2018_基于單模ld泵浦ybkyw激光器的連續(xù)波單諧振腔內光學參量振蕩器

1、Continuous-wave, singly-resonant, intracavity optical parametric oscillator based on asingle-mode-laser-diode-pumped Yb:KYW laserZHENWEN DING,1,2 PEI LIU,1,2 YELI,1,2AND ZHAOWEI ZHANG1,2,*1Mid-Infrared Lasers Group, School of Optical and Electronic Information, Huazhong University of Science and Tec

2、hnology, Wuhan, Hubei 430074, China2National Engineering Research Center for Laser Processing, Wuhan, Hubei 430074, China*Corresponding author: Received 12 April 2018; revised 9 May 2018; accepted 11 May 2018; posted 11 May 2018 (Doc. ID 328258); published 6 June 2018Ti:sapphi

3、re lasers pumped by bulky Ar-ion lasers 9,10. ICSROs were also demonstrated based on PPLN crystals and Nd:YVO4 lasers directly pumped by multimode laser di- odes (LDs), with excellent conversion efficiency and a compact mechanical design 11,12. However, due to the relatively long upper-state lifetim

4、e of Nd:YVO4 crystals compared with the photon life-time of the pump/signal cavity, ICSROs based on PPLN crystals and Nd:YVO4 lasers are highly susceptible to the onset of long-lived relaxation oscillations, which could be triggered by thermal, mechanical, or pump noise 13,14. ICSROs based on Yb:YAG

5、 crystals pumped by multimode di- ode lasers also suffer from large fluctuations in their output power 15.The stability of an ICSRO could be improved by altering its cavity dynamics so that disturbances can be rapidly damped. This was demonstrated by Stothard et al. with two technical routes, either

6、 by utilizing laser gain media with a much shorter gain recovery lifetime 16 or by introducing additional non- linear loss of the pump wave into the ICSRO cavity 17. On the other hand, the stability of an ICSRO can be improved by minimizing the external and internal disturbances so that relaxation o

7、scillations could not be triggered. For instance, Stothard et al. showed that the stability of a periodically poled RbTiOAsO4 (PPRTA) based ICSRO outperformed that of a PPLN-based system 18,19. The relaxation-oscillation-free operation of the PPRTA-based ICSRO was due to the high resistance of PPRTA

8、 to thermal lensing effects, resulting in the OPO being insufficiently perturbed.In this Letter, we report the first, to our knowledge, ICSRO based on a single-mode-LD-pumped Yb:KYW laser. Indeed, the synergy of low-noise single-mode LDs at 981 nm with Yb:KYW crystals 20 were successfully explored f

9、or developing femtosecond lasers at 1 m 21. Compared with previously- demonstrated ICSROs based on Nd:YVO4 lasers pumped by multimode LDs at 808 nm 11,12, the use of Yb:KYW lasers pumped by single-mode LDs at 981 nm has important advan- tages including lower frequency/intensity noise from the pump L

10、D and reduced heat generation due to the smaller quantum defect, and therefore can effectively eliminate the occurrence ofWe report a continuous-wave (cw), intracavity, singly resonant optical parametric oscillator (ICSRO) based on an Yb:KYW laser pumped by a single-mode laser diode (LD). Pumping th

11、e ICSRO by a low-noise single-mode LD, combined with the reduced heat generation due to the lower quantum defect of the Yb:KYW laser system, effectively eliminated the onset of relaxation oscillations, which have been a long-standing prob- lem in previous multimode-LD-pumped ICSROs, and re- sulted i

12、n a cw ICSRO being operated free of relaxation oscillations. At an LD power of 515 mW, the generated idlerpower was 21 mW at 3500 nm. To the best of our knowl- edge, this is the first demonstration of a single-mode-LD- pumped ICSRO. 2018 Optical Society of AmericaOCIS codes: (140.3070) Infrared and

13、far-infrared lasers; (190.4970) Parametric oscillators and amplifiers; (190.4360) Nonlinear optics, devices./10.1364/OL.43.002807Continuous-wave (cw) singly resonant optical parametric oscil- lators (OPOs) are well-established devices capable of generating broadly tunable narrow-linewi

14、dth mid-infrared coherent light emission 13. The oscillation threshold of external-cavity sin- gly resonant OPOs (SROs) based on periodically-poled lithium niobate (PPLN) crystals are usually reached at the level of a few watts 13. It is the relatively high pump threshold that hinders the widespread

15、 implementation of cw OPOs. Thus, doubly resonant 46, pump-resonant 7,8, and intracavity pumping 919 configurations were proposed and demon- strated to reduce the pumping power needed for the oscillation of cw OPOs. However, for doubly resonant or pump-resonant OPOs 48, the length of the resonating

16、cavity needs to be precisely controlled to achieve and maintain oscillation, making them unsuitable for many applications. In intracavity SROs (ICSROs), the nonlinear gain medium was placed within a laser cavity, and by exploiting the high circulating power, a low oscillation threshold can be achiev

17、ed. ICSROs were initially demonstrated based on KTP/KTA crystals and0146-9592/18/122807-04 Journal 2018 Optical Society of AmericaLetterVol. 43, No. 12 / 15 June 2018 / Optics Letters2807the 1.031.07 m range). The useful mid-infrared idler output was extracted via mirrors M3 and M4 and measured afte

18、r col- limation by an uncoated CaF2 lens and after passing through a germanium (Ge) window, which would absorb the leaking pump wave.The cavities were configured so that the parent pump laser could focus on both Yb:KYW and PPLN crystals, and the sig- nal wave had focus on the PPLN crystal. The calcu

19、lated beam radius of the OPO pump wave in the Yb:KYW crystal was16 m. The calculated 1e2 beam radii of the OPO pump and signal waves in the PPLN crystal were 29 and 51 m, respectively.Figure 2 shows the circulating pump power when the OPO was at the off or on state, and the generated mid-infrared id

20、ler power when the OPO was turned on. The circulating pump power was estimated based on the measured leaking pump power from M5 and the coating parameters of M5. The squares describe the performance of the Yb:KYW laser when the OPO was turned off by blocking the light to mirror M7. The parent pumpla

21、serstartedtooscillate at140mWofLDpower, andthe SRO started to oscillate at 365 mW of LD power, correspond- ing to a circulating pump power of 3.8 W. While the LD power was increased, the idler power started to increase, and the power of the circulating pump wave was clamped at the threshold value du

22、e to the parametric downconversion process 13.At a maximum available LD power of 515 mW, the ex- tracted mid-infrared (at 3500 nm) idler power was 9 mW from a single side. The total generated idler power was esti- mated to be 21 mW. Taking into account the idler and signal wavelengths, the losses as

23、sociated with the coatings and also the fact that the idler was generated in both directions, the total downconverted power was estimated to be 70 mW. We defined the downconversion efficiency as the ratio between the down- converted power and the optimum output power of the parent pump laser 13, whi

24、ch was measured by replacing M5 with an optimal output coupling mirror (T 2.5%) of the parent pump laser in the experiment. The optimum output power of the parent pump laser was measured to be 109 mW at the maximum LD power of 515 mW. Thus, the downconverted efficiency was calculated to be 65%. The

25、downconverted power and efficiency could be increased by employing a more powerful LD 13.We characterized the spectrum of the pump wave by using an optical spectrum analyzer, and that of the idler wave with a home-built Fourier transform infrared (FTIR) spectrometer.relaxation oscillations in the IC

26、SRO, resulting in the OPO generating output with excellent power stability. The central wavelength of a 981 nm single-mode LD can be locked by a narrow-band fiber Bragg grating (FBG), which helps to sup- press the mode competition and frequency/intensity noise from the LD. The quantum defect of a 98

27、1 nm-pumped Yb:KYW system is 9%, much smaller than that of an 808 nmpumpedNd:YVO4 laser system (24%).The experiment layout of the Yb:KYW-based ICSRO isshown in Fig. 1. We utilized a 1-mm-thick Brewster-cut Yb:KYW crystal doped with 10 at. % Yb3 . The Yb:KYWplate was oriented in the cavity for propag

28、ation along the b Np axis with polarization parallel to the Nm-axis 20. The pump source was an InGaAs single-transverse-mode LD, with a linearly polarized output. The LD, which was designed as a pump source for erbium-doped fiber amplifier applications, had a fiberized output through a polarization-

29、 maintaining-fiber. An FBG was utilized to lock its central wave- length to 981 nm, ensuring an efficient absorption by the Yb:KYW crystal and enhanced wavelength/power stability. The maximum output power from the LD was 600 mW. The LDoutput was first collimated via an aspheric lens and thenfocused

30、onto the Yb:KYW crystal through a lens with a focal length of 50 mm. A z-fold pump laser cavity was constructed. The curved folding mirrors M1 and M2 with the same radius of curvature (ROC) of 75 mm were designed with high transmission at 981 nm (95%) and high reflection at the Yb:KYW laser (OPO pum

31、p) wavelength (R 99.7% in the 1.031.07 m range). Flat mirrors M5 and M6 were designed with high reflection (R 99.7%) in the 1.031.07 m range. The Yb:KYW laser also shares part of the OPO signal cavity from the beam splitter (BS) to mirror M8.The OPO crystal was 5-mol. % MgO:PPLN (HC Photonics) and w

32、as 25 mm long and 1 mm thick, composed of 10 different gratings with poling periods between 27.58 and31.59 m. The experiment was carried out with a grating having a poling period of 29.98 m. The PPLN crystal was housed in an aluminum heat-sink held at 30C. Both surfaces of the nonlinear crystal were

33、 antireflection coated at the OPO pump, signal, and idler wavelengths. The OPO signal-resonant cavitywascomposed oftheflat mirror M7, thedichroic BS, the curved mirror M4 (with a ROC of 150 mm), the curved mirror M3 (with a ROC of 75 mm), and the flat mirror M8. Mirrors M3, M4, and M8 were based on

34、YAG materials, coated with high reflection at the OPO signal and pump wavelengths (R 99.7% at 1.031.07 m and 1.361.73 m) and low re- flection at the OPO idler wavelength (R 99.7% in the 1.361.73 m range) and low reflection at the OPO pump wavelength (R 2% inFig. 2. Left axis: circulating pump power

35、when the OPO was on and off; right axis: generated idler power from the ICSRO.Fig. 1. Experimental layout of the Yb: KYW-based ICSRO.2808Vol. 43, No. 12 / 15 June 2018 / Optics LettersLetterfromM5Fig. 5(a).Therelaxation-oscillation frequency of the pump laser (i.e., the Yb:KYW laser) was 68 kHz, whi

36、le the damping time was 250 s. The LD itself did not exhibit relaxation oscillation, as shown in the inset of Fig. 5(a).Then we turned on the OPO, and the monitored circulating pump power is shown in Fig. 5(b). Indeed, in this scenario, the circulating pump power exhibited two distinguishable stages

37、 of relaxation-oscillation, at frequencies first at 68 kHz and thenat 200 kHz. The first section is the intrinsic relaxation oscil- lation of the parent pump laser, as already shown in Fig. 5(a).The second section, with a higher frequency, can be attributed to the rapid power flow between the pump a

38、nd signal cavities. The total damping time was 500 s, around two times that of the parent pump laser.Then we characterized the power stability of the ICSRO by monitoring the leaking pumping field through mirror M5 when the LD was turned on and at the maximum power. For comparisons, we also character

39、ized the power stability of the LD and Yb:KYW laser (when the OPO oscillation was in- terrupted throughblocking thelighttomirror M7).An InGaAs photodetector with a response bandwidth of over 35 MHz was employed, and the measured data were recorded by a 16-bit data acquisition (DAQ) card with a maxim

40、um sampling rate of 10 MHz. Initially, for investigating the high-frequency noise, the sampling rate of the DAQ card was set at 10 MHz and the sampling window at 50 ms. The relative intensity noise (RIN) and cumulative power fluctuations are shown in Fig. 6(a). The LD shows excellent power stability

41、, with a power fluctuation of only 0.1%. The cumulative power fluctuations of the Yb:KYW laser was slightly higher, at 0.2%. The cumulative power fluc- tuation of the ICSRO was 0.4%, and the noise accumulatedlargely at 20100 kHz, around the relaxation-oscillation fre-quency of the Yb:KYW laser. For

42、arelatively longer-term mea- surement, the sampling rate and sampling window of the DAQ card were set at 100 kHz and 100 s, respectively, and the RIN and cumulative power fluctuations are shown in Fig. 6(b).Fig. 3. Measured spectra of the (a) pump and (b) idler light.Since no wavelength-selection el

43、ements were inserted inside the pump cavity, it was expected that the Yb:KYW laser could oscillate at any wavelength within the emission band of the laser gain medium. A typical pump spectrum is shown in Fig. 3(a), indicating multilongitudinal-mode oscillation at 10301050 nm. A typical idler spectru

44、m is shown in Fig. 3(b) (measured when the poling period of PPLN was 29.98 m), with a central wavelength at 3510 nm and a full-width athalf-maximum (FWHM) linewidth of 14 nm. The linewidth was very close to the 13 nm phase matching bandwidth ofthe 25-mm-long PPLN crystal utilized in the system.Using

45、 a Spiricon Pyrocam III camera, we measured the spatial beam profile of the OPO signal wave as the LD power was 380 mW (close to the OPO threshold) and 515 mW (at the maximum power level), respectively (Fig. 4). As dem- onstrated by Stothard et al., such measurements offer an ap- proach to evaluate

46、the thermal lensing effect of the ICSRO system at different power levels 19. The beam was slightly elliptical at the maximum power. However, the spatial distribu- tions in both directions were very close to the fundamental Gaussian modes, demonstrating excellent beam quality and in- dicating that th

47、e thermal lensing effects in the system were insignificant even at the maximum power level.Itis well known that the transient dynamics in ICSROs are more complicated than those in simple solid-state lasers and the ICSRO could be plagued by the onset of long-lived burst of relaxation oscillations 111

48、7. In order to investigate the power stability of the ICSRO, we characterized the transient dynamics of the system when the OPO was turned on and off. To induce relaxation oscillations, we operated the LD in a quasi-continuous-wave (QCW) state with a modulation fre- quency of 50 Hz and a duty cycle

49、of 50% by modulating the pump current applied to the LD. First, we turned off the OPO by blocking the signal cavity (through blocking the light to mir- ror M7), and the circulating pump power was monitored by measuring the leaking power (from the Yb:KYW laser cavity)Fig. 5. Circulating pump power of

50、 (a) the parent pump laser and(b) the ICSRO when LD works in a QCW state. Inset in (a): the out- put power of the LD in the QCW state.Fig. 4. Spatial beam profile of the signal wave of the ICSRO when the LD power was (a) 380 mW and (b) 515 mW.LetterVol. 43, No. 12 / 15 June 2018 / Optics Letters2809

51、power. At the maximum available LD power of 515 mW, the extracted mid-infrared idler power was 9 mW from a single side, with a central wavelength at 3500 nm. Over a sampling window of 100 s and at a sampling rate of 100 kHz, thecumulative power fluctuation of the ICSRO was measured to be 0.7%. Cruci

52、ally, we experimentally demonstrated highly sta- ble, relaxation-oscillation-free operation of such a system. Indeed, this scheme represents a promising technical approach to generate cw mid-infrared coherent light sources by frequency conversion from a commercial near-infrared semiconductor LD. Mor

53、eover, implementing frequency selection and wave- length tuning techniques will result in widely tunable, narrow- linewidth, highly stable mid-infrared laser sources ideally suited for many applications including trace gas sensing, high- resolution spectroscopy, and hyperspectral imaging.Funding. Fu

54、ndamental Research Funds for the Central Universities of China (2016YXZD039); Recruitment Program of Global Experts of China.REFERENCES1. J. Courtois, R. Bouchendira, M. Cadoret, I. Ricciardi, S. Mosca, M. De Rosa, P. De Natale, and J.-J. Zondy, Opt. Lett. 38, 1972 (2013).2. X. Hong, X. Shen, M. Gon

55、g, and F. Wang, Opt. Lett. 37, 4982 (2012).3. M. Vainio and L. Halonen, Phys. Chem. Chem. Phys. 18, 4266 (2016).4. C. Drag, A. Desormeaux, M. Lefebvre, and E. Rosencher, Opt. Lett. 27, 1238 (2002).5. J. B. Barria, S. Roux, J.-B. Dherbecourt, M. Raybaut, J.-M. Melkonian,A. Godard, and M. Lefebvre, Op

56、t. Lett. 38, 2165 (2013).6. A. Boucon, B. Hardy-Baranski, and F. Bretenaker, Opt. Commun. 333, 53 (2014).7. G. A. Turnbull, D. McGloin, I. D. Lindsay, M. Ebrahimzadeh, and M. H. Dunn, Opt. Lett. 25, 341 (2000).8. G. V. Basum, D. Halmer, P. Hering, M. Mrtz, S. Schiller, F. Mller, A. Popp, and F. Khne

57、mann, Opt. Lett. 29, 797 (2004).9. F. G. Colville, M. H. Dunn, and M. Ebrahimzadeh, Opt. Lett. 22, 75 (1997).10. T. J. Edwards, G. A. Turnbull, M. H. Dunn, M. Ebrahimzadeh, andF. G. Colville, Appl. Phys. Lett. 72, 1527 (1998).11. D. J. M. Stothard, M. Ebrahimzadeh, and M. H. Dunn, Opt. Lett. 23, 189

58、5 (1998).12. Q. Sheng, X. Ding, C. Shi, S. Yin, B. Li, C. Shang, X. Yu, W. Wen, and J. Yao, Opt. Express 20, 8041 (2012).13. G. A. Turnbull, M. H. Dunn, and M. Ebrahimzadeh, Appl. Phys. B 66, 701 (1998).14. G. A. Turnbull, D. J. M. Stothard, M. Ebrahimzadeh, and M. H. Dunn, IEEE J. Quantum Electron. 35, 1666 (1999).15. O. B. Jensen, T. Skettrup, O. B. Petersen, and M. B. Larsen, J. Opt. A 4, 190 (2002).16. D. J. M. Stothard, J.-M. Hopkins, D. Burns, and M. H. Dunn, Opt. Express 17, 10648 (2009).17. D. J. M. Stothard and M. H.