We demonstrate a highly efficient double pass optical parametric generator based on periodically poled MgO-doped congruent LiNbO3. More than two watts of tunable near-IR radiation (1370-1650 nm) are generated by directly pumping the system with 550 fs pulses from a 42 MHz repetition rate passively mode-locked Yb:KGW oscillator. Pulse durations below 200 fs were achieved without further compression techniques. The system is extremely efficient, compact, cost effective, easy to align and easy to operate, which makes it an interesting alternative to more complex optical parametric oscillators or optical parametric amplifiers.
© 2014 Optical Society of America
Femtosecond laser systems operating in the near infrared are key components for time-resolved optical spectroscopy, multi-photon fluorescence microscopy, and telecommunications [1,2]. For these experiments often a broad tuning range, low noise level and long term stability are mandatory. Furthermore a high repetition rate is desirable to maximize the signal-to-noise ratio of the measurements and a compact and easy alignable system is preferable for practical issues.
As the tunability of lasers is limited by the gain bandwidth of the respective active material, broadband tunable coherent light sources mainly rely on frequency conversion processes. The most widely used scheme is the parametric frequency down conversion, which is the basis for optical parametric oscillators (OPOs), optical parametric amplifiers (OPAs) and optical parametric generators (OPGs) . Intensive research over the last decades finally led to commercially available OPOs and OPAs with femtosecond pulse durations and repetition rates in the megahertz range. Compared with OPOs, OPGs are much more compact, easy to align and cost-effective, as there is no need for a synchronized cavity. The absence of a cavity also makes OPGs less sensitive to external perturbations. While OPAs share these advantages with OPGs they still have to be seeded by an external light source. The seed must cover the desired spectral range, have a low noise level, posses a good spectral stability, and, in case of a pulsed seed, has to be synchronized with the pump laser. These requirements just transfer the complexity of the parametric process to the provision of the seed source [4–8].
So far, the research activities carried out on OPGs are mainly focused on picosecond pulse durations and/or kilohertz repetition rates [9–15]. The reason for this is that OPGs need a high parametric gain, in order to achieve a measurable output and to suppress the intensity fluctuations by driving the parametric process into saturation [15–18]. The high parametric gain is usually achieved by the use of an amplified laser system as a pump source, rather long nonlinear crystals or a strong focusing of the pump light. However, long crystals are usually no option for the generation of femtosecond pulses because of the temporal walk off effects due to the group velocity mismatch (GVM) inside the nonlinear crystal  and strong focusing might damage the nonlinear crystal.
The development of high average power passively mode-locked solid state and fiber lasers [19–21] allowed for the realization of single pass optical parametric generators based on rather short LiNbO3 and LiTaO3 crystals with repetition rates of several tens of megahertz and pulse durations in the femtosecond regime [16,17]. In these experiments the average signal power and the conversion efficiency for long term operation was limited due to the onset of laser damage inside the crystal after a few minutes.
One method which is known from the picosecond regime to increase the parametric gain without using higher pump powers is to send the residual pump and the generated light through the crystal a second time [13,22,23]. This doubles the effective crystal length, but allows for tighter focusing compared to a longer crystal. Furthermore, such a double pass configuration allows for a simple compensation of the temporal walk off . In the following we show that the use of a compact single crystal double pass configuration allows for long term operation with multi-watt output, high conversion efficiency and low output noise in periodically poled MgO:LiNbO3.
2. Experimental setup
The scheme of our OPG is shown in Fig. 1. The OPG is pumped by a solitary mode-locked Yb:KGW oscillator (NT&C Yb:KGW 1040-8) , delivering up to 8 W average power with a pulse duration of 550 fs at a repetition rate of 42 MHz and a center wavelength of λp = 1031 nm. A half-wave plate and a Faraday isolator are used together as a variable attenuator for the pump beam. In order to achieve the optimal focusing condition derived by Boyd et al.  we focused the pump beam to a diameter of 40 µm with a f = 200 mm lens (L1). A dichroic mirror (DCM: Reflection band 920-1160 nm, transmission band 1240-1600 nm) reflects the pump beam into the MgO:LiNbO3 crystal (Covesion Ltd.). The crystal is 10 mm long and 0.5 mm thick, consists of 5% MgO-doped congruent LiNbO3 and has 9 different poling channels (Poling periods: Λ = 27.9-31.6 µm) with respect to the z-axis for quasi phase matching. The end facets are broadband anti reflection coated (R<5%, 1000-5000 nm). The crystal is mounted on an aluminum heat sink whose temperature is controllable from 30 °C up to 200 °C. The crystal holder can be translated with respect to the different poling channels using a computer controlled stage. After passing through the crystal the transmitted pump and generated signal beam are separated by a second dichroic mirror (DCM).
For the second pass the signal beam is collimated by a lens (L2, f = 100 mm), back reflected by a dielectric mirror (DM1: Reflection band 1300-1900 nm) and focused back into the crystal. The transmitted pump beam is collimated by a lens (L3, f = 150 mm), back reflected by a dielectric mirror (DM2: Reflection band 750-1100 nm) and focused back into the crystal. The dielectric mirror (DM2) can be translated parallel to the pump beam to ensure the temporal overlap of the signal and pump beam in the second pass.
After the second pass the signal beam passes the first dichroic mirror, is collected by a lens (L4, f = 100 mm) and sent through two filters in order to suppress any parasitic generated visible light and residual pump light. As all optics are made from BK-7 glass or fused silica the idler light will be absorbed.
The OPG itself (without the pump laser) has a footprint of only 30 cm x 30 cm. For operation and wavelength tuning only the crystal’s position and temperature, the position of DM2 and the input power has to be varied, which can all be done automated.
3. Experimental results
In Fig. 2 the signal power and the conversion efficiency are shown as a function of the incident pump power for the double pass OPG. If not otherwise mentioned, we will always refer to average powers and pump-to-signal conversion efficiencies. These measurements were done for a poling period of 27.9 µm at a temperature of 30 °C, which leads to a signal wavelength of λs = 1372 nm. The signal power was measured using a thermal power meter and corrected for filter losses. The maximum achieved signal power was 2.5 W at 4.6 W incident pump power, which corresponds to a conversion efficiency of nearly 55%. If we calculate the generated idler radiation by using the Manley–Rowe relations  this corresponds to a photon conversion efficiency of 72%. Saturation of the conversion efficiency seems to be reached at 4 W input power. Figure 2 suggests that there is no optical parametric generation below 1.5 W input power, but with the help of a LN2-cooled InGaAs-CCD coupled to a monochromator we were able to record parametric fluorescence at pump powers below 100 µW. For all data points shown the delay between the pump and signal pulses in the second pass was optimized for maximum signal power by translating DM2 (see Fig. 1).
Signal spectra and output power for the different poling periods between 27.9 µm and 30.5 µm, measured at 30 °C and 4 W input power are depicted in the upper part of Fig. 3 and the corresponding intensity autocorrelations in the lower part. The input power was restricted to 4 W to stay well below the damage threshold of the longer poling periods, which should be significantly lower than that of the shorter ones . The spectral positions are in reasonable agreement with numerical values achieved for perfect quasi phase matching, taking into account the temperature dependent Sellmeier equations  and assuming that the temperature inside the crystal should be somewhat higher than on the surface. The broadening of the spectra originates from an increase in the phase matching bandwidth for higher signal wavelengths. We also measured over one watt signal power for the 31 µm poling period. The central wavelength should be approximately 1750 nm, but we are not able to present accurate spectra as the responsivity of our spectrometer strongly decreases at around 1650 nm. This fact might also disturb the measurement of the spectrum for the 30.5 µm poling period. Additionally we have achieved continuous tuning of the signal wavelength by changing the crystal temperature (not shown).
The autocorrelations of the signal pulses can be fitted well assuming a sech2-pulse shape for poling periods up to 29.5 µm. With increasing poling period/signal wavelength a broadening of the side wings appears, which is especially visible for the 30 µm and 30.5 µm poling period. The pulse duration decreases monotonically from 345 fs down to 190 fs with increasing signal wavelength, assuming a sech2-pulse shape. The pulses created with the 30.5 µm poling period have an even smaller FWHM, but this value is obviously not a good figure of merit in the case of such strongly distorted pulses. The time bandwidth product for all poling periods except 30.5 µm is in the range of 0.44-0.53, which is close to the Fourier limit for a sech2-pulse. As can be seen in the upper part of Fig. 3, the signal power is maximal at around 1380 nm signal wavelength and decreases for larger signal wavelengths.
The distortion of the pulse shape and the decrease of the conversion efficiency can be both attributed to a temporal walk off between the pump, signal and idler pulses induced by the GVM inside the crystal, which constitutes a limit for the interaction length inside the crystal [16,17]. However, up to 1450 nm signal wavelength the GVM of the signal and the idler pulses with respect to the pump pulse have different signs and furthermore the same absolute value at around 1380 nm signal wavelength [15,25]. The nonlinear interaction between the pump pulse and the generated signal and idler pulses then leads to a temporal trapping of the generated pulses under the pump pulse. As a consequence both the signal and the idler pulses are nearly transform limited and the length on which conversion takes place is increased [14,15,17,18]. With increasing signal wavelength this trapping effect vanishes and therefore the conversion efficiency decreases and the signal pulses get distorted. We minimized this effect by optimizing the delay between the pump and signal pulses in the second pass to achieve the shortest possible pulse widths and smooth autocorrelations. We also tried to compress the pulses with the help of a SF-10 prism sequence, but this approach gave no satisfying results. This additionally suggests, that the deviation from the Fourier limit is due to the nonlinear pulse interaction. Of course, more elaborate compression techniques might be used.
As an OPG is seeded by vacuum fluctuations one could expect a rather high noise level of the signal output. In contrast to this, it had been shown experimentally as well as theoretically, that the output stabilizes when the OPG process is driven into saturation [14,16–18]. To measure the noise characteristics of our OPG we measured the signal power fluctuations for different conversion efficiencies with the 27.9 µm poling period. Each individual measurement was carried out over 15 minutes and for high conversion efficiencies the results were double-checked by a one hour measurement. In addition we recorded the pulse-to-pulse fluctuation of the OPG signal in the case of low and high conversion efficiency. The noise measurements of the OPG are shown in Fig. 4. As expected we observed a high noise level in the case of low conversion efficiency and a drastic decay down to 1.4% rms (long term) respectively 10.3% rms pulse-to-pulse with increasing conversion efficiency. In addition we performed a long time measurement (one hour) of the spectral stability of our OPG, which showed a fluctuation of the center wavelength of only 97 pm and a fluctuation of the spectral FWHM of 71 pm.
In summary, a high power double pass femtosecond OPG operating with long term stability at 42 MHz has been demonstrated. A signal output of more than two watts and nearly 55% conversion efficiency has been obtained by pumping the MgO:PPLN based setup with an Yb:KGW laser oscillator without any further amplification. The signal is tunable from 1370 nm to at least 1650 nm and pulse durations below 200 fs were achieved without further compression techniques. By using appropriate optics this setup might also be used to generate mid-IR radiation in the wavelength range of 2000 nm up to 4100 nm. Our system is compact, cost efficient, and can be fully automated with very low effort. Hence, this system is an interesting alternative to more complex OPO or OPA systems.
We acknowledge financial support from the Deutsche Forschungsgemeinschaft (SPP1391).
References and links
1. J.-C. Diels and W. Rudolph, Ultrashort Laser Pulse Phenomena (Second Edition) (Elsevier Inc., 2006).
2. O. Wada, “Femtosecond all-optical devices for ultrafast communication and signal processing,” New J. Phys. 6, 183 (2004). [CrossRef]
3. Y. R. Shen, The Principles of Nonlinear Optics (Wiley-Interscience, 2003), Chap. 9.
5. T. W. Neely, T. A. Johnson, and S. A. Diddams, “High-power broadband laser source tunable from 3.0 μm to 4.4 μm based on a femtosecond Yb:fiber oscillator,” Opt. Lett. 36(20), 4020–4022 (2011). [CrossRef]
6. J. Krauth, A. Steinmann, R. Hegenbarth, M. Conforti, and H. Giessen, “Broadly tunable femtosecond near- and mid-IR source by direct pumping of an OPA with a 41.7 MHz Yb:KGW oscillator,” Opt. Express 21(9), 11516–11522 (2013). [CrossRef]
7. T. Steinle, A. Steinmann, R. Hegenbarth, and H. Giessen, “Watt-level optical parametric amplifier at 42 MHz tunable from 1.35 to 4.5 μm coherently seeded with solitons,” Opt. Express 22(8), 9567–9573 (2014). [CrossRef]
8. T. Steinle, S. Kedenburg, A. Steinmann, and H. Giessen, “Combining cw-seeding with highly nonlinear fibers in a broadly tunable femtosecond optical parametric amplifier at 42 MHz,” Opt. Lett. (to be published).
9. T. A. Rabson, H. G. Ruiz, P. L. Shah, and F. K. Tittel, “Stimulated parametric fluorescence induced by picosecond pump pulses,” Appl. Phys. Lett. 21(4), 129–131 (1972). [CrossRef]
10. P. Zhao, B. Zhang, E. Li, R. Zhou, D. Xu, Y. Lu, T. Zhang, F. Ji, X. Zhu, P. Wang, and J. Yao, “Experimental study on a high conversion efficiency, low threshold, high-repetition-rate periodically poled lithium niobate optical parametric generator,” Opt. Express 14(16), 7224–7229 (2006). [CrossRef]
11. M. Levenius, V. Pasiskevicius, F. Laurell, and K. Gallo, “Ultra-broadband optical parametric generation in periodically poled stoichiometric LiTaO3.,” Opt. Express 19(5), 4121–4128 (2011). [CrossRef]
12. S. C. Kumar, M. Jelínek, M. Baudisch, K. T. Zawilski, P. G. Schunemann, V. Kubeček, J. Biegert, and M. Ebrahim-Zadeh, “Tunable, high-energy, mid-infrared, picosecond optical parametric generator based on CdSiP2,” Opt. Express 20(14), 15703–15709 (2012). [CrossRef]
13. F. Seifert, V. Petrov, and F. Noack, “Sub-100-fs optical parametric generator pumped by a high-repetition-rate Ti:sapphire regenerative amplifier system,” Opt. Lett. 19(11), 837–839 (1994). [CrossRef]
14. X. Xie, A. M. Schober, C. Langrock, R. V. Roussev, J. R. Kurz, and M. M. Fejer, “Picojoule threshold, picosecond optical parametric generation in reverse proton-exchanged lithium niobate waveguides,” J. Opt. Soc. Am. B 21(7), 1397–1402 (2004). [CrossRef]
15. M. Marangoni, R. Osellame, R. Ramponi, G. Cerullo, A. Steinmann, and U. Morgner, “Near-infrared optical parametric amplifier at 1 MHz directly pumped by a femtosecond oscillator,” Opt. Lett. 32(11), 1489–1491 (2007). [CrossRef]
16. T. Südmeyer, J. Aus der Au, R. Paschotta, U. Keller, P. G. R. Smith, G. W. Ross, and D. C. Hanna, “Novel ultrafast parametric systems: high repetition rate single-pass OPG and fibre-feedback OPO,” J. Phys. D Appl. Phys. 34(16), 2433–2439 (2001). [CrossRef]
17. S. V. Marchese, E. Innerhofer, R. Paschotta, S. Kurimura, K. Kitamura, G. Arisholm, and U. Keller, “Room temperature femtosecond optical parametric generation in MgO-doped stoichiometric LiTaO3,” J. Appl. Phys. B 81(8), 1049–1052 (2005). [CrossRef]
18. C. Manzoni, G. Cirmi, D. Brida, S. de Silvestri, and G. Cerullo, “Optical-parametric-generation process driven by femtosecond pulses: Timing and carrier-envelope phase properties,” Phys. Rev. A 79(3), 033818 (2009). [CrossRef]
19. F. Brunner, T. Südmeyer, E. Innerhofer, F. Morier-Genoud, R. Paschotta, V. E. Kisel, V. G. Shcherbitsky, N. V. Kuleshov, J. Gao, K. Contag, A. Giesen, and U. Keller, “240-fs pulses with 22-W average power from a mode-locked thin-disk Yb:KY(WO4)2 laser,” Opt. Lett. 27(13), 1162–1164 (2002). [CrossRef]
20. A. Steinmann, B. Metzger, R. Hegenbarth, and H. Giessen, “Compact 7.4 W femtosecond oscillator for white-light generation and nonlinear microscopy,” in Conference on Lasers and Electro-Optics, OSA Technical Digest (CD) (Optical Society of America, 2011), paper CThAA5. [CrossRef]
21. F. Hoos, T. P. Meyrath, S. Li, B. Braun, and H. Giessen, “Femtosecond 5-W Yb:KGW slab laser oscillator pumped by a single broad-area diode and its application as supercontinuum source,” J. Appl. Phys. B 96(1), 5–10 (2009). [CrossRef]
22. G. P. Banfi, C. Solcia, P. Di Trapani, R. Danielius, A. Piskarskas, R. Righini, and R. Torre, “Travelling-wave parametric conversion of microjoule pulses with LBO,” Opt. Commun. 118(3-4), 353–359 (1995). [CrossRef]
23. 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]
24. G. D. Boyd and D. A. Kleinmann, “Parametric Interaction of Focused Gaussian Light Beams,” J. Appl. Phys. 39(8), 3597–3639 (1968). [CrossRef]
25. O. Gayer, Z. Sacks, E. Galun, and A. Arie, “Temperature and wavelength dependent refractive index equations for MgO-doped congruent and stoichiometric LiNbO3,” Appl. Phys. B 91(2), 343–348 (2008). [CrossRef]
26. J. M. Manley and H. E. Rowe, “Some General Properties of Nonlinear Elements – Part I: General Energy Relations,” Proc. IRE44, 904–913 (1956). [CrossRef]