November 1, 2013 / Vol. 38, No. 21 / OPTICS LETTERS
4425
Efficient dual-wavelength Nd:LuLiF4 laser Hongqiang Li,1 Rui Zhang,1 Yulong Tang,1 Shiwei Wang,1 Jianqiu Xu,1,* Peixiong Zhang,2 Chengchun Zhao,2 Yin Hang,2 and Shuaiyi Zhang3 1
Key Laboratory for Laser Plasmas (Ministry of Education) and Department of Physics and Astronomy, Shanghai Jiao Tong University, Shanghai 200240, China 2 Key Laboratory of Materials for High Power Laser, Shanghai Institute of Optics and Fine Mechanics, Chinese Academy of Sciences, Shanghai 201800, China
3
School of Mathematics and Physics, Qiangdao University of Science and Technology, Qingdao 266061, China *Corresponding author: [email protected] Received August 5, 2013; revised September 25, 2013; accepted September 26, 2013; posted September 27, 2013 (Doc. ID 195142); published October 29, 2013
We report an efficient continuous-wave (CW) and passively Q-switched dual-wavelength Nd:LuLiF4 laser at 1314 and 1321 nm for the first time. Maximum CW output power of 6.08 W is obtained, giving an optical-to-optical conversion efficiency of 30.2% and a slope efficiency of 32.1%. Even using high doping V3 :YAG as the saturable absorber to passively Q-switch the laser, stable dual-wavelength operation remains. Maximum pulse energy extracted from the resonator is 108.7 μJ at 17.2 kHz pulse-repetition rate, and maximum peak power is 885 W. © 2013 Optical Society of America OCIS codes: (140.0140) Lasers and laser optics; (140.3070) Infrared and far-infrared lasers; (140.3480) Lasers, diode-pumped; (140.3530) Lasers, neodymium; (140.3540) Lasers, Q-switched. http://dx.doi.org/10.1364/OL.38.004425
Powerful 1.31 and 1.32 μm dual-wavelength lasers have great applications in a silver atom optical clock and terahertz generation by difference-frequency mixing [1,2]. Furthermore, harmonic conversion of 1.3 μm can be used for the generation of red and blue light, which can be utilized, for example, in display technologies, laser therapeutics, and biomedical applications. Especially, the second harmonic of the 1314 nm matches exactly the calcium intercombination line at 657.0 nm required for an optical calcium clock [3], and the fourth-order harmonic generation of the 1312 nm at 328.0 nm can be used for laser cooling [4]. The 4 F3∕2 → 4 F13∕2 transition of Nd3 is an efficient approach for obtaining 1.3 μm laser radiations. Fluoride crystal has laser emission lines simultaneously covering the 1.32 and 1.31 μm bands. Compared with oxides, fluoride crystal usually has long fluorescence, strong spectral anisotropy for emission, and negative thermal dependence of refractive index, which is beneficial for creating a negative thermal lens. Such negative lensing effect can be partly offset by the positive lens due to the thermal bulging of the rod end facet [5]. A well-known representative of fluoride crystal is Nd3 -doped YLiF4 , which has been widely used both in research and commercial devices [6,7]. In practice, 1.32 and 1.31 μm dual-wavelength operation has not been realized in Nd:YLF crystal, and Nd:YLF laser only oscillated at 1.31 μm on σ polarization in a high pump level due to strong negative thermallensing effect associated with the π polarization [8] or 1.32 μm on π polarization with complex cavity setup [9]. Another fluoride crystal Nd:LuLiF4 (Nd:LLF), like its isomorphs YLF, has emerged as a new promising laser crystal for diode pumping. Compared with Nd:YLF, Nd:LLF has several advantages; for instance, larger-emission cross section 5.1 × 10−20 cm2 for π and 2.2 × 10−20 cm2 for σ at 1.3 μm (Nd:YLF, 2–2.5 × 10−20 cm2 for the two polarizations [8]), and comparably long fluorescence lifetime of 489 μs [10]. Usually, the heat generation of 1.3 μm laser is more serious than lasing at 1.0 μm due to larger 0146-9592/13/214425-04$15.00/0
quantum defect. Benefitting from matching in its thermalexpansion coefficient and conductivity along each of the crystal axes [11], Nd:LLF has reduced astigmatic thermallensing effect and is more suitable for high-power 1.3 μm laser operation. However, most current attention was paid to the 1.0 μm laser operation of Nd:LLF [11,12]. In earlier studies, Nd:LLF laser based on 4 F3∕2 → 4 F13∕2 transition was realized with flash-lamp pumping and had low laser efficiency [13]. For passively Q-switching (PQS) operation of solidstate lasers at 1.3 μm, V3 :YAG has some remarkable advantage owing to their excellent physical and optical properties, such as high ground-state absorption cross section (σ gsa ) of 7.2 × 10−18 cm2 and low residual absorption at 1.3 μm [14]. It has been demonstrated that V3 :YAG can be an excellent saturable absorber (SA) for Nd lasers operated in the 1.3 μm [8,15–19]. However, V3 :YAG has not been used to passively Q-switched Nd:LLF at either 1.0 or 1.3 μm. This is the first time, to the best of our knowledge, report of a diode-pumped high-efficient continuous-wave (CW) Nd:LLF laser simultaneously operating at 1314 and 1321 nm with a maximum output power of 6 W. Meanwhile, the characteristics of passively Q-switched Nd:LLF laser with high doping V3 :YAG by adopting different output couplers (OCs) was also investigated. The high-quality Nd:LLF crystal was grown by the Czochralski method. The Nd3 ions concentration is 1.0 at. %, corresponding to a Nd3 ions concentration of 0.706 × 10−20 cm3 . The Nd:LLF crystal was cut along the a axis direction into small pieces of 3 mm × 3 mm × 6 mm. Both end faces of the crystal were polished and antireflection (AR) coated at 792 nm and 1.3 μm. The crystal was wrapped in an indium foil and mounted into a micro-channel copper heat-sink whose temperature was maintained at 12°C to efficiently cool the crystal and avoid thermal fracture. A schematic diagram of the experimental setup is shown in Fig. 1. The pump source was a 30 W 792 nm laser diode with a output © 2013 Optical Society of America
OPTICS LETTERS / Vol. 38, No. 21 / November 1, 2013 Coup
ling f
Laser Diode
iber
M1
Focu
sing
Nd:LLF
optic
s
M2
Output laser 3+
OC V :YAG 3+
Fig. 1. Schematic of the experimental setup for CW and PQS operation. Arm length: M1-M2∼130 mm, M2-V3 ∶YAG∼75 mm, V3 ∶YAG-OC∼5 mm.
pigtail fiber, which has a 200 μm diameter core and a numerical aperture of 0.22. The pump light was launched into the crystal with two 50 mm focal length doublets. The pump beam caustic was also measured, resulting in a pump spot radius (1∕e2 ) of 96 μm and a divergence half-angle of 192 mrad. With this focused pump beam size, the Rayleigh length inside the Nd:LLF crystal was calculated to be ∼1.83 mm. We adjusted the operation temperature of LD to force the pump wavelength to match the Nd:LLF absorption peak at 792.2 nm [10]. The measured single-pass pump absorption under nonlasing was 68%, corresponding to maximum absorbed pump power (P abs ) of 20 W. For negative thermal refractive index laser crystal, whether the thermal lens is positive or negative depends on the competition between bulging of the crystal end facets and the intensity-induced negative thermal refractive index (−dn∕dT) [5]. Usually the thermally induced lens is negative for π polarization (1321 nm laser) and positive for σ polarization (1314 nm laser) [5,20]. In order to realize efficient dual-wavelength laser emission and satisfy the stable-cavity condition for positive and negative thermal lens in high pump power, the resonator was designed to be a three-mirror folded cavity and support the TEM00 oscillation with a laser beam radius of 100 μm, which well-matched the pump beam. M1 is a plane mirror with AR at 792 and 1053 nm, high reflection (HR) at 1.3 μm. M2 is a concave mirror with a curvature radius of 100 mm HR at 1.3 μm and AR at 792 nm. Two OCs with different transmissions of 3% and 8% at 1.3 μm were used. A homemade 0.5 mm thick high doping V3 :YAG crystal was used to investigate the performance of passively Q-switching operation of Nd: LLF. The pulsed-laser signal was recorded with a Tektronix DP04054B digital oscilloscope (500 MHz bandwidth, 5 Gs/s sampling rate) and a photo-detector (PDA30B. Thorlabs Inc.). The laser spectra were measured with an optical analyzer (0.05 nm spectral resolution, AQ6315E. ANDO Inc.).
Fig. 2. Output power versus P abs for different OCs: (a) CW and (b) PQS operations, straight lines are linear fitting.
CW operation of the Nd:LLF laser was first investigated with an optimal cavity length of 210 mm. The evolution of output power with P abs for two different OCs is shown in Fig. 2(a). By adopting the 8% OC, the maximum dual-wavelength laser power of 6.08 W was achieved under the P abs of 20 W, resulting in an optical efficiency of 30.2% and slope efficiency of 32.1% (the threshold P abs was 0.5 W). Reducing the output coupling transmittance to 3%, the maximum output power and slope efficiency decreased to 4.5 W and 25.8%, respectively. The CW laser spectra recorded at different output power levels with the 8% OC is shown in Fig. 3(a). At low pump power, the laser oscillated only at 1321 nm. As soon as the P abs exceeded 2.06 W (output power output power > 0.425 W), the laser stepped into dualwavelength operation. Further increasing the pump power, full width at half-maximum (FWHM) of the laser spectrum at each wavelength became broader, but the center wavelengths did not show obvious dependence on the pump power. Under the maximum output power of 6.08 W, the emission wavelengths were centered at 1314.3 and 1321.2 nm with respective FWHM of 1.2 and 1.5 nm. By using a Glan–Taylor polarizer, we found that both laser emissions at the two wavelengths were linearly polarized. The transmission polarization direction of the Glan–Taylor polarizer was parallel to the c axis of the Nd:LLF crystal. The emission at 1321 nm had the polarization state along the c axis corresponding to the π polarization, while the polarization direction of the 1314 nm emission was along b axis (σ polarization). This shows that a simple polarizer allows us to produce an individual wavelength at 1314 or 1321 nm. Figures 3(b) and 3(c) show the wavelength dependence of the polarized emission cross-sectional spectra and different laser spectra of the Nd:LLF through the Glan–Taylor polarizer. The wavelength did not change with different OCs. Nd:LLF is a typical four-level laser, and the 1.3 μm dualwavelength operation can be understood as follows. The different wavelength laser emissions share the same upper laser level 4 F3∕2 , but their terminal laser levels
Intensity (dBm)
4426
(a)
−20 −40
−20 −40
−60 −60
−80
6.08 4.4
2.8
0.6
Output power (W) 0.4
1310
1314
1318
1322
−80 1325
Wavelength (nm)
Fig. 3. (a) CW Nd:LLF laser spectrum under the different output power. (b), (c) The transmission and reflection laser spectra after Glan–Taylor polarizer.
November 1, 2013 / Vol. 38, No. 21 / OPTICS LETTERS
are two different stark levels split from 4 F13∕2 [inset of Fig. 3(b)] because of the coupling of the lattice field of the LLF crystal to energy levels of Nd3 . The 1321 nm π polarized laser first oscillated because it has larger branching ratio and emission cross section than the 1314 nm σ polarization [10], so it had a lower lasing threshold. With increasing the pump power, a larger population inversion will increase the gain of the 4 F3∕2 → stark level 1 of 4 F13∕2 transition at 1314 nm to overcome its cavity loss [Fig. 3(c)], thus making the 1314 nm laser oscillating as well (while the gain for the 1321 nm emission is clamped). Consequently, dual-wavelength operation came into being. Although the cavity was designed for TEM00 mode with the laser beam radius of 100 μm, different thermal lenses had an effect on the laser-beam radius. The radius of the π polarized 1321 nm laser associated with negative thermal lens is smaller than the σ polarized 1314 nm laser, which has a positive thermal lens. In the stable-cavity condition, assuming the most severe thermal lens, the calculated largest waist radius difference of the two wavelength lasers would be ∼48.5 μm, corresponding to thermal lenses of −15 and 80 mm. However, readjustment of the laser cavity at the maximum pump level did not affect the output power. Therefore, we conjecture that no serious thermal lensing effect existed during the laser operation. By using the 90.0∕10.0 scanning-knife-edge method, the M 2 factor of the laser beam was measured to be around 1.5, indicating fundamental-mode operation of this dual-wavelength laser. The Nd:LLF laser shows good power stability with a root-mean-square fluctuation of about 0.94% in half an hour. To achieve low-frequency and high-energy PQS operation in the Nd:LLF laser, a 0.5 mm thick high doping V3 :YAG (AR at 1.3 μm) was inserted into the laser cavity to act as a SA. The tested initial signal transmission of V3 :YAG was about 82.4% at 1314 nm and 79.3% at 1321 nm, respectively. In our PQS system, the cross-sectional ratio of V3 :YAG to Nd:LLF (σ gsa ∕σ em ) is ∼141 for 1321 nm and ∼288 for 1314 nm. The effective laser-beam area in the Nd:LLF crystal and in the SA was in a ratio (A∕As ) of about 2.8. These conditions facilitate the second threshold condition being easily satisfied [21]. Initial experiments indicated that the PQS operation of the Nd:LLF laser was relatively sensitive to the position of V3 :YAG, and the PQS characteristics were recorded when the V3 :YAG was placed 5 mm away from the OC. Figure 2(b) shows the average output power as a function of P abs with the 3% and 8% OCs. The average output power increased linearly with P abs , and the 3% OC provided much better results than that with the 8% OC. The 3% OC produced a maximum average output power of 1.87 W with a slope efficiency of 11.2%, respecting to P abs of 18.5 W. On the contrary, the 8% OC only provided a maximum average output power of 1.06 W with a slope efficiency of 6.08%. This is mainly because a laser cavity with the 3% OC has higher intracavity laser intensity (compared with the 8% OC cavity), which can bleach the SA more completely; this lead to a lower saturable absorption loss. Therefore the operation efficiency of the case with the 3% OC was higher than that of 8% OC.
4427
Spectral evolution in the PQS operation was different from that in the CW mode. With the 3% OC, dualwavelength operation could only be achieved when P abs > W, and only the 1321 nm laser oscillated at low P abs . Compared to CW laser operation, the 1314 nm laser had much higher threshold due to saturable absorption loss induced by SA. At high pump power, the 1314 nm emission also can have enough gain for compensating the intracavity loss to oscillate; hence the dualwavelength was achieved. With the 8% OC, the dualwavelength operation could not be realized, and only the 1321 nm laser oscillated. V3 :YAG acted as not only a SA but also a frequency selector for the Nd:LLF laser. Figure 4 depicts the dependence of pulse width and repetition rate on P abs with different OCs. It can be found that the pulse width varies only slightly with P abs , and the 3% OC cavity provides narrower pulse width than that of 8% OC at identical P abs . Based on the analysis of coupled rate equations [21,22], the population inversion of the Nd:LLF depleted in a very short time, and the pulse width saturated; thus the pump power has little influence on the pulse width. The average pulse widths were roughly to be 120.1 ns with 3% OC and 129.7 ns with 8% OC. With the 8% OC cavity [Fig. 4(a)], the pulse-repetition rate increases monotonically with P abs , from 0.67 to 22.1 kHz. When the 3% OC was adopted [Fig. 4(b)], the repetition rate increases nonlinearly and saturates around 16.7 kHz when P abs exceeded 9 W, probably because both the large intracavity laser intensity and dual-wavelength oscillation influence the bleaching properties of the V3 :YAG. Meanwhile, the thermal lens of Nd:LLF and V3 :YAG slightly change the laser beam size in the SA. All of these above factors form nonlinear behavior in the pulse-repetition rate [8,23]. According to Fig. 2(b), under the maximum average output power of 1.87 W, the maximum pulse energy from the 3% OC cavity was 108.7 μJ at 17.2 kHz, corresponding to a peak power of 885 W. On the contrary, the 8% OC cavity only provided a maximum pulse energy of 48.3 μJ and peak power of 358 W. Figure 5 shows a pulse train at the repetition rate of 15.1 kHz and a typical pulse shape with a pulse width of 122 ns. The peak-to-peak intensity fluctuation was less than 8%. It is clear that two wavelengths were synchronously Q-switched by the V3 :YAG, and the time jitter is very small. The stable linearly polarized dual-wavelength lasers, which are overlapped in the time domain, can be used as a source for generation of terahertz radiation from nonlinear crystals and in silver atom optical clocks. In this Letter, high performance of CW and PQS operation of 1.3 μm Nd:LLF lasers have been investigated for
Fig. 4. Variation of the pulse width and repetition rate versus the pump power. (a) OC 3%. (b) OC 8%.
4428
OPTICS LETTERS / Vol. 38, No. 21 / November 1, 2013
Fig. 5. Train of the pulses with pulse repetition rate of 15.1 kHz; the inserted figure is the pulse profile with the width of 122 ns.
the first time. Maximum dual-wavelength laser power of 6.08 W was achieved in CW mode with a slope efficiency of 32.1%. The CW laser operated in dual-wavelength regimes (1314 and 1321 nm) with orthogonal polarizations. In pulsed operation, pulse width showed little dependence on pump power and transmission of the OC, and the typical pulse width was 120 ns. The pulse repetition rate can be tuned from 0.5 to 22.1 kHz by changing the pump power and transmission of the OCs, and maximum pulse energy of 108.7 μJ and peak power of 885 W was obtained. The 1.3 μm dual-wavelength Nd:LLF laser has potential applications in a silver atom optical clock and the generation of 1.2 THz radiation by using suitable nonlinear crystals with a Type-II phase-matching scheme. This work was supported by the National Natural Science Foundation of China (No. 61275136) and the Key National Natural Science Foundation of China (No. 61138006). References 1. Y. Louyer, F. Balembois, M. D. Plimmer, T. Badr, P. Georges, P. Juncar, and M. Himbert, Opt. Commun. 217, 357 (2003). 2. A. Saha, A. Ray, S. Mukhopadhyay, N. Sinha, P. K. Datta, and P. K. Dutta, Opt. Express 14, 4721 (2006). 3. Y. Louyer, M. D. Plimmer, P. Juncar, M. E. Himbert, F. Balembois, and P. Georges, Appl. Opt. 42, 4867 (2003).
4. G. Uhlenberg, J. Dirscherl, and H. Walther, Phys. Rev. A 62, 063404 (2000). 5. P. Hardman, W. Clarkson, G. Friel, M. Pollnau, and D. Hanna, IEEE J. Quantum Electron. 35, 647 (1999). 6. U. Keller, D. Miller, G. Boyd, T. Chiu, J. Ferguson, and M. Asom, Opt. Lett. 17, 505 (1992). 7. C. Bollig, C. Jacobs, M. D. Esser, E. Bernhard, and H. M. Von Bergmann, Opt. Express 18, 13993 (2010). 8. R. C. Botha, H. J. Strauss, C. Bollig, W. Koen, O. Collett, N. V. Kuleshov, M. J. D. Esser, W. L. Combrinck, and H. M. von Bergmann, Opt. Lett. 38, 980 (2013). 9. Y. Lü, J. Xia, X. Zhang, A. Zhang, J. Wang, L. Bao, and X. Yin, Appl. Phys. B 98, 305 (2010). 10. C. Zhao, L. Zhang, Y. Hang, X. He, J. Yin, P. Hu, G. Chen, M. He, H. Huang, and Y. Zhu, J. Appl. Phys. D 43, 495403 (2010). 11. M. Wang, S. Zhang, Y. Tang, L. Xu, Y. Wang, C. Zhao, Y. Hang, and W. Chen, Appl. Phys. B 104, 829 (2011). 12. R. Li, T. Yu, L. Zhang, W. Chen, and Y. Hang, Chin. Opt. Lett. 9, 021404 (2011). 13. N. P. Barnes, B. M. Walsh, K. E. Murray, G. J. Quarles, and V. K. Castillo, J. Opt. Soc. Am. B 15, 2788 (1998). 14. A. Malyarevich, I. Denisov, K. Yumashev, V. Mikhailov, R. Conroy, and B. Sinclair, Appl. Phys. B 67, 555 (1998). 15. H. Aman, J. Russ. Laser Res. 34, 1 (2013). 16. C. Xu, G. Li, S. Zhao, X. Li, G. Du, and L. Yin, Appl. Opt. 50, 5099 (2011). 17. F. Liu, J. He, B. Zhang, J. Xu, X. Dong, K. Yang, H. Xia, and H. Zhang, Opt. Express 16, 11759 (2008). 18. S. Zhang, H. Huang, L. Xu, M. Wang, F. Chen, J. Xu, J. He, and B. Zhao, Opt. Express 19, 1830 (2011). 19. A. Podlipensky, K. Yumashev, N. Kuleshov, H. Kretschmann, and G. Huber, Appl. Phys. B 76, 245 (2003). 20. M. Pollnau, P. Hardman, M. Kern, W. Clarkson, and D. Hanna, Phys. Rev. B 58, 16076 (1998). 21. Y. Chen, Y. Lan, and H. Chang, IEEE J. Quantum Electron. 37, 462 (2001). 22. J. Janousek, P. Tidemand-Lichtenberg, J. L. Mortensen, and P. Buchhave, Opt. Commun. 265, 277 (2006). 23. X. Li, G. Li, S. Zhao, K. Yang, T. Li, G. Zhang, K. Cheng, and X. Wang, Opt. Laser Technol. 44, 929 (2012).
4425
Efficient dual-wavelength Nd:LuLiF4 laser Hongqiang Li,1 Rui Zhang,1 Yulong Tang,1 Shiwei Wang,1 Jianqiu Xu,1,* Peixiong Zhang,2 Chengchun Zhao,2 Yin Hang,2 and Shuaiyi Zhang3 1
Key Laboratory for Laser Plasmas (Ministry of Education) and Department of Physics and Astronomy, Shanghai Jiao Tong University, Shanghai 200240, China 2 Key Laboratory of Materials for High Power Laser, Shanghai Institute of Optics and Fine Mechanics, Chinese Academy of Sciences, Shanghai 201800, China
3
School of Mathematics and Physics, Qiangdao University of Science and Technology, Qingdao 266061, China *Corresponding author: [email protected] Received August 5, 2013; revised September 25, 2013; accepted September 26, 2013; posted September 27, 2013 (Doc. ID 195142); published October 29, 2013
We report an efficient continuous-wave (CW) and passively Q-switched dual-wavelength Nd:LuLiF4 laser at 1314 and 1321 nm for the first time. Maximum CW output power of 6.08 W is obtained, giving an optical-to-optical conversion efficiency of 30.2% and a slope efficiency of 32.1%. Even using high doping V3 :YAG as the saturable absorber to passively Q-switch the laser, stable dual-wavelength operation remains. Maximum pulse energy extracted from the resonator is 108.7 μJ at 17.2 kHz pulse-repetition rate, and maximum peak power is 885 W. © 2013 Optical Society of America OCIS codes: (140.0140) Lasers and laser optics; (140.3070) Infrared and far-infrared lasers; (140.3480) Lasers, diode-pumped; (140.3530) Lasers, neodymium; (140.3540) Lasers, Q-switched. http://dx.doi.org/10.1364/OL.38.004425
Powerful 1.31 and 1.32 μm dual-wavelength lasers have great applications in a silver atom optical clock and terahertz generation by difference-frequency mixing [1,2]. Furthermore, harmonic conversion of 1.3 μm can be used for the generation of red and blue light, which can be utilized, for example, in display technologies, laser therapeutics, and biomedical applications. Especially, the second harmonic of the 1314 nm matches exactly the calcium intercombination line at 657.0 nm required for an optical calcium clock [3], and the fourth-order harmonic generation of the 1312 nm at 328.0 nm can be used for laser cooling [4]. The 4 F3∕2 → 4 F13∕2 transition of Nd3 is an efficient approach for obtaining 1.3 μm laser radiations. Fluoride crystal has laser emission lines simultaneously covering the 1.32 and 1.31 μm bands. Compared with oxides, fluoride crystal usually has long fluorescence, strong spectral anisotropy for emission, and negative thermal dependence of refractive index, which is beneficial for creating a negative thermal lens. Such negative lensing effect can be partly offset by the positive lens due to the thermal bulging of the rod end facet [5]. A well-known representative of fluoride crystal is Nd3 -doped YLiF4 , which has been widely used both in research and commercial devices [6,7]. In practice, 1.32 and 1.31 μm dual-wavelength operation has not been realized in Nd:YLF crystal, and Nd:YLF laser only oscillated at 1.31 μm on σ polarization in a high pump level due to strong negative thermallensing effect associated with the π polarization [8] or 1.32 μm on π polarization with complex cavity setup [9]. Another fluoride crystal Nd:LuLiF4 (Nd:LLF), like its isomorphs YLF, has emerged as a new promising laser crystal for diode pumping. Compared with Nd:YLF, Nd:LLF has several advantages; for instance, larger-emission cross section 5.1 × 10−20 cm2 for π and 2.2 × 10−20 cm2 for σ at 1.3 μm (Nd:YLF, 2–2.5 × 10−20 cm2 for the two polarizations [8]), and comparably long fluorescence lifetime of 489 μs [10]. Usually, the heat generation of 1.3 μm laser is more serious than lasing at 1.0 μm due to larger 0146-9592/13/214425-04$15.00/0
quantum defect. Benefitting from matching in its thermalexpansion coefficient and conductivity along each of the crystal axes [11], Nd:LLF has reduced astigmatic thermallensing effect and is more suitable for high-power 1.3 μm laser operation. However, most current attention was paid to the 1.0 μm laser operation of Nd:LLF [11,12]. In earlier studies, Nd:LLF laser based on 4 F3∕2 → 4 F13∕2 transition was realized with flash-lamp pumping and had low laser efficiency [13]. For passively Q-switching (PQS) operation of solidstate lasers at 1.3 μm, V3 :YAG has some remarkable advantage owing to their excellent physical and optical properties, such as high ground-state absorption cross section (σ gsa ) of 7.2 × 10−18 cm2 and low residual absorption at 1.3 μm [14]. It has been demonstrated that V3 :YAG can be an excellent saturable absorber (SA) for Nd lasers operated in the 1.3 μm [8,15–19]. However, V3 :YAG has not been used to passively Q-switched Nd:LLF at either 1.0 or 1.3 μm. This is the first time, to the best of our knowledge, report of a diode-pumped high-efficient continuous-wave (CW) Nd:LLF laser simultaneously operating at 1314 and 1321 nm with a maximum output power of 6 W. Meanwhile, the characteristics of passively Q-switched Nd:LLF laser with high doping V3 :YAG by adopting different output couplers (OCs) was also investigated. The high-quality Nd:LLF crystal was grown by the Czochralski method. The Nd3 ions concentration is 1.0 at. %, corresponding to a Nd3 ions concentration of 0.706 × 10−20 cm3 . The Nd:LLF crystal was cut along the a axis direction into small pieces of 3 mm × 3 mm × 6 mm. Both end faces of the crystal were polished and antireflection (AR) coated at 792 nm and 1.3 μm. The crystal was wrapped in an indium foil and mounted into a micro-channel copper heat-sink whose temperature was maintained at 12°C to efficiently cool the crystal and avoid thermal fracture. A schematic diagram of the experimental setup is shown in Fig. 1. The pump source was a 30 W 792 nm laser diode with a output © 2013 Optical Society of America
OPTICS LETTERS / Vol. 38, No. 21 / November 1, 2013 Coup
ling f
Laser Diode
iber
M1
Focu
sing
Nd:LLF
optic
s
M2
Output laser 3+
OC V :YAG 3+
Fig. 1. Schematic of the experimental setup for CW and PQS operation. Arm length: M1-M2∼130 mm, M2-V3 ∶YAG∼75 mm, V3 ∶YAG-OC∼5 mm.
pigtail fiber, which has a 200 μm diameter core and a numerical aperture of 0.22. The pump light was launched into the crystal with two 50 mm focal length doublets. The pump beam caustic was also measured, resulting in a pump spot radius (1∕e2 ) of 96 μm and a divergence half-angle of 192 mrad. With this focused pump beam size, the Rayleigh length inside the Nd:LLF crystal was calculated to be ∼1.83 mm. We adjusted the operation temperature of LD to force the pump wavelength to match the Nd:LLF absorption peak at 792.2 nm [10]. The measured single-pass pump absorption under nonlasing was 68%, corresponding to maximum absorbed pump power (P abs ) of 20 W. For negative thermal refractive index laser crystal, whether the thermal lens is positive or negative depends on the competition between bulging of the crystal end facets and the intensity-induced negative thermal refractive index (−dn∕dT) [5]. Usually the thermally induced lens is negative for π polarization (1321 nm laser) and positive for σ polarization (1314 nm laser) [5,20]. In order to realize efficient dual-wavelength laser emission and satisfy the stable-cavity condition for positive and negative thermal lens in high pump power, the resonator was designed to be a three-mirror folded cavity and support the TEM00 oscillation with a laser beam radius of 100 μm, which well-matched the pump beam. M1 is a plane mirror with AR at 792 and 1053 nm, high reflection (HR) at 1.3 μm. M2 is a concave mirror with a curvature radius of 100 mm HR at 1.3 μm and AR at 792 nm. Two OCs with different transmissions of 3% and 8% at 1.3 μm were used. A homemade 0.5 mm thick high doping V3 :YAG crystal was used to investigate the performance of passively Q-switching operation of Nd: LLF. The pulsed-laser signal was recorded with a Tektronix DP04054B digital oscilloscope (500 MHz bandwidth, 5 Gs/s sampling rate) and a photo-detector (PDA30B. Thorlabs Inc.). The laser spectra were measured with an optical analyzer (0.05 nm spectral resolution, AQ6315E. ANDO Inc.).
Fig. 2. Output power versus P abs for different OCs: (a) CW and (b) PQS operations, straight lines are linear fitting.
CW operation of the Nd:LLF laser was first investigated with an optimal cavity length of 210 mm. The evolution of output power with P abs for two different OCs is shown in Fig. 2(a). By adopting the 8% OC, the maximum dual-wavelength laser power of 6.08 W was achieved under the P abs of 20 W, resulting in an optical efficiency of 30.2% and slope efficiency of 32.1% (the threshold P abs was 0.5 W). Reducing the output coupling transmittance to 3%, the maximum output power and slope efficiency decreased to 4.5 W and 25.8%, respectively. The CW laser spectra recorded at different output power levels with the 8% OC is shown in Fig. 3(a). At low pump power, the laser oscillated only at 1321 nm. As soon as the P abs exceeded 2.06 W (output power output power > 0.425 W), the laser stepped into dualwavelength operation. Further increasing the pump power, full width at half-maximum (FWHM) of the laser spectrum at each wavelength became broader, but the center wavelengths did not show obvious dependence on the pump power. Under the maximum output power of 6.08 W, the emission wavelengths were centered at 1314.3 and 1321.2 nm with respective FWHM of 1.2 and 1.5 nm. By using a Glan–Taylor polarizer, we found that both laser emissions at the two wavelengths were linearly polarized. The transmission polarization direction of the Glan–Taylor polarizer was parallel to the c axis of the Nd:LLF crystal. The emission at 1321 nm had the polarization state along the c axis corresponding to the π polarization, while the polarization direction of the 1314 nm emission was along b axis (σ polarization). This shows that a simple polarizer allows us to produce an individual wavelength at 1314 or 1321 nm. Figures 3(b) and 3(c) show the wavelength dependence of the polarized emission cross-sectional spectra and different laser spectra of the Nd:LLF through the Glan–Taylor polarizer. The wavelength did not change with different OCs. Nd:LLF is a typical four-level laser, and the 1.3 μm dualwavelength operation can be understood as follows. The different wavelength laser emissions share the same upper laser level 4 F3∕2 , but their terminal laser levels
Intensity (dBm)
4426
(a)
−20 −40
−20 −40
−60 −60
−80
6.08 4.4
2.8
0.6
Output power (W) 0.4
1310
1314
1318
1322
−80 1325
Wavelength (nm)
Fig. 3. (a) CW Nd:LLF laser spectrum under the different output power. (b), (c) The transmission and reflection laser spectra after Glan–Taylor polarizer.
November 1, 2013 / Vol. 38, No. 21 / OPTICS LETTERS
are two different stark levels split from 4 F13∕2 [inset of Fig. 3(b)] because of the coupling of the lattice field of the LLF crystal to energy levels of Nd3 . The 1321 nm π polarized laser first oscillated because it has larger branching ratio and emission cross section than the 1314 nm σ polarization [10], so it had a lower lasing threshold. With increasing the pump power, a larger population inversion will increase the gain of the 4 F3∕2 → stark level 1 of 4 F13∕2 transition at 1314 nm to overcome its cavity loss [Fig. 3(c)], thus making the 1314 nm laser oscillating as well (while the gain for the 1321 nm emission is clamped). Consequently, dual-wavelength operation came into being. Although the cavity was designed for TEM00 mode with the laser beam radius of 100 μm, different thermal lenses had an effect on the laser-beam radius. The radius of the π polarized 1321 nm laser associated with negative thermal lens is smaller than the σ polarized 1314 nm laser, which has a positive thermal lens. In the stable-cavity condition, assuming the most severe thermal lens, the calculated largest waist radius difference of the two wavelength lasers would be ∼48.5 μm, corresponding to thermal lenses of −15 and 80 mm. However, readjustment of the laser cavity at the maximum pump level did not affect the output power. Therefore, we conjecture that no serious thermal lensing effect existed during the laser operation. By using the 90.0∕10.0 scanning-knife-edge method, the M 2 factor of the laser beam was measured to be around 1.5, indicating fundamental-mode operation of this dual-wavelength laser. The Nd:LLF laser shows good power stability with a root-mean-square fluctuation of about 0.94% in half an hour. To achieve low-frequency and high-energy PQS operation in the Nd:LLF laser, a 0.5 mm thick high doping V3 :YAG (AR at 1.3 μm) was inserted into the laser cavity to act as a SA. The tested initial signal transmission of V3 :YAG was about 82.4% at 1314 nm and 79.3% at 1321 nm, respectively. In our PQS system, the cross-sectional ratio of V3 :YAG to Nd:LLF (σ gsa ∕σ em ) is ∼141 for 1321 nm and ∼288 for 1314 nm. The effective laser-beam area in the Nd:LLF crystal and in the SA was in a ratio (A∕As ) of about 2.8. These conditions facilitate the second threshold condition being easily satisfied [21]. Initial experiments indicated that the PQS operation of the Nd:LLF laser was relatively sensitive to the position of V3 :YAG, and the PQS characteristics were recorded when the V3 :YAG was placed 5 mm away from the OC. Figure 2(b) shows the average output power as a function of P abs with the 3% and 8% OCs. The average output power increased linearly with P abs , and the 3% OC provided much better results than that with the 8% OC. The 3% OC produced a maximum average output power of 1.87 W with a slope efficiency of 11.2%, respecting to P abs of 18.5 W. On the contrary, the 8% OC only provided a maximum average output power of 1.06 W with a slope efficiency of 6.08%. This is mainly because a laser cavity with the 3% OC has higher intracavity laser intensity (compared with the 8% OC cavity), which can bleach the SA more completely; this lead to a lower saturable absorption loss. Therefore the operation efficiency of the case with the 3% OC was higher than that of 8% OC.
4427
Spectral evolution in the PQS operation was different from that in the CW mode. With the 3% OC, dualwavelength operation could only be achieved when P abs > W, and only the 1321 nm laser oscillated at low P abs . Compared to CW laser operation, the 1314 nm laser had much higher threshold due to saturable absorption loss induced by SA. At high pump power, the 1314 nm emission also can have enough gain for compensating the intracavity loss to oscillate; hence the dualwavelength was achieved. With the 8% OC, the dualwavelength operation could not be realized, and only the 1321 nm laser oscillated. V3 :YAG acted as not only a SA but also a frequency selector for the Nd:LLF laser. Figure 4 depicts the dependence of pulse width and repetition rate on P abs with different OCs. It can be found that the pulse width varies only slightly with P abs , and the 3% OC cavity provides narrower pulse width than that of 8% OC at identical P abs . Based on the analysis of coupled rate equations [21,22], the population inversion of the Nd:LLF depleted in a very short time, and the pulse width saturated; thus the pump power has little influence on the pulse width. The average pulse widths were roughly to be 120.1 ns with 3% OC and 129.7 ns with 8% OC. With the 8% OC cavity [Fig. 4(a)], the pulse-repetition rate increases monotonically with P abs , from 0.67 to 22.1 kHz. When the 3% OC was adopted [Fig. 4(b)], the repetition rate increases nonlinearly and saturates around 16.7 kHz when P abs exceeded 9 W, probably because both the large intracavity laser intensity and dual-wavelength oscillation influence the bleaching properties of the V3 :YAG. Meanwhile, the thermal lens of Nd:LLF and V3 :YAG slightly change the laser beam size in the SA. All of these above factors form nonlinear behavior in the pulse-repetition rate [8,23]. According to Fig. 2(b), under the maximum average output power of 1.87 W, the maximum pulse energy from the 3% OC cavity was 108.7 μJ at 17.2 kHz, corresponding to a peak power of 885 W. On the contrary, the 8% OC cavity only provided a maximum pulse energy of 48.3 μJ and peak power of 358 W. Figure 5 shows a pulse train at the repetition rate of 15.1 kHz and a typical pulse shape with a pulse width of 122 ns. The peak-to-peak intensity fluctuation was less than 8%. It is clear that two wavelengths were synchronously Q-switched by the V3 :YAG, and the time jitter is very small. The stable linearly polarized dual-wavelength lasers, which are overlapped in the time domain, can be used as a source for generation of terahertz radiation from nonlinear crystals and in silver atom optical clocks. In this Letter, high performance of CW and PQS operation of 1.3 μm Nd:LLF lasers have been investigated for
Fig. 4. Variation of the pulse width and repetition rate versus the pump power. (a) OC 3%. (b) OC 8%.
4428
OPTICS LETTERS / Vol. 38, No. 21 / November 1, 2013
Fig. 5. Train of the pulses with pulse repetition rate of 15.1 kHz; the inserted figure is the pulse profile with the width of 122 ns.
the first time. Maximum dual-wavelength laser power of 6.08 W was achieved in CW mode with a slope efficiency of 32.1%. The CW laser operated in dual-wavelength regimes (1314 and 1321 nm) with orthogonal polarizations. In pulsed operation, pulse width showed little dependence on pump power and transmission of the OC, and the typical pulse width was 120 ns. The pulse repetition rate can be tuned from 0.5 to 22.1 kHz by changing the pump power and transmission of the OCs, and maximum pulse energy of 108.7 μJ and peak power of 885 W was obtained. The 1.3 μm dual-wavelength Nd:LLF laser has potential applications in a silver atom optical clock and the generation of 1.2 THz radiation by using suitable nonlinear crystals with a Type-II phase-matching scheme. This work was supported by the National Natural Science Foundation of China (No. 61275136) and the Key National Natural Science Foundation of China (No. 61138006). References 1. Y. Louyer, F. Balembois, M. D. Plimmer, T. Badr, P. Georges, P. Juncar, and M. Himbert, Opt. Commun. 217, 357 (2003). 2. A. Saha, A. Ray, S. Mukhopadhyay, N. Sinha, P. K. Datta, and P. K. Dutta, Opt. Express 14, 4721 (2006). 3. Y. Louyer, M. D. Plimmer, P. Juncar, M. E. Himbert, F. Balembois, and P. Georges, Appl. Opt. 42, 4867 (2003).
4. G. Uhlenberg, J. Dirscherl, and H. Walther, Phys. Rev. A 62, 063404 (2000). 5. P. Hardman, W. Clarkson, G. Friel, M. Pollnau, and D. Hanna, IEEE J. Quantum Electron. 35, 647 (1999). 6. U. Keller, D. Miller, G. Boyd, T. Chiu, J. Ferguson, and M. Asom, Opt. Lett. 17, 505 (1992). 7. C. Bollig, C. Jacobs, M. D. Esser, E. Bernhard, and H. M. Von Bergmann, Opt. Express 18, 13993 (2010). 8. R. C. Botha, H. J. Strauss, C. Bollig, W. Koen, O. Collett, N. V. Kuleshov, M. J. D. Esser, W. L. Combrinck, and H. M. von Bergmann, Opt. Lett. 38, 980 (2013). 9. Y. Lü, J. Xia, X. Zhang, A. Zhang, J. Wang, L. Bao, and X. Yin, Appl. Phys. B 98, 305 (2010). 10. C. Zhao, L. Zhang, Y. Hang, X. He, J. Yin, P. Hu, G. Chen, M. He, H. Huang, and Y. Zhu, J. Appl. Phys. D 43, 495403 (2010). 11. M. Wang, S. Zhang, Y. Tang, L. Xu, Y. Wang, C. Zhao, Y. Hang, and W. Chen, Appl. Phys. B 104, 829 (2011). 12. R. Li, T. Yu, L. Zhang, W. Chen, and Y. Hang, Chin. Opt. Lett. 9, 021404 (2011). 13. N. P. Barnes, B. M. Walsh, K. E. Murray, G. J. Quarles, and V. K. Castillo, J. Opt. Soc. Am. B 15, 2788 (1998). 14. A. Malyarevich, I. Denisov, K. Yumashev, V. Mikhailov, R. Conroy, and B. Sinclair, Appl. Phys. B 67, 555 (1998). 15. H. Aman, J. Russ. Laser Res. 34, 1 (2013). 16. C. Xu, G. Li, S. Zhao, X. Li, G. Du, and L. Yin, Appl. Opt. 50, 5099 (2011). 17. F. Liu, J. He, B. Zhang, J. Xu, X. Dong, K. Yang, H. Xia, and H. Zhang, Opt. Express 16, 11759 (2008). 18. S. Zhang, H. Huang, L. Xu, M. Wang, F. Chen, J. Xu, J. He, and B. Zhao, Opt. Express 19, 1830 (2011). 19. A. Podlipensky, K. Yumashev, N. Kuleshov, H. Kretschmann, and G. Huber, Appl. Phys. B 76, 245 (2003). 20. M. Pollnau, P. Hardman, M. Kern, W. Clarkson, and D. Hanna, Phys. Rev. B 58, 16076 (1998). 21. Y. Chen, Y. Lan, and H. Chang, IEEE J. Quantum Electron. 37, 462 (2001). 22. J. Janousek, P. Tidemand-Lichtenberg, J. L. Mortensen, and P. Buchhave, Opt. Commun. 265, 277 (2006). 23. X. Li, G. Li, S. Zhao, K. Yang, T. Li, G. Zhang, K. Cheng, and X. Wang, Opt. Laser Technol. 44, 929 (2012).
