Open Access
Add to BibSonomyAbstract
This paper proposes a novel detection technique for wavelength-division multiplexing (WDM) signals that uses the ultrafast field sampling approach. The proposed technique simultaneously samples the total field of the WDM signal with no wavelength-demultiplexing; its electrical post-processing provides a filtering function in the digital domain. As a result, the individual fields of the WDM channels, including mutual phase relationship, can be jointly reconstructed. As a preliminary demonstration, the simultaneous monitoring of a WDM signal composed of two channels with independent polarization states, is successfully performed using a dual-channel field sampling system with polarization diversity. This demonstration can be expanded to the detection of a WDM signal with N channels by using an N-channel field sampling system. It is also experimentally-verified that the reconstructed fields preserve the mutual phase relationship of the original fields. This ‘field sensitive’ WDM detection may open new possibilities for advanced WDM monitoring systems and even the electrical compensation of linear and/or non-linear inter-channel cross-talk in digital coherent detection systems.
©2009 Optical Society of America
1. Introduction
Wavelength-division multiplexing (WDM) technology is currently playing an indispensable role in providing large-capacity fiber-optic communication systems. WDM channel demultiplexing is generally carried out by optical filter devices such as dielectric multilayer filters and arrayed waveguide gratings (AWGs) before the multiple optical signals are detected at their different wavelengths. In this scheme, the optical filter devices need to be more precisely designed and controlled as the channel spacing narrows.
Homodyne coherent detection techniques employing local short-pulses have been studied for characterizing optical fields at ultrafast time scales [1–4]. Particularly in optical communication applications, linear-gating techniques enable optical time-domain demultiplexing [5] or waveform monitoring of ultrafast signals [6, 7]. The capability of linear optical sampling enables us to sample the electrical field of an optical signal by detecting the orthogonal quadratures of the cross-correlation with local short-pulses. Since WDM signals are usually formed by modulating (independently) multiple laser sources with different wavelengths, this total field changes at a speed corresponding to the total wavelength bandwidth. Therefore, it should be possible to reconstruct the individual channels by employing ultrafast sampling with respect to the total field of the WDM signal.
This paper offers a new technique for WDM signal detection that adopts the ultrafast field sampling approach. The proposed technique simultaneously samples the total field of the WDM signal without wavelength-demultiplexing by optical filter devices; subsequent digital signal processing (DSP) provides a filtering function in the digital domain. As a result, the individual fields of the WDM channels, including mutual phase relationship, can be jointly reconstructed. In an experiment based on the results presented in [8], we successfully detect a WDM signal composed of two channels with independent state of polarization (SOP), using a dual-channel field sampling system that incorporates polarization diversity. This demonstration can be expanded to the detection of the WDM signal with N channels by using an N-channel field sampling system. Also, it is verified that the reconstructed fields of the two channels preserve their mutual phase relationship. This result implies that this ‘field sensitive’ WDM detection technique will be useful for measuring or electrically-compensating linear and/or non-linear inter-channel cross-talk [9–13]. The proposed scheme yields easily-designed and controlled digital wavelength-demultiplexing, and furthermore, may open a novel approach to WDM signal detection, especially in combination with digital coherent technology [14, 15].
2. Principle
Figure 1 shows the schematic diagram of the ultrafast field sampling system that incorporates polarization diversity for detecting a WDM signal with N channels; the y-channel configuration for detecting the y-polarization component of the signal is omitted since it is the same as that for x-polarization component. Since each channel has independent SOP, the incoming signals, which have arbitrary SOP and the local short-pulses (sampling pulses) linearly-polarized at 45-degree, are separated into x- and y-polarization components by the polarization beam splitter (PBS), and distributed to N tributaries. The relative time delays T, 2T, ⋯, (N-1)T are then sequentially imposed on each portion of either the signals or the sampling pulses. The definition of delay T is described later. These WDM signals and sampling pulses are combined in phase diversity manner in an optical 90° hybrid, and the in-phase and quadrature components of the interference signal are delivered to two pairs of balanced detectors. All tributaries are then sampled and digitalized in parallel by the analogue-to-digital converters (ADCs), and the sampled data are numerically-processed by the DSP. The N channels in the WDM signal can be detected by the N-channel field sampler incorporating the sequential delay of (N-1)T when each phase of the outputs from the samplers remains completely constant during the measurement.
Figure 2 presents the spectra of the WDM signal and the sampling pulse. The carrier frequency of each WDM channel runs from f1 to fN; ∆f represents the frequency interval. FSR is defined as the free spectral range, which corresponds to N·∆f. The delay T in the field sampling system is set to the inverse of FSR, i.e. N·∆f. Although the channel signals are usually generated by independent lasers, they are mutually coherent within short observation time (N-1)T. Note that, to detect all channels of the WDM signal, the spectrum of the sampling pulse must cover the whole spectrum of the WDM signal.
The following describes the theoretical operation by which a WDM signal with N channels can be detected by the N-channel field sampling system. The x- and y-polarization components of the i-th optical field are written as
where aix(t) and aiy(t) are the amplitudes of the x- and y-polarization components of the i-channel signal, respectively, and δi is the phase difference between them. In the following description, for the sake of simplification, we explain the detection of the x-component. The field amplitude of the total WDM signal can be written as
b(t) varies extremely quickly with time and the time variation is similar to the inverse of total bandwidth (N-1)·∆f. Meanwhile, the x-polarization component of the sampling pulse is written as
where δ(t) represents the pulse waveform, which is assumed to be a delta-function. fs is the center frequency of the sampling pulse and ϕs is the initial phase with respect to the signal. When the WDM signal b(t) and the delayed sampling pulses are input into the field sampling system, the orthogonal quadratures of b(t), b(t+T), ⋯ , b[t+(N-1)T] are sequentially-sampled in accordance with the given delay. The outputs JN=IN+jQN from the N-th channel are given by
Thus, assuming that aix(t) remains unchanged during (N-1)T, aix(0)≈aix[(N-1)T], since the frequency interval ∆f between WDM channels is usually larger than the single bandwidth ∆fsignal in the WDM signals as shown in Fig. 2, the amplitude aix(0) of each channel is reconstructed by computing the Fourier transform of JN as follows;
Similarly, the amplitude aiy(0) of the y-polarization component can be obtained from the other polarization diversity channel. The intensity |Ei|2 of the i-th channel signal can then be obtained by combining the two polarization components as follows;
From the sampling theorem, the sampling interval T must be smaller than 1/2B (B is the total signal bandwidth). Furthermore, the real-time detection of the WDM channel may be realized by utilizing a sampling pulse with repetition rate higher than the Nyquist frequency with respect to the bit rate per channel.
Fig. 1. Schematic diagram of ultrafast field sampling system incorporating polarization diversity for detecting WDM signals with N channels. The above shows the x-channel configuration for detecting the x-polarization component; the same configuration (y-channel) is employed for the y-polarization. T: the relative time delay set to be the inverse of FSR shown in Fig.2. PBS: polarization beam splitter. BPD: balanced photodetector. A/D: analogue-to-digital conversion. DSP: digital signal processing.
Fig. 2. Representation of the spectra of the WDM signal with carrier frequencies from f1 to fN (continuous curves) and the sampling pulse (dashed curve). N: the number of WDM channels. ∆f: frequency interval. FSR: free spectral range and the relation is FSR=N·∆f· ∆fsignal: signal bandwidth per channel.
3. Experiment and results
We conducted a preliminary experiment to confirm the proposed technique using a WDM signal with two channels. Figure 3 illustrates the dual-channel field sampling system incorporating polarization diversity used to detect the 2-channel WDM signal with independent SOPs. The two signals constituting the WDM signal were generated by two independent light sources that had center wavelengths of λ 1=1547.47 nm and λ 2=1548.47 nm. The frequency interval ∆f was 125 GHz, so the FSR was 250 GHz. The time delay was set to 4 ps, which corresponds to the inverse of the FSR. Each of the light sources was independently modulated with a LiNbO3 intensity modulator driven by a 10-Gbit/s pseudo random pattern sequence (PRBS). These signals were multiplexed through a 3-dB fiber coupler. The sampling pulses were generated by a passively mode-locked fiber laser operating at a repetition rate of 10 MHz. The sampling pulses had a pulsewidth of 1.2 ps and spectrum width of 2 nm. We tuned the center wavelength of the sampling pulse so that its spectrum covered that of the WDM signal as shown in the inset of Fig. 3. The average incident power of each signal before multiplexing and that of the sampling pulse were -1 and -4 dBm, respectively. The optical 90° hybrids and optical delay lines consisted of packaged free-space optics components; these devices were connected by polarization-maintaining fibers. We have already described the operation of the dual-channel interferometer [7]. The output signals from each channel were acquired by data acquisition units triggered by the sampling pulse, and the resulting data were digitally-processed.
Figure 4 shows the total field of the WDM signals sampled by the proposed field sampling system. Figures 4(a) and 4(b) are the intensity and the relative phase difference obtained by calculating |J1|2 and arg(J1)-arg(J 2), respectively. These results were subjected to off-line digital Fourier analysis. Figures 5(a) and 5(b) show the intensity waveforms |a1x(0)|2 and |a2x(0)|2, reconstructed by calculating Eq. (5), of the x-polarization component of the signals with wavelengths of λ 1 and λ 2, respectively, while Figs. 5(c) and 5(d) show the intensity waveforms |a1y(0)|2 and |a2y(0)|2 of the y-polarization component of the signals with λ 1 and λ 2, respectively. The total intensity waveforms |E1|2 and |E2|2 of the signals with λ 1 and λ 2 obtained by calculating the sum of x- and y-polarization are shown in Figs. 5(e) and 5(f), respectively, where the two insets show the intensity waveform of each signal measured independently before multiplexing. From these results, we can confirm that the final total intensity waveforms of the two channels were reconstructed accurately, as compared to those measured independently. Also, even though the two channels differ considerably in the ratio of the intensity of the x-polarization and y-polarization components to the total intensity, clear eye diagrams were achieved in both cases. This indicates that the effectiveness of the proposed method is achieved regardless of the SOPs of the input WDM signal.
Fig. 3. Experimental setup. The inset shows the spectra of the WDM signals and the sampling pulses. LD: laser diode. IM: LiNbO3 intensity modulation. PPG: Pulse pattern generator. PMFL: passively mode-locked fiber laser. OBF: optical bandpass filter. Pol.: polarizer. PBS: polarization beam splitter. 90°-H: optical 90-degree hybrid. λ/2, λ/4: λ/2 and λ/4 wavelength plate. HM: half mirror. BPD: balanced photodetector. DAQ: data acquisition unit.
Fig. 4. Sampled intensity (a) and relative phase difference (b) of total field of WDM signals. These are obtained by calculating |J1|2 and arg(J1)- arg(J2).
Fig. 5. Reconstructed intensity waveforms of WDM signals. The left and right panels correspond to the signals with wavelengths of λ 1 and λ 2, respectively. (a) and (b) are x-polarization |a1x(0)|2 and |a2x(0)|2, respectively. (c) and (d) are y-polarization |a1y(0)|2 and |a2y(0)|2, respectively. (e) and (f) are the total intensity waveforms |E1|2 and |E2|2, summation of x- and y-components, respectively. The insets in (e) and (f) show the intensity waveform measured independently before multiplexing.
Figure 6 shows the relative phase difference between the two reconstructed channels, which is given by arg(a1x)-arg(a2x) using the plots filtered out near the intensity level of 0 in Fig. 5(a) and 5(b). The horizontal axis of the Fig. depicts the sampling event whose interval is 100 ns. Although the two wavelengths are from independently oscillating lasers, the observed phase difference remains constant for some finite time. This is because of the long coherent times of the lasers. (The measured linewidths of the two light sources were ~1 MHz: this typically corresponds to the coherent time of several hundred of ns, but the actual coherent time seems to be much longer, except for the long-term drift.) It has been demonstrated that, by using the coherent WDM scheme where all channels are phase-locked to each other, the inter-channel crosstalk induced through linear [10] and nonlinear [11] processes can be compensated. The above feature of the proposed system, where mutual phase difference can be detected, may provide similar compensation even when the phases of the channels are unlocked.
4. Conclusion
We have proposed and demonstrated a novel WDM signal detection technique that uses ultrafast field sampling. The proposed technique simultaneously samples the total field of the WDM signal, no wavelength-demultiplexing is performed with optical filter devices, and its electrical post-processing provides a filtering function in the digital domain. As a result, wavelength-demultiplexing is accomplished in the digital domain and the individual fields of each WDM channel, including the mutual phase relationship, can be jointly reconstructed. In an experiment, we successfully reconstructed the individual intensity waveform of a WDM signal composed of two channels with independent SOP, by using a dual-channel field sampling system with polarization diversity. We also verified that the reconstructed fields preserved the mutual phase relationship between the two channels. The experimental results indicate that this ‘field sensitive’ WDM detection technique will be useful for measuring or electrically-compensating linear and/or non-linear inter-channel cross-talk. Future work includes expanding this demonstration to attempt the detection of a WDM signal with N channels by using an N-channel field sampling system. We believe that the proposed technique may create the possibility for simultaneous WDM signal monitoring (optical filtering devices can be eliminated) and, moreover, a new scheme of WDM signal detection combined with digital coherent receivers.
References and links
1. K. Dou, A. Bébarre, J. L. Le Gouet, I. Lorgeré, and P. Tchénio, “Field cross correlator for analysis of ultrafast signals,” Appl. Opt . 33, 7980–7986 (1994). [CrossRef] [PubMed]
2. M. G. Raymer, J. Cooper, H. J. Carmichael, M. Beck, and D. T. Smithey, “Ultrafast measurement of optical-field statistics by dc-balanced homodyne detection,” J. Opt. Soc. Am. B 12, 1801–1812 (1995). [CrossRef]
3. F. Ito, “Single-shot high-speed signal detection by multiple-angle spectral interferometry,” J. Quantum Electron . 32, 519–524 (1996). [CrossRef]
4. D. F. Mc Alister and M. G. Raymer, “Ultrafast photon-number correlations from dual-pulse, phase-averaged homodyne detection,” Phys. Rev. A 55, 1609–1612 (1997). [CrossRef]
5. F. Ito, “Demultiplexed detection of ultrafast optical signal using interferometric cross-correlation technique,” J. Lightwave Technol . 15, 930–937 (1997). [CrossRef]
6. C. Dorrer, C. R. Doerr, I. Kang, R. Ryf, J. Leuthold, and P. J. Winzer, “Measurement of Eye Diagrams and Constellation Diagrams of Optical Sources Using Linear Optics and Waveguide Technology,” J. Lightwave Technol . 23, 178–186 (2005). [CrossRef]
7. K. Okamoto and F. Ito, “Relatively Delayed Dual-channel Interferometric Sampling for Observing Ultrafast Amplitude and Phase Modulation,” in European Conference on Optical Communications (ECOC2006), Cannes, France, Paper Th2. 4. 4 (2006).
8. K. Okamoto and F. Ito, “Simultaneous WDM Signal Detection Realized by Ultrafast Field Sampling,” in European Conference on Optical Communications (ECOC2008), Brussels, Belgium, Paper We .1.D.4 (2008).
9. M. Shtaif, M. Eiselt, and L. D. Garret, “Cross-phase modulation distortion measurements in multispan WDM systems,” IEEE Photon. Technol. Lett . 12, 88–90 (2000). [CrossRef]
10. A. D. Ellis and F. C. G. Gunning, “Spectral density enhancement using coherent WDM,” IEEE Photon. Technol. Lett . 17, 504–506 (2005). [CrossRef]
11. E. Yamazaki, F. Inuzuka, A. Takada, K. Yonenaga, and T. Morioka, “Inter-channel crosstalk cancellation by encoding with adjacent channels in coherent WDM,” in Optical Fiber Communication Conference (OFC2006), Anaheim, USA, Paper JThB6 (2006).
12. E. Yamazaki, F. Inuzuka, K. Yonenaga, A. Takada, and M. Koga, “Compensation of interchannel crosstalk induced by optical fiber nonlinearity in carrier phase-locked WDM system,” IEEE Photon. Technol. Lett . 19, 9–11 (2007). [CrossRef]
13. F. Inuzuka, E. Yamazaki, K. Yonenaga, and A. Takada, “Nonlinear Inter-channel compensation using electronic pre-distortion in Carrier Phase Locked WDM,” in Optical Fiber Communication Conference (OFC2008), San diego, USA, Paper OTuO5 (2008).
14. M. G. Taylor, “Coherent detection method using DSP for demodulation of signal Demodulation of Signal and Subsequent Equalization of Propagation Impairments,” IEEE Photon. Technol. Lett . 16, 674–676 (2004). [CrossRef]
15. K. Kikuchi, “Phase-Diversity Homodyne Detection of Multilevel Optical Modulation With Digital Carrier Phase Estimation,” J. Sel. Top. Quantum Electron . 12, 563–570 (2006). [CrossRef]
