Laser Mode Locking Using a Saturable Absorber ... - IEEE Xplore

JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 22, NO. 1, JANUARY 2004

51

Laser Mode Locking Using a Saturable Absorber Incorporating Carbon Nanotubes Sze Y. Set, Member, IEEE, Hiroshi Yaguchi, Yuichi Tanaka, and Mark Jablonski, Member, IEEE

Abstract—This paper describes a new class of saturable absorber device based on single-wall carbon nanotube (SWNT)—the saturable absorber incorporating nano tube (SAINT). The device possesses ultrafast optical properties comparable to that of the industrial standard semiconductor saturable absorber mirror (SESAM). Passively mode-locked picosecond fiber lasers in different configurations are demonstrated using SAINTs as mode lockers. This is the first demonstration of optical pulsed lasers based on the carbon nanotube technology, and the first practical application of carbon nanotubes in the field of applied optics. Index Terms—Carbon nanotube, mode locking, picosecond fiber laser, pulsed laser, saturable absorber, soliton.

I. INTRODUCTION

P

ASSIVELY mode-locked fiber lasers are among the best optical pulsed sources available today due to their simplicity and their ability to generate transform-limited optical pulses in the picosecond and subpicosecond regimes [1], [2]. A key device in a passively mode-locked fiber laser is a nonlinear element, which has an intensity-dependent response to favor optical pulse formation over continuous-wave (CW) lasing [3]. This can usually be a saturable absorber, such as a semiconductor saturable absorber [4], [5], or an “effective” saturable absorber, such as a nonlinear polarization switch [6], a nonlinear optical loop mirror [7] (NOLM) and its variants [8], [9]. Among these mode lockers, the semiconductor-based multiquantum-well (MQW) device, commonly referred as the semiconductor saturable absorber mirror (SESAM) [4], has become the main device used in most commercial passively mode-locked fiber lasers. MQW-based saturable absorbers such as SESAMs require relatively complex and costly clean-room-based fabrication systems such as MOCVD or MBE to grow. These devices have a typical recovery lifetime of around a few nanoseconds. The recovery lifetime can be reduced to 5–20 ps by creating defects in the structure using techniques such as high-energy heavy-ion implantation, or a suboptimal low-temperature growth condition. These defects serve as carrier trapping or recombination centers thereby improving the recovery lifetime. Hermetic packaging may also be required for long-term environmental stability. Furthermore, SESAM has a rather low optical damage threshold, therefore requiring special design such as an antiresonant cavity [4], or an expanded input beam spot size, and a cooling heat sink. Manuscript received July 19, 2003; revised October 22, 2003. The authors are with Alnair Labs, Kawaguchi City, Saitama-ken, Japan 332–0015. Digital Object Identifier 10.1109/JLT.2003.822205

The saturable absorption, or optical induced tranparency property of single-wall carbon nanotubes (SWNTs) has been reported in [10] and [11], where their potential application as optical switches was proposed. The recovery time of SWNT using a pump-probe experiment was measured to be [10], [12]. Recently, we have proposed and demonstrated a SWNT-based saturable absorber we call “saturable absorber incorporating nanotube” (SAINT), for optical noise suppression of ultrafast optical pulses in the picosecond regime [13]. Subsequently, the first passively mode-locked fiber lasers using SAINT as a mode locker was demonstrated [14], as well as the first Q-switched laser using SAINT as a Q-switch [15]. In this paper, we present and discuss picosecond fiber lasers mode locked using a SAINT. In Section II, the fabrication and optical properties of carbon nanotubes and SAINT modules are discussed. Experimental results of two different fiber pulsed lasers, in a ring- and a linear-cavity configurations, using a SAINT in transmission and reflection mode, respectively, are presented in Section III. The analyses of these lasers are discussed in Section IV. These results confirm the potential of a SAINT for laser mode-locking applications with performance comparable to that of a conventional ultrafast mode locker.

II. FABRICATION AND OPTICAL PROPERTIES OF SAINT The carbon nanotubes were fabricated using the laser ablation technique [16], which gives a higher purity and controllable tube-diameter distrubution compared to other growth techniques. High energy pulses generated from a frequency-doubled Nd:YAG laser were used to ablate a carbon target placed in a quartz tube flowed with Ar gas. The quartz tube was heated in a high temperature electric furnace. By careful control of the furnace temperature and the adoption of specific catalysts with appropriate relative concentrations in the carbon target, diameter-selective SWNTs can be grown [17]. Through a series of purifying and annealing processes, high purity SWNTs (70%–90%) can be obtained. The atomic-force microscope (AFM) image of a typical nanotube layer in the SAINT shows meshes of SWNT ropes with a (Fig. 1). When observed under a transbundle diameter mission electron microscope (TEM), each individual SWNT strain can be resolved in each nanotube bundles, as shown in Fig. 2, confirming that the sample indeed contains high-quality SWNTs with diameters around 1 nm. The mean nanotube diameter of the samples can be further characterized using a resonant Raman spectroscopy and studying the radial breathing mode (RBM) [18].

0733-8724/04$20.00 © 2004 IEEE

52


Fig. 3. Schematics of SAINT in two different configurations. (a) T-SAINT. (b) R-SAINT.

Fig. 1.

AFM images of the SWNT sample used in the SAINT module.

Fig. 2. TEM image of the SWNT sample showing nanotube bundles.

Two SAINT module configurations are illustrated in the Fig. 3. The transmission-mode SAINT (T-SAINT) is shown in Fig. 3(a). It has a thin layer of purified SWNTs sandwiched between two quartz substrates, which are antireflection (AR) coated on the outer surfaces. The reflective SAINT (R-SAINT), as shown in Fig. 3(b), has a layer of SWNT coated on a highly reflective (HR) mirror, which enables the device to operate in a reflective mode. The SWNT’s layers are typically deposited . with a thickness of The saturable absorption properties of a SAINT module, for example its linear absorption level, threshold power, and center wavelength, can be controlled by varying the thickness and density of the SWNT layer, and the mean diameter of the constituent carbon nanotubes. Since the orientations of the nanotubes are randomly distributed, the optical response of the SAINT device is independent of the polarization of the input light. In the two different SWNT samples, the radial breathing modes (RBMs) measured on their respective Raman specand , troscopy indicated mean diameters of corresponded well to the measured optical absorption peaks at and , respectively, when the Coulomb-interaction blue-shift is accounted for (Fig. 4). It is clear that the operating wavelength of the SWNT is dependent on nanotube diameter, which in turn can be controlled by the nanotube growth conditions, such as furnace temperature and the type of metal catalysts used [17], [18]. The absorption characteristics of one of the SAINT samples, at different incident peak power intensity is shown in Fig. 5.

Fig. 4. Absorption spectra of two SAINT samples with different mean nanotube diameters.

Fig. 5. Absorption characteristics of a SAINT at different incident peak power intensity.

The 10% absorption saturation on-set point of a typical SAINT is measured to be at a peak power intensity of 5.7 , which is relatively high when compared to other conventional saturable absorber. This is particularly favorable for laser mode-locking application, since the power intensity inside a laser cavity is usually very high. Conventional MQW-based saturable absorber, when placed inside a laser cavity, will require measures such as an expanded incident beam spot-size, an AR Fabry–Perot configuration

SET et al.: LASER MODE LOCKING USING A SATURABLE ABSORBER INCORPORATING CARBON NANOTUBES

53

Fig. 6. Experimental setup of a mode-locked fiber ring laser using a T-SAINT.

and/or a water-cooled heat sink, to avoid optical damages caused by intense laser power.

Fig. 7. Waveform of the output pulses from the ring laser at 6.1-MHz repetition rate.

III. LASER MODE LOCKING A. Ring-Cavity Laser Configuration The experimental setup of a fiber ring-cavity laser mode locked using a SAINT is shown in Fig. 6. A 10-m length of erbium-doped fiber (EDF), backward pumped by a 980-nm laser diode (LD) is used as the laser gain medium. Two optical isolators are inserted to prevent backreflection in the cavity and to ensure unidirectional operation. The output light from the EDF is launched through a fiber collimator and a focusing aspherical lens onto the SAINT. The output light from the SAINT is collected and launched back into the fiber cavity via another set of matching aspherical lens and collimator. An angle-tunable thin-film band-pass filter with 7 nm bandwidth is inserted for wavelength tuning. A single-mode fiber (SMF) with a length of 12 m is employed for soliton pulse shaping and dispersion management of the laser cavity. The output of the laser is tapped through a 95% port of a fiber coupler, whereas the other 5% port is used to feed back into the cavity. At a pump power of around 18 mW, the laser starts to mode lock in multiple-pulse mode, where multiple pulses are present within the cavity round-trip time. After that, the pump power can be reduced to a level around 14 mW, and the laser will maintain pulsing in single-pulse mode at a fundamental repetition rate of 6.1 MHz (Fig. 7). when The laser output average optical power is operated in a fundamental mode with a single pulse in a roundtrip. At higher pump powers, the laser will operate in multiplepulse mode resulting in a higher-harmonic repetition rate at the multiple of the fundamental round-trip frequency. The output spectrum of the ring laser is shown in Fig. 8, which , well fitted by a hyperhas a 3-dB spectral width of profile. The SHG autocorrelation bolic secant squared traces of the output pulses, in linear- and logarithmic-scales are measured as shown in Fig. 9. The autocorrelation traces pulse profile [Fig. 9(b)], indiare also well fitted by a cating that soliton pulses are generated. The inferred full-width half-maximum (FWHM) width from the autocorrelation trace is , whereas the resulting time-bandwidth estimated to be product (TBP) of 0.52 indicates that the pulses are chirped,

Fig. 8. Ouput spectrum of the ring laser. (Dashed curve: hyperbolic secant fitting; dotted curve: Gaussian fitting.)

compared to the transform-limited TBP of 0.315 for chirp-free pulses. An FWHM pulsewidth of is expected with proper dispersion compensation. The pulse pedestals shown in Fig. 9(b) are possibly due to imperfect soliton pulse shaping in the length of SMF with a slightly higher nonlinearity over the balancing dispersion. When the SAINT is removed from the laser cavity, it is not possible to mode lock the laser even when the pump power is raised to 100 mW. It is evident that the SAINT provides the mechanism required to initiate and sustain mode-locking operation, particularly with a rather low mode-locking threshold pump power. B. Linear-Cavity Laser Configuration The schematics of a linear-cavity laser mode locked using a reflective SAINT is shown in Fig. 10. All the components have their usual functions as described earlier. At one end of the cavity, a Faraday mirror is used to compensate for birefringence in the cavity. An R-SAINT is employed at the other end of the cavity. Two lenses are employed to focus the incident beam into

54


Fig. 10. Experimental setup of a mode-locked linear cavity laser using an R-SAINT.

Fig. 11. Output spectrum of the linear-cavity laser. (Dashed curve: hyperbolic secant fitting; dotted curve: Gaussian fitting). Fig. 9. SHG autocorrelation traces of the ring laser output pulses. (a) In linear scale. (b) In log scale, with secant hyperbolic (dashed curve) and Gaussian (solid curve) fittings.

a tiny spot on the R-SAINT. In order to maximize the gain bandwidth, an optical band-pass filter is not used. The laser center wavelength is defined by the EDFs gain profile. Pulse shaping SMF is not inserted in the cavity intentionally, to avoid soliton pulse-shaping effects. The output pulses of the linear laser operate at a fundamental cavity repetition rate of 9.85 MHz. The output spectrum of the pulses is shown in Fig. 11, with a 3-dB spectral width of 13.6 nm. The spectral shape is fitted well with a Gaussian profile as expected, due to the absence of soliton pulse shaping. The laser is mode locked solely by the saturable absorbing effect provided by the SAINT. Fig. 12 shows the SHG autocorrelation traces of the modelocked pulses in linear and logarithmic scales. The autocorrelation trace was fitted well with Gaussian profile as shown in Fig. 12(b). The inferred FWHM width is 318 fs, assuming a Gaussian pulse profile, with a TBP of 0.54 (unchirped transform-limited TBP for Gaussian pulses is 0.441). An unchirped FWHM pulsewidth of 258 fs is expected if the chirp induced by the output SMF (connected to the output isolator) is com, when pensated. The output average power is around pumped with 25 mW of pump power at 980 nm.

Since both the ring- and the linear-cavity lasers are operated in single-pulse mode, apart from a slow temperature-induce drift in repetition rate, we do not expect any significant jitter. IV. ANALYSIS A. Ring-Cavity Laser Analysis Here, important parameters of the ring laser are discussed and calculated. Based on the measured average output power , the average power after the SAINT module was of when the losses due to the output estimated to be coupler and isolator were accounted for . This average power for the peak together with a repetition rate of 6.1 MHz gives as a function of the FWHM pulsewidth in (1) as power and follows: (1) By applying (1) and the condition for a first-order soliton (2) a pulsewidth of 0.69 ps is calculated, where a dispersion coefficient of 20.4 , and a nonlinear coefficient of have been assumed. Using the measured output 1.8

SET et al.: LASER MODE LOCKING USING A SATURABLE ABSORBER INCORPORATING CARBON NANOTUBES

55

where is the Gaussian half width. Starting from (5), a simple expression for Gaussian pulse energy can be expressed as (6) is used for both soliton and Note that the same symbol Gaussian half-width, but can be easily distinguished from its context, and should not be confused with each other. From (6), pulse energy of 85.9 picoJoules is calculated assuming an output pulsewidth of 318 fs and peak output power of 253.5 W. V. CONCLUSION In conclusion, optical pulsed lasers employing a SAINT have been demonstrated. The SAINT device offers several , polarizaadvantages such as ultrafast recovery time tion-insensitive operation, exceptionally high optical damage threshold, mechanically and environmentally robust, chemical stability, and the ability to operate both in transmission, reflection and bidirectional modes. Moreover, the fabrication cost and complexity of SAINT devices are potentially much lower than that of the conventional MQW-based devices. The property of the device can be controlled simply by varying the thickness and density of the nanotube layer, as well as selecting nanotubes with diameters matching to the operating wavelength. A wideband operation can also be achieved by mixing together nanotubes with different diameters. This technology is expected to greatly impact future pulsed laser design and development. Furthermore, a SAINT can also be used in various ultrafast photonic applications such as optical noise suppression and pulse shaping in a passive 2R regeneration system. Fig. 12. SHG autocorrelation traces of the linear-cavity laser output pulses. (a) In linear scale. (b) In log scale. (Dashed curve: hyperbolic secant fitting; dotted curve: Gaussian fitting.)

spectral bandwidth of 3.7 nm, a time bandwidth product of 0.319 is calculated, which is close to 0.315, the time bandwidth product of a transform limited soliton pulse. Starting from the expression for soliton energy

ACKNOWLEDGMENT The authors would like to thank H. Kataura and Y. Achiba for fabricating the purified SWNT samples, K. Kikuchi for fruitful discussions, and M. Tokumoto, Y. Sakakibara, and A. Rozhin for sending us our first SWNT sample. REFERENCES

(3) where is the soliton pulsewidth, a simple expression for the soliton energy can be written as (4) Using pulsewidth and the peak power of 84.8 W derived earlier, a pulse energy of 58.0 picoJoules is calculated from (4). B. Linear-Cavity Laser Analysis The Gaussian pulse energy, can be expressed as (5)

[1] M. E. Fermann, “Ultrashort-pulse sources based on single-mode rareearth-doped fibers,” J. Appl. Phys. B, vol. B58, pp. 197–209, 1994. [2] I. N Duling III, Ed., Compact Sources for Ultrashort Pulses. Cambridge, MA: Cambridge University Press, 1995. [3] H. A. Haus, “Mode-locking of lasers,” IEEE J. Select. Topics Quantum Electron., vol. 6, pp. 1173–1185, 2000. [4] U. Keller, D. Miller, G. Boyd, T. Chiu, J. Ferguson, and M. Asorn, “Solid-state low-loss intracavity saturable absorber for Nd:YLF lasers: An antiresonant semiconductor Fabry-Perot saturable absorber,” Opt. Lett., vol. 17, no. 7, pp. 505–507, 1992. [5] E. A. De Souza, C. E. Soccolich, W. Pleibel, R. H. Stolen, J. R. Simpson, and D. J. DiGiovanni, “Saturable absorber modelocked polarization maintaining erbium-doped fiber laser,” Electron. Lett., vol. 29, pp. 447–449, 1993. [6] V. J. Matsas, T. P. Newson, D. J. Richardson, and D. J. Payne, “Selfstarting passively mode-locked fiber ring soliton laser exploiting nonlinear polarization rotation,” Electron. Lett., vol. 28, pp. 1391–1393, 1992. [7] N. J. Doran and D. Wood, “Non-linear optical loop mirror,” Opt. Lett., vol. 14, pp. 56–58, 1988. [8] M. E. Fermann, F. Haberl, M. Hofer, and H. Hochreither, “Nonlinear amplifying loop mirror,” Opt. Lett., vol. 15, pp. 752–754, 1990.

56


[9] W. S. Wong, S. Namiki, M. Margalit, H. A. Haus, and E. P. Ippen, “Selfswitching of optical pulses in dispersion-imbalanced nonlinear loop mirrors,” Opt. Lett., vol. 22, pp. 1150–1152, 1997. [10] Y.-C. Chen, N. R. Raravikar, Y. P. Zhao, L. S. Schadler, P. M. Ajayan, T. M. Lu, G. C. Wang, and X. C. Zhang, “Ultrafast optical switching properties of single-wall carbon nanotube polymer composites at 1.55 ,” Appl. Phys. Lett., vol. 81, pp. 975–977, 2002. [11] Y. Sakakibara, S. Tatsuura, H. Kataura, M. Tokumoto, and Y. Achiba, “Near-infrared saturable absorption of single-wall carbon nanotubes prepared by laser ablation method,” Japan. J. Appl. Phys., vol. 42, pp. 494–496, 2003. [12] S. Tatsuura, M. Furuki, Y. Sato, I. Iwasa, M. Tian, and H. Mitsu, “Semiconductor carbon nanotubes as ultrafast switching materials for optical telecommunications,” Adv. Mater., vol. 15, pp. 534–537, 2003. [13] S. Y. Set, H. Yaguchi, M. Jablonski, Y. Tanaka, Y. Sakakibara, A. Rozhin, M. Tokumoto, H. Kataura, Y. Achiba, and K. Kikuchi, “A noise suppressing saturable absorber at 1550 nm based on carbon nanotube technology,” in Proc. Optical Fiber Communication Conf.’03, Atlanta, GA, 2003, paper FL2. [14] S. Y. Set, H. Yaguchi, Y. Tanaka, M. Jablonski, Y. Sakakibara, A. Rozhin, M. Tokumoto, H. Kataura, Y. Achiba, and K. Kikuchi, “Mode-locked fiber lasers based on a saturable absorber incorporating carbon nanotubes,” in Proc. Optical Fiber Communication Conf.’03, Atlanta, GA, 2003, paper PD44. [15] S. Y. Set, H. Yaguchi, Y. Tanaka, M. Jablonski, Y. Sakakibara, M. Tokumoto, H. Kataura, Y. Achiba, K. Kikuchi, and S. Yamashita, “A dualregime mode-locked/Q-switched laser using a saturable absorber incorporating carbon nanotubes (SAINT),” in Proc. CLEO’03, Baltimore, MD, 2003, paper CThPDA9. [16] A. Thess, R. Lee, P. Nikolaev, H. Dai, P. Petit, J. Robert, C. Xu, Y. H. Lee, S. G. Kim, D. T. Colbert, G. Scuseria, D. Tománek, J. E. Fischer, and R. E. Smalley, “Crystalline ropes of metallic carbon nanotubes,” Science, vol. 273, pp. 483–487, 1996. [17] H. Kataura, Y. Kumazawa, Y. Maniwa, Y. Ohtsuka, R. Sen, S. Suzuki, and Y. Achiba, “Diameter control of single-walled carbon nanotubes,” Carbon, vol. 38, pp. 1691–1697, 2000. [18] H. Kataura, Y. Kumazawa, Y. Maniwa, I. Umezu, S. Suzuki, Y. Ohtsuka, and Y. Achiba, “Optical properties of single-wall carbon nanotubes,” Synth. Met., vol. 103, pp. 2555–2558, 1999.

m

Sze Y. Set (M’99) received the B.Eng. degree with first-class honors in electronics engineering and the Ph.D. degree in electronics engineering and optical fiber communications from the Optoelectronics Research Centre (ORC), Southampton University, U.K., in 1993 and 1998, respectively. During 1996–1998, he was a Research Assistant with the ORC, and was leading a European Union research project, ACTS ’MIDAS’ which resulted in a first successful 40 Gb/s transmission field trial using midspan spectral inversion for dispersion compensation. In 1998, he was awarded a Postdoctoral Research Fellowship from the Japan Society for the Promotion of Science (JSPS) to work at the Research Center for Advanced Science and Technology (RCAST), The University of Tokyo, Japan. He was a Consultant and Senior R&D Engineer with Micron Optics, Inc., Atlanta, GA, during 2001–2002. Currently, he is the R&D General Manager with Alnair Laboratories Corporation, Japan. He has contributed to four patents, and more than 60 conference and journal publications in the areas of short pulse transmission, dispersion compensation, nonlinear optical devices, tunable fiber Bragg gratings, fiber lasers, and nanotube photonics. Dr. Set is the recipient of the Best Paper Awards in OECC’02, two E.E. Zepler prizes, the G.D. Sims prize, and the ORS Awards. He is also a Member of the IEE.

Hiroshi Yaguchi was born in 1976. He received the B.S. degree in mechanical engineering and the M.S. degree in precision system engineering, both from the Tokyo Denki University, Japan, in 1998 and 2000, respectively. He then joined Oyokoden Labs as an engineer working on precision packaging of advanced thin-film devices. Currently, he is an R&D Engineer with Alnair Labs Corp., Japan, where he conducts research and development in thin-film dispersion compensator packaging, dispersion measurement systems, and short-pulse fiber lasers.

Yuichi Tanaka, photograph and biography not available at the time of publication.

Mark K. Jablonski (M’00) was born in Manhattan, New York, NY. He received the B.S. and M.S. degrees in electrical engineering and computer science from the Massachusetts Institute od Technology, Cambridge, in 1986 and 1992, respectively. After being awarded a Monbusho scholarship and a STAR fellowship, he received the Ph.D. degree in electrical engineering in 2000 from the University of Tokyo. He was a Research Engineer with Varian Beverly Microwave Division from 1986 to 1994 working on microwave tubes such as magnetrons and cross-field amplifiers. From 2000 to 2001 he was a Project Director with Oyokoden Labs, Japan, working on thin-film based devices. He has been serving as the Chief Research Officer for Alnair Labs Corporation, Japan, a start-up company he cofounded in 2001, focusing on optical devices, such as thin-film-based dispersion compensators, fiber lasers, and specialized fiber amplifiers. He has contributed to 14 patents and more than 25 technical papers. Dr. Jablonski is a Member of the OSA, the IEE, SPIE, Sigma–Xi, and Zeta Psi.