We demonstrate an experimental study on the influence of the parameters of a graphene-based saturable absorber (SA) on the performance of mode-locked Er- and Tm-doped fiber lasers. We have fabricated a set of saturable absorbers with different number of graphene layers: 9, 12, 24, 37 and 48. Each SA was characterized in terms of nonlinear optical parameters (modulation depth, saturation intensity, saturation fluence) and tested in two state-of-the-art, low-power Er- and Tm-doped fiber lasers. Our results show, that in the Er-laser the broadest output spectrum (11 nm) and shortest pulses (345 fs) are generated using 37 layers of graphene in the SA. In case of a Tm-laser, the best performance (737 fs pulses with 5.82 nm bandwidth) was achieved with 24 layers. Additionally, we show that the modulation depth of a 9-layer SA is insufficient to initiate mode-locking in both lasers. This is the first reported comprehensive study on controlling of the parameters of a SA by scaling the number of graphene layers.
© 2015 Optical Society of America
Graphene is the most popular two-dimensional nanomaterial, widely used as saturable absorber (SA) in fiber laser technology. Its unique optical properties, like broad absorption bandwidth and ultrashort recovery time makes it useful for mode-locking of lasers operating at different wavelengths, ranging from 800 to 2500 nm [1,2]. The most exploited wavelength among the fiber laser community is obviously the 1.56 μm region, which covers the 3rd telecommunication window. The first graphene-based lasers were demonstrated in 2009 and they were based on Er-doped fibers [3,4]. After the first demonstration an enormous number of reports devoted to graphene-based Erbium-doped fiber lasers (EDFLs) appeared in the literature [5–10]. Up to date, the shortest pulse generated from a fiber laser utilizing graphene was 168 fs  at 1555 nm wavelength. Despite the excellent nonlinear optical properties of graphene in the mid-infrared region, there were only few reports of mode-locked Thulium-doped fiber lasers (TDFLs) using a graphene-based SA. The first demonstration was reported by M. Zhang et al. . Later, Q. Wang et al. demonstrated TDFL emitting 2.1 ps pulses using liquid-phase graphene . The shortest pulses from a graphene-based Tm-doped fiber laser (603 fs) were generated by our group in  using graphene/PMMA composite grown by chemical vapor deposition (CVD) technique.
Despite the great number of reports on mode-locked, graphene-based fiber lasers, there were no studies carried out on controlling the nonlinear parameters of the SA. In case of liquid-phase exfoliated (LPE) graphene deposited on fiber connectors (e.g. in [9,10]), the number of layers is absolutely randomized and cannot by controlled in any way. During the drying procedure, the solvent evaporates and the graphene flakes arrange themselves randomly. The obtained layer is thus not uniform and might contain areas with multilayer graphene stacks. A study has been carried out on the influence of the thickness of a mechanically exfoliated graphite layer on the performance of an Er-doped laser . However, this study was limited only to an Er-doped fiber laser. Mechanical exfoliation of graphite, similarly to LPE, is a randomized process and the number of layers cannot be controlled. On the other hand, there are methods of optical deposition of graphene on the optical fiber end facet [16,17]. However, there were no studies carried out on scaling the modulation depth of the SA obtained with this method. The influence of the SA parameters on the laser behavior was also not investigated. The CVD-grown graphene seems to be the most convenient for the use in fiber lasers. First of all, the CVD process allows to control the number of layers, which opens the possibility to create multi-layer graphene stacks . The graphene layers might be transferred from the substrate (usually copper or nickel) onto polymers, e.g. poly(methyl methacrylate) (PMMA), forming free-standing foils. The PMMA is a very convenient polymer, since it is almost fully transparent in the optical C- and L-bands . Such foils might be cut into small pieces and placed on the tip of a fiber connector .
Here, we present an experimental study on the nonlinear optical properties of multi-layer graphene-based saturable absorbers. We have fabricated a set of saturable absorbers with different number of graphene layers: 9, 12, 24, 37 and 48. Each SA was carefully characterized in terms of its nonlinear optical parameters at 1560 nm wavelength and tested in two different, state-of-the-art, low-power Er- and Tm-doped fiber lasers. For this study, an all-fiber, all-polarization maintaining (PM) cavity was chosen. The all-PM design makes the laser independent to any external disturbances (movement of the fibers, etc.). Typically, in non-PM mode-locked lasers it is required to adjust the polarization state using a polarization controller (PC) to initiate the mode-locking. During operation, the parameters of the laser (like spectrum shape, bandwidth, central wavelength, etc.) might be changed by tuning of the PC [20, 21]. Thus, in such cavity it would be impossible to make a fair comparison between the saturable absorbers.
2. Graphene fabrication and characterization
The graphene layers used in this study were grown by CVD method on copper substrates using the Aixtron Black Magic Pro deposition system. The details on the fabrication process of multilayer graphene/PMMA stacks were described in our previous work .
The stacks with different number of layers were achieved by repeating the graphene transfer process from copper onto one Cu/graphene base substrate. The graphene transfer process was performed to achieve the desired number of layers so that the final structure was as follows: copper/graphene × N/PMMA. Afterwards, the graphene layers on copper were delaminated and cleaned, so that free-standing PMMA foils with 9, 12, 24, and 37 layers of graphene were obtained. Figure 1 shows an exemplary Raman spectrum taken from the 12-layer graphene sample, using a Renishaw system with 532 nm laser as an excitation source. The spectrum contains pronounced G (1587 cm−1) and 2D (2702 cm−1) bands, which is typical of the sp2 hybridization of carbon. It is very important to mention, that in our case the Raman spectra of multilayer graphene stacks are not “graphite-like”, since our composites cannot be considered as graphite, despite their multilayer structure. Each layer of the composite was grown separately as monolayer graphene, and afterwards stacked with each other. The layers are not bonded, there are no physical interactions between them (like e.g. in multilayer graphene exfoliated from graphite) and the stacking order is undetermined. Hence, it is impossible to determine the number of layers in such structure using Raman spectroscopy.
Figure 2(a) shows the photographs of the fabricated graphene/PMMA foils used in the experiment. The number of layers in the structure was also confirmed by optical transmittance measurement, which is considered as the most reliable method of defining the number of layers in multilayer graphene stacks . In order to fabricate a saturable absorber suitable for using in a fiber laser, a small piece (around 0.8x0.8 mm) of each multilayer graphene/PMMA composite was cut and placed on the end facet of an angle-polished PM fiber connector (FC/APC). The SA containing 48 layers was made by stacking together two pieces of 24-layer foils. The transmission was measured at 1550 nm wavelength, using a pigtailed laser diode as signal source, with optical power set to 1 mW in continuous-wave (CW) mode. The results were plotted in Fig. 2(b) together with the calculated theoretical transmittance of a multilayer stack.
The transmission T of multilayer graphene can be described with a formula proposed by Zhu et al. :23]. However, this formula is only valid for multilayer graphene with defined stacking sequence (e.g. ABA or ABC), which can be achieved in a CVD growth process on e.g. nickel substrate. This formula takes into account the interactions between the nearest layers, which are present only when the graphene layers are properly stacked. In our case, every graphene layer is grown separately on a copper substrate. Thus, the stacking is undetermined and there are no inter-layer interactions. The passage of a laser beam through such a multilayer graphene stack is illustrated in Fig. 3.
According to Nair et al. , the absorption of a single layer is determined by the fine-structure constant and is equal to ~2.29%. The incident power at the second layer is therefore reduced by 2.29%. The total transmittance T(N) of a stack containing N graphene layers can be expressed as:Fig. 2(b). The measured values are in very good correlation with the calculations.
The nonlinear optical parameters of the fabricated SA were measured in a fiberized power-dependent transmission measurement setup, similar to that presented e.g. in , typically used for characterization of fiber-based saturable absorbers. As an excitation source, a 1560 nm pulsed laser was used, with 2 ps pulse duration and 100 MHz repetition rate. The maximum achievable fluence was at the level of 450 μJ/cm2. The measured saturable absorption curves with the indicated observed transmission change ΔT together with theoretical fitting are plotted in Fig. 4. The fit was calculated using the following formula, valid for fast saturable absorbers [25,26]:Fig. 4, that the modulation depth scales with the number of layers. In our experiments, the used pump laser did not allow to fully saturate the samples, thus, we could not directly measure the exact modulation depth. This is why we refer to “transmission change” (ΔT), which is the contrast between the maximum and minimum measured transmission. It starts from 3% for 9 layers, up to 7.5% for 37 layers. However, the observed ΔT of a 48-layer SA is lower than in case of 37 layers. The expected modulation should be larger, but the sample damages at around 450 μJ/cm2. In consequence, not all of the 48 graphene sheets are saturated below the damage threshold and the SA cannot be fully bleached. Thus, using SAs with such large number of layers seems to be unreasonable in low-power soliton fiber lasers.
3. Fiber laser setup
The influence of the graphene-based SA parameters on the behavior of a fiber laser was investigated in two fully fiberized, all-PM Er- and Tm-doped fiber lasers. A general schematic of the laser design is depicted in Fig. 5. Both resonators comprise an active fiber (37 cm of Liekki Er-80-4/125-PM, and 12 cm of Nufern SM-TSF-5/125 in the EDFL and TDFL, respectively), filter-type wavelength division multiplexer (FWDM), an isolator, an 20% output coupler, and the graphene-based saturable absorber. The Er-doped fiber was counter-directionally pumped by a 980 nm laser diode. In order to pump the TDFL, an Er/Yb-doped fiber amplifier seeded by a 1568 nm laser diode was used. The repetition rates of the Er- and Tm-laser were 54.3 and 28 MHz, respectively. The net dispersion of both lasers was anomalous.
All fibers and components used in both lasers were polarization maintaining. The all-fiber, fully-PM design ensures stable and self-starting mode-locked operation, invulnerable to external disturbances. Each saturable absorber, after confirming the number of graphene layers, was spliced into the cavity and tested.
4. Experimental results
In our experiments, the behavior of both lasers was investigated with all fabricated saturable absorbers, which were replaced in the cavity one after the other. Each SA was spliced (not connected using connectors), in order to provide best reliability and repeatability. Additionally, in order to maintain the same repetition frequency of all lasers, we controlled the length of the fibers carefully as possible, and the differences between subsequent setups were very small. The maximum difference in repetition frequency between the Er-lasers was 560 kHz, which corresponds to a length difference of 3.9 cm (about 1% mismatch). In total, 8 parameters of the laser were recorded: the pump power threshold required for mode-locking (Ppump_thr), maximum pump power with stable mode-locking without any continuous-wave (CW) component in the spectrum (Ppump_max), the central wavelength of the emission (λcenter), the half width at half maximum bandwidth of the spectrum (ΔλFWHM), the pulse duration (τpulse), the average output power (Pout), the time-bandwidth product (TBP), the optical efficiency (η), and the repetition frequency (frep).
a) Er-doped fiber laser
In general, the best performance in terms of pulse duration and spectrum bandwidth was achieved with 37 graphene layers in the SA. The laser efficiency were slightly lower than in case of 12 and 24 layers, due to larger non-saturable losses introduced by the SA (Fig. 6(c)). Increasing the number of layers up to 48 does not cause any improvement in the laser performance – the FWHM bandwidth is narrower and the pulse duration is longer by 14 fs compared to 37 layers (Fig. 6(a)). No mode-locking at any pump power was observed with 9 layers of graphene, presumably due to insufficient modulation depth of the SA. The minimum required number of layers to observe the mode-locking was 12 – however, the 12-layered SA did not provide self-starting of the pulsed operation. As previously investigated by U. Keller , insufficient modulation depth is the main reason of the lack of self-starting. With 24, 37 and 48 layers the laser was always operating in the mode-locked regime by itself, just after turning on the pump above the threshold. During operation, the pump power might be changed without losing the mode-locking in a limited range. The stable operation range is quite constant for 24, 37 and 48 layers. Only for 12 layers the pump power might be changed only by very small values (from 33 to 41 mW, see Fig. 6(d)).
The evolution of the optical spectrum shape and the dependency of the pulse duration on the number of graphene layers is depicted in Fig. 7. It can be nicely seen, that the spectrum shifts towards shorter wavelengths. This effect is caused by larger losses introduced by the SA and was observed by other authors before . The spectrum broadens with the increased number of layers up to 37. After this value, the balance between non-saturable losses, saturation intensity and modulation depth is lost and the overall laser performance worsens. This affects also the pulse duration, which can be observed in Fig. 7(b). The shortest, 345-fs pulses were obtained using 37 layers. In the worst case, with 12 layers, the pulse duration is only 406 fs.
b) Tm-doped fiber laser
The same measurements were conducted with the TDFL. The recorded parameters are summarized in Table 2 and plotted as a function of number of graphene layers (N) in Fig. 8. Again, we have tried to maintain the same repetition frequency of all Tm-lasers lasers. The maximum difference in repetition frequency was 510 kHz, which corresponds to a length difference of 13.1 cm (about 1.78% mismatch).
In this case, mode-locking was observed with three SAs (12, 24 and 37 layers). The best performance in terms of pulse duration, spectrum width, output power and efficiency was achieved with 24-layer SA. The laser generated 737 fs pulses with 5.83 nm bandwidth, centered at 1923.3 nm. Increasing or decreasing the number of graphene layers causes worsening of the performance. No mode-locking was obtained with 9 and 48 layers, most likely because of too low modulation depth (in case of 9 layers), or too high attenuation (with 48 layers). Similarly to the Er-laser, the pump power might be changed during operation in a limited range without losing the mode-locking (Fig. 8(d)).
Analogously to the EDFL, increasing the number of graphene layers causes shifting of the optical spectrum towards shorter wavelengths, which can be nicely seen in Fig. 9(a). The TDFL seems to be more sensitive to additional losses in the cavity, because the spectrum shift is much more significant in comparison to the EDFL (almost 33 nm difference in λcenter between 12 and 37 layers). The measured autocorrelation traces are depicted in Fig. 9(b). In all cases the pulses are nearly transform-limited with TBP values close to 0.315. The shortest, 737-fs pulses were obtained using 24 layers. In the worst case (12 layers) the pulse is almost 100 fs longer.
In conclusion, we have presented a comprehensive study on the performance of two mode-locked fiber lasers, depending on the number of graphene layers in the saturable absorber. Our experiments have shown, that the best performance of a 54 MHz repetition rate EDFL is obtained with 37 layers of graphene. Such number provides the best balance between the modulation depth, non-saturable losses and saturation fluence in a typical, low-power Er-doped fiber laser setup. We have shown, that further increasing the number of layers (up to 48) does not improve the performance of the laser, because of too high non-saturable losses and insufficient fluence in the cavity. On the other hand, the minimal number of graphene layers required for mode-locking is 12. No mode-locking was observed with the SA containing 9 layers. In case of the 28 MHz repetition rate Tm-doped fiber laser, the best performance was achieved with 24 layers of graphene in the SA. The mode-locked TDFL generated 737 fs pulses with 5.83 nm bandwidth, centered at 1923.3 nm. Increasing or decreasing the number of layers in the SA degrades the overall laser performance.
The work on Er-doped fiber laser was supported by the National Centre for Research and Development (NCBiR, Poland) under the project entitled: “Ultrafast graphene-based fiber lasers” (UltraGRAPH). The work on Tm-doped fiber laser was supported by the National Science Centre (NCN, Poland) under the project: “Passive mode-locking in dispersion-managed ultrafast Thulium-doped fiber lasers” (decision no. DEC-2013/11/D/ST7/03138). The research on multilayer graphene fabrication, leading to these results has also received funding from the European Union 7th Framework Programme under grant agreement n°604391 (Graphene Flagship).
References and links
1. I. H. Baek, H. W. Lee, S. Bae, B. H. Hong, Y. H. Ahn, D.-I. Yeom, and F. Rotermund, “Efficient mode-locking of sub-70-fs Ti:sapphire laser by graphene saturable absorber,” Appl. Phys. Express 5(3), 032701 (2012). [CrossRef]
3. Q. L. Bao, H. Zhang, Y. Wang, Z. H. Ni, Z. X. Shen, K. P. Loh, and D. Y. Tang, “Atomic layer graphene as saturable absorber for ultrafast pulsed laser,” Adv. Funct. Mater. 19(19), 3077–3083 (2009). [CrossRef]
4. T. Hasan, Z. Sun, F. Wang, F. Bonaccorso, P. H. Tan, A. G. Rozhin, and A. C. Ferrari, “Nanotube–Polymer Composites for Ultrafast Photonics,” Adv. Mater. 21(38â€“39), 3874–3899 (2009). [CrossRef]
5. Y. M. Chang, H. Kim, J. H. Lee, and Y. Song, “Multilayered graphene efficiently formed by mechanical exfoliation for nonlinear saturable absorbers in fiber mode-locked lasers,” Appl. Phys. Lett. 97(21), 211102 (2010). [CrossRef]
6. N. H. Park, H. Jeong, S. Y. Choi, M. H. Kim, F. Rotermund, and D.-I. Yeom, “Monolayer graphene saturable absorbers with strongly enhanced evanescent-field interaction for ultrafast fiber laser mode-locking,” Opt. Express 23(15), 19806–19812 (2015). [CrossRef] [PubMed]
7. Y. Meng, A. Niang, K. Guesmi, M. Salhi, and F. Sanchez, “1.61 μm high-order passive harmonic mode locking in a fiber laser based on graphene saturable absorber,” Opt. Express 22(24), 29921–29926 (2014). [CrossRef] [PubMed]
8. Z. Sun, T. Hasan, F. Torrisi, D. Popa, G. Privitera, F. Wang, F. Bonaccorso, D. M. Basko, and A. C. Ferrari, “Graphene mode-locked ultrafast laser,” ACS Nano 4(2), 803–810 (2010). [CrossRef] [PubMed]
9. D. Popa, Z. Sun, F. Torrisi, T. Hasan, F. Wang, and A. C. Ferrari, “Sub 200 fs pulse generation from a graphene mode-locked fiber laser,” Appl. Phys. Lett. 97(20), 203106 (2010). [CrossRef]
10. F. Bonaccorso and Z. Sun, “Solution processing of graphene, topological insulators and other 2d crystals for ultrafast photonics,” Opt. Mater. Express 4(1), 63–78 (2014). [CrossRef]
11. J. Tarka, G. Sobon, J. Boguslawski, J. Sotor, J. Jagiello, M. Aksienionek, L. Lipinska, M. Zdrojek, J. Judek, and K. M. Abramski, “168 fs pulse generation from graphene-chitosan mode-locked fiber laser,” Opt. Mater. Express 4(10), 1981–1986 (2014). [CrossRef]
12. M. Zhang, E. J. R. Kelleher, F. Torrisi, Z. Sun, T. Hasan, D. Popa, F. Wang, A. C. Ferrari, S. V. Popov, and J. R. Taylor, “Tm-doped fiber laser mode-locked by graphene-polymer composite,” Opt. Express 20(22), 25077–25084 (2012). [CrossRef] [PubMed]
13. Q. Wang, T. Chen, B. Zhang, M. Li, Y. Lu, and K. P. Chen, “All-fiber passively mode-locked thulium-doped fiber ring laser using optically deposited graphene saturable absorbers,” Appl. Phys. Lett. 102(13), 131117 (2013). [CrossRef]
14. G. Sobon, J. Sotor, I. Pasternak, A. Krajewska, W. Strupinski, and K. M. Abramski, “All-polarization maintaining, graphene-based femtosecond Tm-doped all-fiber laser,” Opt. Express 23(7), 9339–9346 (2015). [CrossRef] [PubMed]
15. A. Martinez, K. Fuse, and S. Yamashita, “Mechanical exfoliation of graphene for the passive mode-locking of fiber lasers,” Appl. Phys. Lett. 99(12), 121107 (2011). [CrossRef]
16. Z.-C. Luo, W.-J. Cao, A.-P. Luo, and W.-C. Xu, “Optical Deposition of Graphene Saturable Absorber Integrated in a Fiber Laser Using a Slot Collimator for Passive Mode-Locking,” Appl. Phys. Express 5(5), 055103 (2012). [CrossRef]
17. A. Martinez, K. Fuse, B. Xu, and S. Yamashita, “Optical deposition of graphene and carbon nanotubes in a fiber ferrule for passive mode-locked lasing,” Opt. Express 18(22), 23054–23061 (2010). [CrossRef] [PubMed]
18. P. L. Huang, S.-C. Lin, C.-Y. Yeh, H.-H. Kuo, S.-H. Huang, G.-R. Lin, L.-J. Li, C.-Y. Su, and W.-H. Cheng, “Stable mode-locked fiber laser based on CVD fabricated graphene saturable absorber,” Opt. Express 20(3), 2460–2465 (2012). [CrossRef] [PubMed]
19. J. Sotor, G. Sobon, I. Pasternak, A. Krajewska, W. Strupinski, and K. M. Abramski, “Simultaneous mode-locking at 1565 nm and 1944 nm in fiber laser based on common graphene saturable absorber,” Opt. Express 21(16), 18994–19002 (2013). [CrossRef] [PubMed]
20. H. Zhang, D. Y. Tang, L. M. Zhao, Q. L. Bao, K. P. Loh, B. Lin, and S. C. Tjin, “Compact graphene mode‐locked wavelength‐tunable erbium‐doped fiber lasers: from all anomalous dispersion to all normal dispersion,” Laser Phys. Lett. 7(8), 591–596 (2010). [CrossRef]
21. S. Huang, Y. Wang, P. Yan, J. Zhao, H. Li, and R. Lin, “Tunable and switchable multi-wavelength dissipative soliton generation in a graphene oxide mode-locked Yb-doped fiber laser,” Opt. Express 22(10), 11417–11426 (2014). [CrossRef] [PubMed]
22. S.-E. Zhu, S. Yuan, and G. C. A. M. Janssen, “Optical transmittance of multilayer graphene,” Europhys. Lett. 108(1), 17007 (2014). [CrossRef]
23. R. R. Nair, P. Blake, A. N. Grigorenko, K. S. Novoselov, T. J. Booth, T. Stauber, N. M. R. Peres, and A. K. Geim, “Fine Structure Constant Defines Visual Transparency of Graphene,” Science 320(5881), 1308 (2008). [CrossRef] [PubMed]
24. G. Sobon, “Mode-locking of fiber lasers using novel two-dimensional nanomaterials: graphene and topological insulators [Invited],” Photon. Res. 3(2), A56–A63 (2015). [CrossRef]
25. C. A. Zaugg, Z. Sun, V. J. Wittwer, D. Popa, S. Milana, T. S. Kulmala, R. S. Sundaram, M. Mangold, O. D. Sieber, M. Golling, Y. Lee, J. H. Ahn, A. C. Ferrari, and U. Keller, “Ultrafast and widely tuneable vertical-external-cavity surface-emitting laser, mode-locked by a graphene-integrated distributed Bragg reflector,” Opt. Express 21(25), 31548–31559 (2013). [CrossRef] [PubMed]
26. T. R. Schibli, E. R. Thoen, F. X. Kärtner, and E. P. Ippen, “Suppression of Q-switched mode locking and break-up into multiple pulses by inverse saturable absorption,” Appl. Phys. B 70(S1), S41–S49 (2000). [CrossRef]
27. U. Keller, K. Weingarten, F. X. Kartner, D. Kopf, B. Braun, I. Jung, R. Fluck, C. Honninger, N. Matuschek, and J. Aus der Au, “Semiconductor saturable absorber mirrors (SESAM’s) for femtosecond to nanosecond pulse generation in solid-state lasers,” IEEE J. Sel. Top. Quantum Electron. 2(3), 435–453 (1996). [CrossRef]
28. Y. Meng, M. Salhi, A. Niang, K. Guesmi, G. Semaan, and F. Sanchez, “Mode-locked Er:Yb-doped double-clad fiber laser with 75-nm tuning range,” Opt. Lett. 40(7), 1153–1156 (2015). [CrossRef] [PubMed]