber-like crystals  as gain media. In the latter work, 14.5 W CW output was obtained with approximately 42 W incident pump power, resulting in an optical–optical efficiency of approximately 35%. Q-switched pulses with 8 mJ pulse energy and 70 ns pulse width were also generated.
Table 1 shows a comparison of LD-pumped Er:YAG lasers in previous works. Other lasers suffer from another problem: strong thermal effects, which eventually call for careful heat treatment with slab geometry [10–12] or cause power drops at high pump powers [13,14]. The optical–optical and slope efficiencies are much lower than the quantum efficiency (~90%). The strong thermal effect in Er:YAG crystals is explained as a result of ETU [3,5,6,15], which is quadratic in the upper laser level population. Therefore, Er:YAG crystals with low Er3+ doping concentration are required. Due to good modal matching, ETU can be solved with 0.5% doped Er:YAG crystals in hybrid fiber-bulk erbium lasers. Further decreases in the Er3+ doping concentration show slight improvements in efficiency [15,16]. LD-pumped 0.5% doped Er:YAG lasers, however, suffer from poorer modal matching, as the result of the much larger pump light divergence. Under lasing conditions, the upper laser level population of the pumped areas mismatched with laser modes is higher than that of matched areas, thereby causing stronger ETU. This phenomenon eventually leads to severe thermal effects.
The current work investigates the performance of LD directly-pumped Er:YAG lasers with a 0.25% Er3+ doped Er:YAG rod. The results are compared with those of a 0.5% Er3+ doped Er:YAG rod with the same crystal length. The performance of the Q-switched 0.25% Er3+ doped Er:YAG rod is also studied and presented.
2. Experimental setup
Figure 1 shows the experimental setup. The pump source used was a 1532 nm fiber-coupled LD with 35 W maximum output, approximately 1 nm bandwidth, 200 μm core diameter, and 0.22 numerical aperture. The fiber was imaged into the crystal by a 1:4 telescope, providing an 800 μm pump spot diameter on the crystal. Three Er:YAG rods with a diameter of 4 mm were investigated. Two of them (3 and 4 cm long) were 0.25% Er3+ doped and the other one (3 cm long) was 0.5% Er3+ doped. The lengths of the rods were chosen to ensure efficient pump absorption for the 0.5% Er3+ doped rod (>80%), while keeping a reasonable modal matching between the pump and laser beams. The 0.25% Er3+ doped, 4 cm rod was chosen to compensate for the relatively lower pump absorption. A plane-concave cavity with 500 mm curvature was adopted for all rods. Temperature for the cooling water of LD was set to 298 K, with an error of 1 K, to ensure efficient absorption of the pump power at maximum level. The temperature of crystals was controlled at 288 K.
3. Results and discussions
In the CW operation, the acousto-optic modulator (AOM) was removed and the cavity length was fixed at 5 cm to 6 cm for all Er:YAG rods. The transmittance of the output couplers (OCs) was set to 5% after comparing the performances of the 0.5% doped, 3 cm rod with different OC transmittances (5%, 10%, and 14%). An additional experiment was conducted for the 0.5% doped, 3 cm rod by reversing the concave output coupling mirror into a convex one to compensate for its serious thermal effect and, therefore, achieve better modal matching. The curvature of the convex mirror was equivalent to approximately –330 mm, taking the refractive index of the OC into consideration. Figure 2 shows the output power as a function of the absorbed pump power.
Using approximately 31 W incident pump power yielded 8 W CW output for the 0.5% Er3+ doped, 3 cm rod; 9.7 W for the 0.25% Er3+ doped, 3 cm rod; and 10.2 W for the 0.25% Er3+ doped, 4 cm rod, with a wavelength of 1645 nm. Optical–optical efficiencies were 26%, 31%, and 33%, respectively. Under non-lasing conditions, the absorption rate at 31 W incident pump power was measured to be 83% for the former rod, but only 38.9% and 45.5% for the latter two. Thus, average slope efficiencies of 40%, 88%, and 83%, respectively, with respect to the absorbed pump power were obtained. Approximately 85% of the unabsorbed pump power was reflected back to the rods by OC and partly absorbed, which increased the calculated slope efficiency. However, this influence is small, considering the much larger spot size of the reflected pump power in the rods. The focal lengths of thermal lens at the highest incident pump power were measured to be approximately 150 mm for the 0.5% Er3+ doped rod and larger than 500 mm for the two 0.25% Er3+ doped rods. The results indicate that the heat generation in the two latter rods was greatly alleviated, which corresponds to the two latter rods’ much higher slope efficiency with respect to absorbed pump power. For focal length measurement, the length of the cavity was increased by moving the OC until the stable cavity becomes unstable at the highest pump power. This method may not be very precise, but the difference between the two focal lengths was large enough for appropriate judgment. Distinct from previous works pumped by narrow-band 1532 nm LDs, in which power drops occur at higher pump powers [13,14], further power scaling in the current work can be realized by adding pump power, considering the unsaturated curve in Fig. 2 and the relatively smaller thermal effect observed when using 0.25% doped rods.
A plane-convex cavity was also used for the 0.5% doped, 3 cm rod for better modal matching, but the results (as in Fig. 2) showed no improvement in efficiency. The measured focal length of the 0.5% doped rod was so short that use of the plane-convex cavity did not result in much improvement in modal matching.
Improvements in the lasers with 0.25% doped rods can be attributed to three reasons. First, the ETU was alleviated by lower doping concentration. Thus, the energy waste caused by the ETU was reduced. Second, alleviation of the ETU resulted in less heat generation, thus lengthening the focal length of the thermal lens. This increased the spot size of the fundamental laser mode and reduced the pump areas mismatched by the laser modes. Therefore, the ETU was reduced further. Third, since the lower laser level was the upper Stark sub-level of 4I15/2 , its population density decreased due to the decrease in the temperature of the 0.25% doped rods. Hence, the re-absorption loss is reduced.
Compared with the work by Chen et al., in which 74% of pump power was absorbed by a 0.25% doped, 4 cm rod , the absorption of the pump power for the same rod in the current work was much lower (44.5%). The relatively broader bandwidth of the LD used (1 nm at maximum power) compared with the fiber laser (0.2 nm) in Ref.  is one of the reasons for the lower pump power absorption, because the highest absorption peak of Er:YAG around 1532 nm is only approximately 1 nm. However, this alone does not explain the much lower absorption coefficient of the 0.25% doped rods compared with that the 0.5% doped rod in the current work. With incident power in a range of 0 to 31 W, the calculated average absorption coefficient of the 0.5% doped, 3 cm rod was approximately 3.5 times higher than that of the 0.25% doped, 3 cm rod. This may due to the varying extent of ETU in these two rods, because ETU will accelerate the recovery of the lower laser level population and, therefore, increase absorption. Additional work is required to understand this process better.
In the Q-switched operation, the 0.25% doped, 3 cm rod was chosen for its high slope efficiency and maximum output power. OCs with three different transmittances of 14%, 21%, and 30%, were used. No laser output was generated with the last transmittance. For the OC with a transmittance of 21%, only the performance under the short cavity length was obtained and recorded. The pulse energy and pulse width as a function of pulse repetition frequency (PRF) is shown in Figs. 3(a) and 3(b), respectively. Under 14% transmittance, 22 cm cavity length, and 8.2 W absorbed power, 7 mJ pulse with a pulse width of 65 ns was obtained.
Figure 4 shows the average output power for both CW and Q-switched operations at T = 14% and L = 20 cm. The energy loss due to inefficient energy storage is negligible at 1 KHz PRF, but it becomes sufficiently strong at a PRF of less than 500 Hz.
The storage efficiency for a CW pumped Q-switch laser can be expressed as17]. Using Eq. (1), we calculated the upper laser level lifetime in the current experiments as a function of absorbed pump power, as shown in Fig. 5 . The experimental condition for the data in Fig. 5 was fixed at 14% OC transmittance and 20 cm cavity length. The upper laser level lifetime declined quickly from 5.5 ms to 2 ms with increasing absorbed power from 4.6 W to 8.2 W. This stimulated lifetime explains the energy loss at 500 Hz PRF in Fig. 4. The published Er radiative lifetime is 6.9 ms , which is far from the actual life time measured in the current work (2 ms at 8.2 W absorbed power) and in many other pervious works (around 2 ms to 3 ms) [9,10,16]. Two distinct explanations were proposed for the decrease in the upper laser level lifetime, namely ETU [3,5,6,15] and amplified spontaneous emission (ASE) . The current experimental results show that the decrease in lifetime is related to the increase in absorbed power, that is, the increase in the upper laser level population. This relativity can be explained by either theory. Nevertheless, the higher upper laser level population caused mass heat generation in the current experiment’s CW operation, and could only be explained by the ETU. Stronger ETU must also occur in the Q-switched operation, which had an increased upper laser level population as well. In fact, ASE is merely suggested as a possible reason and has not been confirmed by experimental data . Moreover, ASE increases the upper laser level lifetime, according to the numerical calculation, considering that the spontaneous emission can be reabsorbed . Thus, the decrease in the upper laser level lifetime may be mainly due to the ETU.
In conclusion, decreases in the Er3+ doping concentration were effective in alleviating the thermal effect and optimizing LD-pumped Er:YAG lasers. When the 0.25% doped, 4 cm Er:YAG rod was used, over 10 W 1645 nm laser output was obtained with 31 W incident pump power. The optical–optical efficiency was approximately 33% and the efficiency with respect to the absorbed pump power was 83%. In a Q-switched operation, 7 mJ pulses with a pulse width of 65 ns were obtained at 100 Hz PRF. The pump absorption in the 0.25% doped rods was notably inefficient, but a higher optical–optical efficiency can be expected if efficient pump absorption can be realized. Due to the large pump light divergence, direct increases in the rod length may not be effective. In fact, a 0.25% doped, 7 cm rod was adopted but lower efficiency was obtained. The combination of novel fiberlike rods and 0.25% Er3+ doping concentration may be a method for higher efficiency, because it allows the use of longer rods with both efficient pump power absorption and good laser modal matching. This will be a goal for future studies.
We acknowledge Jiqiao Liu and Shuaiyi Zhang for their technical support and the Chinese Academy of Sciences Innovation Fund (Grant No: CXJJ-10-M50) for their financial support.
References and links
1. Y. Liu, J. Liu, and W. Chen, “Eye-safe, single-frequency pulsed all-fiber laser for Doppler wind lidar,” Chin. Opt. Lett. 9(9), 090604–090607 (2011). [CrossRef]
2. C. Gao, S. Zhu, W. Zhao, Z. Cao, and Z. Yang, “Eye-safe, high-energy, single-mode all-fiber laser with widely tunable repetition rate,” Chin. Opt. Lett. 7(7), 611–613 (2009). [CrossRef]
4. S. D. Setzler, M. P. Francis, Y. E. Young, J. R. Konves, and E. P. Chicklis, “Resonantly pumped eyesafe erbium lasers,” IEEE J. Sel. Top. Quantum Electron. 11(3), 645–657 (2005). [CrossRef]
6. K. Spariosu, V. Leyva, R. A. Reeder, and M. J. Klotz, “Efficient Er:YAG laser operating at 1645 nm and 1617 nm,” IEEE J. Quantum Electron. 42(2), 182–186 (2006). [CrossRef]
7. J. W. Kim, D. Y. Shen, J. K. Sahu, and W. A. Clarkson, “Fiber-laser-pumped Er:YAG lasers,” IEEE J. Sel. Top. Quantum Electron. 15(2), 361–371 (2009). [CrossRef]
8. D. Garbuzov, I. Kudryashov, and M. Dubinskii, “110W(0.9J) pulsed power from resonantly diode-laser-pumped 1.6-μm Er:YAG laser,” Appl. Phys. Lett. 87(12), 121101 (2005). [CrossRef]
9. I. Kudryashov and A. Katsnelson, “Q-switch resonantly diode-pumped Er:YAG laser,” Proc. SPIE 7578, 75781D (2009).
10. N. W. H. Chang, N. Simakov, D. J. Hosken, J. Munch, D. J. Ottaway, and P. J. Veitch, “Resonantly diode-pumped continuous-wave and Q-switched Er:YAG laser at 1645 nm,” Opt. Express 18(13), 13673–13678 (2010). [CrossRef] [PubMed]
11. I. Kudryashov, A. Katsnelson, N. Ter-Gabrielyan, and M. Dubinskii, “Room temperature power scalability of the diode-pumped Er:YAG eye-safe laser,” in Conference on Lasers and Electro-Optics/International Quantum Electronics Conference, OSA Technical Digest (CD) (Optical Society of America, 2009), paper CWA2.
12. I. Kudryashov, N. Ter-Gabrielyan, and M. Dubinskii, “Resonantly diode-pumped Er:YAG laser: 1470-nm vs. 1530-nm CW pumping case,” Proc. SPIE 7325, 732505 (2009). [CrossRef]
13. M. Eichhorn, “High-power resonantly diode-pumped CW Er3+:YAG laser,” Appl. Phys. B 93(4), 773–778 (2008). [CrossRef]
15. J. W. Kim, J. I. Mackenzie, and W. A. Clarkson, “Influence of energy-transfer-upconversion on threshold pump power in quasi-three-level solid-state lasers,” Opt. Express 17(14), 11935–11943 (2009). [CrossRef] [PubMed]
16. D. W. Chen, M. Birnbaum, P. M. Belden, T. S. Rose, and S. M. Beck, “Multiwatt continuous-wave and Q-switched Er:YAG lasers at 1645 nm: performance issues,” Opt. Lett. 34(10), 1501–1503 (2009). [CrossRef] [PubMed]
17. N. P. Barnes, “Solid-state lasers from an efficiency perspective,” IEEE J. Sel. Top. Quantum Electron. 13(3), 435–447 (2007). [CrossRef]
18. S. A. Payne, L. L. Chase, L. K. Smith, W. L. Kway, and W. F. Krupke, “Infrared cross-section measurements for crystals doped with Er3+, Tm3+, and Ho3+,” IEEE J. Quantum Electron. 28(11), 2619–2630 (1992). [CrossRef]
19. M. Eichhorn, “Numerical modeling of diode-end-pumped high-power Er3+:YAG lasers,” IEEE J. Quantum Electron. 44(9), 803–810 (2008). [CrossRef]