A solid-state-laser based single-frequency 589 nm light source that can be easily used in the laboratory is needed for sodium spectroscopy studies and cold sodium atom experiments. This paper shows that by using a periodically poled Zn-doped LiNbO3 ridge waveguide for sum-frequency generation, we can obtain a high conversion efficiency to 589 nm light from two sub-watt 1064 and 1319 nm Nd:YAG lasers via a simple single pass wavelength conversion process without employing an enhancement cavity. A 494 mW light at 589 nm is generated and achieves overall conversion efficiency from the laser power of 41%. Excellent long-term stability of output power is obtained and its standard deviation is characterized to be 0.09%.
© 2009 OSA
Sodium atoms have long played an important role in spectroscopy and in the study of degenerate quantum gases. Dye lasers with output powers of several hundred milliwatts have mainly been used in such studies since there are no appropriate solid-state lasers for generating continuous wave (CW) at 589 nm, which corresponds to the sodium doublet. These lasers have a problem as regards stability and their operation requires daily maintenance. Moreover, considerable effort is needed to narrow their linewidth. To overcome these difficulties, using sum-frequency generation (SFG) or second harmonic generation (SHG) processes with solid-state lasers has been proposed as summarized in Table 1 [1–10]. SFG has been demonstrated with two Nd:YAG lasers operating at 1064 and 1319 nm [1–3,6,8,9], and two fiber lasers operating at 938 and 1583 nm , and SHG has been realized with a Raman fiber laser [4,5,10]. However, several watts of pump power [3–7,10] or a complex singly [4,10] or doubly [2,3,6,8] resonant enhancement cavity with a feedback locking system had to be used to obtain good conversion efficiency. This makes it difficult to use these light sources in ordinary laboratories. When employing a simple single-pass wavelength conversion configuration with sub-watt pump lasers which can be easily obtained from commercially available products, a 3.4 mW output power was obtained with 589 nm light using a bulk LiNbO3 (LN) crystal with 0.7 and 0.35 W Nd:YAG lasers . And an 18 mW output power was obtained using a bulk periodically poled LiNbO3 (PPLN) crystal with 0.8 and 0.35 W Nd:YAG lasers . These output power levels are insufficient for cold atom experiments.
In this report, we adopted a periodically poled Zn-doped LiNbO3 ridge waveguide for the SFG to obtain a high conversion efficiency to 589 nm light solely with a simple single-pass wavelength conversion process with sub-watt pump lasers. We have developed a direct bonding technique for constructing an effective LN ridge waveguide . Periodic poling relaxes the phase-matching condition and makes it possible to utilize the highest nonlinear coefficient of the crystal. The waveguide structure increases both the optical intensity of the pump beam in the nonlinear crystal and the interaction length. By using these techniques, 494 mW light at 589 nm, which matches the sodium D 2 line wavelength, can be achieved from the SFG of 1064 and 1319 nm Nd:YAG lasers solely with a simple single pass wavelength conversion process without an enhancement cavity. Here, the coupled powers of the Nd:YAG laser to the PPLN waveguide were 389 mW (1064 nm) and 245 mW (1319 nm). The photon conversion efficiency, defined as [output photon numbers at 589nm]/[PPLN coupled pump laser photon number], reached 70% for the 1064 nm photons and 90% for the 1319 nm photons. In the following discussions, we utilize above mentioned definition of the photon conversion efficiency, unless otherwise mentioned.
2. Periodically poled Zn:LiNbO3 ridge waveguide
The PPLN ridge waveguide was manufactured by NTT Electronics using a direct bonding technique that we developed . To obtain sufficient resistivity to photorefractive damage we constructed the waveguide using Zn-doped lithium niobate as the waveguide core material and lithium tantalite (LiTaO3) as a cladding layer. First, a periodically poled structure with a 9.2 μm pitch was formed in advance on the LN substrate by using a conventional electrical poling method. The two wafers were brought into contact in a clean atmosphere and then annealed at 500°C to achieve complete bonding at the atomic level. The waveguide layer thickness was reduced to 10 μm by lapping and polishing. Then 8 μm-wide ridge waveguides were fabricated using a dicing saw. Figure 1 (a) is a cross-sectional scanning electron microscope image of a fabricated ridge waveguide. We obtained smooth ridge side walls without any additional etching. The waveguide was cut to a length of 34 mm. An anti-reflective coating was deposited on both ends of the waveguide for the pump and SFG light, and then it was mounted on a metal carrier and housed in a module package with a Peltier device and thermistor for temperature management as shown in Fig. 1 (b). The optical coupling between the fiber pigtail and the PPLN chip inside the module was achieved by using a collimating lens and a focusing lens.
Since no conventional ion-exchange process is employed in these fabrication processes, our ridge waveguide exhibits strong resistance to photorefractive damage and no degradation of the nonlinear coefficient of the LN crystal. Furthermore, direct bonding and the ridge structure provides strong light confinement owing to the step index profile. Therefore, high wavelength conversion efficiency can be achieved by using our PPLN devices. Our devices are also advantageous as regards long-term reliability because, unlike with conventional proton exchange waveguides, no mobile protons are formed.
3. Experimental setup
Figure 2 shows our experimental setup. The pump light sources are commercially available CW single-frequency diode-pumped InnoLight Mephisto Nd:YAG lasers emitting at wavelengths of 1064 nm (1 W model) and 1319 nm (0.5 W model). Their spectral linewidths are about 1 kHz/100 ms. The emitted elliptical polarized light is converted to linearly polarized light with a quarter-wave plate. Then the polarization direction is adjusted with a half-wave plate. We also install a Faraday isolator between the two wave plates since the laser becomes unstable due to the back reflection of the light. The two Nd:YAG laser beams are combined by using a long-wave-pass dichroic beamsplitter, and then coupled into a polarization maintaining PANDA fiber by using an aspheric lens with a focal length of 13.9 mm. More than 80% of the power is coupled into the single mode fiber, when the position of the focusing lens is optimized for each wavelength. But the optimum lens position differs depending on the laser wavelength. Therefore, we set the lens position to maximize the coupling for a low power laser, i.e. a 1319 nm Nd:YAG laser. Under this condition, around the 60% and 80% of power is coupled into the fiber for 1064 nm and 1319 nm light, respectively. Then this fiber is connected to the fiber pigtail of the PPLN module. The coupling efficiency from the fiber pigtail to the PPLN waveguide inside the module is around 80% for the 1064 nm light and around 70% for the 1319 nm light. Obtained 589 nm emission is separated from residual pump beam with an IR cut color glass filter (HOYA HA-30). The temperature of the PPLN waveguide is monitored with a thermistor and controlled by a Peltier device with electronic feedback circuitry to meet the phase-matching condition.
4. Experimental results
Figure 3 shows the waveguide temperature dependence of the 589 nm SFG power at a low input power level. The highest 589 nm output power was obtained when the waveguide temperature was set at 41.0°C. Since the full-width at half-maximum of the phase-matching temperature was 1.0°C, the temperature control must be better than 0.1°C to obtain a stable output power.
Figure 4(a) shows the measured relationship between the 1319 nm PPLN coupled power and the 589 nm SFG power when the 1064 nm PPLN coupled power was fixed at 360 mW. Figure 4(b) shows the photon conversion efficiency. The PPLN temperature was always adjusted to maximize the 589 nm output power. Since the absorption of infrared light inside a LiNbO3 crystal is negligible , we estimated the PPLN coupled power by measuring the power transmitted from the PPLN module when only one wavelength light was injected. The generated power value at 589 nm was corrected for the loss at the IR cut filter. The 589 nm generated power increased almost linearly as the 1319 nm coupled power increased, and it reached 480 mW when the 1319 nm coupled power was 242 mW. The photon conversion efficiencies were 74% for the 1064 nm photon and 89% for the 1319 nm photon at the maximum output.
Figure 5(a) shows the measured relationship between the 1064 nm PPLN coupled power and the 589 nm SFG power when the 1319 nm PPLN coupled power was fixed at 242 mW. Figure 5(b) shows the conversion efficiency of the PPLN coupled light. When the 1064 nm coupled power was smaller than 350 mW, the SFG power increased almost linearly as the 1064 nm coupled power increased. When the 1064 nm power exceeded 350 mW and became much larger than the fixed 1319 nm coupled power, 88% of the 1319 nm photons were converted, and then the 589 nm output power decreased since the SFG back conversion process occurred. The maximum achieved photon conversion efficiencies were 77% for the 1064 nm coupled photons and 88% for the 1319 nm coupled photons.
Since one 1064 nm photon and one 1319 nm photon are converted to one 589 nm photon in the SFG process, the most efficient total conversion efficiency can be obtained when the ratio between the 1064 and 1319 nm coupled pump powers is proportional to their frequency. We measured the 589 nm generated power by changing the pump power for both 1064 nm and 1319 nm, while keeping the ratio of the 1064 nm coupled power to the 1319 nm coupled power at 1.24:1 as shown in Fig. 6(a) . The SFG power increased as the pump power increased. 76% of the 1064 and 1319 nm coupled photons were converted when their coupled powers were 299 and 241 mW (see Fig. 6(b)). The calculated normalized SFG efficiency, i.e. 100 × [generated SFG power]/([1064 nm coupled power] × [1319 nm coupled power]), is around 1500%/W when the photon conversion efficiency was less than 45% and it decreased to 570%/W in the region of highest photon conversion efficiency.
With our experiments, the maximum 589 nm light power we achieved was 494 mW when the 1064 and 1319 nm Nd:YAG laser powers after the beamsplitter were 772 and 425 mW, the fiber coupled powers were 490 and 334 mW, and the PPLN coupled powers were 389 and 245 mW, respectively. The conversion efficiency from the sum of the PPLN coupled 1064 nm laser power and 1319 nm laser power reached 78%. This means that 70% of the 1064 nm coupled photons and 90% of the 1319 nm coupled photons were converted to 589 nm photons. The overall conversion efficiency from the laser power after the beamsplitter, which includes the coupling loss at the fiber and the PPLN waveguide, is 41%. For producing trapping, cooling, and repumping beams for cold Na atom experiments, at least 400 mW laser output power is required . Therefore, our achieved output power fulfills this requirement. The currently obtainable maximum output power of the 589 nm light was limited by the 1319 nm Nd:YAG laser power. We should be able to obtain a higher output power by replacing the laser with a higher power model.
We also measured the long-term stability of the 589 nm output power. Figure 7 shows the time dependence of the 589 nm generated power and room temperature. Observed temperature oscillation around 1°C with a period of 0.2 hours is due to instability of our air conditioner. The SFG power was synchronized to the room temperature fluctuation. The stability became better after a 4-hour warm-up period. During 8 hours operation after the warm-up period, the fluctuation of the 589 nm output power is evaluated to be less than ± 0.24% and its standard deviation was only 0.09%. A very stable 589 nm power was obtained without any feedback power control. When only one of the driving lasers was injected into the module, a similar oscillatory fluctuation was also observed in the transmission from the module. We confirmed that the output power from each driving laser was stable. Thus, we think that the fluctuation originates from the modulated coupling efficiency into single mode fiber due to the periodic thermal expansion of the components. Better room temperature control and/or using a thermally stable breadboard would probably help to improve the output power stability. We also operated the module continuously for more than three weeks with 400 mW of output power and it showed no signs of degradation.
Figure 8 shows the measured far field transverse mode profile of the output 589 nm beam. A 91% fit to the Gaussian was achieved and its ellipticity was 0.98.
To adjust the SFG wavelength to the sodium D 2 resonance line, it is necessary to change the pump laser wavelength. The wavelength of the InnoLight Mephisto Nd:YAG lasers can be tuned over a 30 GHz range by changing the temperature of the Nd:YAG crystal. The wavelength tuning rage of the commercial model was 1064.12-1064.26 nm (in air) for the 1064 nm Mephisto laser and 1318.74-1318.91 nm for the 1319 nm Mephisto laser. This means they cannot cover the sodium D 2 wavelength (588.9950 nm in air) with SFG. Therefore, we used a custom made 1064 nm Mephisto laser that could operate at a higher crystal temperature. Its wavelength tuning range was 1064.21-1064.40 nm. By combining this laser with a standard 1319 nm Mephisto laser, we can achieve a wavelength tuning range of 588.937-589.032, which covers the sodium D 2 line wavelength. We measured the exact output wavelength with a High Finesse WS-7 wavelength meter and confirmed that we could obtain the sodium D 2 wavelength. The maximum achieved output power at D 2 wavelength was 494 mW. The InnoLight Mephisto laser also has a fast wavelength tuning mechanism that uses a piezoelectric actuator glued to the Nd:YAG crystal to change the resonator optical path length. Its tuning range is 200 MHz, and the response bandwidth is 100 kHz. Therefore, it would be possible to lock the 589 nm output light to the sodium D 2 line by using a feedback circuit. We did not measure the spectral linewidth of the 589 nm output but it must be similar to that of the pump lasers , namely about 1 kHz.
Since the wavelength tuning range of the Nd:YAG laser is limited, it is not possible to cover the sodium D 1 wavelength (589.5924 nm in air) using Nd:YAG lasers. To solve this problem, we also demonstrated the combination of a 1320 nm DFB laser diode (NTT Electronics NLK1B5JAAA) and a InnoLight Mephisto standard 1064 nm Nd:YAG laser (fiber coupling model). The spectrum linewidth of the DFB LD is ≤ 1 MHz. The LD wavelength can be changed from 1318.44 to 1322.16 nm by controlling the operating temperature. Therefore, both the Na D 1 (589.5924 nm in air) and Na D 2 wavelengths can be covered. Here we utilized a WDM fiber coupler to combine the LD light and the Nd:YAG laser light. We obtained a power of 52 mW at Na D 1 wavelength and 68 mW at Na D 2 wavelength when the fiber coupled power of the LD was 60 mW and that of the Nd:YAG laser was 163 mW.
We reported a solid-state-laser based 589 nm wavelength laser that corresponds to the sodium D 2 line. We utilized the periodically poled Zn-doped LiNbO3 ridge waveguide module we developed for SFG from commercially available sub-watt Nd:YAG lasers. With single pass wavelength conversion process, 494 mW output power was efficiently generated with excellent stability. This method is a lot simpler and more reliable compared with enhancement cavity technique that has been widely used. The conversion efficiencies achieved for PPLN coupled photons were 70% for the 1064 nm photons and 90% for the 1319 nm photons. The standard deviation of the output power stability was only 0.09%. The sodium D 1 wavelength light was also obtained with the combination of a 1320 nm DFB laser diode and a 1064 nm Nd:YAG laser. The dye lasers, which are commonly used at this wavelength, can be replaced with this new solid-state-laser based optical source. This source will be very useful for sodium high-resolution spectroscopy, cold atom, and coherent lidar experiments. In a cold Na atom experiment, repumping light whose wavelength is 1.71 GHz away from the D 2 wavelength is also needed, together with a cooling light that matches the Na D 2 line. In our configuration, this repumping light can be generated without an electro-optic modulator simply by adding an extra Nd:YAG laser. The full-width at half-maximum of the phase-matching pump wavelength of our PPLN waveguide is 40 GHz, which is sufficient to generate both wavelengths simultaneously. The possibility to generate two frequencies simultaneously is one of the big advantages of this method, while it is quite challenging when cavity enhanced SFG is utilized. Our PPLN ridge waveguide module is also useful for generating various other wavelengths by SFG, SHG, and difference frequency generation [15,16], which cannot be obtained directly with a solid-state laser. These PPLN ridge waveguide modules are commercially available from NTT Electronics .
T. Nishikawa would like to thank Junji Yumoto for giving him the chance to pursue this research subject and for useful advice. T.W.H. gratefully acknowledges support by the Max-Planck Foundation.
References and links
2. J. D. Vance, C.-Y. She, and H. Moosmüller, “Continuous-wave, all-solid-state, single-frequency 400-mw source at 589 nm based on doubly resonant sum-frequency mixing in a monolithic lithium niobate resonator,” Appl. Opt. 37(21), 4891–4896 (1998). [CrossRef]
3. J. C. Bienfang, C. A. Denman, B. W. Grime, P. D. Hillman, G. T. Moore, and J. M. Telle, “20 W of continuous-wave sodium D2 resonance radiation from sum-frequency generation with injection-locked lasers,” Opt. Lett. 28(22), 2219–2221 (2003). [CrossRef]
4. Y. Feng, S. Huang, A. Shirakawa, and K. Ueda, “589 nm Light Source Based on Raman Fiber Laser,” Jpn. J. Appl. Phys. 43(No. 6A), L722–L724 (2004). [CrossRef]
5. D. Georgiev, V. P. Gapontsev, A. G. Dronov, M. Y. Vyatkin, A. B. Rulkov, S. V. Popov, and J. R. Taylor, “Watts-level frequency doubling of a narrow line linearly polarized Raman fiber laser to 589nm,” Opt. Express 13(18), 6772–6776 (2005). [CrossRef]
6. N. Saito, K. Akagawa, Y. Hayano, Y. Saito, H. Takami, M. Iye, and S. Wada, “Coherent 589-nm-light Generation by Quasi-Intracavity Sum-Frequency Mixing,” Jpn. J. Appl. Phys. 44(47), L1420–L1422 (2005). [CrossRef]
7. J. W. Dawson, A. D. Drobshoff, R. J. Beach, M. J. Messerly, S. A. Payne, A. Brown, D. M. Pennington, D. J. Bamford, S. J. Sharpe, and D. J. Cook, “Multi-watt 589nm fiber laser source,” Proc. SPIE 6102, 61021F1–61021F9 (2006).
8. E. Mimoun, L. De Sarlo, J.-J. Zondy, J. Dalibard, and F. Gerbier, “Sum-frequency generation of 589 nm light with near-unit efficiency,” Opt. Express 16(23), 18684–18691 (2008). [CrossRef]
9. J. Yue, C.-Y. She, B. P. Williams, J. D. Vance, P. E. Acott, and T. D. Kawahara, “Continuous-wave sodium D2 resonance radiation generated in single-pass sum-frequency generation with periodically poled lithium niobate,” Opt. Lett. 34(7), 1093–1095 (2009). [CrossRef]
10. L. Taylor, A. Friedenauer, V. Protopopov, Y. Feng, D. B. Calia, V. Karpov, W. Hackenberg, R. Holzlöhner, W. Clements, M. Hager, F. Lison, and W. Kaenders, “20 W at 589 nm via frequency doubling of coherently beam combined 2-MHz 1178-nm CW signals amplified in Raman PM Fiber Amplifiers,” in The European Conference on Lasers and Electro-Optics and The European Quantum Electronics Conference, Technical Digest (CD) (Institute of Electrical and Electronics Engineers, 2009), paper PDA.7.
11. Y. Nishida, H. Miyazawa, M. Asobe, O. Tadanaga, and H. Suzuki, “Direct-bonded QPM-LN ridge waveguide with high damage resistance at room temperature,” Electron. Lett. 39(7), 609–610 (2003). [CrossRef]
12. Y. Furukawa, K. Kitamura, A. Alexandrovski, R. K. Route, M. M. Fejer, and G. Foulon, “Green-induced infrared absorption in MgO doped LiNbO3,” Appl. Phys. Lett. 78(14), 1970–1972 (2001). [CrossRef]
13. W. Ketterle, K. B. Davis, M. A. Joffe, A. Martin, and D. E. Pritchard, “High densities of cold atoms in a dark spontaneous-force optical trap,” Phys. Rev. Lett. 70(15), 2253–2256 (1993). [CrossRef]
15. F.-L. Hong, H. Inaba, K. Hosaka, M. Yasuda, and A. Onae, “Doppler-free spectroscopy of molecular iodine using a frequency-stable light source at 578 nm,” Opt. Express 17(3), 1652–1659 (2009). [CrossRef]
16. M. Asobe, O. Tadanaga, T. Yanagawa, T. Umeki, Y. Nishida, and H. Suzuki, “High-power mid-infrared wavelength generation using difference frequency generation in damage-resistant Zn:LiNbO3 waveguide,” Electron. Lett. 44(4), 288–289 (2008). [CrossRef]
17. NTT. Electronics, http://www.nel-world.com.