1014 OPTICS LETTERS / Vol. 12, No. 12 / December 1987 Second-harmonic generation of a continuous-wave diode-pumped Nd:YAG laser using an externally resonant cavity W. J. Kozlovsky, C. D. Nabors, and R. L. Byer Edward L. Ginzton Laboratory, Stanford University, Stanford, California 94305 Received June 19, 1987; accepted September 8, 1987 We report 13% second-harmonic conversion efficiency of a 15-mW, cw, diode-laser-pumped Nd:YAG oscillator. 2 mW of single-axial-mode 532-mm radiation was generated by externally resonant second-harmonic generation in a monolithic MgO:LiNbO 3 nonlinear crystal cavity. The measured finesse of 450 for the monolithic external cavity indicated that absorption and scatter losses in the doubler were less than 0.8%. The recent development of high-power laser diodes The nonplanar ring oscillator used for these experi- 8 has renewed interest in diode-laser-pumped solid- ments has been described by Kane et al. It operated state lasers. Diode-pumped lasers offer improved re- TEMOOand in a single axial mode with a linewidth of liability, efficiency, and frequency stability' compared less than 10 kHz and a drift of less than 1 MHz per with lamp-pumped lasers. Second-harmonic genera- minute. It produced 15 mW of cw output power when tion (SHG) of these infrared laser sources is necessary pumped by a 120-mW diode-laser source. for many applications. However, the lower power MgO:LiNbO3 was chosen for the external doubler available from diode-pumped lasers makes efficient because of its low scatter and absorption loss at 1.06 Atm frequency doubling more challenging than for high- (Ref. 9) and its large nonlinear and electro-optic coeffi- power lamp-pumped lasers. cients. Both ends of a 4 mm X 4 mm X 25 mm crystal One demonstrated method for efficient SHG of di- of MgO:LiNbO 3 were polished with 20-mm radii of ode-pumped solid-state lasers is intracavity frequency curvature and then coated to be 99.7% reflecting at 1.06 doubling.2 The nonlinear crystal is placed inside the ,um. This monolithic cavity design provided stability laser resonator to take advantage of the high circulat- of the resonant frequency and also minimized cavity ing intensity to produce second-harmonic output effi- losses; this is essential for large field enhancements in ciently. In these linear laser resonators, however, spa- the external resonator. The cavity geometry with an no tial hole burning usually allows several axial modes to of 2.23 yields a cavity beam waist of 38,4m and an axial- oscillate. Since the losses of these axial modes are mode spacing of 2.7 GHz. The MgO:LiNbO 3 cavity coupled through sum-frequency generation in the resonant frequency was tuned by applying a voltage doubling crystal, the resulting output at the second across the crystal y axis, using the r22 electro-optic harmonic is strongly amplitude modulated.3 coefficient to change no and thereby the optical path An alternative approach is resonant external-cavity length. One free spectral range was scanned with SHG.4 In this method, the nonlinear medium is 1040 V. placed in an external resonator designed to enhance Figure 1 shows a schematic of the experimental set- the fundamental field, the second-harmonic field, or up. For the doubling experiments, the oven contain- both. These resonantly enhanced fields efficiently ing the monolithic MgO:LiNbO3 crystal resonator was generate second-harmonic power without the ampli- heated to the noncritical 900 phase-matching tem- tude instabilities seen in intracavity frequency dou- perature of 1070C. Spatial mode matching into the bling. An additional advantage of external-cavity doubling cavity was accomplished with a 10-cm focal- SHG is that the laser oscillator and the external har- length lens and proper adjustment to the position and monic resonator can be independently optimized. orientation of the harmonic crystal. This was the only This is especially important in low-gain or quasi- mechanical adjustment needed, since the monolithic 5 three-level laser systems. Independent optimization design ensured alignment of the MgO:LiNbO3 resona- also permits the design of a single-axial-mode laser tor itself. A Faraday isolator was placed between the source, ensuring that the output of the external dou- laser and the doubler to collect the fundamental radia- bler is also in a single axial mode. tion backreflected from the doubling crystal, which External-cavity resonant SHG requires the coinci- was a minimum at resonance. On resonance, the high dence of the laser and external resonator frequencies. circulating intensity of the fundamental in the dou- Thus a single-axial-mode, frequency-stable laser oscil- bling crystal efficiently generated the second harmon- lator6 is important. Diode-pumped, monolithic, non- ic. Since the external doubler was a standing-wave planar-ring Nd:YAG lasers7' 8 are ideal laser sources resonator, the second harmonic was generated in both for external resonant SHG because of their single- directions. The use of two dichroic mirrors allowed mode, frequency-stable output power and their good both of the harmonic output beams and the transmit- resistance of feedback. ted fundamental beam to be measured. 0146-9592/87/121014-03$2.00/0 ©D1987, Optical Society of America December 1987 / Vol. 12, No. 12 / OPTICS LETTERS 1015 1 beam." Then tSHG = (1 - YSHGP,) is the fraction of Isleetomodel resonated fundamental not converted in a single pass of doubling. Since the second harmonic is generated Fundamental Fundamental pig, D Detector Detector in both directions, the ABD cavity-reflectance param- eter rm becomes |Faraday|| r. = t2tsHGr2, (1) rotator | Elecra-optc where, following ABD conventions, t is the single-pass Car V '''''~~~~Canre power transmittance of the crystal, t1 is the input- SHG oven SHG mirror power transmittance, and r2 is the backmirror Detector , |___ ____Detector power reflectance. For a given incident fundamental power P1 on resonance, then, Dichroic Duchroic Mirror Mirror p = tip1 (1 - Vrir) 2 (2) Fig. 1. Experimental setup. The two dichroic mirrors al- and the total output at the low the power in both second-harmonic outputs, as well as second harmonic is the transmitted fundamental power, to be measured. The P = 2ySHGPC2. Faraday isolator allows the fundamental power backreflect- 2 (3) ed from the doubling cavity to be detected and therefore to For our crystal, with deff = 5.95 X 10-12 mn/V(Ref. be used for feedback in optimizing resonance. POL, polar- 12) and an effective phase-matching length of only 1.3 izer. 9 cm, 'YSHG = 0.0025/W. Figure 2 shows the numerical solution to the above equations plotted as a curve along with the experimental data points. It is inter- Figure 2 shows the total generated second-harmonic esting to note that the 13% conversion from input power as a function of incident fundamental power. A fundamental to output second-harmonic power corre- commercial scanning confocal interferometer was sponds to an internal conversion efficiency (or effec- used to verify that the second-harmonic output was tive loss to the circulating intensity) of only 0.3% and a single axial mode. At 15 mW of input power, 2 mW of circulating intensity of 620 mW. second-harmonic power was generated, yielding a con- The circulating intensity can also be determined by version efficiency of 13%. The bias voltage necessary measuring the fundamental power transmitted to maintain resonance drifted slowly, probably as a through mirror r . Figure 3 shows the transmitted result of photoconductivity screening the voltage ap- 2 10 fundamental power versus incident power with and plied across the crystal. Resonance was maintained without doubling. The line is the theory, and the by a servo loop based on the detection of the reflected squares are the data. The excellent agreement be- fundamental, with a small dither signal applied to the tween theory and data indicates that all losses of the feedback voltage on the doubler. With the cavity resonator, including doubling, are properly deter- locked, the output at the second harmonic had ampli- mined. The lower circulating intensity while the crys- tude fluctuations of less than 5%, largely because of tal is phase matched corresponds to a reduction in the technical noise in the servo loop. Q of the cavity due to doubling and also to the decrease With the crystal at room temperature, the harmon- of fundamental power coupled into the resonator be- ic-generation process was not phase matched, so no cause of the changing reflectance parameter. During doubling occurred. The crystal resonator then oper- SHG, the measured fundamental power throughput ated as a high-finesse interferometer. The finesse was measured to be greater than 450, indicating that total losses in the resonantor were less than 1.4%. Coating transmittance accounted for 0.6% of this loss, indicat- ing that all other losses were less than 0.8% in a round 0E trip. 0 The output of the doubler may be calculated by ILS2 using the external resonator SHG theory of Askin, 00 4 Boyd, and Dziedzic (ABD). Because scattering and I- c absorption losses in the crystal are so low, depletion of 0ro the resonated fundamental caused by SHG must be 0 An considered. This factor is not considered in the the- -3 ory of ABD but is easily accounted for by an additional a loss term that is proportional to the circulating inten- 0 4 8 12 16 7 Incident sity. Let 7SHG - YSHGPC, where 77SHGis the conversion FundamentalPower (mW) efficiency of the resonated fundamental to the second- Fig. 2. Total second-harmonic power versus incident pow- harmonic power and P, is the circulating intensity of er at the fundamental. The squares are experimental data, the fundamental in the doubling cavity.