We present a compact, all diode-pumped supercontinuum source based on a SESAM mode-locked Yb:glass oscillator at 1040 nm and a tapered fiber. The oscillator has a repetition rate of 20 MHz, a pulse duration of 200 fs, and a maximum pulse energy of about 15 nJ. This system delivers an 1100 nm broad spectrum with an output power of more than 100 mW. Decreasing the repetition rate to 500 kHz by cavity-dumping results in a supercontinuum with a high pulse energy of about 50 nJ. Furthermore, using the frequency-doubled output of this laser at 520 nm with 300 fs pulse duration resulted in supercontinua in the near-UV and visible spectral region. We compare the experimental spectra with theoretical simulations.
©2005 Optical Society of America
The generation of supercontinua in photonic crystal fibers (PCF) with different kinds of laser sources is a very topical subject . It was shown that femto- , pico- [3–4], and even nanosecond pulses  are suitable for continuum generation. Tapered fibers as nonlinear elements are mentioned less frequently in the literature, as there are two seemingly disadvantageous facts. First, a suitable group velocity dispersion (GVD) design is more complicated than in PCFs but in fact is possible . Second, the tapering process restricts the length of the fiber waist  to several ten centimeters, which might limit the applications with commercial low-pulse energy pico- and nanosecond lasers, as they typically require a longer nonlinear interaction length for the generation of a smooth supercontinuum. However, it is feasible to splice several tapered fibers together in order to increase the interaction length. Furthermore, by splicing fibers with diverse diameters and thus different zero dispersion wavelengths together, a flatter and broader spectrum can be achieved . In the femtosecond regime, where only several millimeters of nonlinear fiber are necessary for the continuum generation, tapered fibers are indeed a true competitor to PCFs [9–10]. They can compensate for the challenging GVD design by a much simpler fabrication process, and they possess a large input/output core size that simplifies the incoupling process, offering superior long-term stability.
There is a wide market for white light laser sources, e.g., for optical coherence tomography, for frequency metrology, and in spectroscopy and nonlinear microscopy. But the generally expensive femtosecond laser system limits the number of applications. Our aim was therefore the reduction of the cost of the laser system in combination with a minimization of its size. An Yb:glass oscillator  has the ability to generate femtosecond pulses together with the advantage of being directly diode-pumped, which cuts the cost of this laser system by reducing the price of the pump laser. At the same time, the size of the system becomes very compact and therefore this laser is the optimum choice to achieve our goal.
In this paper, we report on the combination of a diode-pumped Yb:glass oscillator at a wavelength of 1040 nm with tapered fibers of different diameters and thus different zero dispersion wavelengths .
We introduce three different versions of our compact femtosecond white-light source: The standard version has a repetition rate of 20 MHz, a pulse duration of 200 fs, and a maximum utilized pulse energy of about 15 nJ. The cavity dumped version [13–14] has a repetition rate of 500 kHz, a pulse duration of 250 fs, and a maximum utilized pulse energy of 200 nJ. Furthermore, we use the frequency-doubled output of this laser at 520 nm with 300 fs pulse duration and a maximum pulse energy of 20 nJ. In each case, spectrally broad single mode (TEM00) supercontinua were generated in tapered fibers - well suited for the above mentioned applications. Figure 1 outlines the key technical specifications of the three lasers and gives a rough draft of our setup.
In the experiment we used an xyz-stage with a microscope objective to couple the femtosecond pulses into the tapered fiber. By default we used a 0.3 NA objective, but due to a modified beam size, the 0.1 NA objective led to a higher throughput in the cavity dumped version. In the case of the standard version we used a Faraday isolator (EOT 4I 1064) to prevent backreflections from the fiber into the laser. Consequently, this slightly broadened the input pulses. The Faraday isolator was dispensable when working with the cavity dumped version.
For the fabrication of tapered fibers we used single-mode Corning SMF 28 quartz fibers and pulled them in a home-built fiber drawing rig. For drawing, the fibers were heated over a moving propane-butane-oxygen flame with a temperature close to the melting point of quartz. The drawn fibers consisted of a taper region where the outer diameter decreases from 125 µm to a few micrometers over a distance of about 15 mm. This was followed by a waist region with an almost constant diameter in the range from less than one to about five micrometers, followed by another taper region. Variation of the drawing velocity allows the control of the waist diameter in a very reproducible way. The measured waist diameter uncertainty is about 10% . All spectra in this paper were taken with an Ando AQ 6315A optical spectrum analyzer.
3. The Yb:glass oscillator with 20 MHz repetition rate
Figure 2 shows the supercontinuum spectra when pumping with the 20 MHz Yb:glass oscillator. The spectra demonstrate the typical behavior of supercontinua when pumping a tapered fiber with a femtosecond source. Soliton formation and subsequent self-frequency shifting towards the red due to stimulated Raman scattering and due to higher order dispersion in combination with non-solitonic radiation in the blue-green region are the relevant mechanisms in this case [16–19]. The spectra stretch from 400 to more than 1600 nm and display up to three distinct peaks, depending on the input pulse parameter and waist diameter. The broadest spectra are obtained for a waist diameter of 3.3 µm and an input power of 250 mW. A total supercontinuum output power of 111 mW is achieved in this case. For all four waist diameters the non-solitonic radiation in the blue region increases as the pump experiences a stronger anomalous dispersion. Figure 3 elucidates this influence of the waist diameter on the spectral shape. As Fig. 3b shows, a thinner waist diameter corresponds to a stronger anomalous dispersion for the pump at 1040 nm. However, due to phase matching between the self-frequency shifted solitons and the non-solitonic radiation, a stronger anomalous pump leads to a further shift of the non-solitonic radiation towards the blue . This is well documented in Fig. 3a, where the blue peak shifts towards shorter wavelengths when the diameter of the waist is reduced. Simultaneously, the intensity dip around 800 nm increases. A pump that is located closer towards the zero dispersion wavelength will result in a flatter spectrum .
It is worth mentioning that at specific input powers and waist diameter configurations, we obtained an output spectrum with a Gaussian spectral shape around certain wavelengths.
A number of applications such as optical coherense tomography (OCT) require a spectrum with a specifically smooth shape, rather than a predominantly broad spectrum. The two parts of a spectrum presented in Fig. 4 show very smooth Gaussian peaks on a linear scale around 450 and 1100 nm. These pulses might also be compressed to sub-10 fs pulses (5.2 fs in the blue case and 9.2 fs in the IR case) when assuming a time-invariant phase relationship between the different spectral components.
4. Cavity-dumped version of the Yb:glass oscillator with 500 kHz repetition rate as pump source
In our second experiment, we used a cavity-dumped version of the Yb:glass oscillator. This system had the advantage of a reduced repetition rate while keeping the output power high. This technique increased the pulse energies by a factor of 13.
Figure 5 displays the measured supercontinua of four different tapered fibers with a waist length of 90 mm. The cavity-dumped Yb:glass oscillator with a 500 kHz repetition rate and 250 fs pulses at 1040 nm was coupled into the fiber with a 0.1 NA objective. The throughput was 45% in all four fibers at low pulse energies. This throughput dropped significantly to about 30% when increasing the pulse energy. A likely reason is the fact that with higher input power the spectrum broadens towards the blue, where absorption takes place, as the SMF 28 fiber is designed rather for applications in the telecommunication window between 1.3 and 1.55 µm. Below 450 nm the attenuation is supposed to be quite high. The supercontinua stretch from about 400 nm to beyond 1700 nm in the best cases. When comparing the spectral width for different waist diameters, thinner fibers generate broader spectra as expected.
Thicker waist diameters imply pumping closer to the zero dispersion wavelength. This results in flatter spectra, which is quite pronounced. Except for the unconverted pump light around 1040 nm, the continua in Figs. 5(a) and 5(b) (when using a fiber taper diameter around 5 µm) are rather flat. When pumping stronger in the anomalous dispersion regime by decreasing the waist diameter, the non-solitonic radiation shifts further to the blue and increases in such a way that the highest intensities within the spectrum are now located around 500 nm. This leads to the loss of the flatness, and an increasing dip around 800 nm appears. On the other hand, the overall spectral width increases.
The spectral behavior of the four tapered fibers when pumped with approximately the same input power is displayed for comparison in Fig. 6. Just as in the case of pumping with the standard oscillator, the non-solitonic radiation in the blue region shifts further to shorter wavelengths when pumping stronger in the anomalous dispersion regime. This fact and the similar spectral shape of the continua proves that the generation mechanism does not change when increasing the pulse energy by one order of magnitude.
5. Frequency doubled cavity-dumped version of the Yb:glass oscillator with 500 kHz repetition rate as pump source
With such high pulse energies as generated by the cavity-dumped oscillator, it is possible to double the output pulses at 1040 nm very efficiently. However, in order to pump in the anomalous dispersion regime to generate a continuum around 520 nm, the waist diameter of the tapered fiber has to be thinner than 900 nm . Pulling a 90 mm long waist with sub-micron diameter requires an extremely tiny, yet well-controlled flame to ensure that the pressure of the emanating gas does not tear the fiber apart. Figure 7(a) depicts the continua that were generated by the SHG of the cavity-dumped oscillator. It is fascinating to observe how the initially green output of the fiber changes to white when raising the input power. The spectral width stretches from around 370 nm to about 750 nm, and the spectral shape is very flat when coupling an adequate power into the tapered fiber. At high input power the spectrum has three distinct peaks around 410, 520, and 680 nm. These spectral components might be useful for measurement applications and nonlinear microscopy. At all powers the output had a single spatial mode. For reasons that are unclear up to now, the throughput of these very thin fibers was significantly smaller than previously and reached a maximum value of only 25%. This throughput dropped significantly to single digit percentages when increasing the pulse energy. Absorption will certainly influence the throughput again. It is likely that the shape of the tapers has to be improved in order to optimize the transition from core to cladding mode. The stronger blue components in the spectra are hinting that soliton splitting is the reason for the broadening. The blue components are probably non-solitonic radiation in the normal dispersion regime. Simultaneously, the structure of the curve at 0.4 mW input power indicates Stokes and anti-Stokes components , suggesting that four-wave mixing takes place as well. Figure 7(b) shows the corresponding simulation using a modified nonlinear Schrödinger equation. Details of the simulation algorithm may be found in . To mimic the multi-shot nature of the experiment, the simulation spectra represent averages over ten random realizations of the input pulse with a slight spectral phase perturbation. The propagation algorithm in the thin sections of the taper (when the diameter is less than 10 µm) captures exactly the GVD profile over the whole frequency range. The GVD as a function of radius in the input and output tapered sections is modeled by interpolating between the GVD of an untapered SMF28 fiber and a 10 µm thick silica waist in air.
As demonstrated in Fig. 7, at output powers of 0.2 and 0.3 mW, our simulation reproduces the experimental results very well, showing rather flat spectra with the spectral power increasing predominantly at the low-frequency end of the spectrum with increasing pump power. For powers close to threshold, i.e., around 0.1 mW, we see a wider theoretical spectrum compared to the experiment. This is likely due to oversimplified initial condition and due to the approximate treatment of the propagation in the input/output sections of the taper. Note, however, that the low-power simulated spectra indeed exhibit the Stokes/anti-Stokes shoulders similar to those seen in the experiment.
The above comparison between experimental and theoretical supercontinuum spectra shows that the model captures the essential physics and properties of the tapers. However, more quantitative comparison would require to include exactly the nontrivial GVD “evolution” along the input section of the taper , and to use a well-characterized experimental pulse as an input for the simulation. Also, it is important to understand the nature of the rather high losses in these very thin tapers, since they may affect the supercontinuum shaping.
Tapered fibers in combination with a diode-pumped, standard or cavity-dumped Yb:glass oscillator are an ideal combination for the generation of multi-octave femtosecond supercontinua, both when pumping at the fundamental wavelength of 1040 nm using 3 to 5 µm waist diameters, as well as pumping with the second harmonic radiation at 520 nm with sub-micron fiber waists. This combination should find applications in the future in various fields such as spectroscopy, microscopy, and measurement.
The authors thank Rui Zhang for calculating the GVD diagrams. Harald Giessen and Jörn Teipel thank the DFG (FOR557) and BMBF (FKZ 13N8340) for support. Miroslav Kolesik was supported by AFOSR grants F49620-03-1-0194, FA9550-04-1-0355 and NSF grant DMS0335101. Alexander Killi, Uwe Morgner, Max Lederer, and Daniel Kopf thank the European Union contract G1ST-CT-2002-50266 (DACO) for support.
References and links
2. J. K. Ranka, R. S. Windeler, and A. J. Stentz, “Visible continuum generation in air silica microstructure optical fiber with anomalous dispersion at 800nm” Opt. Lett. 25, 25–27 (2000). [CrossRef]
3. S. Coen, A. H. L. Chau, R. Leonhardt, J. D. Harvey, J. C. Knight, W. J. Wadsworth, and P. St. J. Russell, “White-light supercontinuum generation with 60-ps pump pulses in a photonic crystal fiber,” Opt. Lett. 26, 1356–1358 (2001). [CrossRef]
4. S. Coen, A. H. L. Chau, R. Leonhardt, J. D. Harvey, J. C. Knight, W. J. Wadsworth, and P. St. J. Russell, “Supercontinuum generation by stimulated Raman scattering and parametric four-wave mixing in photonic crystal fibers,” J. Opt. Soc. Am. B 19, 753–764 (2002). [CrossRef]
5. W. J. Wadsworth, N. Joly, J. C. Knight, T. A. Birks, F. Biancalana, and P. St.J. Russell, “Supercontinuum and four-wave mixing with Q-switched pulses in endlessly single-mode photonic crystal fibres,” Opt. Express 12, 299–309 (2004), http://www.opticsexpress.org/abstract.cfm?URI=OPEX-12-2-299. [CrossRef] [PubMed]
6. R. Zhang, J. Teipel, X. Zhang, D. Nau, and H. Giessen, “Group velocity dispersion of tapered fibers immersed in different liquids,” Opt. Express 12, 1700–1708 (2004), http://www.opticsexpress.org/abstract.cfm?URI=OPEX-12-8-1700. [CrossRef] [PubMed]
7. S. G. Leon-Saval, T. A. Birks, W. J. Wadsworth, M. W. Mason, and P. St. J. Russell, “Supercontinuum generation in submicron fibre waveguides,” Opt. Express 12, 2864–2869 (2004), http://www.opticsexpress.org/abstract.cfm?URI=OPEX-12-13-2864. [CrossRef] [PubMed]
8. D. Türke, J. Teipel, A. Zintl, B. Braun, and H. Giessen, “Compact high power (5.6 W) picosecond white light source,” CLEO/IQEC and PhAST Technical Digest on CD-ROM (The Optical Society of America, Washington, DC, 2004), CThHH5.
9. T. A. Birks, W. J. Wadsworth, and P. St. J. Russell, “Supercontinuum generation in tapered fibers,” Opt. Lett. 25, 1415–1417 (2000). [CrossRef]
10. J. Teipel, K. Franke, D. Türke, F. Warken, D. Meiser, M. Leuschner, and H. Giessen, “Characteristics of supercontinuum generation in tapered fibers using femtosecond laser pulses,” Appl. Phys. B 77, 245–250 (2003). [CrossRef]
11. J. Aus der Au, D. Kopf, F. Morier-Genoud, M. Moser, and U. Keller, “60-fs pulses from a diode-pumped Nd:glass laser,” Opt. Lett. 22, 307–309 (1997). [CrossRef]
12. J. Teipel, D. Türke, M. Lederer, D. Kopf, and H. Giessen, “Diode-pumped compact broadband femtosecond whitelight laser sources using tapered fibers,” CLEO/IQEC and PhAST Technical Digest on CD-ROM (The Optical Society of America, Washington, DC, 2004), CWA62.
14. A. Killi, J. Dörring, U. Morgner, M. J. Lederer, J. Frei, and D. Kopf, “High Speed Electro-Optical Cavity Dumping of Mode-Locked Laser Oscillators,” submitted to Opt. Express.
16. J. Herrmann, U. Griebner, N. Zhavoronkov, A. Husakou, D. Nickel, J. C. Knight, W. J. Wadsworth, P. St.J. Russell, and G. Korn, “Experimental Evidence for Supercontinuum Generation by Fission of Higher-Order Solitons in Photonic Fibers,” Phys. Rev. Lett. 88, 173901-1–173901-4 (2002). [CrossRef]
17. J. M. Dudley, L. Provino, N. Grossard, H. Maillotte, R. S. Windeler, B. J. Eggleton, and S. Coenet, “Supercontinuum generation in air-silica microstructured fibers with nanosecond and femtosecond pulse pumping,” J. Opt. Soc. Am. B 19, 765–771 (2002). [CrossRef]
18. G. Genty, M. Lehtonen, H. Ludvigsen, J. Broeng, and M. Kaivola “Spectral broadening of femtosecond pulses into continuum radiation in microstructured fibers,” Opt. Express 10, 1083–1098 (2002), http://www.opticsexpress.org/abstract.cfm?URI=OPEX-10-20-1083. [PubMed]
19. A. Ortigosa-Blanch, J. C. Knight, and P. St. J. Russell“Pulse breaking and supercontinuum generation with 200-fs pump pulses in photonic crystal fibers”, J. Opt. Soc. Am. B 19, 2567–2572 (2002). [CrossRef]
20. K. M. Hilligsøe, H. N. Paulsen, J. Thøgersen, S. R. Keiding, and J. J. Larsen, “Initial steps of supercontinuum generation in photonic crystal fibers,” J. Opt. Soc. Am. B 20, 1887–1893 (2003). [CrossRef]
21. W. J. Wadsworth, A. Ortigosa-Blanch, J. C. Knight, T. A. Birks, T. P. M. Man, and P. St. J. Russell, “Supercontinuum generation in photonic crystal fibers and optical fiber tapers: a novel light source,” J. Opt. Soc. Am. B 19, 2148–2155 (2002). [CrossRef]
22. M. Kolesik, E. M. Wright, and J. V. Moloney, “Simulation of femtosecond pulse propagation in sub-micron diameter tapered fibers,” Appl. Phys. B 79, 293–300 (2004). [CrossRef]
23. R. Zhang, X. Zhang, D. Meiser, and H. Giessen, “Mode and group velocity dispersion evolution in the tapered region of a single-mode tapered fiber,” Opt. Express 12, 5840–5849 (2004), http://www.opticsexpress.org/abstract.cfm?URI=OPEX-12-24-5840. [CrossRef] [PubMed]