We report an actively Q-switched Nd:YVO4/YVO4 intracavity Raman laser at second-Stokes wavelength of 1313.6 nm, which is capable of operating efficiently under pulse repetition frequency higher than 80 kHz. A folded coupled cavity is adopted to optimize the fundamental and the Stokes resonators individually and make full use of the high pump intensity on the Raman crystal. With relatively high output coupling of 82% at 1313 nm, the average output power of 5.16 W at 1313 nm is achieved under the incident pump power of 36.7 W. The cascaded Raman emission at both the first- and second-Stokes wavelength of 1176 and 1313 nm is investigated to discuss the optimization of the second-Stokes generation.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Nonlinear optical frequency conversion has long been an essential technique to acquire specific wavelengths that are difficult to access with direct laser generation. Unlike other nonlinear optical processes, stimulated Raman scattering (SRS) is capable of obtaining several spectral lines simultaneously with intervals of certain Stokes shifts when the pump intensity is sufficiently high. The cascade process of SRS builds up a significant basis for multiple-wavelength and wavelength-selectable Raman lasers [1–3]. In order to achieve higher order of Stokes waves efficiently, an intracavity Raman laser where the Raman crystal is placed within the fundamental laser resonator is preferably adopted, taking full advantage of the relatively high intracavity pump intensities. The most commonly used resonator design is a shared linear cavity for all the fundamental and Stokes optical fields, which offers simplicity in resonator alignment [3–8]. In 2012, Shen et al. reported a diode-side-pumped dual-wavelength Nd:YAG/BaWO4 intracavity Raman laser, obtaining the maximum output power of 8.30 W for the first Stokes at 1180 nm and 2.84 W for the second Stokes at 1325 nm with the pulse repetition frequency (PRF) of 15 kHz . However, the round-trip losses for the Raman scattering light are comparatively large, since the oscillating Raman beams would pass through both the laser and Raman crystal back and forth and also the Raman cavity length is inevitably long. The self-Raman laser configuration eliminates the necessity of an additional Raman medium, and is widely used in cascaded Raman lasers as well [9–14]. In 2012, Chen et al. reported the first actively Q-switched second-Stokes Nd:YVO4 self-Raman laser, achieving the output power of 2.34 W at 1313 nm under the pump power of 14.6 W and the PRF of 40 kHz . Recently, Xie et al. employed a suitable output coupler (OC) for a cascaded Nd:YVO4 self-Raman laser and obtained the second-Stokes output of 2.51 W at 1313 nm with the incident pump power of 17.1 W at 50 kHz . Although the insertion loss is lower, self-Raman lasers suffer from severer thermal effects because of the accumulated thermal load from both the laser and SRS processes, especially for higher-order Stokes generation. The heavy thermal deposition causes the onset of laser instability and other detrimental thermal effects, impacting on the power scalability under higher pump [9, 11]. When adopting separate laser and Raman crystals, the thermal load could be distributed and the mode sizes can be optimized individually in the laser and Raman gain media. Meanwhile, the coupled cavity configuration could be employed to eliminate the intracavity losses of the laser crystal and other components for Stokes waves. A flat mirror coated for highly-reflective (HR) at the Stokes wavelengths and highly-transmissive (HT) at the fundamental could be inserted between the two crystals [15, 16], yet the limited adjustable length of Raman cavity makes it difficult to achieve good mode matching with the fundamental beam. Among varieties of higher-order Raman lasers, the 1.3-μm second-Stokes generation based on Nd-doped laser crystals and vanadate or tungstate Raman media is of great interest because of the potential applications in optical communications and medical arena. At present, the output of second-Stokes intracavity Raman lasers operating at 1.3 μm is generally lower than 3 W, with comparatively low PRFs (typically ≤ 50 kHz) [5–7, 11–14]. Factors such as cavity mode and thermal load could be further optimized to improve the performance of higher-order Stokes generation.
In this paper, a coupled cavity with a folded configuration is employed, with which the mode size of the Raman beam can be controlled flexibly to match that of the fundamental laser by adjusting the length of the folded arm. We select a YVO4-Nd:YVO4-YVO4 composite and a pure YVO4 crystal as the laser and Raman gain mediums, respectively. With the adoption of in-band pumping and separate laser and Raman crystals, the thermal load in the cascaded Raman laser is alleviated and hence higher pump is allowed to achieve power scaling. The fundamental beam spot size is designed to be relatively small in the Raman crystal, so that the resultant intense intracavity pump intensity would facilitate the generation of higher-order Stokes waves. Using an OC with transmission of 82% at 1313 nm, the average output power of 5.16W for the second Stokes is achieved under the pump power of 36.7 W and the PRF of 80 kHz. To the best of our knowledge, the output power and the operating PRF are higher than the 1.3-μm second-Stokes generation based on diode-pumped solid-state Raman lasers ever reported. The high-repetition-frequency pulses at 1.3 μm is advantageous in the field of laser remote sensing and optical communications.
2. Experimental setup
The experimental layout of the second-Stokes Nd:YVO4/YVO4 Raman laser with a folded coupled-cavity configuration is depicted in Fig. 1. The Nd-doped laser crystal is in-band pumped by an 880-nm fiber-coupled laser diode (core diameter φ = 400 μm, numerical aperture NA = 0.22), in order to reduce the heat deposition and improve the power scalability of the fundamental laser. A multi-lens coupler focuses the pump beam into the a-cut YVO4-Nd:YVO4-YVO4 composite crystal with a ratio of 1:2, forming a pump spot with a radius of 400 μm near the incident facet of the laser gain medium. The composite laser crystal with dimensions of 3 × 3 × 20 mm3 consists of a 16-mm-long 0.3-at.% Nd-doped segment and two 2-mm-long un-doped segments on each end. The entrance face of the crystal has HT coating at 880 nm (T>99%) as well as 1064 nm (T>99.9%) and the rear face is coated for HT at 1064 nm (T>99.9%). Wrapped in indium foil, the composite crystal is mounted in an aluminum holder with the temperature controlled by circulating water at 12 °C. A 32-mm-long acousto-optic Q-switch (QCQ80-A1.2-L1064-Z32, AA Opto-Electronic) coated for HT at 1064 nm (T>99.6%) on both facets is driven at 80-MHz ultrasonic frequency by 20-W radio-frequency power. We selected an a-cut YVO4 crystal with dimensions of 3 × 3 × 30-mm3 to be the Raman gain medium, which has HT coating at 1064 (T>99.9%), 1176 (T>99.5%) and 1313 nm (T>98%). The crystal length of 30 mm is chosen for the reason that longer interaction length can provide higher Raman gain effectively , which is beneficial for higher-order Stokes generation. The Raman crystal wrapped in indium foil is clamped in an aluminum holder which is cooled by refrigerant water maintained at 19 °C together with the Q-switch crystal.
A concave mirror M1 (radius of curvature ROC = 200 mm) coated for HT (T>99%) at 880 nm and HR (R>99.9%) at 1064 nm, and a flat mirror M2 together constitute the fundamental resonator. The mirror M2 is coated for HR at 1064 and 1176 nm (R>99.9%), and meanwhile partially-transmissive at 1313 nm (T = 81.8%), which also acts as the OC for the second-Stokes wave. A flat mirror M3 coated for HT at 1064 nm (T>97.5%) but HR at 1176 (R>99.5%) and 1313 nm (R>96%) for the p-polarization wave at the angle of 45° is inserted between the acousto-optic Q-switch and the Raman gain medium. The transmittance curve of the mirror M2 and M3 is plotted in Fig. 2(a) and 2(b), respectively. The laser medium, together with the Raman crystal, is thus placed at the direction where the c axis is parallel to the horizontal plane so that the polarization of the oscillating beam is in accordance with the coating requirement of M3. The Stokes resonator is made up of the output coupler M2, the folded-mirror M3 and a concave mirror M4 (ROC = 100 mm) coated for HR at both 1176 and 1313 nm (R>99.5%).
The cavity lengths of the two coupled resonators are estimated on account of the thermal lens effects of the Nd:YVO4 composite and the pure YVO4 crystal. Provided the incident pump power is 36.7 W with a spot radius of 400 μm, the thermal focal length of the laser medium (with the measured absorption coefficient being 0.92 cm−1 at 880 nm) is calculated to be ~185 mm according to Eq. (13) in . The oscillation of the Stokes waves leads to an effective thermal focal power of ~11 diopters in the Raman crystal based on Eq. (30) in . The resultant fundamental beam radius in the Raman crystal is ~137 μm when the fundamental cavity length is set to be 135 mm. The resonator length for the Raman beams is accordingly designed to be 70 mm, resulting in spot radii within the Raman crystal to be ~131 and ~138 μm, respectively, for the first- and second-Stokes waves. As for the traditional 808-nm pumping with a typical 20-mm-long, 0.3-at.% doped Nd:YVO4 crystal as the self-Raman medium to achieve second-Stokes output, like the case in  where the maximum allowed pump power is 17 W with the spot diameter to be 330 μm, the cumulative thermal focal length derived from both the lasing and SRS processes is as short as ∼45 mm. The generated heat could be shared if separate laser and Raman crystals are adopted. In addition, with the employment of the in-band pumping scheme and a relatively large pump spot size, the allowed incident pump could be higher than 36.7 W and the resultant high performance could be expected.
3. Experimental results and discussion
The average output power at 1313 nm as functions of incident pump power under different PRFs is shown in Fig. 3. A flat mirror coated for HR at 880 nm and HT at 1313 nm is placed after the OC of M2 to filter out the residual pump. The second-Stokes average output power at 60 kHz is similar to that at 80 kHz, while the coating of the Raman crystal is broken down when the pump power reaches 36.7 W because of the high peak power intensity at a comparatively low PRF of 60 kHz. So we do not conduct further experiment with lower PRF in case of more coating damage. The second-Stokes threshold is ~7.8 W at 60 kHz and increases to ~8.3 W at 80 kHz and ~9.2 W at 100 and 120 kHz. The average output power exhibits a decrease when the PRF increases from 80 to 100 kHz under lower pump power, since the Raman gain is proportional to the pump intensity  which is larger with lower PRF. However, Raman conversion is a comprehensive result of multiple factors such as peak power and thermal deposition, and thus the output performance at lower PRF may decay with increasing pump because of the stronger thermal effects of lower PRF [20, 21]. Under the incident pump power of 36.7 W, the highest average output power of 5.16 W at 1313 nm is achieved at 80 kHz, which decreases to 5.07 W at 120 kHz and 4.95 W at 100 kHz. The corresponding conversion efficiency is 14.1%, 13.5% and 13.8% at 80, 100 and 120 kHz, respectively. Higher conversion efficiency could be obtained at lower pump power of 22.07 W, when the highest optical efficiency is 16.5% and 16.3%, respectively, at 80 and 100 kHz. The optical efficiency rolls over after the pump power of 22.07 W and thus the output increasing exhibits nonlinearly, which is a result of the thermal effects of YVO4 Raman crystal. On the one hand, the Raman gain is a decreasing function of the Raman crystal temperature under high pump power . On the other hand, the shortened thermal focal length in the Raman crystal would make the fundamental beam radius in the laser gain medium become smaller, and consequently influence the mode matching between the incident LD pump and the fundamental and in turn decay the optical efficiency. According to reported literatures [5–7, 11–14], the higher-order Stokes is always achieved at PRF typically lower than 50 kHz so as to take advantage of the high peak intensity. Yet the experimental results here show that second-Stokes Raman laser can operate efficiently with PRF higher than 100 kHz. We attribute the high performance to the relatively small fundamental beam spot on the Raman crystal which provides sufficient pump intensity. With the employment of the folded coupled cavity, we could optimize and align the fundamental and Raman resonators separately, and meanwhile the mode matching between the cascaded beams is readily achieved. It should be noted that the power saturation is not observed, and the conversion efficiency could be further improved if an 878.7-nm wavelength-locked LD is adopted to increase the in-band pumping absorption. The insets give the beam profiles of the highest second-Stokes output measured at 80 kHz (inset (a) in Fig. 3) and the simultaneous fundamental laser at 1064 nm reflected from the upside of M3 (inset (b) in Fig. 3). The Raman beam cleanup effect is obvious that the spatial beam quality of the Stokes wave (Mx2 = 1.59, My2 = 1.50) is improved compared to that of the fundamental beam (Mx2 = 4.67, My2 = 4.36). The spectrum of the laser beam emitted from the OC of M2 at the pump power of 36.7 W is investigated via an optical spectral analyzer (Yokogawa AQ6370D) with the resolution of 2 nm, shown in Fig. 4. The third-Stokes wavelength at 1487.5 nm is observed because of the relatively high intracavity pump intensity. Given a better coating of the folded mirror M3 with higher transmittance at 1064 and 1487 nm as well as higher reflectivity at 1176 and 1313 nm, the conversion efficiency of the second Stokes at 1313 nm might be promoted. The fine spectrum of the second-Stokes wavelength at 1313.6 nm with the resolution of 0.02 nm is plotted as the inset, where the linewidth is ~0.18 nm.
We used two OCs (ROC = 100 mm) coated for partially-transmissive at the first Stokes wavelength of 1176 nm to substitute the concave mirror M4 with HR coating at 1176 nm to further investigate the cascade process of the Raman laser. The coatings of the two OCs are (1) T = 4.5% at 1176 nm and T = 28% at 1313 nm, and (2) T = 10.2% at 1176 nm and T = 29.1% at 1313 nm, respectively. The laser emission from the OC of M4 includes both the first and second Stokes waves, so we used a flat mirror coated for HT at 1176 nm (T>95%) and HR at 1313 nm (R>99.5%) to separate the Stokes beams. Since the two OCs also have HR coatings at 1064 nm (R>99.9%) and the reflective power of the fundamental laser from the downside of folded mirror M3 is rather limited (basically several watts), the leakage at 1064 nm from M4 is negligible. The Stokes average output power at 1176 and 1313 nm versus incident pump power is depicted in Fig. 5. The recorded 1313-nm power contains the output from mirror M2 as well as the one from mirror M4.
When the OC with lower transmittance of 4.5% at 1176 nm (OC (1)) is used, the highest average output power of 6.73 W at 1176 nm is achieved at a relatively high PRF of 160 kHz, and meanwhile the output at 1313 nm is 1.18 W, as shown in Fig. 5(a). As for the second Stokes, the highest output power of 3.79 W at 1313 nm is achieved at a lower PRF of 80 kHz, while the output power is 4.1 W at 1176 nm. The dependence of the first- and second-Stokes output on the PRF is the same as the OC with transmittance of 10.2% at 1176 nm (OC (2)) is adopted. As shown in Fig. 5(b), the highest output of 8.06 W at 1176 nm is obtained at 160 kHz, when the output at 1313 nm is 0.54 W; the highest output of 2.71 W at 1313 nm is achieved at 80 kHz, while the output at 1176 nm is 5.5 W. The generation of higher-order Stokes requires intense pump intensity with comparatively lower PRF, therefore the optimal PRF for efficient output of second Stokes is lower than that for first Stokes. The change in the slope of the output curves for the cascaded Stokes waves shows the opposite trend, which is an indication of the energy transfer from the first Stokes to the second Stokes. The slope of the first-Stokes output curves become smaller when the second Stokes begins to increase rapidly after the threshold of ~11 W at 80 kHz with both OCs. As Fig. 5(a) shows, the trend is similar with OC (1) at 160 kHz that the increase of 1176-nm output becomes gentle at the second-Stokes threshold of ~15 W. However, the trend is not obvious using OC (2) at 160 kHz in Fig. 5(b), since the transmittance of 10.2% at 1176 nm is not beneficial for the efficient generation of the second Stokes at the PRF of 160 kHz. The output power at 1176 nm is not totally clamped after the oscillation of the second Stokes and rises again with increasing pump, which we attribute to the relatively high output coupling at 1313 nm. The cavity is highly resonant at the fundamental and first-Stokes wavelength but has high output coupling (82%) at the second-Stokes wavelength. The accumulation of the second Stokes would clamp the generation of the first Stokes and thus hinder the depletion of the fundamental pump, which would in turn limit the generation of the second Stokes . So the output coupling could be moderately higher for second Stokes generation, which enables efficient extraction of the second Stokes and realizes favorable pump depletion. It is also clear that the output at 1313 nm is always higher at the same PRF when using the OC with lower transmittance at 1176 nm. Highly resonant at first Stokes provides higher intracavity power intensity that is advantageous to the second Stokes conversion.
As the OC with transmittance of 4.5% at 1176 nm is used, the temporal behavior of the cascaded Stokes waves and the fundamental laser reflected by the folded mirror M3 is monitored with a photoelectric detector (Thorlabs DET08C) and an oscilloscope (Tektronix DPO2024B). Figure 6 shows the typical oscilloscope traces of the three beams at wavelength of 1064, 1176 and 1313 nm under the maximum pump of 36.7 W at 80 kHz. The pulse duration of the fundamental beam is 21.88 ns before SRS taking place and decreases to 8.72 ns as a result of pump depletion due to the Raman conversion. The pulse width of the first and second Stokes is 4.85 and 3.40 ns, respectively. There is a long tail at the falling edge of the first-Stokes temporal profile, as a result of the excessive Raman gain  and the incomplete consumption of the second Stokes by the third Stokes generation. The pulse shortening and the time delay between the pulse peaks reveals the cascade process of higher-order Stokes generation.
In summary, high-repetition-rate second-Stokes generation at 1313 nm is realized based on a Nd:YVO4/YVO4 folded coupled-cavity Raman laser. With the employment of in-band pumping scheme as well as separate laser and Raman crystals, the cavity is capable of accommodating higher incident pump and proves to be effective for power scaling. As the incident pump power increases to 36.7 W, the second-Stokes output power of 5.16 W is achieved at PRF of 80 kHz, which is the highest output amongst 1.3-μm second-Stokes intracavity solid-state Raman lasers ever reported. The adoption of a folded coupled cavity is beneficial for the generation of higher-order Stokes, since the fundamental and Raman resonators can be optimized individually to take full advantage of the high intracavity pump intensity.
National Natural Science Foundation of China (NSFC) (11674242), (11674243); Natural Science Foundation of Tianjin City (16YFZCGX00350), (15JCQNJC02500).
References and links
1. J. A. Piper and H. M. Pask, “Crystalline Raman Lasers,” IEEE J. Sel. Top. Quantum Electron. 13(3), 692–704 (2007). [CrossRef]
2. A. S. Grabchikov, V. A. Lisinetskii, V. A. Orlovich, M. Schmitt, S. Schluecker, B. Kuestner, and W. Kiefer, “Continuous-wave solid-state two-Stokes Raman laser,” Quantum Electron. 39(7), 624–626 (2009). [CrossRef]
3. Y. T. Chang, Y. P. Huang, K. W. Su, and Y. F. Chen, “Diode-pumped multi-frequency Q-switched laser with intracavity cascade Raman emission,” Opt. Express 16(11), 8286–8291 (2008). [CrossRef] [PubMed]
4. S. V. Voitikov, A. A. Demidovich, A. S. Grabtchikov, P. V. Shpak, M. B. Danailov, and V. A. Orlovich, “Two-Stokes generation and effect of multiwave mixing on output pulse parameters of a Q-switched Raman microchip laser,” J. Opt. Soc. Am. B 27(6), 1232–1241 (2010). [CrossRef]
5. H. Shen, Q. Wang, X. Zhang, X. Chen, Z. Cong, Z. Wu, F. Bai, W. Lan, and L. Gao, “1st-Stokes and 2nd-Stokes dual-wavelength operation and mode-locking modulation in diode-side-pumped Nd:YAG/BaWO4 Raman laser,” Opt. Express 20(16), 17823–17832 (2012). [CrossRef] [PubMed]
6. H. Shen, Q. Wang, X. Zhang, X. Chen, F. Bai, Z. Liu, L. Gao, Z. Cong, Z. Wu, W. Wang, Y. Zhang, W. Lan, and C. Wang, “Second-Stokes dual-wavelength operation at 1321 and 1325 nm ceramic Nd:YAG/BaWO4 Raman laser,” Opt. Lett. 37(21), 4519–4521 (2012). [CrossRef] [PubMed]
7. H. Zhang, X. Chen, Q. Wang, and P. Li, “Dual-wavelength actively Q-switched diode-end-pumped ceramic Nd:YAG/BaWO4 Raman laser operating at 1240 and 1376 nm,” Laser Phys. Lett. 11(10), 105806 (2014). [CrossRef]
8. H. Zhu, Y. Ye, Y. Duan, Y. Zhang, D. Zhang, B. Zhao, and A. A. Kaminskii, “Third-Stokes light operation in KTA crystal derived by Nd:Lu0.5Y0.5VO4 crystal laser,” J. Opt. 17(3), 035503 (2015). [CrossRef]
10. C. Du, X. Xie, Y. Zhang, G. Huang, Y. Yu, and D. Wang, “YVO4/NdYVO4/YVO4 self-Raman laser at 1764 nm,” Appl. Phys. B 116(3), 569–574 (2014). [CrossRef]
11. W. Chen, Y. Wei, C. Huang, X. Wang, H. Shen, S. Zhai, S. Xu, B. Li, Z. Chen, and G. Zhang, “Second-Stokes YVO4/Nd:YVO4/YVO4 self-frequency Raman laser,” Opt. Lett. 37(11), 1968–1970 (2012). [CrossRef] [PubMed]
12. C. Du, G. Huang, Y. Yu, X. Xie, Y. Zhang, and D. Wang, “Q-switched mode-locking of second-Stokes pulses in a diode-pumped YVO4/Nd:YVO4/YVO4 self-Raman laser,” Laser Phys. 24(12), 125003 (2014). [CrossRef]
13. H. Zhu, J. Guo, X. Ruan, C. Xu, Y. Duan, Y. Zhang, and D. Tang, “Cascaded Self-Raman Laser Emitting Around 1.2-1.3 μm Based on a c-cut Nd:YVO4 Crystal,” IEEE Photonics J. 9(2), 1500807 (2017). [CrossRef]
14. Z. Xie, Y. Duan, J. Guo, X. Huang, L. Yan, and H. Zhu, “Cascaded a-cut Nd:YVO4 self-Raman with second-Stokes laser at 1313 nm,” J. Opt. 19(11), 115501 (2017). [CrossRef]
15. P. V. Shpak, S. V. Voitikov, R. V. Chulkov, P. A. Apanasevich, V. A. Orlovich, A. S. Grabtchikov, A. Kushwaha, N. Satti, L. Agrawal, and A. K. Maini, “Passively Q-switched diode-pumped Raman laser with third-order Stokes eye-safe oscillation,” Opt. Commun. 285(17), 3659–3664 (2012). [CrossRef]
17. W. Koechner and M. Bass, Solid-State Lasers: a graduate text (Springer Science & Business Media, 2006), p. 376.
18. M. E. Innocenzi, H. T. Yura, C. L. Fincher, and R. A. Fields, “Thermal modeling of continuous-wave end-pumped solid-state lasers,” Appl. Phys. Lett. 56(19), 1831–1833 (1990). [CrossRef]
19. H. M. Pask, “The design and operation of solid-state Raman lasers,” Prog. Quantum Electron. 27(1), 3–56 (2003). [CrossRef]
20. B. Li, P. Lei, B. Sun, and Y. B. Bai, “High-efficiency, high-repetition-rate Nd:YVO4 self-Raman laser pumped by a wavelength-locked 913.9 nm diode laser,” Appl. Opt. 56(5), 1542–1545 (2017). [CrossRef]
21. K. W. Su, Y. T. Chang, and Y. F. Chen, “Power scale-up of the diode-pumped actively Q-switched Nd:YVO4 Raman laser with an undoped YVO4 crystal as a Raman shifter,” Appl. Phys. B 88(1), 47–50 (2007). [CrossRef]
22. R. J. Williams, D. J. Spence, O. Lux, and R. P. Mildren, “High-power continuous-wave Raman frequency conversion from 1.06 µm to 1.49 µm in diamond,” Opt. Express 25(2), 749–757 (2017). [CrossRef] [PubMed]
23. S. H. Ding, X. Y. Zhang, Q. P. Wang, J. Zhang, S. M. Wang, Y. R. Liu, and X. H. Zhang, “Numerical modeling of passively Q-switched intracavity Raman lasers,” J. Phys. D Appl. Phys. 40(9), 2736–2747 (2007). [CrossRef]