Open Access
Add to BibSonomyAbstract
Wavelength conversion of 40 Gb/s and 80 Gb/s return-to-zero on-off-keying signals using a quantum-dot semiconductor optical amplifier in combination with a delay interferometer as subsequent filter is demonstrated. The performance of the 80 Gb/s wavelength converter measured in terms of the bit-error ratio demonstrated here is the highest reported up to now for quantum-dot semiconductor optical amplifiers. The typical fast gain dynamics manifests itself in open eye diagrams of the converted signal. The slow phase dynamics of the carrier reservoir however induces severe patterning and requires compensation. Adaptation of the free-spectral range of the delay interferometer is necessary in order to mitigate these phase effects and to achieve error-free wavelength conversion.
©2011 Optical Society of America
1. Introduction
Wavelength conversion at high bit rates is decisive for future wavelength division multiplexing based networks [1]. The ability of switching between wavelength channels enhances the flexibility of reconfigurable optical networks by prevention of blocking issues due to simultaneous transmission of different information on the same wavelength channel. All-optical wavelength conversion is also advantageous in terms of transparency and enables low switching power. Semiconductor optical amplifiers (SOAs) are particularly suitable for such signal processing, since they feature fast nonlinearities besides offering the possibility of integration with other components and having a small footprint.
Compared to conventional quantum-well or bulk media, a quantum-dot (QD) gain medium offers beneficial properties for SOAs [2,3]. The most prominent advantages are a broad gain bandwidth originating from the distribution in size, composition and strain of the QDs [4], high saturation output power levels [5], and ultra-fast dynamics within the QDs [6,7], enabling amplification as well as signal processing at 40 Gb/s [8–10] and beyond [11,12]. Wavelength conversion via cross-gain modulation (XGM) was demonstrated by eye diagram quality measurements at 40 Gb/s [13], at 80 Gb/s for multicast conversion [14], and at 160 Gb/s [15]. To increase the extinction ratio and to reinvert the cross-gain and cross-phase modulated converted SOA output signal, an optical filtering scheme is needed. One commonly used approach is offset filtering as shown up to 40 Gb/s for QD SOAs [16] and in the range from 40 Gb/s up to 320 Gb/s for conventional SOAs [17–19]. Another filter scheme is based on a delay interferometer (DI) [20–22].
Speed limitations of conventional SOAs are overcome by filtering and the inherent patterning effects resulting from the slow gain recovery time are mitigated by exploitation and balancing of cross-gain and cross-phase modulation. In the case of a QD SOA based wavelength converter the mitigation of patterning effects is not the primary intention, since QD SOAs already have demonstrated their fast dynamics as mentioned above. Instead the filter aims for improvement of the extinction ratio and reinversion of the converted signal, as comprehensively discussed in [23]. If only the gain dynamics of the QDs, i.e. spectral hole burning is taken into account, it is predicted that the best performance in terms of extinction ratio improvement and reinversion is achieved, if the delay or the corresponding free-spectral range (FSR) of the DI equals the bit rate [23].
In this paper, error-free wavelength conversion of 40 Gb/s and 80 Gb/s return-to-zero on-off-keying signals using a QD SOA and subsequent filtering with a DI is demonstrated. A fast gain response in the range of a few ps is found for the QD SOA resulting in open eye diagrams. However, a weak contribution from slow phase effects prevents the use of a one-bit DI to simply reinvert the XGM signal. Instead, the patterning effects induced by cross-phase modulation (XPM) require the use of a DI with a small time offset. Thus, error-free wavelength conversion with a QD SOA combined with a DI is shown for the first time by bit error ratio measurements at 80 Gb/s.
2. Experimental setup
The setup to perform wavelength conversion measurements at 40 Gb/s or 80 Gb/s return-to-zero (RZ) on-off keying (OOK) data signals consists of a tunable mode-locked laser (TMLL) (Fig. 1 ) emitting a pulse train at 9.953 GHz repetition rate at 1310 nm. The pulse train is multiplexed to 39.812 GHz using a dual stage optical time division multiplexer (OMUX). Quantum-well (QW) SOAs boost the pulse train to compensate for the losses induced by the multiplexer stages. A 231-1 pseudo random binary sequence (PRBS) is imposed on the pulse train via a Mach-Zehnder modulator (MZM). A second OMUX can be used to double the bit rate from 39.8 Gb/s (~40 Gb/s) to 79.6 Gb/s (~80 Gb/s). Two praseodymium-doped fiber amplifiers (PDFAs) amplify the data signal to a maximum power of 12.5 dBm at the QD SOA input. The PDFAs ensure pattern-effect free amplification of the modulated data signal. The pulse full width at half maximum (FWHM) is 2.0 ps at the input of the QD SOA. The data signal is combined with a continuous wave (cw) probe signal from a wavelength tunable external cavity laser (ECL). The cw probe power launched into the QD SOA is 9.5 dBm.
Fig. 1 Sketch of the setup for 80 Gb/s RZ OOK wavelength conversion. TMLL: tunable mode-locked laser, QW SOA: quantum well semiconductor optical amplifier, OMUX: optical time division multiplexer, MZM: Mach-Zehnder modulator, PDFA: praseodymium-doped fiber amplifier, VOA: variable optical attenuator, ECL: external cavity laser, DI: delay interferometer, and ETDM DEMUX: electrical time division demultiplexer.
The wavelength conversion takes place in the QD SOA and in consequence the inverted modulation is transferred by means of XGM and XPM onto the probe. The filter after the QD SOA is used to block the pump signal; the probe signal is explicitly not offset filtered. After amplification in a PDFA, the wavelength converted output of the QD SOA is reinverted via a tunable DI and analyzed by eye diagram and bit-error ratio measurements.
The 80 Gb/s receiver consists of two PDFAs in combination with a 50 GHz photo diode (u2t XPDV 2320R). In the case of measuring 80 Gb/s signals the output of the photo diode is demultiplexed in the electrical domain to 40 Gb/s in front of the error analyzer.
The back-to-back eye diagrams for maximum receiver input power are shown in Fig. 2 for (a) 40 Gb/s and (b) 80 Gb/s. The limited temporal resolution of the photo diode and the autocorrelation measurements at 80 Gb/s have proven the pulses to be clearly distinct.
Fig. 2 Back-to-back eye diagrams at 1310 nm of the PRBS 231-1 RZ OOK data signal at (a) 40 Gb/s and (b) 80 Gb/s.
The nonlinear element mediating wavelength conversion is a 2 mm long QD SOA. The device consists of a 4 µm broad and shallow etched ridge waveguide structure. The active region incorporates ten In(Ga)As QD layers and is p-doped. At 1310 nm, a small signal fiber-to-fiber gain of 7 dB and 5 dB at 250 mA and 500 mA, respectively, has been observed (Fig. 3 ). Gain is predominantly provided for the TE polarization (TE/TM ~10 dB). Thus, polarization controllers are required in front of the QD SOA. The QD SOA is operated up to a current of 600 mA, presently limited by heating of the device, which is responsible for the decrease of the small-signal gain, if the current is increased from 250 mA to 500 mA. Nevertheless, a large current is desirable, since it is the decisive parameter to achieve high speed gain dynamics [24] and thus to minimize patterning.
Fig. 3 (a) Gain of the QD SOA versus output power at 250 mA and 500 mA. (b) Pump-probe trace showing gain and phase recovery at 550 mA using 1.6 ps pump pulses at 10 GHz repetition rate.
Pump-probe traces of the gain and the phase recovery are shown in Fig. 3(b). The power of the 10 GHz pump pulse train at 1310 nm is chosen to have roughly the same pulse energy as in the 80 Gb/s wavelength conversion measurement described in section 3. The measurement demonstrates the fast gain dynamics with a 90%-10%-recovery time of less than 10 ps [25]. In contrast to the fast gain recovery, the phase does not fully recover within the pulse repetition period of 100 ps. For similar QD SOAs a 90%-10% phase recovery time as large as 500 ps was observed [6]. A phase change of 0.45 rad is found.
3. Experimental results
Eye diagrams of the converted signal right behind the SOA at 1320 nm are presented in Fig. 4(a) for 40 Gb/s and in Fig. 5(a) for 80 Gb/s, respectively, which are recorded without the DI in use. The fast gain dynamics manifests itself in a low broadening of the mark level of the inverted signal and in a reasonable eye opening, however with a low extinction ratio of approximately 2 dB at 40 Gb/s (1.5 dB at 80 Gb/s) as expected from the gain compression shown in Fig. 3(b). Larger extinction ratios have been published in [15], however for a different SOA structure based on columnar QDs. The eye opening demonstrates that the gain dynamics is fast enough for 80 Gb/s RZ-OOK signal processing.
Fig. 4 (a) Eye diagram of the converted 40 Gb/s signal at 1320 nm right behind the QD SOA without the DI in use. (b) Optical spectrum after the QD SOA (black) and after the DI with 240 GHz FSR (red).
Fig. 5 (a) Eye diagram of the converted 80 Gb/s signal at 1320 nm after the QD SOA without the DI. (b) Normalized optical spectra of the QD SOA output (black) and after the DI (red) for conversion towards 1320 nm. The first right sideband is weaker by 4.3 dB than the first left sideband.
The optical output spectra of the QD SOA and the DI at 40 Gb/s (Fig. 4 (b)) and 80 Gb/s (Fig. 5(b)) provide indication for significant phase effects, since the intensity of sidebands with respect to the cw carrier is asymmetric. At 80 Gb/s, the first left sideband is 4.3 dB larger than the first right sideband. The modulation sidebands are 23 dB below the cw carrier, which explains the poor extinction ratio of the converted signal right at the output of the QD SOA seen in Fig. 4(a) and Fig. 5(a). These observations imply in combination with the phase recovery found in the pump-probe measurement (Fig. 3(b)) that the QD SOA induces in addition to XGM also a significant amount of XPM. As shown in Fig. 3(b), the phase recovery takes place on a time scale exceeding 100 ps (consistent with [6]) which is thus much slower than the gain dynamics. The slow refilling of the reservoir in the higher energy states dominates the phase dynamics rather than the QD ground state itself, as shown by simulations in [26].
The phase patterning becomes important, if the QD SOA output signal is filtered with the DI, since it processes both, amplitude and phase. Setting the FSR of the DI equal to the bit rate results in a strongly deteriorated and closed eye diagram of the DI output as shown in Fig. 6 for the 80 Gb/s data signal. In [23] only amplitude modulation due to spectral hole burning was taken into account, but although phase effects of present QD SOAs are rather weak, they significantly contribute to the filtered output and can obviously not be neglected in this experiment.
The principle of signal processing by the DI is sketched in Fig. 7 . The choice of a suitable FSR enables compensation of the phase patterning.
Fig. 7 Sketch of the signal processing by the QD SOA and the DI. The QD SOA induces XGM and slow XPM. The non-recovered phase results in patterning effects, if the FSR equals the bit rate. Wavelength conversion with low pattering due to compensation of phase patterning is feasible at large FSR values. The dotted line represents the delayed copy of the solid trace.
If the FSR of the DI equals the bit rate (Fig. 7, case a)), successive pulses interfere at the output of the DI. Although the gain is fast enough to recover sufficiently between successive pulses, the phase recovers on a time scale much longer than the bit period. Since the absolute phase and the phase difference between the pulses strongly depend on the preceding data sequence, the uncompensated phase patterning results in a strongly deteriorated output eye diagram.
In contrast to that, a FSR larger than the bit rate (Fig. 7, case b)) causes the pulse to interfere with its time-delayed copy. A larger FSR corresponds to a shorter temporal delay, which in turn determines a temporal switching window. The use of a switching window allows mitigation of the phase induced patterning effects, since the phase difference of the two copies of the same bit is less dependent on the preceding bits than the phase difference of neighboring bits. The corresponding optimum delay in our experiments of 4.2 ps (240 GHz FSR) at 40 Gb/s and 3.1 ps (320 GHz FSR) at 80 Gb/s represents a tradeoff between the formation of sufficiently short pulses and the output power of the DI. Moreover, the difference between the phase of the initial signal and its time delayed copy has to be sufficiently small after the initial switching window (i.e. beyond the delay time) to prevent the generation of a second switching window. Otherwise, trailing pulses would be generated which may cause intersymbol interference. Since the phase recovery time of the QD SOA tested here is found to be slow, a large FSR (short delay) is required.
Autocorrelation measurements of the 80 Gb/s DI output signal for a FSR of 320 GHz exhibit a triangular shape indicating square-like pulses in the time domain.
In the wavelength conversion experiments shown here, the DI is adjusted for maximum suppression of the cw carrier. FSR values of 240 GHz and 320 GHz are chosen to filter the 40 Gb/s and 80 Gb/s data signals, respectively. The spectra of the DI output at 40 Gb/s in Fig. 4(b) (at 80 Gb/s in Fig. 5(b)) prove that the asymmetry of the sidebands is unaffected by filtering, whereas the sideband-to-carrier ratio is tremendously enhanced to 2.3 dB at 40 Gb/s (1.7 dB at 80 Gb/s) by means of the DI. The large FSR values result in a significant improvement of the output eye diagrams of the wavelength converter at 40 Gb/s (Fig. 8(a) ) and 80 Gb/s (Fig. 9(a) ) for conversion from 1310 nm to 1320 nm. Extinction ratios of 10 dB and 9.6 dB are measured at 40 Gb/s and 80 Gb/s, respectively. The tunable band pass filters were not used as offset filters.
Fig. 8 (a) Eye diagram after DI for maximum receiver input power. (b) Bit-error ratio versus received power for 40 Gb/s conversion from 1310 nm to 1320 nm.
Fig. 9 (a) Eye diagram of the converted 80 Gb/s data signal at 1320 nm after the DI with 320 GHz FSR. The extinction ratio is 9.3 dB. (b) BER versus received power for 80 Gb/s RZ-OOK wavelength conversion from 1310 nm to 1320 nm.
Bit-error ratio (BER) measurements of the 40 Gb/s output signal of the wavelength converter (conversion from 1310 nm to 1320 nm) demonstrate error-free performance without an error floor down to 10−12 (Fig. 8(b)). In comparison to the back-to-back measurement a penalty of 2.5 dB is found at a BER of 10−9. The reduced receiver sensitivity at 1320 nm compared to 1310 nm causes 0.9 dB penalty, the remaining penalty of 1.6 dB could be attributed to not fully compensated phase patterning and to the lower extinction ratio of the signal at the output of the wavelength converter.
Wavelength conversion of an 80 Gb/s RZ-OOK signal (Fig. 9(b)) from 1310 nm towards 1320 nm shows error-free (BER < 10−9) operation. A penalty of 4.8 dB is found compared to the back-to-back measurement. However, an error floor appears at a BER of about 10−10. OSNR measurements confirm that this error floor in the BER measurement versus received optical power is caused by a limited OSNR after the wavelength converter. Neglecting the error floor yields an estimated power penalty of about 2.5 dB at 10−9. Similar to the measurements at 40 Gb/s, this includes 0.9 dB penalty from the reduced receiver sensitivity at the wavelength of the converted signal. The remaining penalty of about 1.6 dB can again be attributed to the limited extinction ratio and the uncompensated phase patterning which results in an amplitude patterning effect due to phase-to-amplitude conversion in the DI.
The measurements presented here demonstrate that fast gain dynamics of present QD SOAs enable wavelength conversion at high bit rates with low amplitude patterning. However, a filter, here a DI, is used to reinvert the signal and to improve the extinction ratio. The important finding is, that the DI has to be operated at a large FSR in order to mitigate phase patterning effects resulting from the slow carrier dynamics of the reservoir. Although the phase effects in QD SOAs are commonly found to be smaller than in conventional devices [6], they are still sufficient to allow phase based wavelength conversion. Therefore, for conversion using present QD SOAs the slow phase dynamics cannot be neglected and use of the fast gain dynamics only is not sufficient for high speed wavelength conversion, similar to the conversion using conventional amplifiers [27].
The advantages of the wavelength converter used in this experiment are, that filtering with a DI avoids a format conversion for reinversion of the converted signal and does not require an additional offset filter. A QD SOA and a single DI could easily be integrated into a single chip wavelength converter.
4. Conclusion
Error-free (BER < 10−9) 80 Gb/s RZ-OOK all optical wavelength conversion from 1310 nm to 1320 nm using a QD SOA combined with a DI as subsequent filter is demonstrated for the first time. However, indication of an error floor is found at a BER of 10−10. 40 Gb/s wavelength conversion performs error-free without an error floor down to a BER of 10−12. The fast gain dynamics of the QDs result in open eye diagrams of the converted output signal of the QD SOA in front of the DI. However, the slow carrier dynamics of the reservoir, which dominates the phase dynamics of the QD SOA, requires the application of the DI in a phase scheme, i.e. with large FSR values (320 GHz at 80 Gb/s). The wavelength conversion presented here allows to avoid a format conversion, and the fast gain dynamics enable maximum suppression of the carrier. Offset-filtering with an additional optical band pass filter is not required.
Acknowledgments
This work was supported by the research center SFB 787 of the German Research Foundation (DFG) and the European Commission’s FP7 Network of Excellence EUROfos.
References and links
1. S. J. B. Yoo, “Wavelength conversion technologies for WDM network applications,” J. Lightwave Technol. 14(6), 955–966 (1996). [CrossRef]
2. D. Bimberg, M. Kuntz, and M. Laemmlin, “Quantum dot photonic devices for lightwave communication,” Microelectron. J. 36(3-6), 175–179 (2005). [CrossRef]
3. D. Bimberg, G. Fiol, M. Kuntz, C. Meuer, M. Lammlin, N. N. Ledentsov, and A. R. Kovsh, “High speed nanophotonic devices based on quantum dots,” Phys. Status Solidi., A Appl. Mater. Sci. 203(14), 3523–3532 (2006). [CrossRef]
4. D. Bimberg, M. Grundmann, N. N. Ledentsov, S. S. Ruvimov, P. Werner, U. Richter, J. Heydenreich, V. M. Ustinov, P. S. Kopev, and Z. I. Alferov, “Self-organization processes in MBE-grown quantum dot structures,” Thin Solid Films 267(1-2), 32–36 (1995). [CrossRef]
5. A. V. Uskov, E. P. O'Reilly, M. Laemmlin, N. N. Ledentsov, and D. Bimberg, “On gain saturation in quantum dot semiconductor optical amplifiers,” Opt. Commun. 248(1-3), 211–219 (2005). [CrossRef]
6. T. Vallaitis, C. Koos, R. Bonk, W. Freude, M. Laemmlin, C. Meuer, D. Bimberg, and J. Leuthold, “Slow and fast dynamics of gain and phase in a quantum dot semiconductor optical amplifier,” Opt. Express 16(1), 170–178 (2008). [CrossRef] [PubMed]
7. I. O‘Driscoll, T. Piwonski, J. Houlihan, G. Huyet, R. J. Manning, and B. Corbett, “Phase dynamics of InAs∕GaAs quantum dot semiconductor optical amplifiers,” Appl. Phys. Lett. 91(26), 263506 (2007). [CrossRef]
8. M. Sugawara, N. Hatori, M. Ishida, H. Ebe, Y. Arakawa, T. Akiyama, K. Otsubo, Y. Yamamoto, and Y. Nakata, “Recent progress in self-assembled quantum-dot optical devices for optical telecommunication: temperature-insensitive 10 Gb s-1 directly modulated lasers and 40 Gb s-1 signal-regenerative amplifiers,” J. Phys. D Appl. Phys. 38(13), 2126–2134 (2005). [CrossRef]
9. G. Contestabile, A. Maruta, S. Sekiguchi, K. Morito, M. Sugawara, and K. Kitayama, “Regenerative amplification by using self-phase modulation in a quantum-dot SOA,” IEEE Photon. Technol. Lett. 22(7), 492–494 (2010). [CrossRef]
10. R. Bonk, C. Meuer, T. Vallaitis, S. Sygletos, P. Vorreau, S. Ben-Ezra, S. Tsadka, A. Kovsh, I. Krestnikov, M. Laemmlin, D. Bimberg, W. Freude, and J. Leuthold, “Single and multiple channel operation dynamics of linear quantum-dot semiconductor optical amplifier,” in European Conference on Optical Communications (ECOC 2008) (Brussels, Belgium, 2008), p. Th.1.C.2.
11. C. Schmidt-Langhorst, C. Meuer, A. Galperin, H. Schmeckebier, R. Ludwig, D. Puris, D. Bimberg, K. Petermann, and C. Schubert, “80 Gb/s multi-wavelength booster amplification in an InGaAs/GaAs quantum-dot semiconductor optical amplifier,” in European Conference on Optical Communication (ECOC 2010) (Torino, Italy, 2010), p. Mo.1.F.6.
12. G. Contestabile, A. Maruta, S. Sekiguchi, K. Morito, M. Sugawara, and K. Kitayama, “Regenerative amplification in a quantum dot SOA ” in Optical Fiber Communication Conference (OFC 2010) (San Diego, CA, USA, 2010), p. OMT2.
13. T. Akiyama, N. Hatori, Y. Nakata, H. Ebe, and M. Sugawara, “Pattern-effect-free amplification and cross-gain modulation achieved by using ultrafast gain nonlinearity in quantum-dot semiconductor optical amplifiers,” Phys. Status Solidi, B Basic Res. 238(2), 301–304 (2003). [CrossRef]
14. G. Contestabile, A. Maruta, S. Sekiguchi, K. Morito, and K. Kitayama, “80 Gb/s Multicast Wavelength Conversion by XGM in a QD-SOA,” in European Conference on Optical Communication (ECOC2010) (Torino, Italy, 2010), p. Mo.2.A.3.
15. G. Contestabile, A. Maruta, S. Sekiguchi, K. Morito, M. Sugawara, and K. Kitayama, “160 Gb/s cross gain modulation in quantum dot SOA at 1550 nm,” in European Conference on Optical Communication (ECOC 2009) (Vienna, Austria, 2009), p. PDP 1.4.
16. O. Raz, J. Herrera, N. Calabretta, E. Tangdiongga, S. Anantathanasarn, R. Nötzel, and H. J. S. Dorren, “Non-inverted multiple wavelength converter at 40 Gbit/s using 1550 nm quantum dot SOA,” Electron. Lett. 44(16), 988–989 (2008). [CrossRef]
17. J. Leuthold, R. Ryf, D. N. Maywar, S. Cabot, J. Jaques, and S. S. Patel, “Nonblocking all-optical cross connect based on regenerative all-optical wavelength converter in a transparent network demonstration over 42 nodes and 16800 km,” J. Lightwave Technol. 21(11), 2863–2870 (2003). [CrossRef]
18. M. L. Nielsen, B. Lavigne, and B. Dagens, “Polarity-preserving SOA-based wavelength conversion at 40 Gbit/s using bandpass filtering,” Electron. Lett. 39(18), 1334–1335 (2003). [CrossRef]
19. Y. Liu, E. Tangdiongga, Z. Li, H. de Waardt, A. M. J. Koonen, G. D. Khoe, X. W. Shu, I. Bennion, and H. J. S. Dorren, “Error-free 320-Gb/s all-optical wavelength conversion using a single semiconductor optical amplifier,” J. Lightwave Technol. 25(1), 103–108 (2007). [CrossRef]
20. Y. Ueno, S. Nakamura, K. Tajima, and S. Kitamura, “3.8-THz wavelength conversion of picosecond pulses using a semiconductor delayed-interference signal-wavelength converter (DISC),” IEEE Photon. Technol. Lett. 10(3), 346–348 (1998). [CrossRef]
21. J. Leuthold, B. Mikkelsen, G. Raybon, C. H. Joyner, J. L. Pleumeekers, B. I. Miller, K. Dreyer, and R. Behringer, “All-optical wavelength conversion between 10 and 100 Gb/s with SOA delayed-interference configuration,” Opt. Quantum Electron. 33(7/10), 939–952 (2001). [CrossRef]
22. J. Leuthold, D. M. Marom, S. Cabot, J. J. Jaques, R. Ryf, and C. R. Giles, “All-optical wavelength conversion using a pulse reformatting optical filter,” J. Lightwave Technol. 22(1), 186–192 (2004). [CrossRef]
23. S. Sygletos, R. Bonk, T. Vallaitis, A. Marculescu, P. Vorreau, J. S. Li, R. Brenot, F. Lelarge, G. H. Duan, W. Freude, and J. Leuthold, “Filter assisted wavelength conversion with quantum-dot SOAs,” J. Lightwave Technol. 28(6), 882–897 (2010). [CrossRef]
24. C. Meuer, J. Kim, M. Laemmlin, S. Liebich, G. Eisenstein, R. Bonk, T. Vallaitis, J. Leuthold, A. Kovsh, I. Krestnikov, and D. Bimberg, “High-speed small-signal cross-gain modulation in quantum-dot semiconductor optical amplifiers at 1.3 µm,” IEEE J. Sel. Top. Quantum Electron. 15(3), 749–756 (2009). [CrossRef]
25. C. Schmidt-Langhorst, C. Meuer, R. Ludwig, D. Puris, R. Bonk, T. Vallaitis, D. Bimberg, K. Petermann, J. Leuthold, and C. Schubert, “Quantum-dot semicondcutor optical booster amplifier with ultra-fast gain recovery for pattern-effect free amplification of 80 Gb/s RZ-OOK data signals,” in European Conference on Optical Communication (ECOC 2009) (Vienna, Austria, 2009), p. 6.2.1.
26. J. Kim, C. Meuer, D. Bimberg, and G. Eisenstein, “Role of carrier reservoirs on the slow phase recovery of quantum dot semiconductor optical amplifiers,” Appl. Phys. Lett. 94(4), 041112 (2009). [CrossRef]
27. I. Kang, C. Dorrer, L. Zhang, M. Dinu, M. Rasras, L. L. Buhl (Larry), S. Cabot, A. Bhardwaj, X. Liu, M. A. Cappuzzo, L. Gomez, A. Wong-Foy, Y. F. Chen, N. K. Dutta, S. S. Patel, D. T. Neilson, C. R. Giles, A. Piccirilli, and J. Jaques, “Characterization of the dynamical processes in all-optical signal processing using semiconductor optical amplifiers,” IEEE J. Sel. Top. Quantum Electron. 14(3), 758–769 (2008). [CrossRef]
