576
OPTICS LETTERS / Vol. 36, No. 4 / February 15, 2011
10% Yb3þ -Lu2O3 ceramic laser with 74% efficiency Jas Sanghera,1,* J. Frantz,1 Woohong Kim,1 Guillermo Villalobos,1 Colin Baker,1 Brandon Shaw,1 Bryan Sadowski,2 Michael Hunt,3 Fritz Miklos,2 Austin Lutz,1 and Ishwar Aggarwal1 1
Naval Research Laboratory, Code 5620, 4555 Overlook Avenue, Washington, DC 20375, USA 2
3
GTEC Inc., 2200 Defense Highway, Suite 405, Crofton, Maryland 21114, USA
Institute of University Research Foundation, 6411 Ivy Lane, Suite 110, Greenbelt, Maryland 20770, USA *Corresponding author:
[email protected] Received October 14, 2010; accepted December 9, 2010; posted January 12, 2011 (Doc. ID 136575); published February 15, 2011
We demonstrate laser oscillation at 1080 nm with more than 16 W of output power and with an optical-to-optical slope efficiency of up to 74% using a 10% Yb3þ doped Lu2 O3 ceramic made by hot pressing. This represents the highest output power and efficiency obtained for a Yb3þ doped Lu2 O3 ceramic and demonstrates the feasibility for power scaling. © 2011 Optical Society of America OCIS codes: 140.0140, 140.3380, 140.3580, 140.3615.
Lu2 O3 and the other sesquioxides (Sc2 O3 , Y2 O3 ) possess higher thermal conductivities than yttrium-aluminum garnet [1], which is an important property for rare-earth (e.g. Yb3þ ) doped laser host materials, especially for scaling to high laser power [2]. Lu2 O3 has a thermal conductivity that is predicted from fundamental principles to be almost insensitive to the Yb3þ dopant concentration due to negligible phonon scattering, and measurements bear this out [1,3]. Consequently, Lu2 O3 would be a desirable laser host material to investigate, especially at high rareearth-dopant concentrations. Unfortunately, Lu2 O3 has a very high melting point (>2400 °C) and is difficult to make in large sizes using traditional high-temperature melt-growth techniques. However, vacuum sintering can overcome these limitations and has been used to make transparent ceramic laser materials [4–6]. Lu et al. reported a 0.15% Nd3þ doped Lu2 O3 ceramic that exhibited lasing at 1080 nm with an output power of 10 mW and an efficiency of 12% [4]. Takaichi et al. were the first to demonstrate cw lasing at 1035 nm with an output power of 700 mW and efficiency of 36% using a 3% Yb3þ doped Lu2 O3 ceramic [5]. A similar doped ceramic also exhibited pulsed lasing at 1033:5 nm (pulse width = 357 ns, rep rate ¼ 97 MHz) with an output power of 352 mW and efficiency of 32% [6]. Kaminskii et al. were first to report lasing at around 1079 nm using 3% Yb3þ dopant in lutetia and with an output power of about 250 mW [7]. To date, all the examples highlighted in the literature were made by vacuum sintering rather than hot pressing. Hot pressing could possibly provide a viable alternative pathway to manufacturing large ceramic laser materials. Consequently, we recently used hot pressing to fabricate high optical quality 10% Yb3þ :Lu2 O3 ceramic from coprecipitated powder and demonstrated lasing with 2.5 W of output power and a slope efficiency of 10% [8]. At that time, these results represented the highest doping concentration of Yb3þ used to demonstrate lasing as well as the highest output power obtained from any doped Lu2 O3 ceramic. That was also the first time hot-pressed Lu2 O3 ceramics had demonstrated laser oscillation and was even more remarkable due to the relatively large grain size of 20–50 μm. This grain size is considerably larger than the 1–2 μm size generally believed to be a 0146-9592/11/040576-03$15.00/0
prerequisite for laser oscillation in sintered ceramics as described in the literature [9]. In this Letter we demonstrate lasing with hot-pressed 10% Yb3þ :Lu2 O3 ceramics at 1080 nm with an efficiency of 74% and with more than 16 W output power. Both the efficiency and output power represent record high values for ceramic Yb3þ :Lu2 O3 . The Yb3þ :Lu2 O3 powder was made by coprecipitation following the procedure outlined in [8], but the powder was additionally ball milled [10]. The lasing data in this paper is for a sample with concentration of 10 mol:% Yb3þ relative to Lu3þ , although we did make powder and ceramics with different concentrations of Yb3þ . Ceramics were made by hot pressing the powder at about 1600 °C for 2 h using a uniform coating of a small amount of LiF sintering aid that was eliminated by evaporation prior to full densification. The hot-pressed samples were transparent, with densities greater than 99% of the theoretical value. The samples were subsequently hot isostatically pressed at about 1600 °C for 2 h under an Ar gas pressure of 30,000 psi to produce fully dense and transparent ceramics. The samples were approximately 25 mm in diameter and 3 mm thick. Absorption measurements were performed on polished ceramics using a Fourier-transform IR spectrometer. The fluorescence spectra were recorded with a quarter-meter monochrometer with silicon detector under 940 nm diode pumping. The fluorescence lifetimes of the ceramics and the powders used to fabricate them were measured by pumping at 975 nm using a low-power pulsed diode laser. To minimize absorption and reemission (resonant radiative energy transfer), which can produce artificially longer lifetimes, all measurements were performed on fine powder and ceramics that had been ground to ~250 μm particle size. Small 3 mm diameter samples of 10% Yb3þ :Lu2 O3 ceramic with 2 mm thickness were obtained by core drilling from the large 25 mm diameter samples and polishing both surfaces to a high optical quality (99:9%) at the laser wavelength of 1080 nm and high transmission at the pump wavelength of 975 nm. An antireflective coating for 1080 nm was applied to the sample’s other surface. © 2011 Optical Society of America
February 15, 2011 / Vol. 36, No. 4 / OPTICS LETTERS
The sample was wrapped along its circumference with a thin piece of indium foil and inserted into a copper heat sink that was cooled with chilled water to 15 °C. A fiber-coupled 975 nm diode laser (LIMO GmbH) with a maximum output power of 100 W was used as a pump. The pump beam was collimated and then focused to a spot with a diameter of 290 μm. A dielectric mirror with a radius of curvature of 25 cm was placed approximately 1 cm from the output surface of the sample to act as the laser’s output coupler. Several mirrors, with reflectivities of 90%, 95%, and 98% at 1080 nm were tested in order to find the optimum output coupling. The laser was operated quasi-cw by pumping with a 50% duty cycle at 127 Hz. The absorption and fluorescence spectra for the 10% Yb3þ :Lu2 O3 ceramic are shown in Fig. 1. Both spectra show the characteristic peaks associated with Yb3þ . The additional peak in the fluorescence spectrum at 940 nm is due to the pump laser and absorption/reemission. Most of the lasing data to date has been demonstrated in vacuum-sintered Lu2 O3 ceramic made with a relatively low Yb3þ concentration of 3% [5–7]. However, we recently demonstrated lasing at 1080 nm using ceramics made by hot pressing coprecipitated 10% Yb:Lu2 O3 powder [8]. The efficiency was 10% on pumping at 940 nm, and we obtained an output power of >2 W. However, in this Letter we now demonstrate lasing with an efficiency of 74% at 1080 nm by pumping at 975 nm. Figure 2 shows the laser output power versus absorbed power measured using three different output couplers. The highest slope efficiency of 74% and a maximum output power of more than 16 W were obtained using a 5% output coupler. This represents the highest output power demonstrated to date using any Yb3þ doped Lu2 O3 ceramic. We do observe a small rollover in the output power at these high power levels, indicating the presence of some thermal effects, which we will address shortly. It is also interesting to note that the current efficiency is better than that of the best Nd3þ :YAG ceramic lasers [11]. We performed a Findlay–Clay analysis [12] to estimate the cavity round-trip losses, which include both the ceramic material losses along with the system losses. We obtain a value of 1.5% for the round-trip loss. This value is
Fig. 1. Absorption and fluorescence spectra for 10 mol:% Yb3þ doped Lu2 O3 transparent ceramics.
577
Fig. 2. Output power versus absorbed power for a 10% Yb3þ doped Lu2 O3 ceramic laser using different output couplers.
lower than the 4.5% obtained previously [8], presumably due to the better optical quality of this sample. The Yb3þ ion has a low quantum defect, highlighted by pumping at 975 nm and lasing at 1080 nm, which means the maximum theoretical slope efficiency is higher than 74%. Therefore, there are loss mechanisms that limit the efficiency. At high Yb3þ ion doping levels, the energy associated with an excited ion can resonantly transfer from one ion to another until it transfers to a quenching site. This transfer process would also result in shorter fluorescence lifetimes. Figure 3 shows a plot of measured fluorescence lifetimes for the Lu2 O3 ceramics containing 2, 8, and 10 mol:% Yb3þ as well as the powders used to make the ceramics. The data set highlights two important points. First, the lifetime of the powders appears reasonably constant and insensitive to the dopant concentration from 2% to 10% Yb3þ . This indicates that the dopant is uniformly dispersed and concentration quenching is not an issue in the powder used to make the ceramics. Second, the fluorescence lifetime in the ceramic is similar (within the error bars) to the powder for up to 8% Yb3þ and decreases thereafter, indicative of concentration quenching. This effect might be related to the increased chance of energy transfer to defect sites at high Yb3þ concentration. These defect sites could be either impurity ions or grain boundary defects that not only quench the fluorescence but could also lead to nonradiative decay and therefore thermal effects. Considering that the purity of the ceramic and powder is very similar [8], the difference must be attributed to grain boundary defects. However, grain boundary defects can be quite complex and include dangling bonds and pores. In addition, impurity ions as well as the rareearth-dopant ions could preferentially be localized and enriched at the grain boundaries, whereas the rare-earth ion might be uniformly dispersed at lower concentration within the grain. A better understanding of the grain boundaries will enable us to improve the efficiency to nearly the theoretical value and allow scaling to highpower lasers. Figure 3 also compares the fluorescence lifetime data for our samples with data from the literature on singlecrystal Lu2 O3 doped with different amounts of Yb3þ .
578
OPTICS LETTERS / Vol. 36, No. 4 / February 15, 2011
losses to lower the threshold power and improve the laser performance. In conclusion, we have fabricated high optical quality 10% Yb3þ :Lu2 O3 ceramic that demonstrated lasing with more than 16 W of output power and a slope efficiency of 74%. This result represents the highest efficiency and highest output power obtained from Lu2 O3 ceramics. Future work will focus on performing a detailed analysis of the grain boundary defects, as well as lowering the dopant concentration to ≤8% which will reduce grain boundary clustering, lower the optical losses, and improve the laser efficiency and enable scaling to even higher laser powers.
Fig. 3. Fluorescence lifetimes versus Yb3þ concentration in Lu2 O3 powders and hot-pressed Lu2 O3 ceramics from this work compared with Lu2 O3 single crystals (published data). The error bars are shown for the powders and ceramics fabricated in this work.
Laversenne et al. [13] have shown that the fluorescence lifetime of their single-crystal Yb3þ :Lu2 O3 rises with Yb3þ content and peaks at around 6% concentration, after which the lifetime falls. They attribute the increase in the lifetime to absorption/reemission process we discussed earlier. They also saw a similar effect in the Y2 O3 host. Beyond 6%, their lifetime decreased due to concentration quenching. Since this is a single crystal, we can assume that these defect sites are not associated with grain boundaries. Petermann et al. [14] observed a significant decrease in the lifetime after 3% doping due to concentration quenching, indicating possibly a combination of more clustering of Yb3þ ions and significantly more impurities. However, the quenching effect was not seen by Peters et al. [15] in their Yb3þ :Lu2 O3 single crystals containing up to 10% Yb3þ . The fluorescence lifetime remained almost constant for up to 2% Yb3þ doping, and thereafter increased to almost 1 ms for 10% Yb3þ content, which they attribute to a relatively large amount of the Lu3þ sites now occupied by Yb3þ . We believe this is most likely attributed to the absorption/reemission processes due to the high concentrations of Yb3þ . Overall, there appears to be a large spread in the measured fluorescent lifetime values reported in the literature, and we believe that most of this is due to a combination of effects associated with absorption/ reemission, impurities, and concentration quenching. In our case, it appears that localization and enrichment of Yb3þ occurs at the grain boundary when the concentration is 10%, and this leads to concentration quenching and thermal effects. Therefore, we propose to study lasing in samples containing a Yb3þ concentration of 8% or less. This will be the focus of future work to reduce the
View publication stats
The authors would like to acknowledge the financial support provided by the Joint Technology Office for High Energy Lasers (JTO-HEL) and the Office of Naval Research (ONR). References 1. R. Gaume, B. Viana, and D. Vivien, Appl. Phys. Lett. 83, 1355 (2003). 2. L. Fornaseiro, E. Mix, V. Peters, K. Peterman, and G. Huber, Cryst. Res. Technol. 34, 255 (1999). 3. R. Peters, C. Krankel, K. Petermann, and G. Huber, Opt. Express 15, 7075 (2007). 4. J. Lu, K. Takaichi, T. Uematsu, A. Shirakawa, M. Musha, K. Ueda, H. Yagi, T. Yanagatani, and A. A. Kaminskii, Appl. Phys. Lett. 81, 4324 (2002). 5. K. Takachi, H. Yagi, A. Shirarkawa, K. Ueda, S. Hosokawa, T. Yanagatani, and A. A. Kaminskii, Phys. Status Solidi B 202, R1 (2005). 6. M. Tokurakawa, K. Takaichi, A. Shirakawa, K. Ueda, H. Yagi, S. Hosokawa, T. Yanagatani, and A. A. Kaminskii, Opt. Express 14, 12832 (2006). 7. A. A. Kaminskii, S. N. Bagayev, K. Ueda, K. Takaichi, A. Shirakawa, S. N. Ivanov, E. N. Khazanov, A. V. Taranov, H. Yagi, and T. Yanagitani, Laser Phys. Lett. 3, 375 (2006). 8. J. Sanghera, W. Kim, C. Baker, G. Villalobos, J. Frantz, B. Shaw, A. Lutz, B. Sadowski, R. Miklos, M. Hunt, F. Kung, and I. Aggarwal, Opt. Mater. (Amsterdam) (to be published). 9. S. Hosokawa, H. Yagi, and T. Yanagatani, “Translucent lutetium oxide sinter, and method for manufacturing same,” U.S. patent # 7,597,866 (6 October 2009). 10. W. Kim, G. Villalobos, C. Baker, J. Frantz, B. Shaw, A. Lutz, B. Sadowski, F. Kung, M. Hunt, J. Sanghera, and I. Aggarwal, J. Am. Ceram. Soc. (to be published). 11. J. Lu, M. Prabhu, J. Xu, K. Ueda, H. Yagi, T. Yanagitani, and A. Kaminskii, Appl. Phys. Lett. 77, 3707 (2000). 12. D. Findlay and R. A. Clay, Phys. Lett. 20, 277 (1966). 13. L. Laversenne, Y. Guyot, C. Goutaudier, M. T. Cohen-Adad, and G. Boulon, Opt. Mater. (Amsterdam) 16, 475 (2001). 14. K. Petermann, D. Fagundes-Peters, J. Johannsen, M. Mond, V. Peters, J. J. Romero, S. Kutovoi, J. Speiser, and A. Giesen, J. Cryst. Growth 275, 135 (2005). 15. R. Peters, C. Krankel, K. Petermann, and G. Huber, J. Cryst. Growth 310, 1934 (2008).
