Self-Phase Modulation Effect on Performance of 40 Gbit=s Optical Duty-Cycle Division Multiplexing Technique

Transcription

1 J. Opt. Commun., Vol. (2), xxx xxx Copyright 2 De Gruyter. DOI /joc-2- Self-Phase Modulation Effect on Performance of 4 Gbit=s Optical Duty-Cycle Division Multiplexing Technique Ghafour Amouzad Mahdiraji 1; and Ahmad Fauzi Abas 2 1 Department of Electrical Engineering, University of Malaya, Kuala Lumpur, Malaysia 2 Photonics and Fiber Optic Systems Laboratory, Centre of Excellence for Wireless and Photonics Networks, Engineering and Technology Complex, University Putra Malaysia, Serdang, Malaysia Abstract. Self-phase modulation (SPM) effect on performance of 4 Gbit=s ( 1. Gbit=s) optical duty-cycle division multiplexing (O-DCDM) over 4 km (5 8 km) standard single mode fiber (SSMF) based on three different dispersion compensation schemes is investigated by simulation. The result shows that the SPM effect is very strong in dispersion pre-compensation compared to postcompensation. Further reduction in SPM effect is observed by combination of pre- and post-compensation with the optimum pre-compensation value of around 215 ps=nm. The simulation is performed at different launched power into dispersion compensation fiber (DCF). The range between to 8 dbm is observed as the optimum power for all dispersion compensation schemes. The highest SPM threshold is observed when dispersion pre- and post-compensation are used together, which is around C:7 dbm. In addition, the optimum range of launched power into SSMF for different transmission distances from 8 km to 8 km is also presented. Keywords. Optical fiber communication, fiber nonlinear effect, self-phase modulation (SPM), duty-cycle division multiplexing (DCDM). PACS (21) Sz, 42.5.Jx. 1 Introduction Optical fiber impairments are the main factors that need to be considered in the transmission systems. By developing the advanced Erbium doped fiber amplifiers (EDFAs) * Corresponding author: Ghafour Amouzad Mahdiraji, Department of Electrical Engineering, University of Malaya, 5 Kuala Lumpur, Malaysia; ghafouram@gmail.com. Received: September 8, 211. Accepted: November 24, 211. with low noise figure, large gain, and high amplification bandwidth, the power lost in the optical fiber can be compensated [1 5]. The chromatic dispersion of standard single mode fiber (SSMF) can be managed by employing dispersion compensation fiber (DCF) [ 8]. Thus, the optical nonlinear effects produced by SSMF and DCF are remained as the major impairment that need to be addressed especially in the high bit rate, long-haul transmission systems [9 1]. It has been shown that self-phase modulation (SPM) and cross-phase modulation (XPM) are the major sources of nonlinearities [14, 17 2]. Duty-cycle division multiplexing (DCDM) has been proposed as a multiplexing technique that can multiplex multiple users per wavelength by signing different return-to-zero (RZ) duty-cycles for different users [21 27]. The output of this multiplexing technique is a multilevel step-down shape signal that can be demultiplexed by direct detection. The multiplexing can be done either in the optical domain (referred as O-DCDM [21, 2]) or electrical domain (referred as E-DCDM [22 25, 27]). Generally, the performance of O-DCDM is better than E-DCDM. Similar to M -ary signaling, DCDM signals have higher intensity compared to binary based signals. Thus, investigation of nonlinear effect on performance of DCDM is necessary for further development of the system. Performance of 1 Gbit=s O-DCDM with electroabsorption modulator (EAM) and Mach Zehnder modulator (MZM) over single wavelength system is presented in [21] and [2], respectively, which O-DCDM with MZM outperformed the former. Since the nonlinearity due to XPM is significant only in multi-wavelength system [28], in this paper, only SPM effect is investigated here. Performance is evaluated with three different symmetric dispersion compensation schemes, and different value of launched power into DCFs. In addition, the optimum range of launched power into SSMF over different transmission length (8 km to 8 km) for the best dispersion compensation scheme is presented. 2 Simulation Setup The simulation is performed using the OptiSystem software. Figure 1 shows the simulation setup. At the transmitter side, data of three uses (U1 to U) each with 1. Gbit=s (aggregate bit rate of 4 Gbit=s) at pseudo random bit sequence (PRBS) are curved using three electri- Download Date /2/ 1:21 AM

2 2 G. A. Mahdiraji and A. F. Abas U1 RZ1 U2 RZ2 U RZ MZM1 MZM2 MZM Pre- DCF EDFA 8 km SSMF 1.4 km DCF 4 spans 8 km SSMF Post- DCF Attenuator BPF p-i-n PD LPF U1 U2 U LD Splitter U: User, LD: Laser diode, RZ: Return-to-zero, MZM: Mach-Zehnder modulator, LPF: Low pass filter DCF: Dispersion compensating fiber, SSMF: Standard single mode fiber, PIN PD: PIN photodiode Figure 1. Schematic of the 4 Gbit=s ( 1. Gbit=s) O-DCDM simulation setup for calculating SPM threshold over 4 km (5 8 km) SSMF. cal return-to-zero (RZ) pulse generators. Each of the RZ pulse generators are set at different RZ duty-cycles, which in this study the duty-cycles are uniformly distributed between different users as 25%, 5%, and 75% duty-cycle for RZ1, RZ2, and RZ, respectively. All the RZ pulse generators are assumed to operate synchronously based on a central clock. Output of the RZ pulse generators are then modulated over an optical carrier using three MZMs with db extinction ratio. An optical power splitter is used to distribute the laser diode (LD) signal that operates at 155 nm into the MZMs equally. Outputs of the MZMs are then multiplexed using an optical power combiner and transmitted over 4 km, which is made by 5 spans of 8 km SSMF. Chromatic dispersion of the SSMF is fully compensated using DCF. Parameters of the SSMF and DCF used in the simulation are shown in Table 1. The inline DCFs have a fixed length of 1.4 km with 1 ps=(nm km) dispersion coefficient to fully compensate the total dispersion per SSMF, which is 14 ps=nm (8 km 1.75 ps=(nm km)). An EDFA is used after every DCF and SSMF to compensate the power losses in the link. All the EDFAs used in this study have identical gain except the pre-edfa (EDFA before BPF). To adjust the input launched power into the DCFs and SSMFs, an optical attenuator is located after every EDFA. At the receiver side, a Gaussian optical bandpass filter (BPF) with 1 GHz cut-off frequency is used to eliminate the system noises that mainly produced by optical amplifiers. Then the received signal is detected by a p-i-n photodiode (p-i-n PD) followed by an electrical Gaussian low-pass filter (LPF). The cut-off frequency of the LPF is set at 4 GHz to eliminate the photodetector s noises. Output of the LPF is then fed into the DCDM demultiplexer to recover the original data based on the rules presented in [21, 22, 24]. Result and Discussion For calculating SPM threshold, chromatic dispersion is fully compensated by using DCF. Three symmetric dispersion compensation schemes, namely full pre-compensation, full post-compensation, and combination of pre- and postcompensation were used. Fiber type SSMF DCF Attenuation (db=km).2.5 Dispersion coefficient (ps=(nm km)) Dispersion slope (ps=(nm 2 km)).75.5 PMD coefficient (ps= p km/.5.5 Differential group delay (ps=km).2.2 Effective area (µm 2 / 8 Table 1. Parameters used in the simulations for SSMF and DCF. For pre-compensation, the Post-DCF, attenuator, and the EDFA before the Post-DCF are left out from the simulation setup shown in Figure 1. Thus, the total dispersion per SSMF span, i.e., 14 ps=nm is fully compensated before the SSMF as the pre-compensation. The dispersion map is presented in the Figure 2 (bottom right). The SPM effect is then investigated by increasing the launched power into the SSMFs up to more than C1 dbm. At every sweep, the input power into all SSMF spans was identical, while the input power into all DCFs was fixed at 4 dbm. This process is repeated for three other launched powers into DCFs, whichtheyare, 8, and 1 dbm. Figure2shows the result of SPM effect on performance of the worst user 9 OSNR Launched power to DCF: -4 dbm -8 dbm - dbm -1 dbm Dispersion (ps/nm) Transmission length (km) Figure 2. Effect of SPM and OSNR on performance of 4 Gbit=s O-DCDM over 4 km SSMF based on dispersion pre-compensation at different launched power into DCFs OSNR (db) Download Date /2/ 1:21 AM

3 Self-Phase Modulation Effect on Optical Duty-Cycle Division Multiplexing Technique in three-user O-DCDM system for four different launched powers of 4,, 8, and 1 dbm into DCFs. In addition, the OSNR of the system as a function of the launched power into SSMF is presented for different launched power into DCFs. By increasing the launched power into SSMF, performance is improved and reached to BER 1 9 when the launched power is around 1., 1.9, and 1 dbm for the case of 4,, and 8 dbm launched power into DCFs, respectively. The performance of the worst user never reached to BER 1 9 for the case where the launched power into DCFs is set at 1 dbm. By further increment in the launched power of SSMFs, performance of O-DCDM system is improved up to a certain point. By further increasing the launched power, performance of O-DCDM starts to degrade due to SPM effect. Referring to BER 1 9,the SPM threshold of the worst user in O-DCDM system can be considered at launched power of.95, C:2, and C: dbm, when the launched power of DCFs is fixed at 4,, and 8 dbm, respectively. These results show that by reducing the launched power into DCFs from 4 to dbm, the SPM threshold is improved by 1.15 db, whereas, the SPM threshold increased only by.4 db when the launched power into DCFs is reduced from to 8 dbm. In addition, when the launched power of DCFs is set at dbm, a range of 2.1 db power can be launched into SSMF for BER 1 9. Whereas, in the case of 4 and 8 dbm launched power into DCFs, the range of power is reduced to.5 and 1. db, respectively. This shows that the optimum launched power into DCFs ranges from to 8 dbm. Based on these results, it can be understood that there is a trade off between the nonlinear effect produced in the DCFs and OSNR of the system. When the launched power into the DCFs is kept as high as 4 dbm, even though OSNR of the system is high, the SPM effect dominates, which results in lower SPM threshold. The required OSNR and SPM effect is optimum when the launched power into DCFs is set around dbm. By further reduction in the launched power into DCFs (e.g., 8 dbm), the low OSNR has started to affect the performance of the system. Finally, when the launched power into DCFs is set at 1 dbm, due to very low OSNR, performance of the system never reached to the reference BER. In the second scheme, the total dispersion per SSMF spool is fully compensated after each fiber spool that is called dispersion post-compensation. The dispersion map is shown inside Figure (bottom right). Similar to dispersion pre-compensation, the performance is tested for four different launched powers into DCFs (i.e., 4,, 8, and 1 dbm) as shown in Figure. In this scheme, the minimum launched power required to be launched into SSMF to satisfy the reference BER is around 2.5, 2.2, 1.4, and C:8 dbm when the power launched into DCFs is set at 4,, 8, and 1 dbm, respectively. As explained earlier, by increasing the launched power into SSMF, performance of the system is improved up to a certain point 9 14 Dispersion (ps/nm) 4 Transmission length (km) Figure. Effect of SPM and OSNR on performance of 4 Gbit=s O-DCDM over 4 km SSMF based on dispersion post-compensation at different launched power into DCFs. and then start to reduce due to SPM effect. Referring to BER 1 9, SPM threshold in this compensation scheme is at C1.2, C2.1, C2.5, and C1.9 dbm for the case that the launched power into DCFs is set at 4,, 8, and 1 dbm, respectively. Therefore, a range of.7, 4.,.9, and 1.1 db power can be launched into SSMF to have BER 1 9 when the power launched into DCFs is set at 4,, 8, and 1 dbm, respectively. Similar to dispersion pre-compensation, the optimum launched power into DCFs ranged between to 8 dbm (more towards dbm). In addition, dispersion post-compensation shows better performance compared to pre-compensation for all four different launched powers into DCFs. Considering the launched power of dbm for DCFs, the SPM threshold in the case of post-compensation is improved by around 1.9 db, and the range of launched power into SSMF is also improved by almost twice as compared to the pre-compensation. The simulation is extended to investigate the SPM effect on O-DCDM when the system uses the combination of pre- and post-compensation. For this purpose, the total dispersion in 8 km SSMF spool (i.e., 14 ps=nm) is compensated partly using the Pre-DCF and the balanced with the Post-DCF as shown in Figure 1. This is when the inline DCFs are fixed similar to the two previous approaches presented earlier. Therefore, a pre-simulation is set to find the optimum dispersion value that needs to be compensated as the pre-compensation. This pre-simulation is performed by swiping the length of the Pre-DCF and Post-DCF simultaneously, where the total length in the Pre- and Post-DCFs equals to 1.4 km. This simulation was performed for two different launched powers into SSMFs (i.e., C5 and C7 dbm) and also for four different launched powers into DCFs (i.e., 4,, 8, and 1 dbm). Figure 4 shows performance of the worst user in terms of BER as a function of Pre-DCF length. In this figure, the DCF length is presented in the form of percentage based on the maximum length of 1.4 km. Figure 4 shows that for different launched power OSNR (db) Download Date /2/ 1:21 AM

4 4 G. A. Mahdiraji and A. F. Abas Pre-DCF length (%) Figure 4. Optimum Pre-DCF length in percentage of DCF length. into DCFs and SSMFs, the optimum length of Pre-DCF is around 1% 2%, which is equal to km or ps=nm. Based on Figure 4, the SPM effect on performance of O-DCDM is investigated by setting the Pre- and Post-DCF length to and km, respectively, while all the inline DCFs are fixed at 1.4 km, as the dispersion map shown inside the Figure 5 (top middle). This simulation is also performed for four different launched powers into DCFs as shown in Figure 5. In this compensation scheme, the minimum power required to be launched into SSMF to satisfy BER 1 9 is around 2:8, 2:4, 1:5, and :1 dbm when the launched power into DCFs is set at 4,, 8, and 1 dbm, respectively. Also, the SPM threshold at the reference BER is around C5:1, C:1, C:7, and C:9 dbm for the case that the power launched into DCFs is set at 4,, 8, and 1 dbm, respectively. Therefore, the range of power that can be launched into SSMF to have BER 1 9 is 7:9, 8.5, 8.2, and 7. db for the case that launched power into DCFs is 4,, 8, and 1 dbm, respectively. Even though, the highest SPM threshold is achieved for the launched power of 1 dbm into DCFs, but considering both the SPM and OSNR effect, the optimum launched power into DCFs is around to 8 dbm. This is due to large improvement in the SPM threshold when the launched power into DCFs is reduced from 4 to dbm. In addition, the wider range of power into SSMF can be supported when the launched power into DCFs is set at dbm compared to the other launched power. These results show that the SPM effect is more dominated for the launched power of larger than dbm into DCFs. The effect of low OSNR is more dominated at the launched power of less than dbm into DCFs. Considering the launched power of dbm into DCFs, the SPM threshold in the case of 1% dispersion pre-compensation is improved by around 5.9 and 4 db compared to the fully dispersion preand post- compensation, respectively. In addition, the range of launched power into SSMFs is improved by around four Dispersion (ps/nm) Transmission length (km) Launched 9 OSNR power to DCF: -4 dbm 1 - dbm -8 dbm -1 dbm Figure 5. Effect of SPM and OSNR on performance of 4 Gbit=s O-DCDM over 4 km SSMF based on 1% dispersion pre-compensation at different launched power into DCFs. times and two times in comparison to the dispersion fully pre- and post-compensation, respectively. For comparison purpose, Figures (a), (b) and (c) compare part of the original transmitted bit streams captured after the multiplexer against the received bit streams extracted before the photodetector for the three dispersion compensation scheme, i.e., pre-, post-, and 1% pre-compensation, respectively. The eye diagrams are also presented. All measurements were conducted at C dbm launched power into SSMFs and dbm power into DCFs. As shown in Figure (a), in the pre-compensation, parts of the received signal are strongly shifted and little chirp can be seen. Opposite shifting with strong chirping effect is observed from the received signal is in the case of post-compensation. Both the shifting and chirping effect is significantly reduced in the case of 1% pre-compensation. These effects are due to the interaction between the chromatic dispersion and SPM. In the case of 1% pre-compensation, the jitter is reduced, thereby improved the Q-factor. The effect of SPM on performance of O-DCDM over different length of SSMF is investigated, for the case that the dispersion of the system is compensated by combination of pre- and post-compensation (1% pre-compensation). In this simulation, the launched power into DCFs is fixed at dbm. Figure 7 shows the range of launched power into SSMF, i.e., the lower- and upper-bound of launched power, which result a performance equal or better than BER 1 9 and 1 at different transmission length. Thus, the launched power of less than the lower-bound results in the worse performance than the reference BER (i.e., 1 9 or 1 /, which is due to low OSNR or low received power. On the other hand, applying launched power greater than the upper-bound into SSMF deteriorates the system performance, which this is due to very strong SPM effect. In addition, the launched power at the knee point of the 4 2 OSNR (db) Download Date /2/ 1:21 AM

5 Self-Phase Modulation Effect on Optical Duty-Cycle Division Multiplexing Technique 5 Power (mw) Power (mw) Power (mw) Time (ns) Figure. Comparison between the original transmitted and received bit streams, (a) pre-, (b) post-, and (c) 1% precompensation. Transmission distance (km) Range of LP to get BER < E- over 9 km SSMF Range of LP that SPM effect is dominant to the LP Range of LP that the effect of LP is dominant to SPM Lower-bound of launched power (LP) BER E- BER E-9 Knee point Upper-bound of LP Range of LP to get BER < E-9 over 8 km SSMF Figure 7. The range of launched power to SSMF for different transmission length for the performance at BER 1 9 and 1. SPM effect is shown in the figure. The effect of launched power or OSNR is more dominant to the SPM effect for the launched power below the knee points, whereas for the launched power higher than the knee points, the SPM effect becomes more dominant. This means that by increasing the launched power into SSMF, performance of the system improved up to the knee point, whereas, increasing the launched power more than the knee point reduce the system performance due to the SPM effect. Thus, the best range of the launched power into SSMF is around the knee point. Except for the case of 8 km transmission link, the optimum launched power into SSMF is slightly increased by increasing the transmission distance. The optimum launched power into SSMF for 1 km transmission link is around 1. dbm and it goes towards 2.5 and.7 dbm for 4 and 8 km distance, respectively. The other observation that can be discussed from this figure is that the lower- and upper-bound of the launched power are getting closer to each other with the increase of the fiber length. This is because after every 8 km SSMF, two EDFAs are added to the system (one for SSMF and the other for DCF), which added the amplified spontaneous emission (ASE) noise to the system, thereby, reduce the OSNR. Therefore, to achieve the reference BER at the longer distance, higher launched power is required to maintain the OSNR of the system. This is the main reason that the lower-bound of launched power increases by increasing the transmission length. Thus, the lower-bound of the launched power represents the effect of EDFA s ASE noise at different transmission lengths or at different number of EDFAs. On the other hand, by increasing fiber length, optical nonlinearities in fiber (here SPM effect) become stronger, thereby, a smaller intensity can produce higher nonlinear effect compared to the shorter fiber. Therefore, by increasing the transmission distance, the upper-bound of launched power is also reduced. This means that the upperbound of the launched power can be used to represent the SPM threshold over different fiber length. The results from Figure 7 confirm the ability of O- DCDM to be used as an alternative multiplexing technique. For the case of error free communication (BER 1 9 /, the technique can be used for the applications in the range Download Date /2/ 1:21 AM

6 G. A. Mahdiraji and A. F. Abas of metropolitan area network (MAN) and wide area network (WAN). Considering BER of 1, the technique can be used for the long-haul transmission applications, which in this case forward error correction (FEC) is necessary. For comparison purposes, the performance of several other techniques is presented. For example, according to [1], 4 Gbit=s RZ differential quadrature-phase-shiftkeying (DQPSK) has the range of launched power into SSMF from C2 to C8:5 dbm and 1: to C11:7 dbm at BER 1 9 and 1, respectively, over 1 km link. The link is made by 1 1 km SSMF and dispersion of fiber is compensated using DCF with 25 ps=(nm km) dispersion coefficient implementing the combination of preand post-compensation. In this paper, the launched power into SSMF at BER 1 and 1 km transmission distance ranges from C:1 to C7:4 dbm. In other reports, for the test over 8 km, the SPM threshold at BER 1 9 for 4 Gbit=s 4-ary, absolute polar DCDM (AP-DCDM), and NRZ on-off-keying (OOK) is around C7, C11:1, and C dbm, respectively [2], where, the dispersion of fiber is compensated using DCF with 1 ps=(nm km)) dispersion coefficient in post-compensation arrangement. 4 Conclusion The effect of SPM on performance of 4 Gbit=s O-DCDM is investigated using three symmetric dispersion compensation schemes over 4 km. The results show that the combination of both dispersion post-compensation and precompensation is the best scheme to be used in this system. From our observation it can be concluded that by considering a correct dispersion compensation scheme with optimum parameters, O-DCDM can be used in the MAN and WAN applications with minimum deterioration from SPM effect for error free communication. References [1] P. Hajireza, N. S. Shahabuddin, S. Abbasi-Zargaleh, S. D. Emami, H. A. Abdul-Rashid and Z. Yusoff, Performance of a high-concentration erbium-doped fiber amplifier with 1 nm amplification bandwidth, in AIP Conference Proceedings, 21, pp [2] L. Huang, G. Yang, Y. Li, J. Gao and Z. Li, L-band EDFA with high saturation output power and low noise figure, in 21 International Conference on Computer Design and Applications, ICCDA 21, 21, pp. V457 V459. [] J. Sun, X. R. Ma, Q. Ji, H. Zhang and J. H. Luo, Study on a novel low-noise erbium-doped fiber amplifier, Guangdianzi Jiguang/Journal of Optoelectronics Laser, Vol. 21 (21), pp [4] R. I. Laming, M. N. Zervas and D. N. Payne, Erbiumdoped fiber amplifier with 54 db gain and.1 db noise figures, Photonics Technology Letters, IEEE, Vol. 4 (1992), pp [5] R. Freund, L. Molle, F. Raub, C. Caspar, M. Karkri and C. Weber, Triple-(S/C/L)-band WDM transmission using erbium-doped fibre amplifiers, in Optical Communication, 25. ECOC 25. 1st European Conference on 25, Vol. 1 (25), pp [] W. Chen, S. Y. Li, D. Y. Lei, D. X. Wang, W. Y. Luo and W. J. Huang, Development of wide-band dispersion compensation module for high-speed communication systems, Guangdianzi Jiguang/Journal of Optoelectronics Laser,Vol. 21 (21), pp [7] L. Zong, F. Luo, Y. Wang and X. Cao, Dispersion compensation module for 1 Gbit=s optical system and beyond, Optical Fiber Technology, 211 (In press). [8] Y. A. Yaroshenko, Dispersion compensation with use of DCF fiber, in KpbiMuKo 21 CriMiCo 21 2th International Crimean Conference Microwave and Telecommunication Technology, Conference Proceedings, 21, pp. 2. [9] D. Marcuse, A. R. Chraplyvy and R. W. Tkach, Effect of fiber nonlinearity on long-distance transmission, Lightwave Technology, Journal of, Vol. 9 (1991), pp [1] A. R. Chraplyvy and R. W. Tkach, Optical fiber nonlinear effects in lightwave communication systems, in Nonlinear Optics: Materials, Fundamentals, and Applications, NLO 94 IEEE, 1994, p. 2. [11] E. Iannone, F. Matera, A. Mecozzi and M. Settembre, Nonlinear optical communication networks: John Wiley and Sons, Inc., [] J. Toulouse, Optical nonlinearities in fibers: review, recent examples and systems applications, Lightwave Technology, Journal of, Vol. 2 (25), pp [1] C. Weber, C. A. Bunge, M. Winter and K. Petermann, Fibre nonlinearities in 1 and 4 Gbit=s electronically dispersion precompensated WDM transmission, in Optical Fiber Communication - incudes post deadline papers, 29. OFC 29. Conference on, 29, pp. 1-. [14] F. Yaman and L. Guifang, Nonlinear Impairment Compensation for Polarization-Division Multiplexed WDM Transmission Using Digital Backward Propagation, Photonics Journal, IEEE, Vol. 2 (21), pp [15] Y. Zhao, Y. J. Qiao and Y. F. Ji, Study on fiber nonlinearities in 1 Gbit=s and 4 Gbit=s RZ-DPSK electrical predistortion systems, Guangdianzi Jiguang/Journal of Optoelectronics Laser, Vol. 21 (21), pp , 17. [1] S. Susskind and E. A. de Souza, 4 Gbit=s RZ DQPSK transmission with SPM and ASE suppression by dispersion management, in Microwave and Optoelectronics Conference (IMOC), 29 SBMO/IEEE MTT-S International, 29, pp [17] B. Chatelain, C. Laperle, D. Krause, K. Roberts, M. Chagnon, X. Xian, A. Borowiec, F. Gagnon, J. C. Cartledge and D. V. Plant, SPM-Tolerant Pulse Shaping for 4- and 1-Gbit=s Dual-Polarization QPSK Systems, Photonics Technology Letters, IEEE, Vol. 22 (21), pp [18] E. Yamazaki, A. Sano, T. Kobayashi, E. Yoshida and Y. Miyamoto, SPM compensation of no-guard-interval PDM Co-OFDM for optical transport network, in Photonics Society Summer Topical Meeting Series, 21 IEEE, 21, pp [19] A. Y. Yang and Y. N. Sun, "Comprehensive research on self phase modulation based optical delay systems," Chinese Physics B, vol. 19, 21. Download Date /2/ 1:21 AM

7 Self-Phase Modulation Effect on Optical Duty-Cycle Division Multiplexing Technique 7 [2] A. Malekmohammadi, G. A. Mahdiraji, A. F. Abas, M. K. Abdullah, M. Mokhtar and M. F. A. Rasid, Effect of selfphase-modulation on dispersion compensated absolute polar duty cycle division multiplexing Transmision, IET Optoelectronics, Vol. (29), pp [21] G. A. Mahdiraji, A. F. Abas, M. K. Abdullah, A. Malekmohammadi and M. Mokhtar, Duty-Cycle Division Multiplexing (DCDM): Alternative for High Speed Optical Networks, Japanese Journal of Applied Physics (JJAP), Vol. 48 (29), p.. [22] G. A. Mahdiraji, M. K. Abdullah, A. M. Mohammadi, A. F. Abas, M. Mokhtar and E. Zahedi, Duty-Cycle Division Multiplexing (DCDM), Optics and Laser Technology, Vol. 42 (21), pp [2] G. A. Mahdiraji, M. K. Abdullah, M. Mokhtar, A. M. Mohammadi, A. F. Abas, S. M. Basir and R. S. A. R. Abdullah, 7 Gbit=s Amplitude-Shift-Keyed System with 1 GHz Clock Recovery Circuit using Duty Cycle Division Multiplexing, Photonic Network Communications, Vol. 19 (21), pp [24] G. A. Mahdiraji, M. K. Abdullah, M. Mokhtar, A. M. Mohammadi, R. S. A. R. Abdullah and A. F. Abas, Duty- Cycle-Division-Multiplexing: Bit Error Rate Estimation and Performance Evaluation, Optical Review, Vol. 1 (29), pp [25] G. A. Mahdiraji, A. Malekmohammadi, A. F. Abas, M. Mokhtar and M. K. Abdullah, A novel economical duty cycle division multiplexing with electrical multiplexer and demultiplexer for optical communication systems, Int. J. Information and Communication Technology, Vol. 2 (29), pp [2] G. A. Mahdiraji, A. M. Mohammadi, A. F. Abas, M. K. Abdullah, M. Mokhtar, S. M. Basir, M. F. A. Rasid, R. S. A. R. Abdullah and E. Zahedi, Duty-Cycle Division Multiplexing (DCDM): towards the largest optical networks capacity, International Review on Modelling and Simulations (I.RE.MO.S.), Vol. 1 (28), pp [27] G. A. Mahdiraji, A. M. Mohammadi, A. F. Abas, M. K. Abdullah, M. Mokhtar and E. Zahedi, Performance Analysis of Duty Cycle Division Multiplexing Technique with Electrical Multiplexer and Demultiplexer in Fiber Optic Communication System, International Engineering and Technology (IETECH) Journal of Communication Techniques, Vol. (29), pp [28] J. Huang and J. Yao, Analyses of the performances of 1 Gbs 1 time-division multiplexing and wavelengthdivision multiplexing signals in single-mode fibers and nonzero dispersion-shifted fibers, Journal of Optics A: Pure andapplied Optics, Vol. (21). Download Date /2/ 1:21 AM