We report on performance studies of high-average-power single-pass picosecond optical parametric generation (OPG) and amplification (OPA) tunable near 2 µm in MgO:PPLN pumped by an Yb-fiber laser at 1.064 µm and 80 MHz pulse repetition rate. The simple setup based on two identical crystals, and without the need for an intermediate delay line for synchronization, delivers up to 6.3 W of average power at an overall conversion efficiency of ∼50% and is tunable across 1902–2415 nm. We present systematic characterization of OPG and OPA stages to compare their performance and investigate the effect of parametric generation in the high-gain limit, enabling high output power and full-width-half-maximum (FWHM) spectral bandwidths as large as 189 nm. The OPG-OPA output exhibits excellent passive power stability better than 0.3% rms and central wavelength stability better than 0.03% rms over 1 hour, in high spatial beam quality with M2<2. The OPG output pulses have duration of 5.2 ps with a FWHM spectral bandwidth of 117 nm at 2123 nm, resulting in a time-bandwidth product of ΔτΔν∼40, indicating ∼4 times temporal compression compared to the input pump pulses. Theoretical simulations confirm the effect of pump beam divergence on the observed shift in wavelength tuning with respect to temperature, while the exponential gain in the parametric process is identified as playing a key role in the resulting pulse compression.
© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Tunable laser sources providing high-repetition-rate picosecond pulses with high average power in the infrared (IR) are desirable for a variety of scientific and technological applications in spectroscopy , imaging , material characterization  and processing . In particular, the strong water absorption in the 2-µm wavelength range makes them attractive for a wide range of applications such as the study of large temperature variations during melting and freezing processes in water . More importantly, laser-assisted surgical procedures greatly benefit from compact 2-µm sources . While significant progress has been made in the development of ultrafast mode-locked lasers near 2 µm , traditionally, the most common approach to generate picosecond and femtosecond pulses in this spectral range has relied on nonlinear frequency conversion techniques [8–19]. Various nonlinear optical sources based on optical parametric generation (OPG) [11–14], amplification (OPA) [15–18], and synchronously pumped optical parametric oscillators (SPOPOs) [8–10] have been extensively investigated. Each of these processes has a specific set of requirements with regard to the choice of pump laser and nonlinear material candidate. In particular, SPOPOs driven by mode-locked Yb-fiber lasers at 1064 nm have been established as reliable sources of high-repetition-rate picosecond pulses across ∼1.45–4.2 µm in the mid-IR . Due to the high pump pulse repetition rate (∼80 MHz), long pulse duration (∼20 ps), and low pump pulse energy (<250 nJ), the SPOPO approach has been established as the most viable method for picosecond pulse generation in the near-IR to mid-IR. However, a major challenge in the development of ultrafast SPOPOs is the requirement for synchronization between the OPO cavity length and the input pump pulse repetition rate, which also dictates the minimum system size. Moreover, as a resonant device, the SPOPO requires a cavity comprising specially coated mirrors to provide optical feedback at the signal and/or idler wavelengths, and of suitable focal length to enable optimum spatial mode-matching between the resonant parametric wave(s) and the input pump beam, which result in relatively high overall system complexity and cost. For many applications, SPOPOs also require active cavity length control for maximum output power and spectral stability , further increasing operating demands. On the other hand, OPG schemes, which are typically based on single-pass architectures, represent a far simpler approach, but generally require high-energy, low-repetition-rate pulsed lasers to generate mJ-class output pulse energies [15,16]. Moreover, while in ultrafast SPOPOs the low pump pulse energies and the losses associated with the optical cavity and the nonlinear crystal have major impact on operation threshold, such challenges are easily circumvented in single-pass OPG schemes by using high-energy pump pulses, leading to relatively low parametric generation thresholds. On the other hand, implementation of single-pass OPG approach in the high-repetition-rate picosecond pulse regime with very low pulse energies presents significant challenges, due to the requirement for sufficiently high pumping intensity under strong focusing, which can lead to catastrophic damage to the nonlinear gain crystal before any practical macroscopic output power can be generated from the device.
Progress in quasi-phase-matched (QPM) nonlinear materials such as MgO-doped periodically poled lithium niobate (MgO:PPLN), offering improved optical quality, together with high nonlinearity (deff∼16 pm/V) under type-0 (e→ee) noncritical phase-matching and long available interaction lengths (>50 mm), and further advances in high-average-power, MHz-repetition-rate picosecond pump lasers near 1 µm, of high beam quality enabling the attainment of high peak intensities under strong focusing, present new opportunities for the development of tunable parametric sources near 2 µm in simple, robust, and reliable single-pass architecture. It is thus important to investigate such single-pass approaches as practical alternatives to SPOPOs for the generation of high-repetition-rate picosecond pulses in the near-IR and mid-IR, without the need for synchronous cavities and associated complexities, yet providing comparative average powers and other practical performance capabilities. In this context, the ongoing advances in mode-locked picosecond fiber pump lasers providing high average powers at 1064 nm, combined with near-degenerate parametric interaction for ∼2 µm generation in MgO:PPLN, opens up new avenues for the deployment of cascaded OPG-OPA schemes, where both power and spectral characteristics can be tailored in a simple single-pass architecture with compact footprint. Such schemes involving OPG and OPA have been investigated in the high-pulse-energy regime using birefringent nonlinear materials such as BBO , BIBO , as well as QPM monolithic PPLN . The low pulse repetition rates in these devices require lengthy delay lines between successive OPG and OPA stages to achieve amplification. In the low-pulse-energy regime, such amplification has been demonstrated by injection-seeding the OPG stage with a narrowband continuous-wave (cw) laser [14,15]. Single-pass OPG pumped at 1064 nm in a 55-mm-long PPLN crystal provided tunable near-IR radiation over 1571–1642 nm in the signal, with an idler tuning across 3026–3302 nm. The source delivered a maximum total average power of 8.9 W in 3.6 ps pulses at 82 MHz repetition rate for a pump power of 24 W, with an OPG threshold of 7 W , and an output pulse compression factor of ∼2 was also observed relative to the input pump. The output power was further boosted to 9.5 W by injection-seeding the OPG with a cw diode laser at 1580 nm . In another report, a 40-mm-long MgO:PPLN crystal was used for single-pass OPG, providing tunable radiation over 1450–3615 nm, with a maximum total power of 3.9 W for a pump power of 12 W in 130 ps pulses at 1 MHz repetition rate, with OPG threshold of ∼5 W, and a pulse compression of <1.15 times compared to the input pump. Subsequent OPA using a tunable cw seed laser improved the output power to 4.3 W, however, resulted in catastrophic damage to the MgO:PPLN crystal at a pump power of 9.2 W . All of the above demonstrations have been focused on OPG away from degeneracy. However, a single-pass OPG source based on a 20-mm-long PPLN crystal pumped by 20 ps pulses at 20 MHz from a mode-locked Yb-fiber laser at 1064 nm was also reported, providing tunable output across 2000–2200 nm, with a nominal average output power of 550 mW and full-width at half-maximum (FWHM) spectral bandwidth of 250 nm centered at 2128 nm .
Recently, we demonstrated a cascaded single-pass OPG-OPA source based on two identical MgO:PPLN crystals pumped by an Yb-fiber laser at 1064 nm, providing high-repetition-rate picosecond pulses at high average power and conversion efficiency near the ∼2 µm wavelength range . Here we report detailed and systematic characterization of this OPG-OPA source, where we investigate various performance parameters including the effects of high parametric gain and beam divergence on tuning as well as the output spectral bandwidth. Output pulses as short as 5.2 ps with a compression factor of ∼4 relative to the input pump pulses and FWHM spectral bandwidths as large as 189 nm have been generated. Using theoretical simulations, we also confirm the role of exponential gain in temporal pulse compression from the pump to OPG output, where close agreement between calculations and experimental data is obtained. The OPG-OPA system provides up to 6.3 W of average output power at an overall conversion efficiency of ∼50%. The source is tunable across 1902–2415 nm and exhibits excellent passive power stability better than 0.3% rms and a central wavelength stability better than 0.03% rms over 1 hour, in high spatial beam quality with M2<2.
2. Experimental setup
The schematic of experimental setup for the single-pass high-repetition-rate picosecond OPG-OPA system is the same as that used in our previous report , and is shown in Fig. 1. The input pump source is an Yb-fiber laser (Fianium, FP1060-20) delivering an average power of up to 14 W at 1064 nm in 21 ps pulses at 80 MHz repetition rate. The laser has a FWHM spectral bandwidth of ∼1.5 nm, with a corresponding time-bandwidth product of ∼8.4, which is ∼19 times the transform-limit, assuming Gaussian pulse shape. The pump beam has a TEM00 mode profile with a quality factor of M2∼1. A combination of a half-wave plate (λ/2) and a polarizing beam-splitter (PBS) is used to control the pump power, while a second λ/2-plate provides the required polarization for optimum phase-matching in the nonlinear crystals used in the OPG- OPA setup comprising two separate single-pass stages. The first stage is based on a 50-mm-long, 2-mm-wide, 1-mm-thick MgO:PPLN crystal, C1, with a single QPM grating period of Λ=32.16 µm for OPG under type-0 (e→ee) phase-matching. Using the lens, L1, the pump beam is focused to a waist radius of w01∼52 µm at the center of the crystal, C1, with a corresponding focusing parameter of ξ1∼1.4 . A dichroic mirror, M1, with R>99.9% at 1064 nm, T>99% over 1605–2128 nm, and T>86% over 2128–2415 nm is used to separate the OPG output from the undepleted pump to perform complete characterization after the first stage (not shown in Fig. 1). The output from the OPG stage together with the undepleted pump are then collimated and refocused using the lens, L3, to a waist radius of w02∼39 µm at the center of the crystal, C2, in the OPA stage, with a corresponding focusing parameter of ξ2∼2.6. The nonlinear crystal for OPA is identical to that used for OPG. Both crystals are antireflection (AR)-coated (R<0.5%) at 1064 nm and across 2050–2150 nm, with high transmission (R<10%) over 2250–4300 nm. The output from the OPA stage is collimated using a lens, L4, before extraction from the residual pump. It is to be noted that the AR-coatings on the crystal facets as well as collimation and focusing optics were optimized for the signal, resulting in higher losses at the idler wavelengths. The group velocity mismatch (GVM) between the pump and signal in MgO:PPLN crystal is calculated to vary from 116.5 fs/mm at 1900 nm to 113.3 fs/mm at 2128 nm. Therefore, across the operating wavelength range of 1900–2418 nm, it is estimated to be δps=1/vp–1/vs=115 ± 1.6 fs/mm, resulting in a temporal walk-off length, Leff=τp/δps>185 mm. This is much greater than the length of both C1 and C2 crystals, resulting in a temporal walk-off of Δps<5.8 ps over the 50-mm-long crystal length. Further, the GVM between the pump and idler is estimated to be δpi=1/vp–1/vi=107 ± 6.3 fs/mm across the wavelength range of 2128–2400 nm, resulting in temporal walk-off of Δpi<5.7 ps over a 50-mm-long MgO:PPLN crystal . Moreover, due to operation close to degeneracy, the temporal walk-off between the signal and idler is also very low (Δsi<0.7 ps), enabling us to implement a successive OPA stage without the need for a delay line for synchronization between the OPG and OPA stages.
3. Wavelength tuning and output power
Initially, we investigated the wavelength tuning performance of the OPG stage by varying the temperature of the nonlinear crystal, C1. The calculated temperature tuning curves for a single grating period of Λ=32.16 µm in MgO:PPLN, using the relevant Sellmeier equations , is shown in Fig. 2. The normalized single-pass parametric gain provided by the 50-mm-long MgO:PPLN crystal in the low-gain limit is shown in Fig. 2(a). In the low-gain limit (Γ0L ≲ 1), the single-pass nonlinear gain is given by 
The spectral evolution of signal and idler as a function of the temperature after the OPG stage is presented in Fig. 3(a). The spectra were measured using a spectrometer for the 800–2600 nm wavelength range with a resolution of <0.5 nm. By varying the temperature of C1, the OPG source provides central wavelength tuning over 1902–2415 nm. As the temperature is increased from 30 °C to 55 °C, the generated signal and idler spectra tune as independent branches. With further increase in temperature above 55 °C, the two branches merge into a single broad spectrum, owing to the large parametric gain bandwidth of the MgO:PPLN crystal, as we approach degeneracy. Similar measurements at the output to the OPA stage, obtained by varying the temperature of C2, are shown in Fig. 3(b), where broadband signal and idler generation is achieved above 50 °C. The white solid curves in Figs. 3(a) and 3(b) are the theoretical calculations in the high-gain limit with beam divergence, confirming good agreement with the measurements. Figure 3(c) displays the generated total average power at the output to the OPG stage varying from 2.8 W at 30 °C to 140 mW at 80 °C, with a maximum of 3.6 W at 57 °C. For a fixed pump power of 13 W, the OPG output power remains >2.5 W up to a C1 crystal of temperature of 70 °C. Similarly, for a pump power of 12 W at the input to the OPG stage, the total average power after the OPA stage remains constant at ∼6.5 W for C2 crystal temperatures over 30–60 °C, with corresponding signal central wavelengths ranging over 1902–2087nm. Above 60 °C, the OPA output power drops to 520 mW at 84 °C, at a corresponding central wavelength of 2124 nm, as presented in Fig. 3(d).
4. Spectral characterization
The representative spectra recorded at three different temperatures of 30 °C, 60 °C and 80 °C at the output of the OPG stage in comparison with those at the OPA output, measured at a pump power of 12 W at the input to the OPG stage, are presented in Fig. 4. The amplification provided by the OPA stage results in a slight increase in bandwidths compared to the OPG stage. This could be attributed to the increase in parametric gain and bandwidth in the OPA stage in the presence of seed from the OPG stage. With both crystals at a temperature of 30 °C, the output signal from the OPA stage has a spectrum centered at 1902 nm with a FWHM bandwidth of 50 nm. The corresponding idler branch centered at 2415 nm can also be seen in Fig. 4(a). It is to be noted that the central wavelengths were estimated using a center-of-mass algorithm, accounting for the multi-peak structure on the generated spectra. The lower idler power relative to the signal in Fig. 4(a) is partly due to the lower idler photon energy, but mainly attributed to the higher transmission losses through the optical components whose AR coatings were optimized for signal wavelengths. At an operating temperature of 60 °C for both crystals, the output from the OPG as well as OPA stage has a single broadband spectral feature. The spectrum is centered at 2090 nm with a FWHM bandwidth of 160 nm after the OPA stage. Similar measurements at 80 °C resulted in FWHM spectral bandwidth of 133 nm centered at 2119 nm from the OPA stage.
We also compared the generated spectral bandwidths with theoretical calculations, as presented in Fig. 5. The normalized single-pass parametric gain as a function of wavelength is shown in Fig. 5(a). In the absence of the pump beam divergence and in the low-gain limit, independent signal and idler branches are expected. However, in the high-gain limit, the intensity-dependent parametric gain increases the overall bandwidth. Moreover, the pump beam divergence together with high gain results in a FWHM gain bandwidth of ∼260 nm. The comparison of this simulated gain bandwidth with the experimentally measured spectra is presented in Fig. 5(b). The upper abscissa of Fig. 5(b) shows the deviation of the generated wavelength from that expected in the low-gain limit. While the generated FWHM bandwidth of ∼189 nm centered at 2111 nm after the OPA stage is larger than that after the OPG stage (∼122 nm), it is still below the theoretically estimated bandwidth in the high-gain limit. This indicates that the generated bandwidth is not limited by the gain bandwidth. Further, the pump acceptance bandwidth in an ultrafast femtosecond or picosecond parametric process is determined by the effective interaction length limited by the GVM between the pump and signal. Due to the relatively long pump pulse duration of 21 ps and a temporal walk-off-length much greater than the crystal length (Leff>185 mm), the entire pump bandwidth can efficiently contribute to the parametric generation process. On the other hand, the GVM between the signal and idler, δsi, rapidly approaches zero as we tune close to degeneracy, resulting in a large gain bandwidth. Hence, the generated bandwidths are neither limited by Leff, nor by GVM between the interacting beams. Although wider spectral bandwidths are expected due to the equally large parametric gain in the OPA stage, the nominal increase in the spectral bandwidth after the OPA stage is attributed to the significant pump depletion in the OPG stage. We found that independent optimization of the temperature of C1 and C2 resulted in central wavelength tuning with modified spectral properties, but with minimal variation in the generated spectral bandwidth.
5. Power scaling
The power scaling results for the OPG and OPA output as a function of pump power at the input to the OPG stage are presented in Fig. 6(a). As can be seen, the output powers in both OPG and OPA stage increase linearly with pump power. For a maximum available input power of 12.7 W, we were able to generate up to 3.6 W from the OPG stage at a slope efficiency of ∼71%, while operating at a central wavelength of 2085 nm. This corresponds to an optical conversion efficiency of ∼28%. The pump depletion in the OPG stage was recorded to be ∼35% at the highest input pump power of 12.7 W. Similar power scaling measurements at the output of the OPA stage resulted in a maximum total power of 6.3 W at a slope efficiency of ∼80%. The corresponding optical conversion efficiency was measured to be ∼50%, indicating ∼1.8 times enhancement compared to OPG. The corresponding pump depletion at the maximum input pump power also increased by >100% to ∼73% after the OPA stage, resulting in an internal conversion efficiency of ∼68%. While the pump depletion was measured to rise linearly in the OPG stage, a saturation behavior in the pump depletion was observed after the OPA stage above ∼8 W of input pump power. However, the output powers in OPG and OPA stages increased to practical levels beyond 8 W and 5 W of input pump power, respectively, as evident in Fig. 6(a). It is interesting to note that there is no saturation in the output power in either stage, pointing to the feasibility of further power scaling. Using a power meter with a sensitivity better than 1 µW, we measured the threshold pump power of ∼880 mW for OPA, as shown in the inset of Fig. 6(a). The corresponding measurement for OPG resulted in a threshold of 1.9 W. In order to study the effect of pump power on parametric gain, we recorded the OPG output spectrum as a function of input pump power beyond 5 W. The independently normalized spectral evolution of OPG with input pump power is shown in Fig. 6(b), indicating higher output power in the signal compared to the idler. As stated earlier, the relatively low idler power compared to signal is partly due to the quantum defect, but mainly due to the higher transmission loss of optics at the longer idler wavelengths.
6. Power and spectral stability
We performed long-term power and spectral stability measurements of the output from both OPG and OPA stages, with the results presented in Figs. 7(a)–7(c). The output power was recorded to exhibit excellent passive stability better than 0.4% rms from the OPG and 0.3% rms from the OPA stage, over a period of 1 hour, compared to 0.1% rms for the input pump. The long-term spectral stability of the output from the OPG stage, measured at a central wavelength of 2129 nm, is shown in Fig. 7. The measurement resulted in a central wavelength stability better than 0.01% rms and a FWHM bandwidth stability better than 4% rms over 1 hour. Similar measurements of the OPA output resulted in a central wavelength stability better than 0.03% rms at 2146 nm and a FWHM bandwidth stability of 7.4% rms over 1 hour. It is to be noted that the 17 nm difference in the central wavelengths between the OPG and OPA is due to the optimization of C2 crystal temperature for a broader output spectrum.
7. Spatial beam quality
The spatial beam profile of the output from the OPA stage at a central wavelength of 2114 nm, measured using a pyroelectric camera, is shown in Fig. 8(a), confirming single-peak Gaussian distribution and excellent circularity of >95%. Using a lens of focal length, f=50 mm, and scanning beam profiler, we measured the radius of the generated beam over the Rayleigh range, while translating the beam profiler through the focus. By analyzing the variation of the recorded beam radius, we estimated beam quality parameters of Mx2<1.3 and My2<1.3 for the OPG, as presented in Figs. 8(c) and 8(d). Similarly, the spatial beam profile of OPA output is shown in Fig. 8(b), also confirming single-peak Gaussian distribution with circularity of >88%. The beam quality factors in this case were measured to be Mx2<2 and My2<1.9, as shown in Figs. 8(e) and 8(f). The slight asymmetry in the OPA output beam may be attributed to the distortion of the residual pump beam at the output of the OPG stage due to the 35% depletion.
8. Temporal characterization
Finally, we performed temporal characterization of the input pump as well as the generated output pulses from the OPG and OPA using an interferometric autocorrelator. The typical autocorrelation trace of the pump pulse, together with the corresponding spectrum centered at 1064 nm with an FWHM bandwidth of 1.5 nm, is shown in Figs. 9(a) and 9(b). The pump pulses have FWHM duration of 30 ps, resulting in an estimated pulse duration of 21 ps, assuming Gaussian pulse shape. The output pulse duration from the OPG stage, measured at a central wavelength of at 2123 nm, was determined to be ∼5.2 ps, as presented in Figs. 9(c) and 9(d). With a FWHM spectral bandwidth of 117 nm, this results in a time-bandwidth product of ΔτΔν∼40, which is ∼4.8 times that of the input pump pulses, and >90 times the transform-limit for a Gaussian pulse. Moreover, the measured signal pulse duration of 5.2 ps indicates a pulse compression in the OPG stage by a factor of ∼4 with respect to the input pump. Similar autocorrelation measurements of OPG output at other wavelengths resulted in pulse durations ranging from 5.6 ps at 1902 nm to 6.6 ps at 2106 nm. The autocorrelation measurements of OPA output resulted in Gaussian pulse duration of ∼11 ps with a slightly larger FWHM spectral bandwidth of ∼133 nm, as presented in Figs. 9(e) and 9(f), with a corresponding time-bandwidth product of ΔτΔν∼98 and a pulse compression factor of ∼2. The time-bandwidth product can be further improved by controlling the bandwidth using spectral narrowing techniques such as spectral filtering, injection seeding, or by deploying a diffraction grating for selecting a portion of the OPG output spectrum before amplification in the OPA stage. We also measured the idler pulse duration at the longest wavelength at the OPA output, where we were able to separate the signal and idler using a dichroic mirror. The measurement resulted in an idler pulse duration of ∼5.9 ps at 2425 nm. To the best of our knowledge, the shortest pulse duration of 5.2 ps in Fig. 9(c), corresponds to a record pulse compression in a high-average-power, high-repetition-rate, single-pass picosecond OPG scheme [14,18]. This strong pulse compression is achieved due to the exponential gain available for OPG in MgO:PPLN in the high-gain limit. It is to be noted that all autocorrelation measurements were performed for a pump power of 12 W at the input to the OPG stage.
In order to understand the origin of the observed pulse compression in our picosecond single-pass OPG, we performed theoretical simulations corresponding to the experimental conditions using the real laboratory values as input parameters. A Gaussian pump pulse with a 21 ps duration at 80 MHz and an input average pump power of 12 W at 1064 nm, focused to a beam waist of w0∼52 µm in a 50-mm-long MgO:PPLN crystal with a deff∼16 pm/V. The time-dependent parametric gain for OPG at a crystal temperature of 72 °C was then estimated, considering the time-varying input pump intensity leading to a Γ0L∼19, corresponding to a peak pump intensity of 85 MW/cm2. The resulting temporal gain profile contributing to OPG in the high-gain limit then determines the generated output pulse duration. In Fig. 10(a), the calculated normalized time-dependent parametric gain contributing to signal and idler pulse generation as a function of the phase-mismatch parameter, ΔkL, is shown in a color-scale plot. The resulting temporal profile of the signal pulse in the OPG, for a phase-mismatch parameter, ΔkL=0, corresponding to the dotted line in Fig. 10(a), is presented in Fig. 10(b). This indicates an FWHM duration of ∼6.8 ps, in close agreement with the experimental value of 7.4 ps, thus confirming the role of exponential parametric gain in the pulse compression observed in our experiments.
In summary, we have presented theoretical study and experimental characterization of a high-average-power, Yb-fiber-pumped picosecond OPG-OPA system at 80 MHz pulse repetition rate tunable in the 2 µm wavelength range. The source is based on cascaded single-pass OPG and OPA in two identical MgO:PPLN crystals with a single grating period, without the need for an intermediate delay line for temporal synchronization, and can provide a combined signal plus idler tuning over >500 nm, across 1902–2415 nm. We have performed systematic characterization of OPG and OPA stages to compare their performance and investigated the effect of parametric generation in the high-gain limit, enabling high output power and broad spectral bandwidths. Using this simple setup, we have generated 3.6 W and 6.3 W of average power from the OPG and OPA stages, respectively, at an overall conversion efficiency of ∼50% at a central wavelength of 2085 nm, corresponding to a photon conversion efficiency of >25%. The feasibility of power scaling is also confirmed by the absence of any saturation in the output power from the OPA stage. We have further studied the influence of pump beam divergence on the shift in the measured wavelengths towards the lower temperature. Theoretical calculations also indicate that the generated spectral bandwidth are neither limited by the effective interaction length nor by temporal walk-off effects. The generated power and spectral bandwidths are limited by the significant pump depletion in the OPG stage. The final output from OPA-OPG source exhibits excellent passive power stability better than 0.4% rms over 1 hour, in Gaussian spatial profile with >88% circularity and M2<2, and the central wavelength and FWHM bandwidth exhibit excellent passive stability better than 0.03% and 7.4% rms over 1 hour. Owing to the exponential parametric gain provided by the high nonlinearity and strong pump intensity in the 50-mm-long MgO:PPLN crystal, the source generates signal pulses of 5.2 ps duration from the OPG stage, corresponding to a ∼4 times temporal compression with respect to the pump across the tuning range. By deploying pump pulses of higher peak power and independent pump beam lines for the OPG and OPA stages, further enhancement of overall conversion efficiency and parametric gain can be achieved, leading to stronger pulse compression. Moreover, the ability to independently vary the phase-matching conditions in the OPG and OPA stage enables fine control of the central wavelength as well as the accessible spectral bandwidths. The use of further successive amplification stages could not only benefit power scaling, but also help relax the strong focusing condition, while operating in the high-gain limit. The development of this simple and compact high repetition-rate picosecond single-pass OPG-OPA source alleviates the need for complex synchronous cavities associated with SPOPOs, delivering multi-Watt average powers with short pulse duration, providing a cost-effective alternative for a variety of applications.
Ministerio de Ciencia, Innovación y Universidades (TEC2015-68234-R); Severo Ochoa (CEX2019-000910-S); Generalitat de Catalunya (CERCA Programme); Fundación Cellex (Fundación Cellex); Fundació Mir-Puig; State Research Agency (AEI) and the European Social Fund (RYC2019-027144-I/10.13039/501100011033).
The authors declare no conflicts of interest.
1. M. Towriey, A. W. Parker, W. Shaikh, and P. Matousek, “Tunable picosecond optical parametric generator-amplifier system for time resolved Raman spectroscopy,” Meas. Sci. Technol. 9(5), 816–823 (1998). [CrossRef]
2. S. Junaid, S. Chaitanya Kumar, M. Mathez, M. Hermes, N. Stone, N. Shepherd, M. Ebrahim-Zadeh, P. Tidemand-Lichtenberg, and C. Pedersen, “Video-rate, mid-infrared hyperspectral upconversion imaging,” Optica 6(6), 702–708 (2019). [CrossRef]
3. H. Kämmer, G. Matthäus, S. Nolte, M. Chanal, O. Utéza, and D. Grojo, “In-volume structuring of silicon using picosecond laser pulses,” Appl. Phys. A 124(4), 302 (2018). [CrossRef]
4. X. Zhu and S. Jiang, “2 Micron Fiber Laser Enable Versatile Processing of Plastics”, Industrial Laser Solutions, 17–23 (May/June 2016).
5. M. Citroni, S. Fanetti, B. Guigue, P. Bartolini, A. Taschin, A. Lapini, P. Foggi, and R. Bini, “Picosecond optical parametric generator and amplifier for large temperature-jump,” Opt. Express 22(24), 30047–30052 (2014). [CrossRef]
6. M. Durbec, E. Dinkel, S. Vigier, F. Disant, F. Marchal, and F. Faure, “Thulium-YAG laser sialendoscopy for parotid and submandibular sialolithiasis,” Lasers Surg. Med. 44(10), 783–786 (2012). [CrossRef]
7. J. Ma, Z. Qin, G. Xie, L. Qian, and D. Tang, “Review of mid-infrared mode-locked laser sources in the 2.0 µm –3.5 µm spectral region,” Appl. Phys. Rev. 6(2), 021317 (2019). [CrossRef]
8. M. Ebrahim-Zadeh and S. Chaitanya Kumar, “Yb-fiber-laser-pumped ultrafast frequency conversion sources from the mid-infrared to the ultraviolet,” IEEE J. Sel. Top. Quantum Electron. 20(5), 624–642 (2014). [CrossRef]
9. S. Chaitanya Kumar and M. Ebrahim-Zadeh, “Yb-fiber-based, high-average-power, high-repetition-rate, picosecond source at 2.1 µm,” Laser Photonics Rev. 10(6), 970–977 (2016). [CrossRef]
10. S. Chaitanya Kumar, A. Esteban-Martin, T. Ideguchi, M. Yan, S. Holzner, T. W. Hänsch, N. Picqué, and M. Ebrahim-Zadeh, “Few-cycle, broadband, midinfrared optical parametric oscillator pumped by a 20-fs Ti:sapphire laser,” Laser Photonics Rev. 8(5), L86–L91 (2014). [CrossRef]
11. B. Nandy, S. Chaitanya Kumar, J. C. Casals, H. Ye, and M. Ebrahim-Zadeh, “Tunable high-average-power optical parametric oscillators near 2 µm,” J. Opt. Soc. Am. B 35(12), C57–C67 (2018). [CrossRef]
12. A. Laubereau, L. Greiter, and W. Kaiser, “Intense tunable picosecond pulses in the infrared,” Appl. Phys. Lett. 25(1), 87–89 (1974). [CrossRef]
13. H. He, Q. Zhao, J. Guo, Y. Lu, Y. Liu, and J. Wang, “Picosecond optical parametric generation in flux grown KTP,” Opt. Commun. 87(1-2), 33–35 (1992). [CrossRef]
14. B. Köhler, U. Bäder, A. Nebel, J. P. Meyn, and R. Wallenstein, “A 9.5-W 82-MHz-repetition-rate picosecond optical parametric generator with cw diode laser injection seeding,” Appl. Phys. B 75(1), 31–34 (2002). [CrossRef]
15. J. Y. Zhang, J. Y. Huang, Y. R. Shen, and C. Chen, “Optical parametric generation and amplification in barium borate and lithium triborate crystals,” J. Opt. Soc. Am. B 10(9), 1758–1764 (1993). [CrossRef]
16. Z. Sun, M. Ghotbi, and M. Ebrahim-Zadeh, “Widely tunable picosecond optical parametric generation and amplification in BiB3O6,” Opt. Express 15(7), 4139–4148 (2007). [CrossRef]
17. A. Chiang, T. Wang, Y. Lin, C. Lau, Y. Chen, B. Wong, Y. Huang, J. Shy, Y. Lan, Y. Chen, and P. Tsao, “Pulsed optical parametric generation, amplification, and oscillation in monolithic periodically poled lithium niobate crystals,” IEEE J. Quantum Electron. 40(6), 791–799 (2004). [CrossRef]
18. L. Xu, H. Chan, S. Alam, D. J. Richardson, and D. P. Shepherd, “High-energy, near- and mid-IR picosecond pulse generated by a fiber-MOPA-pumped optical parametric generator and amplifier,” Opt. Express 23(10), 12613–12618 (2015). [CrossRef]
19. Y. Isyanova, W. Tian, and P. F. Moulton, “High-repetition rate, picosecond-pulse, tunable, mid-IR PPLN OPG source,” Proc. SPIE 9731, 97310W (2016). [CrossRef]
20. B. Nandy, S. Chaitanya Kumar, and M. Ebrahim-Zadeh, “Fiber-laser-pumped picosecond optical parametric generation and amplification in MgO:PPLN,” Opt. Lett. 45(22), 6126–6129 (2020). [CrossRef]
21. G. D. Boyd and D. A. Kleinman, “Parametric interaction of focused Gaussian light beams,” J. Appl. Phys. 39(8), 3597–3639 (1968). [CrossRef]
22. P. Di Trapani, A. Andreoni, C. Solcia, P. Foggi, R. Danielius, A. Dubietis, and A. Piskarskas, “Matching of group velocities in three-wave parametric interaction with femtosecond pulses and application to traveling-wave generators,” J. Opt. Soc. Am. B 12(11), 2237–2244 (1995). [CrossRef]
23. O. Paul, A. Quosig, T. Bauer, M. Nittmann, J. Bartschke, G. Anstett, and J. A. L’Huillier, “Temperature-dependent Sellmeier equation in the MIR for the extraordinary refractive index of 5% MgO doped congruent LiNbO3,” Appl. Phys. B: Lasers Opt. 86(1), 111–115 (2006). [CrossRef]
24. M. Ebrahim-Zadeh and M. H. Dunn, “Optical Parametric Oscillators,” In OSA Handbook of Optics, Vol. 4, pp. 2201–22712 (McGraw-Hill, New York2000).
25. D. Kleinman, “Theory of optical parametric noise,” Phys. Rev. 174(3), 1027–1041 (1968). [CrossRef]
26. N. P. Barnes and V. J. Corcoran, “Parametric generation processes: spectral bandwidth and acceptance angles,” Appl. Opt. 15(3), 696–699 (1976). [CrossRef]