An ultrafast all-optical switching using pulse trapping by orthogonally polarized soliton pulse in birefringent fiber is investigated both experimentally and numerically. The signal pulse, which is polarized along the fast axis, is trapped by the control soliton pulse along the slow axis; they subsequently copropagate along the fiber. Their wavelengths are red-shifted by the soliton self-frequency shift. The trapped pulse is shaped as a sech2 ultrashort pulse. It is also amplified by the Raman gain of the control pulse. Temporal response of the pulse trapping is estimated as 250 fs for 150 fs signal pulses. Ultrafast all-optical switching is demonstrated for the 0.84 THz pulse train.
©2005 Optical Society of America
Ultrafast all-optical switching is an important technique for future ultrahigh speed optical communication systems, optical computing, and all-optical data processing [1–3]. Recently, novel fibers, such as photonic crystal fibers  and highly nonlinear fibers , have been demonstrated and we can use dispersion managed, highly nonlinear fibers. Using such optical fibers and ultrashort pulses, we can obtain drastic nonlinear phenomena [6,7]. So far, a few groups have discovered the interesting phenomena of pulse trapping in optical fibers [8–10]. An ultrashort pulse captures another optical pulse. Then they copropagate along the fiber. In 1989, Islam discovered the soliton-trapping phenomenon , by which orthogonally polarized soliton pulses trap each other and copropagate along the fiber. In 2002, we also discovered the novel phenomenon of pulse trapping across a zero-dispersion wavelength . A signal pulse at the normal dispersion region is captured by an ultrashort soliton pulse at the anomalous dispersion region. They then copropagate along the fiber. In pulse propagation, the soliton pulse and trapped pulse are shifted respectively toward the longer and shorter wavelength sides. Ultrafast all-optical switching for a signal pulse in normal dispersion region is demonstrated using this phenomenon of pulse trapping across a zero-dispersion wavelength . In terms of application for actual optical communication, ultrafast switching at the anomalous dispersion region is both important and desired.
In 2002, we also discovered a novel phenomenon of trapped pulse amplification in birefringent optical fibers . Optical pulse is trapped by the orthogonally polarized ultrashort soliton pulse. As the wavelength of the soliton pulse is shifted toward the longer wavelength side, the wavelength of the trapped pulse is also shifted toward the longer wavelength to satisfy the condition of group velocity matching. The trapped pulse is amplified through the Raman gain of the soliton pulse.
In this study, we demonstrate high-performance ultrafast all-optical switching using the novel phenomenon of trapped pulse amplification at the anomalous dispersion region for the first time. An orthogonally polarized ultrashort pulse captures the arbitrary signal pulse in the pulse train. The captured pulse is amplified by the control pulse and is shaped to be an ultrashort soliton pulse. Merely varying the power of the control pulse can control the wavelength of the trapped pulse.
In Section 2, characteristics of pulse trapping are analyzed both numerically and experimentally. Ultrafast all-optical switching for pulse trains using pulse trapping is demonstrated in Section 3. Output pulses are observed directly using cross-correlation frequency resolved optical gating (X-FROG) technique .
2. Analysis of pulse trapping by orthogonally polarized ultrashort soliton pulse
First, we analyze characteristics of pulse trapping in a birefringent optical fiber. Figure 1 shows the experimental setup for pulse trapping. As the pump pulse source, a passively mode-locked Er-doped fiber laser is used. It generates about 100 fs sech2 like pulses at the 1.55 μm wavelength. The repetition frequency is 50 MHz. The output pulses are coupled into polarization maintaining fiber (PMF, 3M FS-PM-7811). The polarization direction of the input pulses is inclined from the birefringent axis of this fiber. Thus, the orthogonally polarized two pulses propagate independently and the wavelength tunable two-colored soliton pulses are generated [13,14]. They are used respectively as the signal pulse and control pulse. The wavelengths of the two pulses are shifted continuously by varying the fiber input power and the polarization direction of the pump pulse. The temporal widths of the generated pulses are 150 fs. The two pulses are generated from one pump pulse in the same fiber. Therefore, the timing jitter is sufficiently small to analyze pulse trapping.
The control and signal pulses are divided into two optical axes using a polarization beam splitter. Then they are temporally overlapped after adjustment of the time delay; they are coupled into highly nonlinear polarization maintaining fibers (HN-PMF). As the fiber parameters, β2=-10.9 ps2/km, β3=0.0648 ps3/km, the mode field diameter is 4.4 μm, nonlinear coefficient γ=12 W-1 km-1, and the magnitude of birefringence is 4.8 × 10-4 at the wavelength of 1.55 μm.
Figure 2 shows the principle of pulse trapping in a birefringent fiber. The polarization directions of the control and signal pulses are adjusted respectively along the slow and fast axes. The center wavelengths of the control and signal pulses are set respectively as 1600 and 1734 nm. The fiber has an anomalous dispersion property in this wavelength. Therefore, the group velocity of the control pulse is greater than that of the signal pulse. The pulse wavelengths are adjusted so that the group velocity difference from the chromatic dispersion compensates that from the birefringence. Consequently, the group velocity matching is satisfied between control and signal pulses and they copropagate along the fiber. The output pulses are observed using an optical spectrum analyzer and X-FROG technique .
Figure 3 shows the experimentally observed variation of optical spectra for pulse trapping. In Fig. 3, the average optical powers in front of the fiber are 5 mW and 1.9 mW respectively for control and signal pulses. The coupling efficiency into HN-PMF is 50% in this experiment. The corresponding pulse energy in the fiber is 50 pJ for the control pulse and 19 pJ for the signal pulse. The control pulse traps the signal pulse with high efficiency through the cross phase modulation when the initial temporal separation is adjusted and the signal pulse and control pulse are temporally overlapped. Thereafter, they copropagate along the fiber. In propagation along the fiber, the wavelengths of the two pulses are shifted toward the longer
wavelength side because the intra-pulse Raman scattering and the group velocity matching condition is always satisfied. The temporal shape is retained as sech2 because of the soliton effect. The overlapped pulses behave as a single soliton pulse. The control pulse energy is also transferred into the signal pulse through the Raman gain of the control pulse. After propagation along the 100-m-long fiber, almost all energy of the control pulse is transferred into the signal pulse .
Figure 4(a) shows characteristics of the wavelength shift of the trapped pulse as a function of the power of the control pulse when the fiber length is 20 m. Symbols represent experimental results and the red curve shows the calculated ones. To analyze the interaction between two orthogonally polarized pulses, a coupled nonlinear Schrödinger equation is used for this analysis . The parameters are adjusted to those in the experiments. Figure 4(a) shows that the wavelength of the trapped pulse is shifted continuously by varying the power of the control pulse. The calculated results are well in agreement with the experimental ones.
Figure 4(b) shows the calculated results of the characteristics of trapping efficiency when the initial temporal separation between the signal and control pulses is changed. The wavelength of the signal pulse is fixed as 1734 nm, whereas the wavelength of the control pulse is varied from 1.55 to 1.75 μm. The group velocity matching condition is satisfied and the maximum trapping efficiency reaches ~100% when the wavelength of the control pulse is 1600 nm. In this case, the temporal width for the trapping efficiency is about 250 fs at full width at half maximum (FWHM). As the wavelength of the control pulse is shifted, the group velocity mismatch is increased; thereby, the trapping efficiency magnitude is decreased. These calculated results agree well with the experimental ones.
The trapped pulse is also amplified through the Raman gain of the control pulse. The maximum amplification ratio is as high as 360% in Fig. 3 when the fiber length is 100 m. As the power of the control pulse increases, the maximum amplification ratio also increases.
Numerical analysis shows that the temporal width for trapping efficiency is decreased as the temporal widths of the signal and control pulses are decreased. The temporal width is also decreased as the power of the control pulse is increased. The temporal response of pulse trapping is limited by the temporal width of the optical pulses.
3. Ultrafast all-optical switching for pulse train using pulse trapping
Results shown in Fig. 4 suggest that ultrafast, high-performance all-optical switching can be demonstrated using pulse trapping. We demonstrate ultrafast all-optical switching for a high repetition rate pulse train in an anomalous dispersion region using pulse trapping in birefringent fiber.
Figure 5 shows the calculated results of ultrafast all-optical switching using pulse trapping. As the signal pulse train, we prepared a train of four pulses with temporal separation of 1.5 ps, which corresponds to a 0.67 THz ultrahigh repetition rate pulse train. The respective energies of signal pulses and control pulses were 30 pJ and 60 pJ. The wavelength and temporal width of each pulse were set as identical to those in Fig. 3. The HN-PMF was assumed as the fiber for pulse trapping. When the control pulse (soliton pulse) overlaps with a signal pulse and the group velocity matching condition is satisfied, the overlapped pulse is trapped by the control pulse with almost perfect efficiency; they copropagate along the fiber. The trapped pulse is shifted toward the longer wavelength side to satisfy group velocity matching. It is also amplified by the Raman gain of the control pulse. When the control pulse is overlapped with the third and fourth pulses, because the control pulse is shifted toward the longer wavelength side in the pulse propagation, group velocity matching is not satisfied and the control pulse does not trap them. For that reason, we can trap only one arbitrary pulse using pulse trapping. Only the trapped pulse is shifted toward the longer wavelength side, so we can pick off only the trapped signal pulse using a wavelength filter. The magnitude of the wavelength shift is increased monotonously as the power of the control pulse increases.
Next, we demonstrate the ultrafast all-optical switching experimentally. Figure 6 shows the experimental setup for all-optical switching, which is the advanced scheme of Fig. 1. A part of the output pulse from a fiber laser was picked off. It was amplified using a fiber amplifier. Then three ultrashort signal pulses were generated and were used as the signal pulse train. Adjusting the delay line controlled the temporal separation between signal pulses. The X-FROG technique was used for observation of all-optical switching. Using this technique, we can directly observe the temporal distribution of the spectral components.
Figure 7 shows the observed spectrogram of ultrafast all-optical switching. Figure 7(a) shows the spectrogram of output pulse train without pulse trapping and 7(b) and 7(c) show those at the outputs of 4 and 20 m long fibers when the second pulse is trapped by the control pulse. The pulse energy was 50 pJ for control pulse and 8 pJ for each signal pulse. The wavelength and temporal width were identical to those in Figs. 3 and 5. Three equally separated pulses were observed clearly when pulse trapping was not applied. Temporal separation between pulses was 1.8 ps. The second pulse component was trapped and it copropagated with the control pulse when the pulse trapping was applied. We can easily select the trapped signal pulse using a wavelength filter. In Fig. 7(c), because the signal pulses had insufficient power to form the fundamental solitons, they were broadened temporally by the effect of chromatic dispersion. Consequently, the interference pattern was observed by the temporal overlapping of signal pulses. For the trapped pulse, the pulse energy is amplified through the Raman gain of the control pulse and the magnitude of wavelength shift is increased. The temporal width of the trapped pulse is also narrowed. It is shaped as a sech2 ultrashort pulse. Nothing was trapped by the control pulse when the second pulse was “off”. These results show that we have achieved ultrafast all-optical switching for a 0.56 THz pulse train.
Finally, we examined all-optical switching’s fastest switching speed. Figure 8 shows the spectrogram of output pulses when the temporal separation of each signal pulse is 1.2 ps. The corresponding repetition frequency is 0.83 THz. At first, the train of three signal pulses was prepared and only the second signal pulse was trapped by the control pulse. Then we made the second signal pulse “off” and the extinction ratio was examined. In Fig. 8(b), when the second pulse was “off”, nothing was trapped by the control pulse, meaning that only the second pulse is trapped in Fig. 8(a). The extinction ratio was estimated as more than 25 dB. All-optical switching was shown when the repetition frequency was increased to 1 THz, but the extinction ratio was degraded to 10 dB. By increasing the power or by narrowing the temporal width of control pulse, it is expected that we can demonstrate much higher speed and high performance ultrafast all-optical switching.
We can demonstrate higher performance ultrafast all-optical switching using pulse trapping than that by other all-optical switching techniques. We can trap and amplify an arbitrary pulse using the ultrashort soliton pulse. The temporal shape of the trapped pulse is narrow and it is shaped as an ultrashort ideal sech2 pulse. Varying the control-pulse power can control the wavelength of the trapped pulse. Although pulses of a few tens of picojoule were used in this work, this technique is applicable to small-energy signal pulses that are used in practical optical communications. The pulse-trapping phenomenon offers great potential for many applications: ultrafast optical communication, optical computing, quantum optics, etc.
In this paper, we have investigated ultrafast all-optical switching in the anomalous dispersion region using pulse trapping between orthogonally polarized ultrashort pulses in birefringent fibers. Characteristics of pulse trapping were investigated both experimentally and numerically. The maximum trapping efficiency was ~100% and the temporal width of the trapping efficiency was 250 fs at FWHM for 150-fs signal pulses. The trapped pulse is shaped as a sech2 ultrashort pulse and it is amplified by the Raman gain of the control pulse. Varying the control-pulse power changes the trapped-pulse wavelength.
Subsequently in this paper, high performance ultrafast all-optical switching was demonstrated. The control pulse traps one arbitrary pulse from the pulse train, then they copropagate along the fiber. The wavelength of the trapped pulse is red-shifted so that the trapped pulse can be selected using the wavelength filter. The trapped pulse is also amplified and shaped as a sech2 ultrashort pulse. Ultrafast all-optical switching was demonstrated for 0.83 THz pulse train with an extinction ratio of more than 25 dB when 150 fs ultrashort pulses were used.
We would like to thank The Furukawa Electric Co. Ltd. for providing us HN-PMF. This work was supported through a Grant-in-Aid for Young Scientists (A) from the Ministry of Education, Culture, Sports, Science, and Technology, Japan.
References and links
1. J. P. Sokoloff, P. R. Prucnal, I. Glesk, and M. Kane, “A terahertz optical asymmetric demultiplexer (TOAD),” IEEE Photon. Technol. Lett. 5, 787–790 (1993). [CrossRef]
2. S. Nakamura, Y. Ueno, and K. Tajima, “Ultrafast (200-fs switching, 1.5 Tb/s demultiplexing) and high-repetition (10 GHz) operations of a polarization-discriminating symmetric Mach-Zehnder all optical switch,” IEEE Photon. Technol. Lett. 10, 1575–1577 (1998). [CrossRef]
3. B. E. Olsson and D. J. Blumenthal, “All-optical demultiplexing using fiber cross-phase modulation (XPM) and optical filtering,” IEEE Photon. Technol. Lett. 13, 875–877 (2001). [CrossRef]
4. J. C. Knight, T. A. Birks, P. St. J. Russell, and P. J. Roberts, “Photonic crystal fibers, All-silica single-mode optical fiber with photonic crystal cladding,” Opt. Lett. 21, 1547–1549 (1996). [CrossRef] [PubMed]
5. T. Okuno, M. Onishi, T. Kashiwada, S. Ishikawa, and M. Nishimura, “Silica-based functional fibers with enhanced nonlinearity and their applications,” IEEE J. Sel. Top. in Quantum Electron. 5, 1385–1391 (1999). [CrossRef]
6. J. K. Ranka, R. S. Windeler, and A. J. Stentz, “Visible continuum generation in air-silica microstructure optical fibers with anomalous dispersion at 800 nm,” Opt. Lett. 25, 25–27 (2000). [CrossRef]
7. N. Nishizawa and T. Goto, “Widely broadened super continuum generation using highly nonlinear dispersion shifted fibers and femtosecond fiber laser,” Jpn. J. Appl. Phys. 40, L365–L367 (2000). [CrossRef]
9. N. Nishizawa and T. Goto, “Pulse trapping by ultrashort soliton pulses in optical fibers across zero-dispersion wavelength,” Opt. Lett. 27, 152–154 (2002). [CrossRef]
10. N. Nishizawa and T. Goto, “Trapped pulse generation by femtosecond soliton pulse in birefringent optical fibers,” Opt. Express 10, 256–261 (2002). http://www.opticsexpress.org/abstract.cfm?URI=OPEX-10-5-256. [PubMed]
11. N. Nishizawa and T. Goto, “Ultrafast all optical switching by use of pulse trapping across zero-dispersion wavelength,” Opt. Express 11, 359–365 (2003) http://www.opticsexpress.org/abstract.cfm?URI=OPEX-11-4-359. [CrossRef] [PubMed]
12. N. Nishizawa and T. Goto, “Experimental analysis of ultrashort pulse propagation in optical fibers around zero-dispersion region using cross-correlation frequency resolved optical gating,” Opt. Express 8, 328–335 (2001) http://www.opticsexpress.org/abstract.cfm?URI=OPEX-8-6-328. [CrossRef] [PubMed]
13. N. Nishizawa and T. Goto, “Compact system of wavelength-tunable femtosecond soliton pulse generation using optical fibers,” IEEE Photon. Technol. Lett. 11, 325–327 (1999). [CrossRef]
14. N. Nishizawa, R. Okamura, and T. Goto, “Simultaneous generation of wavelength tunable two-colored femtosecond soliton pulses using optical fibers,” IEEE Photon. Technol. Lett. 11, 421–423 (1999). [CrossRef]