We conceived a unique fully parametric source based on two independent cylindrical OPOs simultaneously pumped by the same Nd:YAG laser. Each OPO delivers more than 2 mJ and is continuously tunable between 1.41 µm and 4.3 µm. This source is of particular interest for the study of the generation of infrared parametric light in nonlinear crystals. It was validated by performing difference frequency generation experiments in CdSe crystals with output in the range 8 – 10 µm.
© 2013 Optical Society of America
High energy coherent sources emitting two wavelengths λ1 and λ2 close to each other have attracted much interest recently because of the prospect of infrared (IR) to terahertz (THz) wavelengths through Difference Frequency Generation (DFG) processes at λDFG defined by 1/λ1 – 1/λ2 = 1/ λDFG in a nonlinear crystal. For example, dual frequency operation of an optical parametric generator (OPG) made of two PPSLT crystals at different temperatures had led to a set of signal wavelengths close to 1.6 µm and separated by less than 2 THz . Another solution consisted in the operation of a slightly misaligned KTP optical parametric oscillator (OPO) cavity: a dual wavelength emission of two idler wavelengths close to 1.51 µm was achieved . However in both cases, the tunability of the two wavelengths λ1 and λ2 was limited.
Recently we built one OPO based on a 5-mm-thick and 38-mm-long 5%MgO:PPLN crystal cut as a partial cylinder enabling a wide and continuous tunability of the generated beams from 1.41 µm to 4.3 µm and delivering more than 2 mJ . In the present paper, we describe the design of a tunable source using two such cylindrical OPOs running in parallel and pumped simultaneously with the same Nd:YAG laser. This dual wavelength source was specifically designed for the characterization of nonlinear crystals having potential DFG properties in the mid-IR and THz regions. It is especially suited for birefringence phase-matched (BPM) or quasi-phase-matched (QPM) nonlinear crystals that require the use of pump wavelengths above 1.4 µm in order to avoid detrimental two-photon-absorption (TPA). We report the validation of this unique dual wavelength source with two kinds of DFG experiments up to λDFG = 10 µm in the well-known BPM CdSe crystal .
2. Architecture and performances of the dual-wavelength source
Our source is made of two identical OPOs based on two 38 x 16 x 5 mm3 5%MgO:PPLN crystals with a single grating of periodicity Λ0 = 28 µm, cut as partial cylinders with the z-axis as the revolution axis and polished on the side. There is no antireflection (AR) coating. Each partial cylinder is inserted between two plane mirrors forming a cavity that is singly resonant on the signal with one recycling of the pump. The pump beam is emitted by a 10-Hz repetition rate and 10-ns (full width at 1/e2) Nd:YAG laser Surelite from Continuum. Adequate cylindrical focusing of the pump beam is required to ensure its parallel propagation in the partial cylinder over the first pass in the cavity. A schematic view of the dual source is shown in Fig. 1.
Independent and continuous tuning of the OPOs is achieved at room temperature, i.e. T = 21°C without any temperature control, through the independent rotations of the 5%MgO:PPLN partial cylinders. When the rotation angles Φ1 and Φ2 are varied over an angular range of 30°, the continuous variation of the effective grating periodicity, defined as Λeff(Φ1,2) = Λ0/cos(Φ1,2), leads to the following tunability of the OPOs: from λ1s,2s = 1.41 µm to 2.128 µm for the signal, and from λ1i,2i = 2.128 µm to 4.3 µm for the idler, as shown in Fig. 2.This implies that the two “arms” of the dual-wavelength source can be independently and continuously tuned over the entire wavelength range 1.41 µm-4.3 µm. The signal wavelength up to 1.7 µm was measured directly using a NIRQuest spectrometer, the idler one being deduced by the energy conservation law. For longer wavelengths, we used the occurrence in the 5%MgO:PPLN crystal of non-quasi-phase-matched second harmonic generation of the signal as well as the recombination of the signal or idler with the 1.064 µm beam leading to visible wavelengths. In that case, the spectrometer was an OCEAN optics HR 2000.
Depending on the signal wavelength, the oscillation threshold of each OPO ranges between 1.2 mJ and 1.9 mJ. The “signal plus idler” energy conversion efficiency increases with the signal wavelength, from 20% at 1.41 µm up to 33% close to degeneracy, so that the output energy of each OPO reaches up 2 mJ over the entire tuning range when the pump input energy is at 10 mJ. The linewidth of the signal resonating in each OPO increases with the signal wavelength as well as the pump input energy: it takes the values of 2.5 nm, 8.0 nm and 46.0 nm at 1.43 µm, 1.63 µm, and 2.0 µm respectively for a pump energy of 2 mJ. For 10 mJ of pump, the linewidth is consistently 2.5 times broader . The generated beams exhibit astigmatism due to the cylindrical geometry of the nonlinear crystals: they have an elliptical cross-section with an anisotropic M2 factor. In the case of a signal wavelength at 1.67 µm for example, the M2 factor is measured to be 8.5 ± 1.0 in the rotation plane of the partial cylinder, i.e. the x-y plane, and 28 ± 2 along the z-axis.
3. Test of the dual-wavelength source with DFG measurements in CdSe
The two “arms” of the dual source are fully independent one to each other, since the wavelengths, polarization states as well as energies of the input beams can be freely adjusted. It is then possible to study any BPM or QPM nonlinear crystal in order to estimate its potentialities for DFG in the mid- to far-infrared. Two configurations of DFG experiments can be carried out: determination of the BPM or QPM wavelengths λDFG in a given direction of propagation, or direct measurement of the DFG phase-matching angles (or QPM period) corresponding to a given set of input wavelengths λ1 and λ2. We tested these two configurations in the case of BPM by studying the well-known crystal CdSe .
A first DFG experiment was performed in a 40-mm-long uncoated CdSe slab crystal from Moltech, oriented at θ = 90° for angular noncritical Type III DFG BPM, i.e. 1/λ1o – 1/λ2e = 1/ λDFGo, where the indices o and e denote the “ordinary” and “extraordinary” polarization states respectively. The DFG power was measured using a PEM MCT detector from Vigo whose spectral response extends up to 11 μm. Besides a vanishing walk-off angle, the effective coefficient for this interaction has the strongest magnitude in this plane, i.e. deff = d15 = 18 pm/V ), which makes this DFG scheme especially appealing. The spatial overlap between the two incident beams at λ1 and λ2 is ensured by maximizing the generated SFG (1/λ1e + 1/λ2e = 1/ λSFGe) in a 4-cm-long 5%MgO:PPLN crystal with a grating period of 32.3 µm. When the wavelength λ1 is set at 2.72 µm for example, the DFG BPM wavelength λ2,BPM was found to be 4.10 µm, instead of 4.15 µm as predicted from calculations based on Sellmeier equations of reference , and as shown in Fig. 3. The corresponding wavelength of the beam generated through DFG is λDFG = 8.08 ± 0.08 µm, while the calculation gives 7.88 µm. The corresponding full width at half maximum (FWHM) spectral acceptance of this DFG was measured to be 150 nm instead of 20 nm for the ideal case of a monochromatic pump wave. Actually, in our case, the FWHM linewidths of the input beams are Δλ1 = 41 nm and Δλ2 = 21 nm for the beams at λ1 = 2.72 µm and λ2 = 4.10 µm, respectively. However, Fig. 3 indicates that this broadening is not detrimental for an accurate determination of the phase-matching wavelength.
We performed the same kind of measurements at λ1 = 2.79 µm, and a similar discrepancy between the measured and calculated λ2,PM and λDFG wavelengths was found, as shown on the tuning curve of Fig. 4.This discrepancy may be attributed to the presence of sulfur impurities in our sample, as suggested by the strength of the CdSO4 absorption band at 18.6 µm , to the divergence of the beams, or to a non-collinearity between the two incoming beams at λ1 and λ2. About the last point, an angle of about 0.5° between the input wave vectors could explain such a splitting. However, values of noncritical DFG wavelength were previously reported in the case of a pump tunable synchronously-pumped OPO using CdSe , and they were also found to overestimate the calculations from  as shown in Fig. 4. Note that even if CdSe can generate up to 25 µm, we cannot perform our measurements above 10 µm due to the detectors we used.
When the two input wavelengths are set to their phase-matching values λ1 = 2.72 µm and λ2 = 4.10 µm, the type III DFG energy was measured as a function of the energy E2 of the beam at λ2 for three values of the energy E1 of the beam at λ1. The corresponding data are shown in Fig. 5..
Note that the input energies E1 and E2 are increased independently by adjusting separately the pump level of each OPO. The highest DFG energy reached 0.5 µJ for E1 + E2 = 275 µJ. The DFG energy conversion efficiency η = EDFG / (E1 + E2) is the highest when the energies of the two input beams are nearly balanced (i.e. E1 ≈E2): it reached consistently η = 0.3% corresponding to a quantum conversion efficiency of 0.9%. The saturation-like effect observed when the energy E2 is increased could be attributed to either the spectral broadening of the input beams above the spectral acceptances of the crystal at higher pumping levels, or to a slight decrease of the idler wavelengths emitted by the OPOs through a thermal heat-up of the 5%MgO:PPLN crystals.
We also performed a type III DFG experiment in a CdSe crystal from Cleveland Crystals cut as a 5-mm-diameter cylinder, and polished to optical quality as shown in Fig. 6.The cylinder was oriented with the x-axis as the revolution axis so that any phase-matching direction in the y-z plane can be measured. The wavelength λ1 was set at 2.79 µm. This beam was mixed with the tunable beam at λ2. The corresponding tuning curve is given Fig. 6, including the comparison with calculations from available Sellmeier equations , as well as previous phase-matching measurements obtained in CdSe-based OPO pumped at 2.79 µm [10, 11]. As in the case of the angular non critical phase-matching DFG described previously in the CdSe slab, a non-collinearity of 0.5° enables to achieve the agreement between the experimental data and the calculations.
We conceived a unique dual wavelength source emitting two independent and tunable beams in the near and mid-IR. It is suitable to perform accurate characterizations of any DFG processes in crystals with band gaps falling below 1.8 eV. We validated the potentialities of this source by performing two separate DFG experiments in CdSe crystals up to 10 µm. The measurements of the phase-matching angles and wavelengths as well as the conversion efficiencies were carried out. A slight difference between the calculated and measured phase-matching angles was found. Despite a good agreement with earlier reports, this difference may be attributed to either the presence of sulfur impurities in the CdSe crystals that were studied, to the divergence of the beams, or to a tilt angle of 0.5° between the two incident beams. Such a small angle is very difficult to check experimentally yet, since it falls within the experimental error of the current alignment procedure of the beams.
Based on the full independence of the two wavelengths of the dual source, similar DFG studies can be foreseen in other materials or wavelength ranges up to the THz range.
Besides the purpose of nonlinear crystal characterization, it is clear that such a dual wavelength source could be used as a versatile pump stage of a DFG device for applications.
References and links
1. M. Koichi, K. Miyamoto, S. Ujita, T. Saito, H. Ito, and T. Omatsu, “Dual-frequency picosecond optical parametric generator pumped by a Nd-doped vanadate bounce laser,” Opt. Express 19(19), 18523–18528 (2011). [CrossRef] [PubMed]
2. M. Tang, H. Minamide, Y. Wang, T. Notake, S. Ohno, and H. Ito, “Dual-wavelength single-crystal double-pass KTP optical parametric oscillator and its application in terahertz wave generation,” Opt. Lett. 35(10), 1698–1700 (2010). [CrossRef] [PubMed]
3. V. Kemlin, D. Jegouso, J. Debray, P. Segonds, B. Boulanger, B. Menaert, H. Ishizuki, and T. Taira, “Widely tunable optical parametric oscillator in a 5 mm thick 5% MgO:PPLN partial cylinder,” Opt. Lett. 38(6), 860–862 (2013). [CrossRef] [PubMed]
4. G. D. Boyd, E. Buehler, and F. G. Storz, “Linear and nonlinear optical properties of ZnGeP2 and CdSe,” Appl. Phys. Lett. 18(7), 301–304 (1971). [CrossRef]
5. 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,” J. Appl. Phys. B. 86, 111–115 (2007).
6. R. L. Herbst and R. L. Byer, “Efficient parametric mixing in CdSe,” Appl. Phys. Lett. 19(12), 527–530 (1971). [CrossRef]
7. G. C. Bhar, D. C. Hanna, B. Luther-Davies, and R. C. Smith, “Tunable down-conversion from an optical parametric oscillator,” Opt. Commun. 6(4), 323–326 (1972). [CrossRef]
8. A. Burger, D. O. Henderson, S. H. Mogan, and E. Siberman, “Purification, crystal growth and characterization of CdSe single crystals,” J. Cryst. Growth 109(1-4), 298–308 (1991). [CrossRef]
9. M. A. Watson, M. V. O’Connor, D. P. Shepherd, and D. C. Hanna, “Synchronously pumped CdSe optical parametric oscillator in the 9-10 µ region,” Opt. Lett. 28(20), 1957–1959 (2003). [CrossRef] [PubMed]
10. T. H. Allik, S. Chandra, D. M. Rines, P. G. Schunemann, J. A. Hutchinson, and R. Utano, “Tunable 7 12-µ optical parametric oscillator using a Cr,Er:YSGG laser to pump CdSe and ZnGeP2 crystals,” Opt. Lett. 22(9), 597–599 (1997). [CrossRef] [PubMed]
11. D. Rines, G. Rines, and P. Moulton, “CdSe OPO Pumped by a 2.79 µm Cr,Er:YSGG Laser,” in Advanced Solid State Lasers, Vol. 24 of OSA Proceedings Series, paper PO7.