An ultra-broadband supercontinuum was generated in a short piece of step-index germania-core fiber using a fiber laser with a peak power of 4.4 kW. The pure germania core made this fiber capable of propagating light towards the desirable mid-infrared region. The spectral broadening characteristics towards the mid-infrared region under different lengths of germania-core fiber were investigated using pump pulses of 4.4 kW and 1.1 ns at 1550 nm. The large nonlinear refractive index of germania and the small core size of germania-core fiber produced a nonlinear coefficient as high as 11.8 (W km)−1 at 1550 nm, which was beneficial for supercontinuum generation. The pump wavelength was located in the anomalous dispersion regime and close to the zero dispersion wavelength of this germania-core fiber, 1.426 μm. Eventually, an ultra-broadband supercontinuum source with a spectrum spanning from 0.6 to 3.2 μm was obtained and had a total output power of 350 mW at an optimized germania-core fiber length of 0.8 m. This work is the first demonstration, to the best of our knowledge, of a germania-core fiber-based ultra-broadband supercontinuum source that spans from the visible region to the mid-infrared region.
© 2016 Optical Society of America
Supercontinuum (SC) sources with broadband spectrum ranges are useful in many scientific applications, such as environmental sensing [1,2], bio-medicine [3,4], infrared spectroscopy [5,6], defense and security [7,8]. Comparing to traditional light sources, a fiber-based SC source provides many advantages, including a broadband spectrum, high brightness, spatial coherence, and compactness . Silica fiber has been widely investigated to for SC generation and generally supports the generation of SCs with spectra ranging from 0.4 to 2.4 μm . SC sources operating in the visible and near-infrared (NIR) spectral regions have been developed, reaching the stage of commercialization. However, due to the abrupt increase in phonon resonance absorption loss, it is very difficult to generate SC beyond 2.7 μm within silica fibers . To extend the SC spectrum into longer wavelengths, such as the mid-infrared region, soft-glass materials with lower phonon energy, including fluoride [12–14], telluride  and chalcogenide  fibers, have been invented and researched. Unfortunately, these fibers are often expensive, fragile and rather difficult to handle. In particular, fluoride fibers are highly reactive to water vapor; consequently, any SC source based on fluoride fibers suffers from output power decrease due to fiber-end degradation , thus requiring special protections.
Germania is another type of glass material with similar physical characteristics to silica. Nevertheless, it has a lower phonon energy of ~820 cm−1, which is lower than the phonon energy of silica, ~1100 cm−1 ; thus, fibers made of germania have a longer transmission window extending towards the mid-infrared region. Recently, germania-core fiber (GCF) has attracted great attention in continuous-wave Raman lasers in the near-infrared  and short-wavelength infrared regions . There have also been a few recent SC investigations using GCFs pumped by different laser pulses at 1.5  and 2 μm . In , by pumping a piece of GCF using 1.5 μm pulses with 35 ns pulse width and 6 kW peak pump power, a SC with a spectrum spanning from 1.5 to 2.8 μm was achieved. The GCF has a core doped with 64 mol.% GeO2 and a zero dispersion wavelength (ZDW) of 1.6 μm. The experiment in  utilized a chirped pulse amplification (CPA) system delivering 2 μm pulses of 850 fs, 12 kW, to pump a piece of fiber whose core was doped with 75 mol.% GeO2 (with ZDW in the range 1.8~1.9 μm). The obtained spectrum spanned from 1.9 to 3.0 μm. An all-fiber SC source based on pure germania-core fiber has also been reported, with a 10 dB spectral bandwidth ranging from 1.93 to 3.18 μm . The pumping source was a SC source based on a Tm-doped fiber amplifier (TDFA), with a spectrum ranging from 1.95 to 2.5 μm, and the GCF had a ZDW of 1.74 μm. In the literature mentioned above, because the pump wavelengths were well beyond the ZDWs of the GCFs utilized in the experiments, the spectra mainly extended towards the long-wavelength side, with little expansion towards the short-wavelength side. By carefully designing the ZDW of GCF to match the pump wavelength, the spectrum could hopefully be extended towards the short-wavelength side. Although the scattering loss of germania in visible region is higher than that of silica , it is still acceptable. Therefore, an ultra-broadband SC could be expected.
In this paper, an all-fiber ultra-broadband SC spanning from 0.6 to 3.2 μm was demonstrated based on a short piece of GCF and a fiber laser with a 4.4 kW peak power. The GCF is designed with a small pure germania core with a diameter of 3 μm, resulting in a high nonlinear coefficient of 11.8 (W km)−1 at 1550 nm and a ZDW of 1426 nm. In general, a peak pump power level of ten kilowatts is required for SC generation. However, due to the high nonlinear coefficient of this GCF, a much lower peak power of < 10 kW would be sufficient for a similar result. By injecting ~4.4 kW, 1.1 ns laser pulses at 1550 nm into the GCF, the SC spectra and output power were investigated. In the experiments, the length of the GCF was optimized to 0.8 m to maximize the SC’s bandwidth. The resulting SC has a spectral bandwidth of 2.6 μm ranging from 0.6 to 3.2 μm and an output power of 350 mW. The adopted all-fiber configuration guarantees the high robustness and reliability of the SC source. To the best of the authors’ knowledge, this work is the first to report a GCF-based SC source with such a broadband spectrum, ranging from the visible to the mid-infrared region.
2. Experimental setup
Figure 1 shows the layout of the experimental setup of the GCF-based SC source. It has a very simple configuration that consists of a pulsed seed laser at 1550 nm, a backward pumped Erbium/Ytterbium co-doped fiber amplifier (EYDFA) and a piece of GCF. The pulsed seed laser provides seed pulses with 1.1 ns pulse width at a repetition rate of 100 kHz. The average power of the seed pulses is 5 mW, which could be scaled up to 506 mW by the EYDFA, corresponding to a peak power of 4.6 kW. The output fiber of the combiner is a piece of standard single-mode fiber (SMF-28). The amplified laser pulses are then coupled into the GCF with a permanent low-loss fusion splicing joint between the SMF-28 and the GCF fiber to make the whole SC system all-fiber structured, guaranteeing its robustness and reliability. The splicing loss is 5%, measured with a cut-back method utilizing a low-power continuous-wave laser at 1550 nm. The end of the GCF was cleaved at 8° to reduce the back-reflected light coupled into the fiber core.
Figure 2 shows the optical spectrum of the EYDFA operating at 4.6 kW peak power. The spectrum spans from 1525 to 1608 nm (30 dB decrease of the peak intensity) and is slightly broadened due to nonlinear effects in the EYDFA. Thus, the 30 dB optical bandwidth of the pump laser was ~83 nm. More specifically, a spectral density integral shows that ~94.5% of the total output power is located in the wavelength range from 1525 to 1608 nm.
The GCF used has a pure germania core surrounded by a silica cladding. The core/cladding diameters are 3/125 μm, respectively. The core numerical aperture (NA) of the GCF is ~0.65. The nonlinear refractive index of germania is ~4.5 times  larger than that of silica. The small core and large nonlinear refractive index result in a calculated high nonlinear coefficient of 11.8 (W km)−1 for this GCF at 1550 nm, which is ~6.5 times larger than that of SMF-28 fiber. Figure 3(a) plots the numerically calculated group velocity (GV) and the group velocity dispersion (GVD) curves of the GCF. Compared to the ZDW of bulk germania at 1.738 μm , this GCF has a numerically calculated ZDW of 1.426 μm due to the positive waveguide dispersion. The unique characteristic of dispersion benefits SC generation in this GCF towards both the long- and short-wavelength sides. There are two marked points on the GV curve, 3020 nm and 700 nm, which represent the matched wavelengths of soliton pulses at the long-wavelength side and the wavelength of the trapped dispersion wave (DW) at the short-wavelength side, respectively. The cut-off wavelength of the fundamental mode of this GCF is calculated to be ~2.55 μm.
Figure 3(b) plots the spectral loss profile of GCF measured in the experiment together with the theoretically calculated loss profile of bulk silica  and germania . Compared with pure silica, the introduction of a germania core is shown to have greatly reduced the infrared loss at the long-wavelength side, i.e., for wavelengths beyond 2.2 μm. However, because the core diameter of GCF is only 3 μm, the ability to constrain long-wavelength light in the fiber core is limited. Thus, a portion of long-wavelength light leaks into the silica cladding of GCF and is then attenuated. As a result, the long-wavelength transmittance of GCF is lower than that of pure germania material. The rapidly increasing absorption loss of GCF beyond 3.0 μm indicated that an optimal GCF length was critical to balance nonlinear gain and absorption loss.
3. Results and discussion
In the experiments performed to optimize the length of GCF, the peak power of the pump source was fixed at 4.6 kW. Therefore, the peak pump power injected into the GCF was estimated to be ~4.4 kW, considering a fusion splicing loss of 5%. The initial length of the GCF was 4.0 m and was cut back stepwise to optimize the SC’s performance. Both the spectra and the output power of the SC were recorded. To obtain a SC with the longest wavelength, SC spectra above 2600 nm with different fiber lengths were measured. As shown in Fig. 4(a), when the GCF length increases from 0.1 to 4.0 m, the long-wavelength edge at the −10 dB level of the SC spectrum first increases from ~2900 nm to > 3000 nm and then decreases to < 2900 nm. When the GCF is short, nonlinear gain rather than absorption loss is the dominant factor, so the long-wavelength edge at the −10 dB level continually extends towards the long-wavelength region. However, as the length of the GCF increases, the absorption loss at the long-wavelength edge at the −10 dB level gradually increases to exceed the nonlinear gain, and therefore the long-wavelength edge of the SC spectrum at the −10 dB level shrinks slightly. It is easy to determine that the SC has its long-wavelength edge at the −10 dB level of ~3200 nm at the length of 0.8 m. Compared with 0.8 m, shorter lengths of GCF (0.1 and 0.4 m) result in insufficient nonlinear spectral broadening, whereas longer lengths (2.4 and 4.0 m) cause extra absorption loss at the long-wavelength side.
Figure 4(b) plots the results of the measured 10 dB decrease in the peak intensity on the long-wavelength side and the output power at different GCF lengths. The long-wavelength edge at the −10 dB level first increases and then decreases as the GCF length increases, peaking at a GCF length of 0.8 m. The output power of the SC source decreases gradually from 430 mW to 280 mW as the GCF length increases from 0.1 m to 4.0 m because of the absorption loss. At the optimized GCF length of 0.8 m, the SC wavelength is maximally extended to 3.2 μm, and an output power of 350 mW is obtained. The evolution of both the long-wavelength edge at the −10 dB level and the output power in Fig. 4(b) agrees well with the explanation of Fig. 4(a).
Figure 5(a) plots the measured full spectra of the SC source at the initial fiber length of 4.0 m and at the optimized length of 0.8 m, showing one full spectral bandwidth of 2.2 μm ranging from 0.75 to 2.95 μm, and of 2.6 μm, ranging from 0.6 to 3.2 μm (excluding the pump residual at 1550 nm). More specifically, the 10 dB spectral bandwidth of the resultant supercontinuum at the optimized GCF length of 0.8 m was 2281 nm (spanning from 717 to 2998 nm), excluding the pump residual. The main nonlinearities responsible for this SC generation are the Raman soliton self-frequency shift towards the long-wavelength side and the matched dispersion wave generation on the short-wavelength side. Figure 5(b) and 5(c) depict the measured beam profiles of the SC source filtered at 700 nm and beyond 2700 nm, which were captured by a CCD camera and an InAs camera, respectively, indicating Gaussian-like beam output. In particular, considering the cut-off wavelength of the fundamental mode of this GCF, one can infer fundamental mode profiles for both soliton pulses in the long-wavelength region of > 2.55 μm and the resulting dispersive wave in the short-wavelength region, as shown in Fig. 5(b) and Fig. 5(c). It is believed that most of the pump power propagates in the fundamental mode. The fundamental-mode propagation of the pump light was guaranteed by fine cleaving and careful fusion splicing. During the fusion splicing process, precise alignment between fiber cores was needed.
To estimate the portion of the pump residual near 1550 nm, a spectral density integral of the SC spectrum ranging from 1545 to 1555 nm (10 nm bandwidth centered at 1550 nm) was calculated, showing that the pump residual is only ~8.6%, which means 320 mW of the total 350 mW is the SC component. It is worth noting that the 30 dB spectral bandwidth of the output laser of the EYDFA (83 nm) was far less than the 10 dB spectral bandwidth (2281 nm) of the resultant SC spectrum. Therefore, it is reasonable to conclude that the main spectral broadening process was due to the nonlinear effects in the GCF rather than in the EYDFA.
More importantly, the whole SC range covers the abundant absorption lines of many gases and molecules in the infrared region, allowing it to be used in spectroscopy. To preliminarily verify the validity of the SC source, a water-vapor absorption experiment was conducted. The atmosphere in the laboratory has a high relative humidity (~60%) due to the humid climate. Therefore, the absorption of water vapor is easy to observe. The output SC was first collimated by a CaF2 lens, then recorded by a grating-based optical spectrum analyzer (OSA) at a distance of 3 m. A 0.05 nm-resolution sweep of the SC spectrum from 1300 to 2000 nm showed two clear absorption bands due to water vapor in the laboratory atmosphere. Figure 6(a) shows the absorption spectrum of water vapor from the HITRAN database  and the measured data at the 1.4 μm waveband, while Fig. 6(b) shows the corresponding information for the 1.9 μm waveband. The HITRAN data and experimental data are plotted in the same frame of axes, and the measured data are mirrored on the baseline for clarity. The measured data in Fig. 6 was the result of a single sweep. The absorption spectra of water vapor has little overlap with those of other gas components of the atmosphere (e.g., CO2, O2) at the 1.4 and 1.9 µm wavebands . Consequently, due to the minimal absorption disturbance from other gas components, high similarity is obtained between the measured data and the HITRAN data.
Because the whole SC spectrum covers the important 1, 1.5, 2 and 2.7-3 μm spectral bands, it could be widely used to test in-laboratory optical components at these wavebands. In addition, the high robustness and compactness of the SC, owing to its simple structure and all-fiber configuration, indicate its excellent feasibility for various applications, for instance, serving as a broadband light source in gas detection and hyperspectral imaging systems. Finally, considering the excellent power handling ability of GCF, similar to that of silica fiber, power scaling could be readily achieved by increasing the pulse width or repetition rate while keeping the peak pump power constant, which could further extend the application scenarios.
In conclusion, to the best of our knowledge for the first time, an all-fiber ultra-broadband SC spanning from 0.6 to 3.2 μm was demonstrated based on a short piece of GCF and a fiber laser with 4.4 kW peak-power. By pumping the GCF with a 4.4 kW peak-power fiber laser at 1550 nm, an ultra-broadband SC source was obtained with a spectrum spanning from 0.6 to 3.2 μm. The ultra-broadband output spectrum ranging from the visible region to the mid-infrared region has a wide range of applications in areas such as spectroscopy (e.g., gas detection) and environmental sensing (e.g., hyperspectral imaging systems).
This work is supported by the National Natural Science Foundation of China (NSFC) (Grant No. 61235008, 61435009 and 11504424), Hunan Provincial Natural Science Foundation of China (Grant No. 14JJ3001) and Graduate Student Innovation Foundation of National University of Defense Technology (Grant No. B150703).
References and links
1. G. Méjean, J. Kasparian, E. Salmon, J. Yu, J.-P. Wolf, R. Bourayou, R. Sauerbrey, M. Rodriguez, L. Wöste, and H. Lehmann, “Towards a supercontinuum-based infrared lidar,” Appl. Phys. B 77(2–3), 357–359 (2003). [CrossRef]
2. N. Cezard, A. Dobroc, G. Canat, M. Duhant, W. Renard, C. Alhenc-Gelas, S. Lefebvre, and J. Fade, “Supercontinuum laser absorption spectroscopy in the mid-infrared range for identification and concentration estimation of a multi-component atmospheric gas mixture,” in SPIE Remote Sensing, (International Society for Optics and Photonics, 2011), 81820V.
3. A. B. Seddon, “Potential for using mid-infrared light for non-invasive, early-detection of skin cancers in vivo,” in 2013), in SPIE BiOS, (International Society for Optics and Photonics, 2013), 85760V.
4. A. Labruyère, A. Tonello, V. Couderc, G. Huss, and P. Leproux, “Compact supercontinuum sources and their biomedical applications,” Opt. Fiber Technol. 18(5), 375–378 (2012). [CrossRef]
6. L. J. Medhurst, “FTIR determination of pollutants in automobile exhaust: an environmental chemistry experiment comparing cold-start and warm-engine conditions,” J. Chem. Educ. 82(2), 278 (2005). [CrossRef]
7. H. T. Bekman, J. Van Den Heuvel, F. Van Putten, and R. Schleijpen, “Development of a mid-infrared laser for study of infrared countermeasures techniques,” in European Symposium on Optics and Photonics for Defence and Security, (International Society for Optics and Photonics, 2004), 27–38. [CrossRef]
8. F. K. Tittel, D. Richter, and A. Fried, “Mid-infrared laser applications in spectroscopy,” in Solid-state Mid-infrared Laser Sources (Springer, 2003), 458–529.
9. J. M. Dudley and J. R. Taylor, Supercontinuum Generation in Optical Fibers (Cambridge University, 2010).
11. J. Swiderski, “High-power mid-infrared supercontinuum sources: Current status and future perspectives,” Prog. Quantum Electron. 38(5), 189–235 (2014). [CrossRef]
12. J. Swiderski, F. Théberge, M. Michalska, P. Mathieu, and D. Vincent, “High average power supercontinuum generation in a fluoroindate fiber,” Laser Phys. Lett. 11(1), 015106 (2014). [CrossRef]
13. C. Xia, M. Kumar, M. N. Islam, A. Galvanauskas, F. L. Terry, and M. J. Freeman, “All-fiber-integrated mid-infrared supercontinuum system with 0.7 watts time-averaged power,” in Conference on Lasers and Electro-Optics, (Optical Society of America, 2007), CThH1. [CrossRef]
14. W. Yang, B. Zhang, G. Xue, K. Yin, and J. Hou, “Thirteen watt all-fiber mid-infrared supercontinuum generation in a single mode ZBLAN fiber pumped by a 2 μm MOPA system,” Opt. Lett. 39(7), 1849–1852 (2014). [CrossRef] [PubMed]
15. R. Thapa, D. Rhonehouse, D. Nguyen, K. Wiersma, C. Smith, J. Zong, and A. Chavez-Pirson, “Mid-IR supercontinuum generation in ultra-low loss, dispersion-zero shifted tellurite glass fiber with extended coverage beyond 4.5 μm,” in SPIE Security and Defence, (International Society for Optics and Photonics, 2013), 889808.
16. R. R. Gattass, L. B. Shaw, V. Nguyen, P. Pureza, I. D. Aggarwal, and J. S. Sanghera, “All-fiber chalcogenide-based mid-infrared supercontinuum source,” Opt. Fiber Technol. 18(5), 345–348 (2012). [CrossRef]
17. K. Yin, B. Zhang, J. Yao, L. Yang, S. Chen, and J. Hou, “Highly stable, monolithic, single-mode mid-infrared supercontinuum source based on low-loss fusion spliced silica and fluoride fibers,” Opt. Lett. 41(5), 946–949 (2016). [CrossRef] [PubMed]
18. X. Zou and T. Izumitani, “Spectroscopic properties and mechanisms of excited state absorption and energy transfer upconversion for Er3+-doped glasses,” J. Non-Cryst. Solids 162(1), 68–80 (1993). [CrossRef]
19. E. M. Dianov and V. M. Mashinsky, “Germania-Based Core Optical Fibers,” J. Lightwave Technol. 23(11), 3500–3508 (2005). [CrossRef]
20. B. A. Cumberland, S. V. Popov, J. R. Taylor, O. I. Medvedkov, S. A. Vasiliev, and E. M. Dianov, “2.1 µm continuous-wave Raman laser in GeO2 fiber,” Opt. Lett. 32(13), 1848–1850 (2007). [CrossRef] [PubMed]
21. V. A. Kamynin, A. E. Bednyakova, M. P. Fedoruk, I. A. Volkov, K. N. Nishchev, and A. S. Kurkov, “Supercontinuum generation beyond 2 µm in GeO2 fiber: comparison of nano- and femtosecond pumping,” Laser Phys. Lett. 12(6), 065101 (2015). [CrossRef]
22. M. Zhang, E. J. Kelleher, T. H. Runcorn, V. M. Mashinsky, O. I. Medvedkov, E. M. Dianov, D. Popa, S. Milana, T. Hasan, Z. Sun, F. Bonaccorso, Z. Jiang, E. Flahaut, B. H. Chapman, A. C. Ferrari, S. V. Popov, and J. R. Taylor, “Mid-infrared Raman-soliton continuum pumped by a nanotube-mode-locked sub-picosecond Tm-doped MOPFA,” Opt. Express 21(20), 23261–23271 (2013). [CrossRef] [PubMed]
23. V. V. Dvoyrin and I. T. Sorokina, “All-fiber optical supercontinuum sources in 1.7-3.2 μm range,” in Proc. of SPIEVol. 896189611C–2, 2014), 89611C. [CrossRef]
28. “"HITRAN” 2012 database, http://vpl.astro.washington.edu/spectra/h2opnnlimagesmicrons.htm.
29. “"HITRAN” database, http://hitran.iao.ru/.