Open Access
Add to BibSonomyAbstract
There is an increasing demand for all-fiber passively mode-locked lasers with pulse repetition rates in the order of gigahertz for their potential applications in fields such as telecommunications and metrology. However, conventional mode-locked fiber lasers typically operate at fundamental repetition rates of only a few megahertz. In this paper, we report all-fiber laser operation with fundamental repetition rates of 4.24 GHz, 9.63GHz and 19.45GHz. This is, to date and to the best of our knowledge, the highest fundamental repetition rate reported for an all-fiber laser. The laser operation is based on the passive modelocking of a miniature all-fiber Fabry-Pérot laser (FFPL) by a carbon nanotube (CNT) saturable absorber. The key components for such device are a very high-gain Er:Yb phosphosilicate fiber and a fiber compatible saturable absorber with very small foot print and very low losses. The laser output of the three lasers was close to transform-limited with a pulsewidth of approximately 1ps and low noise. As a demonstration of potential future applications for this laser, we also demonstrated supercontinuum generation with a longitudinal mode-spacing of 0.08nm by launching the laser operating at 9.63GHz into 30m of a highly nonlinear dispersion shifted fiber.
©2011 Optical Society of America
1. Introduction
Fiber lasers offer high beam quality, reliability and efficient heat dissipation in a compact size [1]. As a result, they are often preferred to the bulkier conventional solid state lasers for many commercial applications including material processing, supercontinuum generation, optical frequency metrology and biomedical applications. In addition, passively modelocked fiber lasers produce highly stable, close to transform-limited pulses of subpicosecond duration in a simple configuration [2]. Typically fiber lasers operate at low pulse repetition frequencies in the range of 10s of megahertzs. This is due to the long laser cavities (generally in the order of meters) required to attain sufficient gain in erbium doped fibers. There is, however, a demand for passively mode-locked lasers with a high pulse repetition frequency in research fields such as optical communications, microscopy and metrology. In order to push the operation into the gigaherz regime, one can operate a pulsed fiber laser in a higher harmonic of its fundamental repetition rate [3, 4]. However, this approach relies on having multiple pulses at one time in the laser cavity which leads to pulse to pulse jitter and supermodes, thus, mode suppression and jitter control are needed adding complexity and cost to the device [4]. In light of this, there is still a need in the market for high repetition rates while maintaining fundamental pulse operation in an all-fiber configuration. Recently, researchers have achieved repetition rates as high as 447MHz while preserving the ring cavity fiber laser format by using a simplified cavity with integrated components and a carbon nanotube based saturable absorber (CNT-SA) [5]. In that case, the laser cavity was less than 50cm-long and the laser pulsewidth was 270fs. While the results shown in [5] are impressive, it is complicated to scale down the cavity length to increase the repetition rate any further in an all-fiber ring cavity format. In order to reach multi-gigahertz fundamental repetition frequencies, we previously proposed a fiber Fabry-Pérot laser (FFPL) configuration where passively modelocked laser operation at 5GHz repetition rate was demonstrated making use of a CNT-SA [6]. CNTs offer several advantages for these applications since, in addition to their well documented absorption properties and fast recovery time; they are easily integrated into a fiber configuration and have very small dimensions and low losses [7]. A similar approach employing a semiconductor saturable absorber mirror (SESAM) has been reported where a maximum fundamental repetition frequency of 2GHz was demonstrated [8]. However, in order to use a SESAM, free space coupling was required and this added complexity to the laser cavity and limited the laser’s repetition rate. More recently, researchers have achieved 1GHz repetition rate Er-doped fiber laser generating pulses as short as 187fs using a saturable Bragg reflector [9]. In the last few years, CNT have been making the transition from a material with very interesting optical properties and potential from a scientific point of view to a position where CNT-SA are an important device in the development of real technological applications and their advantages over competing techniques are utilized. In addition to the results reported in this paper, a number of reports apply CNT to the development of optical sampling systems [10], frequency combs [11], octave spanning supercontinuum generation [12], few-cycle pulses fiber lasers [13] and tunable, ultra-broadband fiber laser operation [14, 15].
In this paper, we demonstrate all-fiber passively modelocked lasers producing subpicosecond pulses with fundamental repetition rates as high as 19.45GHz, corresponding to a laser cavity of less than 5 millimeters. This represents, to the best of our knowledge the fastest fundamental repetition rate for an all fiber laser. We “sandwiched” a section of a highly doped Er3+:Yb3+ fiber with two fiber ferrules coated with highly reflective dielectric mirrors one of which was coated with a thin layer of CNT to provide the saturable absorption required for self-started modelocking. We work with three different lengths of Er3+:Yb3+ fiber of approximately 25mm, 10mm and 5 mm length, which lead to fiber lasers operating at fundamental repetition rates of 4.24GHz, 9.63GHz and 19.45GHz, respectively with pulsewidth of approximately 1ps. Furthermore, we generate supercontinuum with a mode spacing as high as 0.08nm by amplifying the output of the laser operating at a repetition rate of 9.63GHz with an erbium doped fiber and launching it into 30m of a highly nonlinear dispersion shifted fiber (HNL-DSF). Supercontinuum sources operating at high frequencies are desirable frequency comb sources for frequency metrology applications since the large mode spacing would allow absolute wavelength resolution of the frequency comb and such high repetition frequencies increase the power per individual mode [1, 16, 17]. For these applications to be realized, we require phase-stabilized, coherent supercontinuum. Hence, the upcoming challenge for this technology is to reduce the pulsewidth of these laser sources from the current 1ps to less than 100fs [16, 17].
2. Experimental set up
We achieve fiber laser operation with a fundamental repetition rates reaching and exceeding 10GHz by implementing an all-fiber Fabry-Pérot laser (FFPL) such as depicted in Fig. 1 . The repetition frequency of a FFPL is determined by the cavity length, L, and the refractive index of the gain material, n, as shown in Eq. (1).
where c is the velocity of light in vacuum. For example, for a 10GHz repetition rate, the laser cavity must have a total length of approximately 10 millimeters. Therefore, the key requirements to implement such device are; a gain fiber with a very high pump light absorption and gain, a fiber-compatible saturable absorber with a small footprint and a high finesse laser cavity.
Fig. 1 Scheme of a fiber Fabry- Pérot laser (FFPL). WDM - wavelength division multiplexer, HR - highly reflective mirror CNT-SA - carbon nanotube saturable absorber.
Stringent requirements must be satisfied to achieve passive modelocking at such fundamental repetition rates. In order to reduce the cavity length, hence increasing the repetition rate, the gain fiber must be highly doped and must absorb efficiently the pump light. Here we use a phosphosilicate fiber heavily codoped with erbium (Er3+) and ytterbium (Yb3+) as the gain fiber. Erbium doped fibers are routinely used for fiber laser applications however erbium has low pump absorption per unit length, it also has a low absorption cross-section compared to other rare-earth materials. Increasing the concentration levels is not a suitable solution since it leads to clustering and ion pair interactions. The low pump absorption of erbium doped fibers means that relatively long lengths of erbium-doped fiber are required to provide sufficient gain impeding the implementation of short cavity fiber laser configurations such as the FFPL proposed here for high pulse repetition operation. The use of ytterbium (Yb3+) as a co-dopant with erbium offers a number of advantages over solely erbium doped fiber [18]. Unlike erbium, the ytterbium concentration in a fiber is not restricted by concentration quenching, thus, higher concentration levels are available. In addition, the absorption cross-section for ytterbium is up to two orders of magnitude larger than that of erbium, leading to a highly efficient absorption of the pump. The fiber used in this work is a soft phosphate glass fiber heavily doped with 1wt.% of Er+3 and 8wt.% of Yb+3 with a core diameter of 14μm and low NA of ~0.08. The absorption at the pump wavelength is 2.4dB/mm, the fiber is single-mode around the wavelength of 1560nm [19]. In this paper, we use sections of this fiber of approximately 25mm, 10mm and 5mm which roughly correspond to fundamental repetition rates of 4GHz, 10GHz and 20GHz based on Eq. (1).
The saturable absorber used to achieve pulse operation consists of a CNT film deposited on one of the fiber ferrule mirrors that forms the laser cavity. The CNTs used in this work were commercial available CNTs, made by the high-pressure CO conversion (HiPCO) method. CNTs were independently dispersed into Dimethylformamide (DMF) solvent by ultrasonification in order to separate individual CNTs and breaking of bundles of CNT that are formed due to van der Waal forces. Efficient dispersion was achieved after 30 minutes of ultrasonification. The solution is then subjected to centrifugation in order to separate the remaining agglomerated CNTs. The CNTs were then sprayed into the mirror surface. The sprayed CNT films exhibit estimated insertion losses lower than 0.5dB. Modulation depth and saturation fluence were measured by using a test fiber pigtail sample that was sprayed at the same time as the mirrors used in the experiment, the measured values were approximately 5% and 2.5MW/cm2, respectively.
The laser cavity is completed with two highly reflective dielectric mirrors deposited in the polished end faces of two single mode fibers, therefore maintaining the all-fiber configuration. The mirrors used in this work have a reflectivity at the emission wavelength higher than 99% and transmittance at the pump wavelength (980nm) higher than 90% The laser is pumped using a 980nm laser diode through a 980nm/1550nm wavelength division multiplexing (WDM) coupler.
3. Experimental results and analysis
Self-starting mode-locked laser operation was observed for all three lasers of different cavity lengths. The longest laser (with an approximate length of 25mm) operated at a central wavelength of 1600nm with an optical spectral bandwidth at full-width half maximum of approximately 4.1nm. Mode-locking started at a relatively low pump power of approximately 100mW at this pump power the output power was −2dBm. The 10mm-long laser operated at a central wavelength of 1560nm and the optical spectral bandwidth was 3.3nm. In this case, the pump power necessary for self modelocking was approximately 300mW and the output power was 4dBm. Finally, the 5mm-long laser operated at a central wavelength of 1563nm with an spectral bandwidth of 4.2nm. The shorter laser required the highest pump power for self-modelocking, approximately 600mW and the output power was 8dBm. The optical spectrum for the three lasers is shown in Fig. 2(a) , Fig. 2(b) and Fig. 2(c). In Fig. 2(a), we fitted the output spectrum of the laser to a hyperbolic secant squared (sech2) profile and a Gaussian profile, from which we can see that the output spectrum is well fitted by the sech2 profile for the three lasers. Stable pulsed operation was observed for the three lasers for a range of pump powers centered in the self-starting pump powers for modelocking mentioned above, however this range was significantly larger for the 25mm-long, and 10mm-long lasers, than for the 5mm-long laser. While the two shorter lasers operate at central wavelength of approximately 1560nm, the longer laser operates at 1600nm. This is due to the pump light being absorbed fully in the first section of the fiber combined with the higher absorption of the Er3+:Yb3+ fiber at 1560nm than at 1600nm. By increasing the pump power we also observed laser operation at 1560nm; however it was always in coexistence with the 1600nm peak.
Fig. 2 Optical spectrum of the (a) 25mm-long laser, (b) 10mm-long laser and (c) 5mm-long laser. (d) pulsed laser scheme. Optical spectrum showing the longitudinal modes and mode-spacing for the (e) 25mm-long laser, (f) 10mm-long laser and (g) 5mm-long laser.
Figure 2(d) shows the basic pulsed laser scheme. When a pulsed laser is operating at its fundamental repetition rate, only one pulse exist at any given time in the laser cavity, likewise it produces one pulse per round trip, hence the repetition rate is inversely proportional to the cavity length. In addition, the separation between adjacent longitudinal modes (mode-spacing) is inversely proportional to the cavity length, following Eq. (2).
where Δλ is the longitudinal mode-spacing, n is the refractive index of the fiber and L is the cavity length. In Fig. 2(e), Fig. 2(f) and Fig. 2(g), we can clearly observe the mode-spacing of the three lasers which is 0.036nm, 0.076nm and 0.175nm for each of the lasers. From this we can extract the actual cavity lengths of each of the lasers to be 24.2mm, 10.9mm and 4.75mm, respectively.The pulse duration of the three lasers was measured using an autocorrelator. Assuming a sech2-pulse shape, the autocorrelator trace yielded pulse durations of 0.68ps, 0.94ps and 0.79ps for the 25mm, 10mm and 5mm-long lasers. From this, we calculated the time-bandwidth products (TBP) of the three lasers to be 0.33, 0.38 and 0.41 respectively, close to the transform-limited value for sech2-pulse shapes of 0.315. The TBP of the three lasers indicates a slight chirp that can be attributed to the extra couple of meters of SMF fiber used between the laser cavity and the measuring devices.
We measured the repetition rate using an RF spectrum analyzer. The measured fundamental repetition rates are 4.237GHz, 9.64GHz and 19.45GHz. Figure 4 shows the RF signal of the 25mm-long laser at its fundamental frequency and its higher order frequency components multiples of the fundamental frequency up to the RF spectrum analyzer range limit (26GHz). The RF signal to noise ratio of the fundamental frequency is 56dB and the frequency width is 120Hz with a resolution bandwidth of 30Hz. The time-jitter of this laser is analyzed latter on in this section.
Fig. 4 RF signal for the laser operating at a fundamental repetition rate of 4.24 GHz, (a)fundamental repetition rate, (b) 2nd, (c) 3rd, (d) 4th, (e) 5th and (f) 6th order.
In Fig. 5 , we can see the fundamental and second order frequency components of the 10mm-long laser operating at a fundamental frequency of 9.63GHz. In this case, the signal to noise ratio is 45dB with frequency width of 60Hz and resolution bandwidth 30Hz. In Fig. 6 , we can see the RF signal at the fundamental frequency of the 5mm-long laser operating at 19.45GHz with a signal to noise ratio of 27 dB frequency width of 400Hz and a resolution bandwidth of 100Hz.
Fig. 5 RF signal for the laser operating at a fundamental repetition rate of 9.63 GHz fundamental repetition rate laser, (a) fundamental frequency, (b) 2nd order frequency.
We also investigated the jitter for the laser operating at a fundamental frequency of 4.237GHz, following the method described in [20]. We measured the RF power peaks of the signal (PA), and the narrow-band (PB) and wide-band components (PC) and the frequency bandwidth of the noise bands for the fundamental frequency and higher order frequency components. Figure 7(a) shows the scheme of the measured values of the various frequency components (nf1), and Fig. 7(b) shows the measured power ratios PB/PA and the PC/PA as a function of the order of the frequency component (n). n2-functions were fitted to the measured PB/PA and the PC/PA power ratios for the laser operating at 4.24GHz for a frequency range of 20kHz and 1MHz respectively and from this we estimated low and high timing jitters of 58fs and 372fs.
Fig. 7 Time-jitter analysis for the three lasers. (a) Schematic of the measured parameters corresponding to the signal (PA), and the narrow-band (PB) and wide-band components (PC) measured at the multiples of the fundamental frequency. (b)Measured values for the three lasers, fitted n2-function for the PB/PA and PC/PA, for the 25mm-long laser.
While the time-jitter and noise level for the laser operating at 9.63GHz and 4.24GHz are similar, the laser noise and time-jitter for the 19.45GHz were significantly higher and the laser stability was lower. The lower performance observed in the smallest laser is due to the significantly higher pump power required for modelocking and the higher frequency which lead to increased heating of the gain fiber. Furthermore, it is likely that the pump light in the 5-mm long laser was not fully absorbed by the fiber contributing to additional heating and degradation of the saturable absorber, while the longer lasers exhibit good stability and low jitter.
Finally, as demonstration of potential applications for this laser, we generated supercontinuum with a 9.63GHz mode spacing. Supercontinuum generation is arguably one of the most promising applications for this type of lasers since supercontinuum sources operating at high frequencies are desirable frequency comb sources for frequency metrology applications where the large mode spacing allows absolute wavelength resolution of the frequency comb with an optical meter [1]. By amplifying the FFPL output to an average power of approximately 25dBm using an erbium doped fiber and launching the amplified pulses into 30m of highly nonlinear dispersion shifted fiber (HNL-DSF) with a nonlinear coefficient of 20W−1.km−1, zero dispersion at 1576nm and normal dispersion at the seed wavelength (1560nm). In Fig. 8(a) the experimental set-up employed for the supercontinuum generation is shown. Figure 8(b) shows the optical spectra of the seed FFPL and the generated supercontinuum.
Fig. 8 Supercontinuum generation. (a) Set-up used for supercontinuum generation, HNLF-Highly Nonlinear fiber and EDFA (Erbium doped fiber amplifier. (b) seed laser source (f1-9.63GHz) and generated supercontinuum.
Further work is required before these laser sources can be employed as supercontinuum sources for many real applications in particular for optical frequency metrology where highly coherent, octave spanning supercontinuum is required. Our current work focuses on improving the laser stability, reducing the noise as well as optimizing the characteristics of the laser cavity as well as managing the dispersion outside the cavity in order to reduce the pulsewidth of the laser output.
4. Conclusions
In this paper, we demonstrate self-starting, passively mode-locked laser operation in a fiber Fabry-Pérot laser configuration. The lasers operated at fundamental repetition frequencies of 4.24GHz, 9.63GHz and 19.45GHz, the highest ever reported fundamental repetition rate of an all-fiber laser. The pulsewidth of the lasers was approximately 1ps. Finally, we generated supercontinuum with a 9.63GHz mode-spacing by amplifying the corresponding laser output into launching it into a highly nonlinear dispersion shifted fiber.
Acknowledgements
The authors would like to thank Dr. S.Y. Set and Dr. C.S. Goh from Alnair Labs for the helpful discussions and technical support.
References and links
1. M. E. Fermann and I. Hartl, “Ultrafast fiber laser technology,” IEEE J. Sel. Top. Quantum Electron. 15(1), 191–206 (2009). [CrossRef]
2. E. P. Ippen, “Principle of Passive Mode Locking,” Appl. Phys. B 58(3), 159–170 (1994). [CrossRef]
3. A. B. Grudinin and S. Gray, “Passive harmonic mode locking in soliton fiber lasers,” J. Opt. Soc. Am. B 14(1), 144 (1997). [CrossRef]
4. M. Nakazawa, K. Tamura, and E. Yoshida, “Supermode noise suppression in a harmonically modelocked fibre laser by selfphase modulation and spectral filtering,” Electron. Lett. 32(5), 461 (1996). [CrossRef]
5. J. W. Nicholson and D. J. DiGiovannni, “High-repetition-frequency low-noise fiber ring lasers mode-locked with carbon nanotubes,” IEEE Photon. Technol. Lett. 20(24), 2123–2125 (2008). [CrossRef]
6. S. Yamashita, Y. Inoue, K. Hsu, T. Kotake, H. Yaguchi, D. Tanaka, M. Jablonski, and S. Y. Set, “5-GHz pulsed fiber Fabry–Pérot laser mode-locked using carbon nanotubes,” IEEE Photon. Technol. Lett. 17(4), 750–752 (2005). [CrossRef]
7. S. Y. Set, H. Yaguchi, Y. Tanaka, and M. Jablonski, “Ultrafast Fiber Pulsed Lasers Incorporating Carbon Nanotubes,” IEEE J. Sel. Top. Quantum Electron. 10(1), 137–146 (2004). [CrossRef]
8. J. J. McFerran, L. Nenadovic, W. C. Swann, J. B. Schlager, and N. R. Newbury, “A passively mode-locked fiber laser at 1.54 mum with a fundamental repetition frequency reaching 2 GHz,” Opt. Express 15(20), 13155–13166 (2007). [CrossRef] [PubMed]
9. H. Byun, M. Y. Sander, A. Motamedi, H. Shen, G. S. Petrich, L. A. Kolodziejski, E. P. Ippen, and F. X. Kärtner, “Compact, stable 1 GHz femtosecond Er-doped fiber lasers,” Appl. Opt. 49(29), 5577–5582 (2010). [CrossRef] [PubMed]
10. S. Y. Set, C. S. Goh, D. Wang, H. Yaguchi, and S. Yamashita, “Non-synchronous Optical Sampling and Data-Pattern Recovery Using a Repetition-Rate-Tunable Carbon-Nanotube Pulsed Laser,” Jpn. J. Appl. Phys. 47(8), 6809–6811 (2008). [CrossRef]
11. J. Lim, K. Knabe, K. A. Tillman, W. Neely, Y. Wang, R. Amezcua-Correa, F. Couny, P. S. Light, F. Benabid, J. C. Knight, K. L. Corwin, J. W. Nicholson, and B. R. Washburn, “A phase-stabilized carbon nanotube fiber laser frequency comb,” Opt. Express 17(16), 14115–14120 (2009). [CrossRef] [PubMed]
12. K. Kieu, R. J. Jones, and N. Peyghambarian, “High power femtosecond source near 1 micron based on an all-fiber Er-doped mode-locked laser,” Opt. Express 18(20), 21350–21355 (2010). [CrossRef] [PubMed]
13. K. Kieu, R. J. Jones, and N. Peyghambarian, “Generation of Few-Cycle Pulses From an Amplified Carbon Nanotube Mode-Locked Fiber Laser System,” IEEE Photon. Technol. Lett. 22(20), 1521–1523 (2010). [CrossRef]
14. F. Wang, A. G. Rozhin, V. Scardaci, Z. Sun, F. Hennrich, I. H. White, W. I. Milne, and A. C. Ferrari, “Wideband-tuneable, nanotube mode-locked, fibre laser,” Nat. Nanotechnol. 3(12), 738–742 (2008). [CrossRef] [PubMed]
15. S. Kivistö, T. Hakulinen, A. Kaskela, B. Aitchison, D. P. Brown, A. G. Nasibulin, E. I. Kauppinen, A. Härkönen, and O. G. Okhotnikov, “Carbon nanotube films for ultrafast broadband technology,” Opt. Express 17(4), 2358–2363 (2009). [CrossRef] [PubMed]
16. G. Genty, S. Coen, and J. M. Dudley, “Fiber supercontinuum sources (invited),” J. Opt. Soc. Am. B 24(8), 1771–1785 (2007). [CrossRef]
17. T. R. Schibli, K. Minoshima, F.-L. Hong, H. Inaba, A. Onae, H. Matsumoto, I. Hartl, and M. E. Fermann, “Frequency metrology with a turnkey all-fiber system,” Opt. Lett. 29(21), 2467–2469 (2004). [CrossRef] [PubMed]
18. J. E. Townsend, W. L. Barnes, K. P. Jedrzejewski, and S. G. Grubb, “Yb3+ sensitised Er3+ doped silica optical fibre with ultrahigh transfer efficiency and gain,” Electron. Lett. 27(21), 1958–1959 (1991). [CrossRef]
19. P. Polynkin, A. Polynkin, D. Panasenko, N. Peyghambarian, M. Mansuripur, and J. Moloney, “All-fiber passively mode-locked laser oscillator at 1. 5 µm with watts-level average output power and high repetition rate,” Opt. Lett. 31(5), 592–594 (2006). [CrossRef] [PubMed]
20. D. von der Linde, “Characterization of the noise in continuously operating mode-locked lasers,” Appl. Phys. B 39(4), 201–217 (1986). [CrossRef]
