We describe a 1.55 μm hydrogen cyanide (HCN) optical frequency and repetition rate stabilized mode-locked fiber laser, where the optical frequency was locked to the P(10) HCN absorption line and the repetition rate was locked to 9.95328 GHz by using a microwave phase-locked loop. The optical frequency stability of the laser reached 5 x 10−11 with an integration time τ of 1 s. With a bidirectional pumping scheme, the laser output power reached 64.6 mW. To obtain a short pulse train, the average dispersion in the cavity was managed so that it was zero around 1.55 μm, resulting in a 0.95 ps pulse train. In addition, the stabilization of the optical frequency and the repetition rate, meant that the entire spectral profile remained the same for 24 hours.
© 2016 Optical Society of America
Ultrashort optical pulse sources with repetition rates of a few tens of GHz have been receiving a lot of attention with respect to their application to ultrahigh-speed optical communication [1–3] and optical metrology . The optical frequency stabilization of a laser is particularly important in terms of applying it to coherent optical pulse transmission [5,6], where both the amplitude and phase of the optical pulse are used to encode information. The one-octave method is well known as an optical frequency stabilization technique for mode-locked lasers . However, it is difficult to apply the method to a high repetition rate mode-locked laser because the peak power is insufficient to generate a one octave spectrum. On the other hand, we have already proposed a new technique designed to simultaneously stabilize the repetition rate and optical frequency of a laser using regenerative mode-locking . With this technique, we demonstrated experimentally the first simultaneous stabilization of the optical frequency and repetition rate of a 40 GHz mode-locked fiber laser, in which we used an acetylene (C2H2) gas cell as a frequency reference . Since the C2H2 gas has large absorption lines in the shorter wavelength region in the C band, the center wavelength of the laser was set at 1.54 μm, where the EDF gain is relatively small. This was a factor that limited the laser output power.
Recently, we reported a mode-locked fiber laser with a stabilized hydrogen cyanide (HCN) optical frequency and a stabilized 10 GHz repetition-rate . By using HCN instead of C2H2 absorption lines, the center wavelength of the laser oscillation can be shifted to 1.55 μm, where the EDF has a flat high gain. As a result, we obtained a laser output power of as high as 60 mW. Ref , describes the basic output characteristics of the laser. In this paper, important improvements as regards increasing the output power and shortening the pulse width of the laser are newly described in detail. Furthermore, the long-term stability of the laser spectrum is discussed. By simultaneously stabilizing the optical frequency and repetition rate, the spectral profile of the laser remained unchanged during 24 hours of continuous operation.
2. 10 GHz regeneratively mode-locked fiber laser
Figure 1(a) shows the configuration of our mode-locked fiber laser, where the thick and thin lines are optical and electrical paths, respectively. The repetition rate of the laser is locked to 9.95328 GHz, which is the OC-192 standard, for digital coherent optical pulse transmission. All the fibers in the cavity are polarization maintaining to prevent polarization fluctuation. The cavity length is 11 m, which corresponds to a cavity mode spacing of 18 MHz. A 3 m-long polarization-maintaining erbium-doped fiber (PM-EDF) is used as the laser gain medium and part of the EDF is wound onto a PZT to change the cavity length, thus enabling the optical frequency to be tuned by applying a voltage signal. A bidirectional pumping scheme with 1.48 μm InGaAsP laser diodes is adopted to increase the output power. Furthermore, to obtain a shorter pulse train, a lithium niobate (LN) phase modulator is employed as a mode locker. A 4.5 m-long dispersion compensation fiber (DCF) is also installed to adjust the average dispersion of the cavity to zero at 1.55 μm as shown in Fig. 1(b), where the average dispersion is estimated from the relationship between the repetition rate and the center wavelength of the laser oscillation . We install an optical etalon with a finesse of 400 and keep the resonance peak of the etalon at the longitudinal-mode frequency of the laser by controlling the etalon temperature . This allows us to realize mode-hop-free laser operation in the harmonic mode locking . The transmittance characteristics of the etalon are shown in Fig. 2. The full-width at half maximum (FWHM) of the pass band and the insertion loss of the etalon are 25 MHz and 3 dB, respectively. Here, the free spectral range (FSR) of the etalon is adjusted exactly to the laser repetition rate of 9.95328 GHz by controlling the etalon temperature. By using the etalon, we can completely suppress the supermodes of the laser.
To obtain a sinusoidal harmonic beat signal between the longitudinal modes for regenerative mode locking, part of the output beam is coupled into a clock extraction circuit consisting of a high-speed photodetector, a 9.95328 GHz high-Q dielectric filter (Q~1400), and a high-gain electrical amplifier. The beat signal is amplified and then fed back to the LN phase modulator in the cavity, resulting in regenerative mode locking . The repetition rate of the laser is stabilized by changing the microwave phase delay in the regenerative feedback loop. With this technique, the laser repetition rate can be controlled without disturbing the optical frequency .
To clarify the pulse shortening effect realized by dispersion management in the cavity, we numerically evaluated the transient pulse evolution in the laser cavity from the ASE noise as an initial condition by using the split-step Fourier method (SSFM) with an FFT size of 1024 and a step size of 0.02 m. Figure 3 shows a numerical result for the transient pulse evolution in the laser cavity. A steady-state pulse is obtained within 300 round trips. Figure 4 shows a numerical result for the relationship between the spectral width of the laser output and theaverage dispersion in the laser cavity. Here, the average dispersion without a DCF corresponds to 14.7 ps/nm/km. By compensating for the average dispersion within +/− 1 ps/nm/km with a DCF, the spectral width was broadened from 1.7 to 3.8 nm. Figures 5(a-1), 5(a-2) and 5(b-1), 5(b-2) show numerical results for the pulse waveforms and optical spectra without and with dispersion compensation. These figures indicate that the pulse width can be shortened to less than 1 ps by compensating so that the average dispersion is around zero. In the laser cavity, the self phase modulation effect is negligible because the cavity length is much shorter than the soliton period of more than 400 m for an average dispersion of less than 1 ps/nm/km. As a result, a Gaussian pulse train with a time-bandwidth product of 0.44 is generated from the laser. In the present laser, the average dispersion is compensated to 0.4 ps/nm/km at 1.55 μm as shown in Fig. 1(b).
Figure 6 shows the output characteristics of the laser. Figure 6(a) shows the relationship between the laser output power and the pump power. The horizontal axis corresponds to he total pump power. The threshold pump power for the laser oscillation was 20 mW, and the slope efficiency was 11.3%. By employing the bidirectional pumping scheme, the maximum pump power was increased from 300 to 600 mW, resulting in a maximum output power of as high as 64.6 mW. Figure 6(b) shows the relationship between the output pulse width and the pump power. A pulse width of less than 1 ps was obtained for a pump power of more than300 mW, and was saturated for a pump power range of between 300 and 600 mW. Here, the minimum achievable pulse width is determined by the modulation depth of the LN phase modulator and the bandwidth of the optical filter installed in the laser cavity . Figures 6(c) and 6(d) show an autocorrelation waveform and an optical spectrum, respectively, for a total pump power of 400 mW. The output pulse width was 0.95 ps and the spectral width was 3.8 nm (470 GHz). The time-bandwidth product was 0.45, indicating that the output pulse is a transform-limited Gaussian pulse. Figure 6(e) shows the RF spectrum of a 10 GHz clock signal with a span of 100 MHz. There was only one clock component at 10 GHz, and the supermode noise was suppressed by 80 dB.
The precise control of the etalon FSR plays an important role as regards broadening the spectral width of the laser. Since the finesse of the etalon is 400, the 3-dB bandwidth of the transmission peak Δfetalon is 25 MHz. If there is a frequency misalignment of Δf between the FSR of the etalon and the repetition rate of the laser, a frequency detuning of NΔf occurs in a longitudinal mode separated by N modes from the center of the laser oscillation. Therefore, to obtain a laser oscillation with a spectral width of 500 GHz (2N = 50), Δf should be much smaller than Δfetalon/2N = 500 kHz. Figure 7 shows the change in the generated spectral width of the laser as a function of the etalon temperature. Here, the upper horizontal axis corresponds to Δf. The spectral width of the laser was less than half when Δf was 200 kHz. This indicates that precise FSR control of the etalon is very important if we are to obtain a short pulse train. We controlled the etalon temperature with a precision of 0.01 degrees, corresponding to a Δf of 0.5 kHz. As a result, we could successfully generate a 0.95 ps pulse train at a widest spectral width of 3.8 nm.
3. Optical frequency stabilization of mode-locked fiber laser
Figure 8 shows the configuration of an optical frequency stabilization circuit with an HCN gas cell as a frequency reference. The HCN gas cell is 165 mm long and filled at a pressure of 1 Torr. The spectral width of the HCN absorption line is 500 MHz due to the Doppler effect. The optical frequency of the laser is tuned by changing the cavity length with the PZT. After amplifying the output optical power from the laser with an erbium-doped fiber amplifier (EDFA), one of the longitudinal modes is extracted by using a fiber Bragg grating (FBG) with a spectral width of 1.2 GHz. We mounted the FBG on a voltage-controlled multi-layer piezo (MLP) actuator, which enables us to tune the optical filter frequency. We use an optical frequency component near the P(10) absorption line (1549.7 nm) of the HCN molecules. A frequency deviation from the HCN absorption line is detected with a phase sensitive detection (PSD) circuit . Then the error signal, which is proportional to the frequency deviation, is fed back to the PZT for optical frequency stabilization. In our previous work , optical frequency stabilization was achieved by controlling the temperature of the entire fiber cavity for coarse tuning and with the pump power control for fine tuning. This time, both tuningapproaches were replaced with a voltage-controlled PZT, which could simultaneously compensate for the high and low frequency fluctuations in the optical frequency. The frequency stabilization performance was the same, but the PZT is much simpler to use and easier to construct without the temperature control. On the other hand, optical frequency control that is realized by changing the cavity length simultaneously causes a microwave phase change. However, the amount of the phase change is quite small. For example, when the cavity length is changed by one optical wavelength (1.55 μm), the optical phase and frequency are changed by 2π and an FSR of 18 MHz, respectively. On the other hand, the amount of microwave phase change is given by,Fig. 6.
Figures 9(a) and 9(b) show the frequency fluctuation of the laser estimated from the error signal in the PSD circuit and its Allan standard deviation, respectively. The deviation was 5 x 10−11 for an integration time τ of 1 s, which was 2.5 times larger than that of the C2H2 frequency-stabilized mode-locked fiber laser (2 x 10−11 for τ = 1 s) . This was due to the relatively weak HCN absorption depth, which results in a PSD signal with a lower SNR than that of C2H2.
Next, we consider the stability of the laser spectrum. In a regeneratively mode-locked laser, the modulation frequency of the mode-locker is automatically changed in accordance with the change in the cavity length, resulting in stable long-term mode-locking operation. When the repetition rate is stabilized by controlling the microwave phase in the regenerative feedback loop, there is a timing shift between the modulation signal and the optical pulse which passes through the mode-locker due to the change in the cavity length. To reduce the timing shift, the group velocity of the optical pulse is changed automatically by shifting the center wavelength of the laser oscillation , resulting in a fluctuation of the laser spectrum peak. In this case, simultaneous optical frequency stabilization plays an important role in keeping the cavity length constant and therefore there is no fluctuation in the spectral profile in principle. Figures 10(a) and 10(b) show changes in the oscillation spectrum observed every 6 hours without and with optical frequency stabilization. Without optical frequency stabilization, the spectral peak was changed to a short wavelength over time. On the other hand, there was no change in the spectral profile over the long term because of the optical frequency stabilization. Figure 11 shows the changes in the center wavelength of the laseroscillation without and with optical frequency stabilization for 24 hours of continuous operation. Without optical frequency stabilization, the center wavelength of the laser oscillation changed periodically with a day cycle. In Fig. 11, the room temperature exhibited its maximum value at around 15 hours, resulting in a shorter wavelength shift. With the optical frequency stabilization, the center wavelength change was zero (within the resolution of the spectrum analyzer). This result indicates that long-term stable laser oscillation can also be obtained by realizing the simultaneous stabilization of the optical frequency and the repetition rate.
We demonstrated a 10 GHz, 1.55 μm mode-locked fiber laser in which the optical frequency was stabilized to an HCN absorption line, resulting in an optical frequency stability of 5x10−11 for an integration time of 1 s. In this laser, an etalon filter with an FSR of exactly 9.95328 GHz was employed to suppress mode hopping. The output power was increased to more than 60 mW by employing a bidirectional pumping scheme. The pulse width was reduced to less than 1 ps by employing dispersion management in the laser cavity. The pulse shortening effect was also revealed numerically by using the SSFM method. Furthermore, a stable spectral profile was obtained over a long period with the simultaneous stabilization of the optical frequency and repetition rate. This laser is applicable to ultra-high-speed digital coherent optical pulse transmission [17–19].
This work was supported by the JSPS Grant-in-Aid for Specially Promoted Research (26000009).
References and links
1. M. Nakazawa, T. Yamamoto, and K. R. Tamura, “1.28 Tbit/s-70 km OTDM transmission using third- and fourth-order simultaneous dispersion compensation with a phase modulator,” Electron. Lett. 36(24), 2027–2029 (2000). [CrossRef]
2. H. G. Weber, S. Ferber, M. Kroh, C. Schmidt-Langhorst, R. Ludwig, V. Marembert, C. Boerner, F. Futami, S. Watanabe, and C. Schubert, “Single channel 1.28 Tbit/s and 2.56 Tbit/s DQPSK transmission,” Electron. Lett. 42(3), 178–179 (2006). [CrossRef]
3. K. Harako, D. Seya, D. Suzuki, T. Hirooka, and M. Nakazawa, “2.56 Tbit/s/ch (640 Gbaud) polarization-multiplexed DQPSK non-coherent Nyquist pulse transmission over 525 km,” Opt. Express 23(24), 30801–30806 (2015). [CrossRef] [PubMed]
4. The 14th Int. Conf. Ultrafast Phenomena (2004), Part XI and XIII.
5. K. Kasai, T. Omiya, P. Guan, M. Yoshida, T. Hirooka, and M. Nakazawa, “Single-channel 400 Gbit/s, OTDM-32 RZ/QAM coherent transmission over 225 km using an optical phase-locked loop technique,” IEEE Photonics Technol. Lett. 22(8), 562–564 (2010). [CrossRef]
6. D. O. Otuya, K. Kasai, T. Hirooka, and M. Nakazawa, “Single-channel 1.92 Tbit/s, 64 QAM coherent Nyquist orthogonal TDM transmission with a spectral efficiency of 10.6 Bit/s/Hz,” J. Lightwave Technol. 34(2), 768–775 (2016). [CrossRef]
7. D. J. Jones, S. A. Diddams, J. K. Ranka, A. Stentz, R. S. Windeler, J. L. Hall, and S. T. Cundiff, “Carrier-envelope phase control of femtosecond mode-locked lasers and direct optical frequency synthesis,” Science 288(5466), 635–640 (2000). [CrossRef] [PubMed]
8. M. Nakazawa and M. Yoshida, “Scheme for independently stabilizing the repetition rate and optical frequency of a laser using a regenerative mode-locking technique,” Opt. Lett. 33(10), 1059–1061 (2008). [CrossRef] [PubMed]
9. M. Nakazawa, K. Kasai, and M. Yoshida, “C2H2 absolutely optical frequency-stabilized and 40 GHz repetition-rate-stabilized, regeneratively mode-locked picosecond erbium fiber laser at 1.53 μm,” Opt. Lett. 33(22), 2641–2643 (2008). [CrossRef] [PubMed]
10. M. Yoshida, K. Yoshida, K. Kasai, and M. Nakazawa, “A 0.95 ps, 10 GHz, 60 mW HCN frequency-stabilized and mode-locked fiber laser at 1.55 μm,” inProceedings of Conference on Lasers and Electro-Optics (Optical Society of America, 2016), paper STu1P.3. [CrossRef]
11. M. Yoshida, T. Yaguchi, S. Harada, and M. Nakazawa, “A 40 GHz regeneratively and harmonically mode-locked erbium-doped fiber laser and its longitudinal-mode characteristics,” IEICE Trans. Electron. E87-C(7), 1166–1172 (2004).
12. M. Yoshida, K. Kasai, and M. Nakazawa, “Mode-hop-free, optical frequency tunable 40-GHz mode-locked fiber laser,” IEEE J. Quantum Electron. 43(8), 704–708 (2007). [CrossRef]
13. M. Nakazawa, E. Yoshida, and Y. Kimura, “Ultrastable harmonically and regeneratively modelocked polarization-maintaining erbium-doped fiber ring laser,” Electron. Lett. 30(19), 1603–1605 (1994). [CrossRef]
14. D. J. Kuizenga and A. E. Siegman, “FM and AM mode locking of the homogeneous laser - Part I: Theory,” IEEE J. Quantum Electron. 6(11), 694–708 (1970). [CrossRef]
15. K. Kasai, A. Suzuki, M. Yoshida, and M. Nakazawa, “Performance improvement of an acetylene (C2H2) frequency-stabilized fiber laser,” IEICE Electron. Express 3(22), 487–492 (2006). [CrossRef]
17. K. Kasai, T. Omiya, P. Guan, M. Yoshida, T. Hirooka, and M. Nakazawa, “Single-channel 400-Gb/s OTDM-32 RZ/QAM coherent transmission over 225 km using an optical phase-locked loop technique,” IEEE Photonics Technol. Lett. 22(8), 562–564 (2010). [CrossRef]
18. E. Palushani, C. Schmidt-Langhorst, T. Richter, M. Nölle, R. Ludwig, and C. Schubert, “Transmission of a serial 5.1-Tb/s data signal using 16-QAM and coherent detection,” in Proceedings of the European Conference on Optical Communication (ECOC, 2011), paper We.8.B.5. [CrossRef]
19. D. O. Otuya, K. Harako, K. Kasai, T. Hirooka, and M. Nakazawa, “Singlechannel 1.92 Tbit/s, 64 QAM coherent orthogonal TDM transmission of 160 Gbaud optical Nyquist pulses with 10.6 bit/s/Hz spectral efficiency,” in Proceedings of the Optical Fiber Communication Conference (OFC, 2015), paper M3G.2.