Open Access
Add to BibSonomy
Abstract
A self-seeded reflective semiconductor optical amplifier (RSOA) has been viewed as an attractive candidate for the low-cost colorless transmitter in wavelength-division-multiplexed passive-optical networks (WDM-PON). In this paper, we propose and demonstrate the use of two-section RSOA with segmented electrodes to simultaneously improve the modulation bandwidth and extinction ratio in a self-seeded operation. We first find the optimal driving condition to induce strong modulation cancellation by analyzing the gain dynamics inside the RSOA. In addition, we experimentally confirm the improvement in direct modulation bandwidth by segmenting the electrodes. Finally, we employ the two-section RSOA in a self-seeded system and successfully demonstrate transmission over a 25-km standard single-mode fiber at 5 Gb/s as well as improved back-to-back modulation at 10-Gb/s in C-band by using the optimal driving condition.
© 2017 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Wavelength-division-multiplexed passive optical network (WDM-PON) has attracted special attention for next generation access network due to its extremely large data capacity, excellent network security, and flexibility [1,2]. For WDM-PON to be deployed in practical systems, it is strongly desired to have colorless transmitters to cut down the manufacturing and maintenance cost [3]. Among various colorless transmitters, such as wavelength-tunable laser [4], injection-locked Fabry-Perot laser [5], and externally seeded reflective semiconductor optical amplifier (RSOA) [6], self-seeded RSOA scheme has attracted special attention due to its simplicity [7–14]. As shown in Fig. 1(a), a portion of light generated from the RSOA at the optical network unit (ONU) is reflected at the remote node (RN) and fed back into RSOA. As a result, a resonant cavity is formed between the RSOA and the RN, where the channel wavelength is determined automatically by the corresponding arrayed waveguide grating (AWG) port. Thus, identical RSOA can be applied to any ONU without the need for precise wavelength locking.
Fig. 1 (a) Self-seeded RSOA scheme in WDM-PON (FR: Faraday rotator, FRM: Faraday rotator mirror, RRH: remote radio head). (b) Modulation canceling effect of RSOA under DC bias.
The essential mechanism behind the self-seeded RSOA scheme is the strong gain saturation inside the RSOA, which allows us to erase the previous intensity-modulated data as shown in Fig. 1(b) and remodulate with new data at a bit rate much faster than the round-trip time (i.e., photon lifetime) of the entire cavity [11,12]. As a result, successful transmission experiments at up to 10 Gb/s in O-band have been reported [10]. On the other hand, it has also been revealed that there is an intrinsic tradeoff between the extinction ratio and modulation bandwidth. In order to induce strong gain saturation, relatively long (around 1 mm) RSOA is desired [13]. Such a long device, however, generally suffers from limited direct modulation bandwidth due to the large RC constant and slow carrier lifetime. To solve this tradeoff, we have recently proposed the use of two-section RSOA in self-seeded configuration and demonstrated back-to-back characteristics [14]. By using the long section under DC for modulation cancellation and the short section for modulation, we have demonstrated enhanced modulation canceling effect without sacrificing the modulation bandwidth.
In this paper, we provide detailed numerical analysis of carrier and photon distributions inside the two-section RSOA to elucidate the physical mechanism of enhancing the modulation-canceling effect under optimal driving condition. Moreover, we experimentally compare the direct modulation characteristics of the two-section and conventional single-section RSOAs to confirm the effect of segmented electrodes in extending the modulation bandwidth. Finally, by using an optimized self-seeded setup, we demonstrate, for the first time, transmission over 25-km standard single mode fiber (SSMF) at 5 Gb/s, as well as improved BTB modulation at 10 Gb/s, both in C band without electric pre-compensation.
The paper is organized as follows. Section 2 presents the numerical analysis and electro-optic (E/O) characterization of two-section RSOA. We then provide the results of self-seeded transmission experiments in Section 3, and finally conclude the work in Section 4.
2. A two-section RSOA for a self-seeded transmitter
Figure 2 shows the two-section RSOA that we employ in this work. The electrode is split into two sections with respective length of L1 and L2. When an intensity-modulated signal is incident on RSOA, the original data is erased by the strong cross-gain saturation (XGS) between the counter-propagating waves inside Section 2, which is driven with a DC current. Before exiting the device, the light is remodulated by a new data at Section 1.
Fig. 2 Device structure and top view of two-section RSOA (AR: anti-reflection coating, HR: high-reflection coating.
Segmentation of electrodes offers three important factors to improve the signal quality. First, the short section (Section 1) near the output facet, which is used for modulation, provides larger modulation bandwidth due to small RC delay. Second, the long section (Section 2) can enhance the modulation-canceling effect because of longer interaction length for counter-propagating waves, which is critical for successful self-seeded operation [13]. Third, the electrodes-segmentation design itself can dramatically extend the modulation bandwidth [15,16]. This is attributed to the enhanced modulation cancellation, which generally has high-pass filtering function. Moreover, by tuning the current density at each section, we can optimize the carrier distribution inside the RSOA to maximize the modulation cancellation.
For proof-of-concept demonstration, we prepared two-section RSOA samples based on InGaAlAs multi-quantum-well active layers as shown in Fig. 2. Similar to our previous work [13], 1.5-μm-wide ridge waveguide was employed with curved geometry and anti-reflection coating at the facet to suppress reflection. From the measured gain properties of a single section device [13], we set L1 = 400 μm and L2 = 800 μm. The gap between two electrodes was 6 μm, where the InGaAs contact layer was also removed to improve electrical isolation.
2.1 Numerical analysis of the modulation-cancelling effect
To confirm the effect of electrode segmentation, we first simulate the static gain properties by using a numerical model based on rate equations and wave-propagation equations. We employ the material parameters derived experimentally in our previous work [13]. For quantitative evaluation of the effect of modulation cancellation obtained by the RSOA, we define modulation cancellation dynamic range (MCDR) as the range of input power, within which the variation of output power is kept below1 dB as shown in Fig. 1(b). It is important to obtain large MCDR and optical gain for efficient self-seeded operation.
Figure 3(a) and 3(b) shows the calculated MCDR and optical gain as a function of injected current I1 and I2 to Section 1 and Section 2, respectively. Note that I1 and I2 axes are scaled equally in terms of current density. From Fig. 3(a) and 3(b), we can confirm that the MCDR and optical gain of more than 20 dB can be obtained at wide range of current conditions. On the other hand, there exists asymmetry with respect to current density in Fig. 3(a) and 3(b), indicating that non-uniform current densities at two sections are preferable to maximize the modulation cancellation and gain. For clarity, we plot in Fig. 3(c) the gain as a function of J2/J1 for fixed total current of 110 mA and 170 mA, where J1,2 = I1,2/(wL1,2) (w: waveguide width) are the current densities at respective sections. Comparing to the case of uniform current density (J1 = J2), which corresponds to the conventional single-section RSOA, larger gain can be obtained when J2/J1 is around 0.4~0.6, implying that larger current density is desired at Section 1.
Fig. 3 (a) MCDR and (b) optical gain of two-section RSOA, calculated as a function of current I1and I2 to respective sections. (c) Optical gain as a function of current density ratio (J2/J1) for the fixed total current of 110 mA (blue line) and 170 mA (red line).
The physical mechanism behind this trend can be explained as a result of superior overlap between the carrier and photon distributions inside the RSOA [17]. To confirm this, Fig. 4 shows the simulated carrier and photon distributions along the longitudinal position in RSOA for various cases of current density (J2/J1 = 0.6, 1 and 1.5) under a fixed total current of 110mA. For the case of uniform current density (J2/J1 = 1), corresponding to the single-section RSOA, carrier density is depleted more heavily near the output facet of RSOA due to the larger optical power [see Fig. 4(b)]. When a larger current is injected to Section 1 (J2/J1 = 0.6), it compensates for the carrier depletion and hence results in larger optical gain. In contrast, for the opposite case of larger current in Section 2 (J2/J1 = 1.5), optical gain is reduced due to the larger carrier depletion near the output facet.
Fig. 4 Calculated (a) carrier and (b) photon distributions inside RSOA for J2/J1 = 0.6 (black), 1 (red), and 1.5 (blue) and total current of 110 mA. Dashed and solid curves in (b) denote the optical modes propagating in forward and backwards directions, respectively.
2.2 Modulation bandwidth of the two-section RSOA
To confirm the effect of extending the modulation bandwidth by the electrode segmentation, we measured the small-signal E/O response of the two-section RSOA by sending a continuous-wave (CW) light at 1550-nm wavelength into the device and detecting the directly modulated light from the RSOA. The input optical power was set to −25dBm. We applied only DC voltage without modulation at Section 2, while a small-signal modulation with DC bias was applied at Section 1. We adjusted the DC currents at both sections, so that the current densities were uniform along the device in this measurement.
Figure 5 shows the measured E/O response at the current density of 4.7 kA/cm2 (dashed line) and 6.0 kA/cm2 (solid line). For reference, we also show the E/O response of a single-section 1000-μm RSOA used in our previous work [13] under the same current densities. While the 3-dB modulation bandwidth of single-section RSOA is 3 GHz, that of two-section RSOA is increased to more than 10 GHz at the same current density of 6.0 kA/cm2. We note that although a 1200-μm-long single-section RSOA (which was not available) should be used for fair comparison, it would have even smaller bandwidth than the 1000-μm-long device. We can therefore confirm that the modulation bandwidth can be clearly extended by segmenting the electrode.
Fig. 5 Small-signal E/O response of two-section RSOA (blue) and 1000-μm single-section RSOA (red) at current density of 6.0 kA/cm2 (solid) and 4.7 kA/cm2 (dashed).
There are mainly two reasons to explain the remarkable improvement of modulation bandwidth of the segmented RSOA. First, the short section, used for modulation, reduces the capacitance and consequently enhances the RC bandwidth. Second, the longer section driven by DC induces strong gain saturation primarily to the low-frequency components, and thus functions effectively as a high-pass filter under DC bias [15,16].
3. Self-seeded experiment using two-section RSOA
To verify the numerical prediction and effect of broadened modulation bandwidth presented in the previous section, self-seeded experiment was carried. Figure 6 shows the experimental setup. The DC bias I1 and 231-1 pseudorandom-bit-sequence (PRBS) data were applied on Section 1, whereas Section 2 was driven by only DC current I2. For simplicity, optical circulator and polarization controller (PC) were used for managing the polarization state instead of Faraday rotator (FR) and Faraday rotator mirror (FRM). The round-trip loss of entire self-seeded cavity was measured to be around 12 dB. The transmitted signal was received by a pre-amplified receiver before or after 25-km SSMF. The peak-to-peak modulation voltage VPP was optimized to obtain lowest BER in each case. All characterizations were performed at 25°C.
Fig. 6 Setup for self-seeded RSOA experiment. (VOA: variable optical attenuator, EDFA: erbium-doped fiber amplifier, PD: photodetector).
Figure 7 shows the eye diagrams and corresponding optical spectra measured at 2.5 Gb/s under three different bias conditions (I1,I2) = (50 mA,60 mA), (37 mA,73 mA), and (27 mA,83 mA), corresponding to A, B, and C in Fig. 3(c), with respective current density ratio J2/J1 = 0.6, 1.0, and 1.5. The optical power launched into the fiber from the RSOA is around 4 dBm. Clear eye diagram with a largest extinction ratio of 5.9 dB is obtained both before and after 25-km transmission when J2/J1 = 0.6 [Fig. 7(a)], which is consistent with simulation [Fig. 3(c)]. Thanks to the large MCDR, we could increase VP-P to 2.8 V to have higher extinction ratio. In contrast, when J2/J1 = 1 or 1.5, extinction ratio is decreased to 5.4 dB and 5.1 dB due to reduced MCDR that allows VPP of only 2.4 V and 2.0 V, respectively [Fig. 7(b) and 7(c)]. Figure 8 shows the BER curves before and after transmission for three cases. As expected, we achieve the lowest BER of 2.0 × 10−10 and 3.2 × 10−7 before and after 25-km transmission when J2/J1 = 0.6.
Fig. 7 Measured optical spectra and eye diagrams before (BTB) and after 25 km transmission at 2.5 Gb/s on three different DC bias conditions: (I1,I2) of (a) (50 mA,60mA), (b) (37mA,73mA), and (c) (27 mA,83mA).
Fig. 8 BER curves measured at 2.5 Gb/s before (BTB) and after 25-km transmission on three different DC bias conditions. The letters A, B, and C in the legend corresponds to the operating bias points shown in Fig. 3.
Finally, we test the self-seeded transmission experiment at higher bit rate. Figure 9 shows the BER characteristics and eye diagrams measured at 5 Gb/s and 10 Gb/s when (I1, I2) = (90 mA,80 mA), corresponding to the point D (J2/J1 = 0.6) in Fig. 3(c). Note that the total current is increased to 170 mA to achieve higher optical gain and broader bandwidth in these experiments. As a result, the launched optical power into the fiber is also increased to around 7 dBm. Owing to the optimized driving condition and short section used for modulation, BER of 1.0 × 10−8 and 4.3 × 10−3 are achieved at 5 Gb/s and 10 Gb/s for back-to-back (BTB), respectively. In addition, transmission over 25-km SSMF is achieved with BER of 7.8 × 10−4at 5 Gb/s. On the other hand, transmission at 10 Gb/s is strictly hindered by the large group velocity dispersion (GVD) of SSMF in C band. For comprehensive comparison, Fig. 10 shows the measured eye diagrams at 10 Gb/s when (I1, I2) = (57 mA, 113 mA) and (43 mA, 127 mA), corresponding to the operation points E (J2/J1 = 1) and F (J2/J1 = 1.5) in Fig. 3(c). It is confirmed that the eyes are distorted seriously due to insufficient modulation cancellation. BER below 1 × 10−2is hardly achieved when J2/J1 = 1.0 or 1.5 at 10 Gb/s.
Fig. 9 Measured BER curves and eye diagrams at 5 Gb/s and 10 Gb/s before (BTB: back-to-back) and after 25-kmtransmission for (I1,I2) = (90 mA,80mA), corresponding to D in Fig. 3.
4. Conclusion
We have numerically and experimentally demonstrated the use of two-section RSOA for colorless self-seeded transmitter to enhance the modulation cancellation effect and extend the modulation bandwidth. Through detailed numerical analyses based on the rate equations and wave-propagation equations, we have revealed the existence of optimal bias condition to provide enhanced modulation cancellation. Increase in modulation bandwidth beyond 10 GHz was also confirmed experimentally by the use of two-section RSOA. With the derived optimal bias condition and extended modulation bandwidth, we have successfully demonstrated self-seeded modulation up to 10 Gb/s as well as transmission over 25-km SSMF at 5 Gb/s. While data transmission at 10 Gb/s was hindered by the GVD of SSMF at C band, the demonstrated scheme of segmenting the electrode should be applicable in extending the bandwidth of O-band RSOAs [10], which should enable >10-Gb/s self-seeded transmitter for the next-generation low-cost WDM access networks.
References and links
1. I. Kaminow and T. Li, Optical Fiber Telecommunications Volume VIB, Sixth Edition, 1st ed. (Academic, 2013), Chap. 23.
2. D. Nesset, “NG-PON2 technology and standards,” J. Lightwave Technol. 33(5), 1136–1143 (2015).
3. Y. Horiuchi, “Economical solutions of the WDM-PON system,” in Optical Fiber Communication Conference,2012, paper OW3B.6.
4. J. Buus and E. Murphy, “Tunable lasers in optical networks,” J. Lightwave Technol. 24(1), 5–11 (2006).
5. S. G. Mun, J. H. Moon, H. K. Lee, J. Y. Kim, and C. H. Lee, “A WDM-PON with a 40 Gb/s (32 x 1.25 Gb/s) capacity based on wavelength-locked Fabry-Perot laser diodes,” Opt. Express 16(15), 11361–11368 (2008). [PubMed]
6. P. Healey, P. Townsend, C. Ford, L. Johnston, P. Townley, I. Lealman, L. Rivers, S. Perrin, and R. Moore, “Spectral slicing WDM-PON using wavelength-seeded reflective SOAs,” Electron. Lett. 37(19), 1181–1182 (2001).
7. E. Wong, K. Lee, and T. Anderson, “Directly modulated self-seeding reflective semiconductor optical amplifiers as colorless transmitters in wavelength division multiplexed passive optical networks,” J. Lightwave Technol. 25(1), 67–74 (2007).
8. P. Parolari, L. Marazzi, M. Brunero, A. Gatto, M. Martinelli, P. Chanclou, Q. Deniel, F. Saliou, S. Le, R. Brenot, S. Barbet, F. Lelarge, S. Gebrewold, S. O’Duill, D. Hillerkuss, J. Leuthold, G. Gavioli, and P. Galli, “Self-tuning transmitter for fibre-to-the-antenna PON networks,” Opt. Switching Networking 14(1), 25–31 (2014).
9. F. Saliou, G. Simon, P. Chanclou, A. Pizzinat, H. Lin, E. Zhou, and Z. Xu, “WDM PONs based on colorless technology,” Opt. Fiber Technol. 26(A), 126– 134 (2015).
10. P. Parolari, L. Marazzi, M. Brunero, M. Martinelli, R. Brenot, A. Maho, S. Barbet, G. Gavioli, G. Simon, S. Le, F. Saliou, and P. Chanclou, “C- and O-band operation of RSOA WDM PON self-seeded transmitters up to 10 Gb/s,” J. Opt. Commun. Netw. 7(2), A249–A255 (2014).
11. S. Ó. Dúill, L. Marazzi, P. Parolari, R. Brenot, C. Koos, W. Freude, and J. Leuthold, “Efficient modulation cancellation using reflective SOAs,” Opt. Express 20(26), B587–B594 (2012). [PubMed]
12. S. A. Gebrewold, L. Marazzi, P. Parolari, R. Brenot, S. P. Ó. Dúill, R. Bonjour, D. Hillerkuss, C. Hafner, and J. Leuthold, “Reflective-SOA fiber cavity laser as directlymodulated WDM-PON colorless transmitter,” IEEE J. Sel. Top. Quantum Electron. 20(5), 3100409 (2014).
13. W. Zhan, P. Zhou, Y. Zeng, M. Mukaikubo, T. Tanemura, and Y. Nakano, “Optimization of modulation-cancelling reflective semiconductor optical amplifier for colorless WDM transmitter applications,” J. Lightwave Technol. 35(2), 274–279 (2017).
14. P. Zhou, W. Zhan, T. Tanemura, M. Mukaikubo, and Y. Nakano, “Two-section RSOA with enhanced modulation-cancelling effect for self-seeded colorless WDM transmitter,” in Proceedings of European Conference on Optical Communication (ECOC), paper M.2.E (2016).
15. R. Brenot, J. Provost, O. Legouezigou, J. Landreau, F. Pommereau, F. Poingt, L. Legouezigou, E. Derouin, O. Drisse, B. Rousseau, F. Martin, F. Lelarge, and G. Duan, “High modulation bandwidth reflective SOA for optical access networks,” in Proceedings of European Conference on Optical Communication (ECOC), paper Thur.10.3.6 (2007).
16. H. S. Kim, B. S. Choi, K. S. Kim, D. C. Kim, O. K. Kwon, and D. K. Oh, “Improvement of modulation bandwidth in multisection RSOA for colorless WDM-PON,” Opt. Express 17(19), 16372–16378 (2009). [PubMed]
17. L. Huang, W. Hong, and G. Jiang, “All-optical power equalization based on a two-section reflective semiconductor optical amplifier,” Opt. Express 21(4), 4598–4611 (2013). [PubMed]
