We report on efficient single-pass, high-power second-harmonic generation in a periodically poled MgO-doped planar waveguide using a distributed Bragg reflector tapered diode laser as a pump source. A coupling efficiency into the planar waveguide of 73% was realized, and of visible laser light at was generated. Corresponding optical and electro-optical conversion efficiencies of 26% and 8.4%, respectively, were achieved. Good agreement between the experimental data and the theoretical predictions was observed.
© 2011 Optical Society of America
High-power laser light sources in the visible spectral range with good beam quality () and narrow spectral linewidth () are desired for many applications, e.g., spectroscopy or biotechnology. Such powerful lasers can also be used in display technology. The frequency doubling of diode laser near-IR (NIR) radiation in a single-pass configuration is a promising concept for realizing efficient and miniaturized devices with the aforementioned specifications, as was shown in [1, 2].
In recent years, second-harmonic generation (SHG) in channel waveguides [3, 4], bulk crystals [5, 6], as well as in a planar waveguide , all pumped by diode lasers, has already been reported. SHG in planar waveguides is most promising for frequency doubling of diode laser radiation. It bypasses the limitation of the nonlinear material at high intensities present in channel waveguides  and enables more efficient utilization of diode laser radiation during the SHG process than a bulk crystal. Hence, it enables a compromise between high output power and high conversion efficiency and therefore needs to be studied accurately with respect to optimal laser beam parameters.
SHG in planar waveguides has already been investigated theoretically  and experimentally [7, 10, 11, 12]. Until now only at with an optical conversion efficiency of 7.6% has been reported  when a diode laser was applied as an NIR source. Maximal SH powers of and at with an optical conversion efficiency of 40% and 60%, respectively, were reached with a solid-state laser and a fiber amplifier .
In this Letter we report on efficient high-power single-pass SHG in a periodically poled MgO-doped lithium niobate (ppMgO:LN) planar waveguide using a distributed Bragg reflector (DBR) tapered diode laser at as a pump source. First we present the experimental setup and describe the diode laser and the nonlinear crystal. In the following part we present the experimental results and compare them with theoretical predictions. Finally, we conclude the Letter with a summary.
The experimental setup for frequency doubling is presented in Fig. 1. As an NIR light source, a 6-mm-long DBR tapered diode laser  was chosen. The diode laser consisted of a 1-mm-long passive DBR section and two separately contacted active sections: a 1-mm-long index-guided ridge-waveguide (RW) section and a 4-mm- long gain-guided tapered (TA) section. The RW was wide, and the taper full angle was . The laser was mounted on a CuW heat spreader, p side up on a conduction-cooled packaging and operated at .
The laser beam was collimated in the fast axis with an aspheric lens, L1 (). An additional cylindric lens, L2 (), was used to collimate the laser radiation in the slow axis and compensate the astigmatism induced by the tapered section. The first half-wave plate (HWP), H1, and the optical isolator made it possible to adjust the NIR power in front of the crystal without changing the properties of the laser beam. The second HWP, H2, was used to adjust the polarization of the NIR radiation. The laser beam was focused with a cylindric lens, L3 (), and an acylindric lens, L4 (cylindric lens with corrected surface profile for minimal cylindrical aberration, ), in the slow and fast axes, respectively. All optics applied in front of the crystal had an antireflective (AR) coating for .
The waveguide-integrated quasi-phase-matching (QPM) structure for SHG was obtained from a 500-μm-thick z-cut MgO:LN wafer ( at ). The wafer had one side polished down to and prepared for periodical poling, as described in . The periodical poling was performed in a poling jig with electrolyte electrodes  at room temperature by applying an external electric field generated by a voltage of over . Sample examination with a polarization microscope and chemical etching of identical samples made it possible to determine the structure duty cycle to be very close to 50%. Subsequently, the polished surface was bonded to an LN substrate using an epoxy, which was the bottom cladding layer in the final waveguide structure. The other side of the bonded wafer was secondarily polished to have a thickness of , with a thickness uniformity better than . Subsequently, the wafer was deep dry-etched by a modified reactive ion etching (RIE) technique, forming ridge waveguides with different widths and a thickness of . The residual slab thickness of on both waveguide sides results in better sample reliability and a lower propagation loss for channel waveguides. Finally, another LN substrate was bonded to the structure’s top layer with an epoxy, which formed the upper cladding layer.
In this work, a 190-μm-wide waveguide was applied for SHG. The propagation loss was estimated by a simulation to be and for 1064 and , respectively. The periodically poled region and the crystal were 12 and long, respectively. Both waveguide facets were optically polished under and AR-coated for 1064 and .
The radiation behind the crystal was collimated with an aspheric lens, L5 (). Two dichroic mirrors were used to separate the second-harmonic and fundamental waves. The applied filter system featured a total transmission of 0.88 and for the SH and NIR radiation, respectively.
During the experiment, the RW section and the TA section of the DBR tapered laser were biased with and , respectively. The laser beam waist radius was and in the fast and slow axes, respectively. An astigmatism of relative to air was measured. The NIR output power was , which resulted in behind the optical isolator and the collimating and focusing lenses. The laser emitted at a peak wavelength of in a longitudinal single mode. The beam quality parameter, , was estimated through a caustic measurement in front of the crystal according to intensity levels and resulted in 1.2 and 1.1 for the fast and slow axes, respectively. The lateral central lobe contained 83% of the total NIR power.
Initially, the laser beam was focused onto the waveguide facet. The focusing lenses were chosen according to a simulation in order to achieve a maximal coupling efficiency into the waveguide ground mode and a maximal SHG conversion efficiency . The measured beam waist radii of and showed good agreement with the simulated optimal counterparts. The coupling efficiency at the extraordinary polarization of 73% was determined by measuring the NIR power behind and in front of the crystal and incorporating the propagation loss. The deviation from the theoretical coupling efficiency for a Gaussian beam of 99% could result from the non-Gaussian beam in the experiment, the vertically asymmetric beam profile in focus due to the asymmetric diode laser waveguide core, and the angled crystal facets.
The dependence of the SH power on the crystal temperature is presented in Fig. 2. At an NIR power level of the maximal SH power was reached at a crystal temperature of with a corresponding acceptance bandwidth of (FWHM). The experimental results deviate only slightly from the results of the simulation performed according to . This confirms the relatively high homogeneity of the periodical poling and of the waveguide thickness, which has an influence on the shape of the temperature curve and on the conversion efficiency of the SHG process .
Subsequently, the lateral beam waist was placed in the middle of the crystal in order to maximize the conversion efficiency of the SHG process . The SH power dependence on the NIR power in front of the crystal is presented in Fig. 3, where the experimental data (filled circles) are compared with the theoretical results. We calculated the normalized conversion efficiency, , for SHG in a planar waveguide according to Table 1 in  with , , , simulated , , and . Because the theoretical model was developed for a Gaussian beam, in the simulation only the NIR power content with () coupled into the waveguide () was employed: . The quadratic dependence (Fig. 3, dashed curve) determined in  shows good agreement with the experimental data only in the low NIR power range (). An empirically derived dependence (solid curve), which takes the NIR power depletion into consideration, shows good agreement with the experiment in the whole range. The small deviation from the theoretical predictions in the experiment could result from a minor inhomogeneity in the periodical poling or in the waveguide thickness (Fig. 2). Although the temperature of the crystal mount had to be gradually reduced with the increasing SH output power by in order to satisfy the QPM condition, no saturation of the generated SH power in the investigated NIR power range could be observed. A maximal SH power of was generated in the planar waveguide structure, which resulted in a maximal SH power of measured behind the dichroic mirrors. At the maximal power level of , corresponding optical and electro-optical conversion efficiencies of 26% and 8.4%, respectively, were achieved. Higher efficiencies are expected if a longer planar waveguide is applied.
The SH spectrum at the maximal power of is presented in Fig. 4. The central wavelength, the spectral linewidth, and the side-mode suppression ratio (SMSR) were determined to be , (FWHM), and , respectively.
In conclusion, with the presented setup, an efficient high-power frequency doubling of a DBR tapered diode laser in a ppMgO:LN planar waveguide was carried out. A coupling efficiency of 73% was achieved, and of laser radiation at was generated. This is, to the best of our knowledge, the highest power level generated by means of single-pass SHG of NIR diode laser radiation in a waveguide structure. The corresponding optical and electro-optical conversion efficiencies of 26% and 8.4% were realized, respectively. Good agreement between the experimental data and the theoretical predictions was observed.
The authors are grateful to K.-H. Hasler and B. Sumpf for fruitful discussions on diode laser properties. This work was supported by the German Federal Ministry of Education and Research in the project InnoProfile (grant 03IP613).
2. C. Fiebig, A. Sahm, M. Uebernickel, G. Blume, B. Eppich, K. Paschke, and G. Erbert, Opt. Express 17, 22785 (2009). [CrossRef]
3. H. K. Nguyen, M. H. Hu, Y. Li, K. Song, N. J. Visovsky, S. Coleman, and C.-E. Zah, Proc. SPIE 6890, 68900I-1 (2008).
6. C. Fiebig, G. Blume, M. Uebernickel, D. Feise, C. Kaspari, K. Paschke, J. Fricke, H. Wenzel, and G. Erbert, IEEE J. Sel. Top. Quantum Electron. 15, 978 (2009). [CrossRef]
8. D. Jedrzejczyk, R. Güther, K. Paschke, B. Eppich, and G. Erbert, IEEE Photon. Technol. Lett. 22, 1282 (2010). [CrossRef]
9. D. Fluck, IEEE J. Quantum Electron. 35, 53 (1999). [CrossRef]
10. D. Fluck, B. Binder, M. Küpfer, H. Looser, C. Buchal, and P. Günter, Opt. Commun. 90, 304 (1992). [CrossRef]
11. L. Zhang, P. J. Chandler, P. D. Townsend, Z. T. Alwahabi, S. L. Pityana, and A. J. McCaffery, J. Appl. Phys. 73, 2695 (1993). [CrossRef]
12. K. Sakai, Y. Koyata, S. Itakura, and Y. Hirano, J. Lightwave Technol. 27, 590 (2009). [CrossRef]
13. B. Sumpf, K.-H. Hasler, P. Adamiec, F. Bugge, J. Fricke, P. Ressel, H. Wenzel, G. Erbert, and G. Tränkle, Proc. SPIE 7230, 72301E-1 (2009).
14. S. W. Kwon, W. S. Yang, H. M. Lee, W. K. Kim, H. Y. Lee, W. J. Jeong, M. K. Song, and D. H. Yoon, Appl. Surf. Sci. 254, 1101 (2007). [CrossRef]
15. S. W. Kwon, Y. S. Song, W. S. Yang, H. M. Lee, W. K. Kim, H. Y. Lee, and D. Y. Lee, Opt. Mater. 29, 923 (2007). [CrossRef]