We demonstrate a compact mid-infrared (mid-IR) radiation source based on difference frequency generation (DFG) in periodically poled lithium niobate (PPLN) crystal. The system incorporates a dual-wavelength master oscillator power amplifier (MOPA) source capable of simultaneous amplification of 1064 nm and 1548 nm signals in a common active fiber co-doped with erbium and ytterbium ions. Two low-power seed lasers were amplified by a factor of 14.4 dB and 23.7 dB for 1064 nm and 1548 nm, respectively and used in a nonlinear DFG setup to generate 1.14 mW of radiation centered at 3.4 μm. The system allowed for open-path detection of methane (CH4) in ambient air with estimated minimum detectable concentration at a level of 26 ppbv.
© 2013 Optical Society of America
Mid-infrared (mid-IR) lasers are being rapidly developed all over the world. This fact is mainly due to the excellent usability of those sources in laser spectroscopy and remote gas sensing applications. The mid-IR wavelength region is covered with fundamental vibration and associated rotational-vibrational absorption bands of many molecules . It is worth noting, that absorption bands located in the mid-IR are mostly several orders of magnitude more intense than the ones found in the near infrared wavelength region (0.8 µm – 2 µm). This fact, along with recent technological progress in mid-IR laser sources, allowed ultra-sensitive and selective detection of number of molecules [2–5]. Amongst many other trace gas sensing techniques laser absorption spectroscopy provides many advantageous solutions. Compactness and low power consumption of new type of laser sources allowed to design and build small and portable detection systems which are capable of detecting slight fluctuations of various molecules, including bio-markers which aid in early disease recognition [6,7]. Commonly available mid-IR laser sources, suitable for laser spectroscopy, can be split into four main groups: gas lasers, quantum cascade lasers (QCLs), solid state lasers and frequency conversion-based laser sources . Each approach has its advantages and disadvantages. The main pros of mid-IR sources based on DFG effects are narrow linewidth, capability of long-term Continuous Wave (CW) operation and wide and non-complex tuning . Several approaches to DFG are commonly known in the laser community. One can find systems using separate solid state lasers [10,11] or fiber based radiation sources [12,13]. This report will focus on using an all-fiber MOPA source in a DFG setup. Fiber MOPA sources can provide multi-watt-level output powers while maintaining excellent beam quality. Cascaded design of the amplifier allows for easy control of the amplified radiation through controlling the parameters of low power, relatively cheap, off-the-shelf seed diodes . The combination of the MOPA wide amplification bandwidth and DFG systems allows for very wide tuning of the generated mid-IR radiation. Moreover, all-fiber design of the amplifiers assures robustness and compactness of the design, which in turn could lead to development of small, transportable mid-IR sources. Recently, we have demonstrated a compact, all-fiber setup capable of generating two wavelengths (1064 and 1550 nm) simultaneously in one active fiber with watt-level output power and single-mode beam . Such system seems to be an ideal pumping source for PPLN crystals in order to generate mid-IR radiation in the 3.38 – 3.40 µm range via DFG nonlinear processes. In this work we demonstrate the first attempt, to the best of our knowledge, of utilizing the dual-wavelength fiber MOPA as a simple and efficient source for DFG process, allowing generation of more than 1 mW of laser radiation centered at 3.4 µm. The overall performance and stability of the system was determined by performing simple methane (CH4) measurements in ambient air using Tunable Diode Laser Spectroscopy (TDLAS) technique.
2. Dual-wavelength amplifier
The pumping source for the PPLN crystal in our experiments is designed in a MOPA scheme and consists of an Er:Yb co-doped fiber amplifier (EYDFA) seeded by two lasers: operating at 1.55 µm and 1.06 µm wavelengths. The dual-seeding technique of EYDFAs was firstly proposed by Kuhn et al. as an efficient method of suppressing the amplified spontaneous emission (ASE) and parasitic lasing from the Yb ions in the amplifiers operating at nominal 1550 nm wavelength [16,17]. Our further work revealed, that certain commercially available Er:Yb co-doped fibers (EYDFs) are suitable for very efficient dual-wavelength amplification and the gain at 1.06 µm might be only 6 dB lower than at the nominal 1.55 µm wavelength . Such MOPA might serve as an ideal dual-wavelength source for pumping of PPLN crystals and generating radiation in the mid-IR, centered near 3.4 µm. The setup of the double-seeded Er:Yb co-doped fiber amplifier (DS-EYDFA) is shown in Fig. 1. The amplifier is seeded by two signals from DFB laser diodes operating at 1548 nm and 1064 nm. The 1548 nm signal, was additionally pre-amplified in an erbium-doped fiber amplifier (EDFA), which provides ~11 dB gain. Both signals are combined in a wavelength division multiplexing (WDM) coupler and directed to the active fiber (Nufern SM-EYDF-7/130). The EYDF is backward-pumped by two 975 nm laser diodes, delivering in total 20 W of power. The pumping beam is coupled into the EYDF through a multimode pump-combiner. The pump power, which is not absorbed by the active fiber is rejected by a cladding mode stripper (CMS). After amplification, both signals are separated and directed to fiber isolators, in order to avoid unwanted back-reflections and parasitic lasing.
Several measurements of the dual-wavelength amplifier were performed in order to determine its usefulness for mid-IR generation via nonlinear DFG processes occurring in PPLN crystals.
Output power for both wavelengths as a function of pump power along with their 20 minute stability are depicted in Fig. 2. At maximum pump power the amplifier delivered more than 1.1 W of 1.06 μm radiation and 4 W of 1.55 μm radiation. Output power fluctuations were registered at half of the maximum pump capability due to power limitation of measuring device used (Accelink, PMSII-B). 20 minute power instabilities did not exceed 3.8% and 1.7% for 1.06 μm and 1.55 μm wavelength, respectively. Slow oscillations along with a gentle drift was recorded. This was probably caused by insufficient temperature stabilization of the pump diodes. Slow temperature drift of the diode structure caused shifts of the pumping wavelength, which in return produced variations in active fiber absorption efficiency and therefore in output power of both wavelengths. The pump diodes were mounted on one Peltier-cooled heat-sink attached to a water-cooled aluminum block. Sinusoidal oscillations were caused by PID controller overshooting the current delivered to the Peltier elements, whereas the drift was caused by slow heating-up of the liquid in the closed-circuit water-cooling system. This problem will be addressed in the future by using a more sophisticated heat dissipation system for the multimode pump diodes.
Optical spectra of both amplified signals registered using 0.01 nm resolution and high sensitivity at maximum pump power are depicted in Fig. 3.
Both signals had a very good signal to noise ratio (SNR), reaching 57dB and 60dB for 1 μm and 1.5 μm wavelength, respectively. No signs of Q-switching were observed in the amplifier. In order to characterize the designed amplifier as an dual-wavelength source for DFG processes additional polarization properties of the output radiation were registered. Degree of polarization (DOP) of both wavelength at maximum output power are depicted in Fig. 4.
It should be noted that although no polarization maintaining fiber elements were used to construct this amplifier, the DOP of both wavelength is surprisingly good. 60 minute stability measurements show that the DOP experiences a drift after the amplifier has been turned on. This effect slows down after 20 minutes, what can be connected with stabilization of the temperature related bifringence changes in the heavily pumped active fiber and passive fibers. Azimuth and ellipticity time-dependent fluctuations are depicted in Fig. 5.
For 1 μm wavelength the 60 minute fluctuations did not exceed 1.7 and 3 degrees for azimuth and ellipticity, respectively. The 1.5 μm radiation performed better, yielding variations less than 1.6 and 2.3 degrees for azimuth and ellipticity, respectively.
3. DFG experimental setup and performance
The next step was to use the dual-wavelength amplifier in a nonlinear frequency generation process, where the influence of all of the registered instabilities would determine overall performance and stability of the system. The experimental difference frequency generation setup is depicted in Fig. 6.
The dual-wavelength source delivers the radiation through a standard FC/APC SMF-28 fiber connector. In order to minimize the number of elements incorporated in the system we used a simplified optical setup. Both laser beams (1 μm and 1.5 μm) emerging from the fiber tip are collimated using an Anti-Reflection coated (AR-coted) standard achromatic doublet lens (Thorlabs AC080-010-C-ML). The beams are then focused by means of an AR-coated achromatic doublet lens (Thorlabs AC254-150-C) with a focal length of 150 mm to spots of diameter equal to 70 μm and 145 μm, for 1 μm and 1.5 μm wavelength, respectively. The profiles of the focused beams can be seen in Fig. 7.
The registered M2 parameters were 1.21 and 1.19 for 1 μm and 1.5 μm wavelength, respectively. Although using an achromatic-doublet-based lens system, the offset between the focusing points of both beams was 9.1 mm (measured in air). The beams were focused into a temperature stabilized, 40 mm long, multi-band AR-coated, Magnesium-doped PPLN crystal (Covesion MOPO1-0.5-40). The crystal offers 9 channels with apertures of 0.5x0.5 mm and periods ranging from 27.91 to 31.59 μm. With adequate period selection, temperature and signal tuning, idler generation stretching from 2128 – 4335 nm is possible. In our case, for Pump and Signal wavelength centered around 1064 nm and 1548 nm, respectively, the 29.98 µm period was used along with temperature stabilization of the crystal at 169°C ( ± 0.01). Relatively high crystal temperature was chosen to reduce the effects of photorefractive damage at high incident powers . At the PPLN crystal output a f = 75 mm, AR-coted, Calcium-Fluoride lens was used (Thorlabs, LA5817-E) to collimate the generated mid-IR radiation. The additional germanium filter was integrated in the system to filter out the un-absorbed Pump and Signal radiation. The collimated mid-IR beam is focused on a three-stage Thermoelectrically cooled (TEC) MCT detector (VIGO, PVI-3TE4) via an off-axis parabolic mirror (Thorlabs, MPD254762-90-P01) after traveling 8 meters in ambient air under atmospheric pressure. Signal recovery was performed with a Lock-in Amplifier (Stanford Research Systems, SR830, LIA) and the signal was observed and registered using an oscilloscope (AT DSO9254A). Figure 8 shows generated Idler power as a function of power delivered to the PPLN crystal.
While running both pumping diodes at maximum current the DFG system was capable of generating more than 1.14 mW of mid-IR radiation centered around 3.39 μm, measured with a 10 μW resolution using a pyroelectric power meter (LabMax Top with PowerMax PS10 measuring head). The calculated conversion efficiency was approximately 71 μW W−2 cm−1 (taking in account 4% and 6% insertion losses of the Ge filter and CaF2 lens, respectively), which is equivalent with experiments performed by Richter et al., in a very similar setup . The conversion efficiency was limited mostly by use of a single collimator setup scheme, which introduced severe offset in the focusing position of the Pump and Signal beams. On top of that, very narrow PPLN channel dimensions (0.5 x 0.5 mm) introduced beam clipping at higher intensities. Although single collimator setup provides lower maximum mid-IR output powers, this approach, due to limited number of elements used, tends to be less sensitive to environmental disturbances, like mechanical vibrations or temperature drift related optics misalignment, making the system more robust and cost-effective. Nevertheless, the achieved mid-IR output power is far in excess of that necessary to compensate the insertion loss of multipass cells commonly used in absorption spectroscopy systems.
4. TDLAS 2f ambient methane measurements
The overall performance of the DFG system was evaluated by carrying out spectroscopic measurements of methane concentrations in ambient air at an mid-IR beam power of approximately 60 μW. Low mid-IR power was deliberately used not to saturate the detector. We targeted a strong methane absorption line centered at 3403 nm (~2938.3 cm−1) which partially overlaps with a weak water absorption line, but at the same time is interference free from other molecules normally present in ambient air. Assuming the concentration of ambient methane will not drop below ~500 ppm, the sensor will still measure the targeted absorption line with little influence of water background. Hitran  spectral simulation of methane and water absorption lines around 2938 cm−1 for an open-path interaction length of 8 m can be seen in Fig. 9.
In order to detect methane absorption lines we used TDLAS technique. In our case the wavelength of the first seed diode was kept stable at a wavelength of 1064 nm and the mid-IR Idler frequency tuning was achieved by varying the injected current to the 1.5 μm seed diode. According to (1) , assuming a fixed Pump wavelength – λp, a linear change of the Signal wavelength – λs will result in a linear change of the generated Idler wavelength – λi:Fig. 10(a). Figure 10(b) shows a 2f signal with signs of detector saturation after rising the methane concentration in the laboratory by fanning-in pure methane released from a cylinder.
The 2f signal was registered with a single scan using a 30 ms time constant on the LIA and no additional sample averaging done by the oscilloscope. The retrieved signal is very similar to the one presented by Armstrong et al. , and consists of several separate methane absorption lines merged together. Unfortunately, at atmospheric pressure the lines are too broad to be able to clearly separate their 2f side-lobes. The baseline in our case was recorded after off-tuning the Idler wavelength from the methane absorption lines to a region where only weak water absorption lines can be found (2940.2 – 2944.5 cm−1). This was achieved by changing the temperature of the 1.5 μm seed diode (additional PPLN crystal temperature offset was administered in order to compensate for the Idler temperature-bandwidth related amplitude changes). Because the measurements were taken in an open-path scheme, we were unable to eliminate the ambient CH4 in order to perform base-line measurements at the exact same wavelength scan range. Assuming an average concentration of methane in ambient air at a level of 1.8 ppmv  and a measured 1σ signal to noise ratio (SNR) of 69, we determined a minimum detectable concentration of methane at a level of 26 ppbv, which is better than obtained by Armstrong et al. in a similar setup .
We presented a novel, up to our best knowledge, unpublished method of generating mid-IR radiation by means of DFG processes in PPLN crystals using a dual-wavelength amplifier. The amplifier design allows for simultaneous amplification of both wavelengths (1 μm and 1.5 μm) essential for the nonlinear process. Not only does it provide more than 14 dB of amplification for the 1 μm seed signal, form energy which under normal circumstances is unexploited, but also prevents from Yb-ASE related instabilities that influence the amplified 1.5 μm signal. The amplifier was characterized by a set of parameters critical for DFG processes. The overall performance of the setup was determined by carrying out simple 2f ambient methane detection using TDLAS. Preliminary results suggest quite bearable minimum detectable concentration at a level of 26 ppbv, considering an open path of 8 m under ambient pressure. Further work will address optimizing the optics used for delivering the beams to the PPLN crystal, additional amplifier pump thermal stabilization, system miniaturization and incorporating a Herriot multipass cell in the system.
The work presented in this paper was supported by the National Science Centre (NCN, Poland) under the project entitled “Generation of mid-infrared radiation using novel dual-wavelength all-fiber laser sources” (decision no. DEC-2012/07/B/ST7/01476). The research fellowship of two of the authors (K.K. and G.S.) is co-financed by the European Union as part of the European Social Fund.
References and links
1. R. F. Curl and F. K. Tittel, “Tunable infrared laser spectroscopy,” Annu. Rep. Prog. Chem., Sect. C 98, 219–272 (2002).
2. K. Krzempek, R. Lewicki, L. Nähle, M. Fischer, J. Koeth, S. Belahsene, Y. Rouillard, L. Worschech, and F. K. Tittel, “Continuous wave, distributed feedback diode laser based sensor for trace-gas detection of ethane,” Appl. Phys. B 106(2), 251–255 (2012). [CrossRef]
3. C. A. Zaugg, R. Lewicki, T. Day, R. F. Curl, and F. K. Tittel, “Faraday rotation spectroscopy of nitrogen dioxide based on a widely tunable external cavity quantum cascade laser,” Proc. SPIE 7945, 79450O (2011). [CrossRef]
4. V. Spagnolo, A. A. Kosterev, L. Dong, R. Lewicki, and F. K. Tittel, “NO trace gas sensor based on quartz enhanced photoacoustic spectroscopy and external cavity quantum cascade laser,” Appl. Phys. B 100(1), 125–130 (2010). [CrossRef]
5. A. A. Kosterev, Y. A. Bakhirkin, F. K. Tittel, S. McWhorter, and B. Ashcraft, “QEPAS methane sensor performance for humidified gases,” Appl. Phys. B 92(1), 103–109 (2008). [CrossRef]
6. M. R. McCurdy, A. Sharafkhaneh, H. Abdel-Monem, J. Rojo, and F. K. Tittel, “Exhaled nitric oxide parameters and functional capacity in chronic obstructive pulmonary disease,” J Breath Res 5(1), 016003 (2011). [CrossRef]
7. R. Lewicki, A. A. Kosterev, D. M. Thomazy, T. H. Risby, S. Solga, T. B. Schwartz, and F. K. Tittel, “Real time ammonia detection in exhaled human breath using a distributed feedback quantum cascade laser based sensor,” Proc. SPIE 7945, 79450K–2 (2011). [CrossRef]
8. F. K. Tittel, D. Richter, A. Fried, I. T. Sorokina, and K. L. Vodopyanov, “Mid-Infrared Laser Applications in Spectroscopy,” Top. Appl. Phys. 89, 458–516 (2003) (Solid-State Mid-Infrared Laser Sources). [CrossRef]
9. D. Richter, D. Lancaster, R. Curl, W. Neu, and F. K. Tittel, “Compact mid-infrared trace gas sensor based on difference-frequency generation of two diode lasers in periodically poled LiNbO3,” Appl. Phys. B 67(3), 347–350 (1998). [CrossRef]
10. S. Stry, P. Hering, and M. Mürtz, “Portable difference-frequency laser-based cavity leak-out spectrometer for trace-gas analysis,” Appl. Phys. B 75(2-3), 297–303 (2002). [CrossRef]
11. V. L. Kasyutich, R. J. Holdsworth, and P. A. Martin, “Mid-infrared laser absorption spectrometers based upon all-diode laser difference frequency generation and a room temperature quantum cascade laser for the detection of CO, N2O and NO,” Appl. Phys. B 92(2), 271–279 (2008). [CrossRef]
12. L. Goldberg, D. G. Lancaster, J. Koplow, R. F. Curl, and F. K. Tittel, “Mid-IR DFG source pumped by a 1.1 μm/1.5 μm dual wavelength fiber amplifier for trace gas detection,” Opt. Lett. 23, 1517–1519 (1998). [CrossRef]
14. G. Sobon, P. Kaczmarek, A. Antonczak, J. Sotor, A. Waz, and K. M. Abramski, “Pulsed dual-stage Fiber-MOPA source operating at 1550 nm with arbitrarily shaped output pulses,” Appl. Phys. B 105(4), 721–727 (2011). [CrossRef]
15. G. Sobon, P. Kaczmarek, A. Antonczak, J. Sotor, and K. M. Abramski, “Controlling the 1 μm spontaneous emission in Er/Yb co-doped fiber amplifiers,” Opt. Express 19(20), 19104–19113 (2011). [CrossRef]
16. V. Kuhn, P. Wessels, J. Neumann, and D. Kracht, “Stabilization and power scaling of cladding pumped Er:Yb-codoped fiber amplifier via auxiliary signal at 1064 nm,” Opt. Express 17(20), 18304–18311 (2009). [CrossRef]
17. V. Kuhn, D. Kracht, J. Neumann, and P. Wessels, “Dependence of Er:Yb-codoped 1.5 μm amplifier on wavelength-tuned auxiliary seed signal at 1 μm wavelength,” Opt. Lett. 35(24), 4105–4107 (2010). [CrossRef]
18. Y. L. Lee, C. Jung, Y. Noh, D. Ko, and J. Lee, “Photorefractive Effect in a Periodically Poled Ti:LiNbO3 Channel Waveguide,” J. Korean Phys. Soc. 44, 267 (2004).
19. D. Richter, B. Wert, A. Fried, P. Weibring, J. Walega, J. White, B. Vaughn, and F. K. Tittel, “High precision carbon dioxide isotope spectrometer with a difference frequency generation laser source,” Opt. Lett. 34, 172–174 (2009). [CrossRef]
21. I. Armstrong, W. Johnstone, K. Duffin, M. Lengden, A. Chakraborty, and K. Ruxton, “Detection of CH4 in the mid-IR using difference frequency generation with tunable diode laser spectroscopy,” J. Lightwave Technol. 28(10), 1435–1442 (2010). [CrossRef]
22. S. D. Bridgham, H. Cadillo-Quiroz, J. K. Keller, and Q. Zhuang, “Methane emissions from wetlands: biogeochemical, microbial, and modeling perspectives from local to global scales,” Glob. Change Biol. 19(5), 1325–1346 (2013). [CrossRef]