Open Access
Add to BibSonomy
Abstract
We present a tunable (6.62-11.34 µm), singly-resonant, cascade optical parametric oscillator with intracavity pumping of BaGa2GeSe6 in the second stage and spectral narrowing realized by a Volume Bragg Grating acting on the signal wave of the first stage which serves as a pump for the second stage. The maximum energy achieved near 8 µm in the narrowband regime is 1.1 mJ at 100 Hz (spectral width: $\sim$20 cm-1, pulse duration: $\sim$7 ns). The overall conversion efficiency from 1 to 8 µm for broadband and narrowband operation is 4.0% and 3.1%, respectively, corresponding to quantum efficiencies of 31% and 23%.
© 2024 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
A new class of non-oxide Barium compounds with different noncentrosymmetric structures is coming of age recently for frequency conversion by three-wave interactions in the mid-IR part of the spectrum [1]. As nonlinear optical crystals, these chalcogenides offer few essential advantages over the classical and commercially available chalcopyrites AgGaS2 (AGS) and AgGaSe2 (AGSe), such as surface chemical stability, no need of post-growth annealing, and better thermo-mechanical properties [1,2]. However, their characterization and wide application are hampered by the lower crystallographic symmetry. Out of the four such Ba chalcogenides reviewed in [1], the monoclinic (biaxial) ternary selenide BaGa4Se7 (BGSe) is by far the most thoroughly studied and the one used in more than 90% of the demonstrated applications, in particular in optical parametric oscillators (OPOs).
Due to difficulties in finding the optimum orientation for maximum effective nonlinearity, its quaternary counterpart, the trigonal (uniaxial) BaGa2GeSe6 (BGGSe) has been rarely employed so far, with the most successful applications for few-cycle femtosecond pulse generation in the mid-IR. In [3], a seeded BGGSe optical parametric amplifier (OPA) was pumped at 1.96 µm, generating sub-160-fs idler pulses between 4 to 12 µm with a maximum energy of 1.05 µJ (105 mW at 100 kHz) at 9 µm. The same type-I BGGSe sample was employed in [4] for difference-frequency generation (DFG) between the signal and idler outputs of a synchronously pumped near-IR OPO, to produce tunable femtosecond pulses in the 3.88-17.65 µm range with a maximum energy of 1.34 nJ (∼54 mW at 40 MHz) at 4.8 µm [4]. The advantageous dispersive properties of BGGSe enabled the DFG of sub-4 optical cycle (91 fs) carrier-envelope-phase (CEP) stable pulses at 7 µm with a spectrum spanning over 2400 nm at −20 dB [5]. In this case the two output beams from an Er-fiber laser system (Er- and Tm/Ho-fiber amplified outputs) were mixed producing a mid-IR pulse energy of 21 pJ (2.1 mW at 100 MHz). The same type-I BGGSe sample was employed also for intra-pulse DFG producing 86-fs pulses (2.5 optical cycles) at 10.3 µm [6] as a key component of a novel high-brightness seven-octave and few-cycle CEP stable light source. Finally, mixing of discrete lines of Q-switched CO and CO2 lasers was also studied in type-I BGGSe [7–10], with the DFG providing discrete tuning in the 11-17 µm spectral range [10].
BGGSe can be considered nowadays to be a relatively well characterized nonlinear optical crystal [1], in particular the components of its nonlinear susceptibility tensor have been fully resolved including relative signs and the dependence on the octant selection [11], and the Sellmeier equations have been refined using phase-matched three-wave interactions [12]. More recently, also its thermo-optic [13], and thermo-mechanical [14] properties have been accurately described and it has been included in the widely used SNLO software package [15]. However, it shall be outlined that in all the above demonstrated applications the BGGSe samples were cut for type-I phase-matching (PM) at φ = 30°, utilizing only the d11 tensor component, because optimization by the azimuthal angle would have required knowledge of the relative signs which were only subsequently established [11].
Selenides such as AGSe, BGSe, and BGGSe are transparent up to $\sim$18 µm and possess substantially higher nonlinearity compared to the respective sulfide compounds. While the larger band-gaps of BGSe and BGGSe, in contrast to AGSe, still enable pumping near 1 µm without the onset of two-photon absorption, longer pump wavelengths as demonstrated in the above applications ensure damage-free long-term operation. Recently, it was also established that BGGSe is free of the growth striae observed in BGSe and seems to be more promising for further research. In the present work, we describe a narrowband, nanosecond BGGSe OPO pumped intracavity by an oxide-based OPO, in turn pumped by a Nd:YAG laser. This scheme profits from the higher intracavity pump intensity and the compact and robust design.
2. Experimental setup
The experimental setup is shown in Fig. 1(a). The beam of a multimode (M2 $\sim$2, measured after an expanding vacuum telescope with a diamond pinhole), multifrequency (Δν $\sim$1 cm-1), diode-pumped, Q-switched, 1.064 µm Nd:YAG laser-amplifier at 100 Hz is reimaged by a lens telescope (L1 and L2) to a waist diameter of 3.26 mm (almost equal Gaussian fits in the two planes) in the position of the Rb:PPKTP (Rb-doped, periodically-poled KTiOPO4) crystal employed in the first OPO stage. The pump pulse duration is $\sim$8 ns.
Fig. 1. (a) Experimental setup of the cascade OPO and (b) unpolarized transmission of the AR-coated BGGSe sample (shown as an inset). -1 and -2 in (a) refer to the 1st and 2nd OPO stage.
The Rb:PPKTP sample is 8-mm wide (along y-axis) × 5-mm thick (along z-axis) in aperture and 20-mm along the x-axis (propagation direction). The partial substitution of K by Rb was shown to overcome the major limitations of KTP with respect to quasi-phase-matching (QPM) device fabrication for such thick samples while at low doping levels (< 1 mol. %) linear and nonlinear optical properties are expected to remain unchanged [16]. Since the stoichiometric Rb isomorph of KTP is not readily available from commercial vendors with acceptable quality for high-yield electric-field periodic poling, Rb-doped KTP, is a very promising alternative for this [16]. The QPM part of our crystal is 5-mm wide, 5-mm thick, and 15-mm long, with a period of Λ = 38.75 µm. These dimensions are smaller than the physical ones due to the requirements of electric isolation during the poling process [16]. The Rb:PPKTP crystal is antireflection (AR)-coated at 1.064 µm and 1.91–2.4 µm. The 45° dichroic pump mirrors DM1 and DM2 have high reflection (HR) at 1.064 and 0.532 µm, and high transmission (HT) at 1.9–2.5 µm. DM3 is HR at 1.064 µm for a double pump pass and HT at 0.532 µm, to eliminate the feedback of parasitically generated second harmonic in the QPM process.
Apart from the different Rb:PPKTP sample, there are two significant modifications compared to the preliminary setup presented in [17]: the PPKTP OPO is singly-resonant for the signal (1.946 µm) and the double pass for its idler (2.349 µm) is eliminated. Thus, the resonator mirror (RM) reflects only the signal while the idler exits the cavity both through RM and the ZnSe dichroic mirror DM4 which also reflects only the signal, see Fig. 1(a). This suppresses in general back conversion in the first stage. DM4 is HT for the signal (2.553 µm) and idler (8.185 µm) of the second BGGSe OPO stage. The shared Au (on Cu substrate) total reflector (TR) serves as an end mirror for both cavities and ensures a double pass for the mid-IR idler before extracting it through the HT ZnSe output coupler (OC) which is HR for the 2.553 µm signal. The physical length of the PPKTP OPO cavity is ∼100 mm, and that of the BGGSe OPO is ∼45 mm. The separation between RM and DM3 is ∼55 mm.
The BGGSe crystal, pumped intracavity by the 1.946 µm signal from the first stage, has an aperture of 7.15 × 6 mm2 and a length of 9.33 mm. It is cut at θ = 27.5° and φ = 0° for type-II (oeo) PM and AR-coated near 2 µm for both signal waves which results in somewhat improved transmission also in the mid-IR. The measured transmission of the AR-coated sample is shown in Fig. 1(b). For point group 3, the effective nonlinearity amounts to 29.4 or 12.5 pm/V depending on the octant [11]. In fact higher effective nonlinearity is achievable for BGGSe in type-II PM because three instead of two tensor components are involved [11]. Our choice of type-II PM was, however, primarily dictated by its narrower spectral acceptance and the azimuthal angle of the sample used was in fact not optimized for maximized effective nonlinearity.
We established that the above modifications and the purely singly-resonant operation of both stages contributed to almost doubling the overall conversion efficiency and output energy compared to [17], and greatly improved the stability of the BGGSe OPO mid-IR idler output. Moreover, this design enables seamless transition to narrowband operation by replacing RM by a Volume Bragg Grating (VBG) which represents an essential novelty in the present work. The VBG employed here is 9-mm thick and AR-coated on both sides (aperture: 5 × 23 mm2) with the grating slightly tilted (about 2°) from normal incidence to the AR-coated front glass surface in order to avoid parasitic subcavity effects.
3. Results and discussion
Using two additional TRs, not shown in Fig. 1(a), one instead of RM/VBG to reflect both Signal-1 and Idler-1, and another one to retro-reflect Idler-1 after DM4, i.e. making the first stage doubly-resonant, we first established that the mid-IR Idler-2 was strongly fluctuating (±50%) at a pump level of 22 mJ. Without such feedback for Idler-1, i.e. for the setup as depicted in Fig. 1(a), the mid-IR output energy greatly stabilized and increased more than twice. Under these conditions we measured the tuning of the OPO which is shown in Fig. 2(a). The overall tuning for the mid-IR Idler-2 extended from 6.62 to 11.34 µm. The mid-IR wavelengths (Idler-2) were calculated from the measured pump (Signal-1) and signal (Signal-2) wavelengths. Comparing with [17] it can be seen that the agreement with the refined Sellmeier equations [12] is much better. In fact the deviation within their range of validity (up to 10.591 µm) is almost independent of the wavelength and can be attributed to a cut angle deviation (roughly 0.5°) which is an acceptable tolerance in crystal specifications.
Fig. 2. (a) Angle tuning and (b) mid-IR idler output energy versus wavelength of the type-II BGGSe broadband OPO stage at a pump energy of 24 mJ at 1.064 µm. The calculated curves in (a) are based on the Sellmeier equations from [12] for a pump (Signal-1) wavelength of 1.946 µm.
Figure 2(b) shows the obtained mid-IR idler energy versus wavelength of the broadband BGGSe OPO at a fixed (1.946 µm) signal wavelength of the PPKTP OPO using RM as an end mirror which is HT for Idler-1 and avoids its double pass. The lower limit of the tuning range (6.62 µm) is defined by the transmission characteristics of DM4. The enhancement close to normal incidence (near 8 µm) is due to subcavity effects. Decreasing energy towards yet longer wavelengths corresponds to decreasing nonlinear coupling away from degeneracy, however, it shall be outlined that the upper limit is also approaching a potentially retracing point outside the validity of the existing Sellmeier equations [12].
The energy characteristics in Fig. 3 are recorded close to normal incidence for the BGGSe crystal which permits TR in Fig. 1(a) to be translated by $\sim$5 mm, shortening the length of both cavities. The maximum energies achieved in the mid-IR amount to 1.45 mJ (145 mW at 100 Hz) at 8.185 µm in the broadband mode with RM and 1.10 mJ (110 mW at 100 Hz) in the narrowband mode with the VBG, for a pump energy of 36 mJ at 1.064 µm. The VBG was centered at 1.9445 µm, it was HT for Idler-1, and the mid-IR Idler-2 occurred in this case at $\sim$8 µm.
Fig. 3. Input-output dependence of the cascade OPO in the broadband (BB) and narrowband (NB) regimes of operation, i.e. RM and VBG used in Fig. 1(a). Insets: far-field mid-IR idler output beam profiles.
The corresponding conversion efficiencies for the broadband and narrowband cases from the 1.064 µm pump pulses are 4.0% and 3.1%, or 31% and 23% in terms of quantum efficiencies. The higher values for the broadband case are attributed to the different saturation behavior in the absence of spectrally selective elements for Signal-2. This behavior is, however, very complex in the coupled OPO stages and we observed saturation also for the first stage output related to increased depletion, by monitoring the Idler-1 energy behind DM4. The extrapolated threshold in the narrowband mode is lower ($\sim$8 mJ) compared to the broadband case (roughly 10 mJ). The pulse-to-pulse energy stability (RMS) for the idler in the narrowband mode was measured to be 2.3% for 500 pulses at maximum pump level (36 mJ).
The spectral, temporal and spatial properties are characterized at a pump energy of 24 mJ at 1.064 µm. The measured bandwidth of the broadband Signal-1 at 1.946 µm using RM is 141 cm-1 (53.3 nm), practically coinciding with the calculated spectral acceptance for a DFG process in the PPKTP crystal. With this PPKTP OPO signal pulse serving as a pump for BGGSe, its bandwidth plays a dominant role for the output of the second stage since the acceptance bandwidth for type-II PM in BGGSe is much lower (38 cm-1). The VBG narrows the bandwidth of the PPKTP OPO signal at 1.9445 µm down to less than 1.6 cm-1 or 0.6 nm, measured with a resolution of 0.3 nm, see Fig. 4(a). The first stage idler has a very similar bandwidth of 1.7 cm-1, or 0.9 nm near 2.35 µm, see Fig. 4(a), which is reasonable having in mind the $\sim$1 cm-1 bandwidth of the 1.064 µm pump. The measured bandwidth of the BGGSe signal near 2.57 µm is 6.7 cm-1 or 4.5 nm, see Fig. 4(a). The calculated output bandwidth for a DFG process in the BGGSe crystal assuming narrowband 1.9445 µm pump amounts to 38.4 cm-1. This translates to roughly 25 nm at 2.57 µm (Signal-2) and 245 nm near 8 µm (Idler-2). Direct measurement of the Idler-2 bandwidth near 8 µm gave a value of 132 nm (roughly 20 cm-1), see Fig. 4(b). The narrower measured bandwidths compared to the calculation for DFG are normally attributed to multiple pass OPO spectral narrowing effect which is more pronounced for the resonated signal at 2.57 µm. The measured bandwidths from the second stage depend, however, on the pump level. Close to the threshold they were roughly two times narrower so that a corresponding increase can be expected far above threshold at the maximum applied pump energy of 36 mJ.
Fig. 4. (a) Recorded Signal-1, Idler-1, and Signal-2 spectra, and (b) Idler-2 spectrum centered at 7.99 µm.
The mid-IR idler bandwidth when using RM (broadband case) was too large to be directly measured. However, judging from the spectral widths measured for Signal-2 in the broadband and narrowband cases, the spectral narrowing factor shall exceed 20. This more than compensates for the roughly 25% lower output energy in Fig. 3.
The measured pulse durations are shown in Fig. 5. They were similar in the narrowband and broadband regimes of operation. The mid-IR pulse duration near 8 µm amounts to $\sim$7 ns, somewhat shorter than the pump pulse for the second stage (8.3 ns).
Fig. 5. Measured pulse durations of Signal-1 and Idler-1 by a InGaAs photodiode with a 70-ps response time, and Idler-2 by a (HgCdZn)Te detector with a rise time of 2 ns.
The beam propagation factor (M2) for the mid-IR idler beam was also not very different in broadband and narrowband operation. It was estimated recording the beam profile with a pyroelectric camera behind a CaF2 focusing lens. In terms of second moments the average value (for the two planes) is 27 for narrowband operation, see Fig. 6. It is slightly higher in the uncritical for BGGSe (vertical) plane. In broadband operation the average value was somewhat lower (21). As can be seen from the insets in Fig. 3, an interference pattern occurs in the case of narrowband operation though its origin could not be identified since the construction of the PY-III-C-B camera (Spiricon) window was unknown.
Fig. 6. Measured and fitted Idler-2 beam diameters at $\sim$8 µm after a 75-mm lens to estimate the beam propagation factor M2 in the narrowband case.
Comparing the performance of this BGGSe sample with AGSe we are confident that, fortunately, it was cut in an octant where the contributions of the tensor components add up, with an effective nonlinearity as high as 29.4 pm/V [11]. However, this applies only for the two tensor components d11 and d31 because at φ = 0°, the contribution of d22 is suppressed. Further increase of the effective nonlinearity by more than 10% will be possible at an optimum azimuthal angle of φ = 12.7°, which is weakly dependent on the interacting wavelengths [11], by involving all three nonlinear coefficients which is possible only for type-II interaction in BGGSe. The present experiment in fact presents the only application of BGGSe in type-II phase-matching [17] which in general is capable of providing higher effective nonlinearity, higher compared also to the classical AGSe crystal [11].
4. Conclusion
Cascade type mid-IR OPOs have been initially realized using AGSe in the second stage but subsequently the interest shifted to ZnGeP2 (ZGP) due to its superior nonlinear-optical and thermo-mechanical properties [2]. However, selenides such as AGSe, BGSe, and BGGSe provide access to much longer wavelengths than phosphides. Comparing the achieved results with previous realizations of the cascade scheme with intracavity pumping, the improvement in terms of mid-IR energy / conversion efficiency relative to [17] (preliminary work) and [18] (using AGSe), is roughly 2 and 8 times, respectively. Moreover, comparing with AGSe for which optical damage was observed even when reducing the repetition rate down to 10 Hz [18], no degradation of the performance of BGGSe was encountered in the present experiment at 100 Hz under similar pump conditions. For the first time we employed a VBG for narrowing the mid-IR output spectral bandwidth and increasing its spectral density.
In conclusion, BGGSe seems very promising for parametric down-conversion of 2-µm laser radiation (Tm- and Ho-lasers) to the mid-IR. While it shares a similar transmission range with AGSe, its full tuning potential in the mid-IR cannot be exploited with the narrowband type-II PM due to the retracing behavior. Pump wavelengths near 1.56 µm from Er-lasers or oxide based OPOs like the one employed in the present work might prove more feasible for tunable narrowband operation to longer wavelengths but accurate estimates require refinement of the dispersion relations well above 10.6 µm [12].
Acknowledgments
We acknowledge the contributions of A. Boyko (Fig. 1(b)), V. Badikov and D. Badikov (resources) in an early stage of this work, see [17].
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. V. Petrov, V. V. Badikov, D. V. Badikov, et al., “Barium nonlinear optical crystals for the mid-IR: characterization and some applications,” J. Opt. Soc. Am. B 38(8), B46–B58 (2021). [CrossRef]
2. V. Petrov, “Frequency down-conversion of solid-state laser sources to the mid-infrared spectral range using non-oxide nonlinear crystals,” Prog. Quantum Electron. 42(1), 1–106 (2015). [CrossRef]
3. M. Baudisch, M. Beutler, M. Gebhardt, et al., “Sub-8 optical cycle, 4-12 µm tunable, µJ-level pulse generation via a BaGa2GeSe6-based, 1.96 µm pumped OPA at 100 kHz,” XXIst International Conference on Ultrafast Phenomena, Hamburg, Germany, July 15-20, 2018, paper WED.3B.2.
4. G. Stibenz, M. Beutler, I. Rimke, et al., “Femtosecond mid-IR difference-frequency generation in BaGa2GeSe6 from a 40 MHz optical parametric oscillator pumped at 1035 nm,” Conference on Lasers and Electro-Optics (CLEO), OSA Technical Digest (Optical Society of America, 2018), paper STh4F.5.
5. U. Elu, L. Maidment, L. Vamos, et al., “Few-cycle mid-infrared pulses from BaGa2GeSe6,” Opt. Lett. 45(13), 3813–3815 (2020). [CrossRef]
6. U. Elu, L. Maidment, L. Vamos, et al., “Seven-octave high-brightness and carrier envelope phase-stable light source,” Nat. Photonics 15(4), 277–280 (2021). [CrossRef]
7. D. V. Badikov, V. V. Badikov, A. A. Ionin, et al., “Sum-frequency generation of Q-switched CO laser radiation in BaGa2GeSe6 and GaSe nonlinear crystals,” Opt. Quantum Electron. 50(6), 243 (2018). [CrossRef]
8. A. A. Ionin, D. V. Badikov, V. V. Badikov, et al., “Sum frequency generation of multi-line slab radio frequency discharge carbon monoxide laser system with intracavity nonlinear BaGa2GeSe6 crystal,” Opt. Lett. 43(18), 4358–4361 (2018). [CrossRef]
9. A. A. Ionin, I. O. Kinyaevskiy, Y. M. Klimachev, et al., “Broadband (1.7–6.0 µm) multiline CO laser system with intra- and extracavity sum frequency generation in BaGa2GeSe6 crystals,” Opt. Laser Technol. 115, 205–209 (2019). [CrossRef]
10. A. A. Ionin, I. O. Kinyaevskiy, Y. M. Klimachev, et al., “Frequency conversion of mid-IR lasers into the long-wavelength domain of 12-20 µm with AgGaSe2, BaGa2GeSe6 and PbIn6Te10 nonlinear crystals,” Opt. Express 27(17), 24353–24361 (2019). [CrossRef]
11. K. Kato, V. V. Badikov, L. Wang, et al., “Effective nonlinearity of the new quaternary chalcogenide crystal BaGa2GeSе6,” Opt. Lett. 45(8), 2136–2139 (2020). [CrossRef]
12. K. Kato, K. Miyata, V. V. Badikov, et al., “Phase-matching properties of BaGa2GeSe6 for three-wave interactions in the 0.778-10.5910 µm spectral range,” Appl. Opt. 57(26), 7440–7443 (2018). [CrossRef]
13. K. Kato, K. Miyata, T. Okamoto, et al., “Thermo-optic dispersion formula for BaGa2GeSe6,” Appl. Opt. 58(35), 9709–9711 (2019). [CrossRef]
14. G. Exner, M. A. Susner, J. Murray, et al., “Thermo-mechanical properties of BaGa2GeS(e)6 nonlinear optical crystals,” Opt. Mater. Express 13(5), 1335–1344 (2023). [CrossRef]
15. https://as-photonics.com/products/snlo/.
16. A. Zukauskas, N. Thilmann, V. Pasiskevicius, et al., “5 mm thick periodically poled Rb-doped KTP for high energy optical parametric frequency conversion,” Opt. Mater. Express 1(2), 201–206 (2011). [CrossRef]
17. A. Boyko, V. Badikov, G. Shevyrdyaeva, et al., “Intracavity-pumped, cascaded optical parametric oscillator based on BaGa2GeSе6,” High-Brightness Sources and Light-Driven Interactions, Mid-Infrared Coherence Sources (MICS), OSA Technical Digest (Optical Society of America, 2018), paper MW2C.3.
18. A. A. Boyko, G. M. Marchev, V. Petrov, et al., “Intracavity-pumped, cascaded AgGaSe2 optical parametric oscillator tunable from 5.8 to 18 µm,” Opt. Express 23(26), 33460–33465 (2015). [CrossRef]
