10 Gb/S Bidirectional Transmission with an Optimized SOA and a SOA-EAM Based ONU

applied sciences

Article 10 Gb/s Bidirectional Transmission with an Optimized SOA and a SOA-EAM Based ONU

Xin Rui Chen 1 and Guang Yong Chu 1,2,*

1 School of Science, Jiangnan University, Wuxi 214122, China; [email protected] 2 Jiangsu Provincial Research Center of Light Industrial Optoelectronic Engineering and Technology, Wuxi 214122, China * Correspondence: [email protected]

Received: 23 October 2020; Accepted: 13 December 2020; Published: 15 December 2020

Abstract: We investigated the application of a semiconductor optical amplifier (SOA) and an SOA electro-absorption modulator (SOA-EAM) as attractive, low-cost solutions in passive optical networks (PONs). The main characteristics of an SOA with optimal performance for phase and amplitude modulation were tested. Additionally, a 10 Gb/s bidirectional transmission with an optical network unit (ONU) transmitter integrated with a distributed feedback (DFB) laser, electro-absorption modulator (EAM), and SOA was designed. The upstream (US) and downstream (DS) receiver sensitivities at the forward error correction (FEC) level reached 29.5 dBm and 33.5 dBm for − − back-to-back (BtB) fiber and 28.9 dBm and 33.1 dBm for 20 km fiber. For multichannel transmission, − − the US receiver sensitivities reached 28.8 dBm and 28.2 dBm for BtB and 20 km fibers, and the DS − − receiver sensitivities reached 33 dBm and 32.6 dBm for BtB and 20 km fibers, respectively. − − Keywords: semiconductor optical amplifier (SOA); semiconductor optical amplifier electro-absorption modulator (SOA-EAM); phase modulation (PM); amplitude modulation (AM); optical network unit (ONU)

1. Introduction The explosive growth of network data requires higher transmission capacity and flexibility from optical communication networks, which is accelerating the development of next-generation passive optical networks (PONs) [1,2]. Thus, research on the application of passive devices is gradually advancing, including monolithically integrated reflective semiconductor amplifiers (RSOAs) [3], dual electro-absorption modulated lasers (DEMLs) [4], distributed feedback semiconductor optical amplifiers (DFB-SOAs) [5], dual-output-DEMLs [6], and electro-absorption modulated laser semiconductor optical amplifiers (EML-SOAs) [7], which have been applied to 1.25 Gb/s, 2.5 Gb/s, or beyond for next-generation PON2s (NGPON2s) [8,9]. This means that in facing the needs of future users, it is imperative to deploy a small-footprint, large-capacity, low-budget, and dense network. In the actual design of optical network units’ (ONUs) structure, the performance of the light source, modulator, and power amplifier determine the transmission capacity and signal quality of the system [1]. It is desired that the whole system applies amplifiers and modulators with a limited volume and affordable costs, such as semiconductor optical amplifiers (SOAs) and semiconductor optical amplifier electro-absorption modulators (SOA-EAMs). Unstable received output is an issue at the ONU. Facing the existing issue, gain-control would be considered [10]. Compared with an erbium-doped fiber amplifier (EDFA) [11], an SOA is an integrated device with little volume, a long service life, low cost, and a mature manufacturing process. Thus, an SOA (as the amplifier of an ONU) can better meet the requirements of ONU design for an NGPON2 [12]. Additionally, an SOA can be used for all-optical wavelength conversions (AOWC) [13]

Appl. Sci. 2020, 10, 8960; doi:10.3390/app10248960 www.mdpi.com/journal/applsci Appl. Sci. 2020, 10, 8960 2 of 11 and all-optical switches [14], which means that it is necessary to make an SOA choose the required proper condition to fit different functions. Hence, we show the performance of an SOA under different bias currents and input powers for selecting the operating environment in distinct applications. Reflective SOAs (RSOAs) [3], electro-absorption modulators (EAMs) [15], RSOA-EAMs [16,17], and SOA-EAMs have been used in actual optical links as the modulators [18]. However, ONUs composed of reflective modulators are affected by inherent interference noise from backward reflected light [19], and the high insertion loss of EAMs will lead to the serious degradation of an optical signal-to-noise ratio (OSNR). Therefore, this paper mainly discusses the modulation performance of SOA-EAMs, which can compensate the loss from EAMs by SOAs. Additionally, the overshoot mode caused by an SOA saturation effect can compensate for the undershoot mode of an EAM to output an undistorted waveform. Additionally, the chirp of an SOA can also be compensated for by the chirp of an EAM [17,18]. The integrated SOA–EAM combination currently employed in NGPON2-ONUs has a wider modulation bandwidth and a lower chirp than common modulators. This can increase system tolerance, improve optical signal quality in an optical access network, and enable a PON’s configuration to reach a longer range. In this paper, we first characterize principal SOA performance via a small-signal model analysis method, obtaining the appropriate operating point of the input optical power and the bias current-carrying signal in Section2. We then construct a carrier-distributed dense wavelength division multiplexed PON (DWDM-PON), discussing the modulation characteristics of the SOA-EAM and obtaining the sensitivities in single-ONU and multi-ONU transmissions for 10 Gb/s at a standard forward error correction (FEC) level [20] in Section3. The paper ends with the conclusion in Section4.

2. Optimizing the Performance of an SOA This section shows the eﬀects of input power and bias current on the gain and noise ﬁgure (NF) of an SOA in detail, via a small-signal model analysis. By adjusting the variable optical attenuator (VOA) and direct current (DC), the input power and bias current of the SOA are transformed, respectively. The equation for NF measurement is as follows [21]:

NF = 10 log10[1/Gain(PASE/hνB0 + 1)](dB) (1) where h is the Planck constant, ν is the center frequency of the optical signal, and B0 is the frequency bandwidth of a spectrometer; they remain unchanged under constant external conditions. We can infer that NF is proportional to the ratio of ampliﬁed spontaneous emission (ASE) power to gain. The parameters used for the simulation are listed in Table1.

Table 1. Parameters chosen for the simulation of the semiconductor optical ampliﬁer (SOA).

Parameters Value Active length 0.0011 m Active layer width 0.4 µm Active layer height 0.4 µm Optical confinement factor 0.45 Recombination coefficient A 3.6 108 s 1 × − Recombination coefficient B 5.6 10 16 m3/s × − Recombination coefficient C 3 10 41 m6/s × − Group velocity 7.5 107 m/s ×

Figure1a depicts the e ffect of input power on the characteristics of the SOA. The increasing power of gain goes to the nearly flat region between 45 dBm and 20 dBm and the linear descent region − − from 20 dBm to 0 dBm. The increasing power of NF goes to the almost flat region between 45 dBm − − and 20 dBm, the linear descent region from 20 dBm to 10 dBm, and the linear ascent region from − − − 10 dBm to 0 dBm. − Appl. Sci. 2020, 10, 8960 3 of 11

Appl. Sci. 2020, 10, x FOR PEER REVIEW 3 of 11 The SOA consumes a spot of carriers at a low level of input power, which can fully amplify the −45 dBmoptical and signal. −20 Therefore,dBm, the thelinear gain remainsdescent a region constant from substantially −20 dBm because to −10 the dBm, output and power the increases linear ascent with the input power correspondingly. With the input power increases, the rate of carrier consumption region from −10 dBm to 0 dBm. caused by stimulated radiation increases, leading to a new balance of carrier concentration in the Theactive SOA region consumes at a lower a spot level. of Accordingly, carriers at thea low gain level of the of SOA input decreases power, with which the reductioncan fully ofamplify the the opticalmagnification signal. Therefore, ability. Under the thegain gain-flat remains region, a theconstant NF remains substant the sameially at firstbecause since thethe stimulated output power increasesand with spontaneous the input radiation power increase correspondingly. accordingly. With Then, the the input NF decreases power increases, in that the outputthe rate ASE of carrier consumptiondecreases andcaused the output by stimulated OSNR increases radiation under the increases, gradual saturation leading of theto SOA.a new Lastly, balance the NF risesof carrier concentrationwith the decreasesin the active in the region OSNR asat a a result lower of thelevel. intense Accordingly, gain saturation the gain of the of SOA. the SOA decreases with the reductionFigure of1 theb depicts magnification the influence ability. of bias Under current the on the gain-flat characteristics region, of the the NF SOA. remains The gain the function same at first is almost in linear ascent in the region of the bias current from 60 to 80 mA and nearly flat in the region since the stimulated and spontaneous radiation increase accordingly. Then, the NF decreases in that between 80 mA and 200 mA. The NF with the input power from 15 dBm to 45 dBm is nearly flat − − the outputand from ASE 0 decreases dBm to 15 and dBm the isalmost output linear OSNR in descent increases in the under region the of gradual the bias currentsaturation from of 60 the to SOA. − Lastly,100 the mA NF and rises nearly with flat the in the decreases region between in the 100 OSNR mA and as a 200 result mA. Theof the bias intense current hasgain nothing saturation to do of the SOA. with NF approximately at low input power.

(a)

(b)

Figure 1. Cont. Appl. Sci. 2020, 10, x FOR PEER REVIEW 4 of 11 Appl. Sci. 2020, 10, 8960 4 of 11

(c)

(d)

Figure 1. Characteristic analysis diagram of an SOA: (a) Gain and NF with input power; (b) Gain and Figure 1. Characteristic analysis diagram of an SOA: (a) Gain and NF with input power; (b) Gain and NF withNF withbias biascurrent; current; (c, (dc,)d Gain) Gain and and NF withwith wavelength wavelength (the (the left left input input power power is 25 is dBm, −25 the dBm, right the right − inputinput power power is −5 is dBm).5 dBm). − FigureThere 1b isdepicts a little gain the due influence to the low of carrier bias concentrationcurrent on ofthe the characteristics SOA, and there isof a highthe SOA. NF since The gain the low carrier concentration makes the stimulated radiation weaker at small bias current. As shown function is almost in linear ascent in the region of the bias current from 60 to 80 mA and nearly flat in Figure1b, the gain increases monotonically at first and then maintains the same level as the bias in thecurrent region level—which between 80 is mA because and the 200 carrier mA. density The NF of thewith SOA the increases, input power and the from system −15 is restraineddBm to − by45 dBm is nearlythe fl recoveryat and from time of0 thedBm carrier to − density.15 dBm The is almost increase linear in carrier in concentrationdescent in the enhances region the of stimulated the bias current from emission,60 to 100 which mA and leads nearly to the decreaseflat in the in region NF. Finally, between NF remains 100 mA stable and with 200 bias mA. current The bias since current the has nothingstimulated to do with radiation NF approximately tends to be unaltered at low at theinput adequate power. bias current. ThereFigure is a little1c,d showsgain due that to the the gain low flatness carrier of concen the SOAtration meets theof the standard SOA, betweenand there 1540 is a nm high and NF since the low1560 carrier nm. Theconcentration output performance makes the can stimulated basically meet radiation the requirements weaker at small of various bias applications,current. As shown and different applications can be selected according to the performance of gain and NF in different in Figure 1b, the gain increases monotonically at first and then maintains the same level as the bias operating regions, as summarized in Table2. current level—which is because the carrier density of the SOA increases, and the system is restrained by the recovery time of the carrier density. The increase in carrier concentration enhances the stimulated emission, which leads to the decrease in NF. Finally, NF remains stable with bias current since the stimulated radiation tends to be unaltered at the adequate bias current. Figure 1c,d shows that the gain flatness of the SOA meets the standard between 1540 nm and 1560 nm. The output performance can basically meet the requirements of various applications, and different applications can be selected according to the performance of gain and NF in different operating regions, as summarized in Table 2.

Table 2. Areas of operation for an SOA and its exploited performance for their application. Appl. Sci. 2020, 10, x FOR PEER REVIEW 5 of 11 Appl. Sci. 2020, 10, 8960 5 of 11 Input Power SOA Low Medium High Table 2. Areas of operation for an SOA and its exploited performance for their application. low gain, low gain, no gain, low high NF highInput NF Power high NF SOA moderateLow gain, moderate Medium gain, low High gain, medium low gain, low gain, no gain, low Bias moderate–highhigh NF NF moderate–low high NF NF high high NFNF current highmoderate gain, gain, high moderate gain, gain, low low gain, Bias current medium moderate–highhigh NF, NF moderate–low low NF, NF high high NF, NF high suitablehigh for gain, phase suitable high for gain, optical suitable lowfor gain, amplitude high high NF, low NF, high NF, suitablemodulation for phase modulation suitableamplification for optical ampliﬁcation suitable formodulation amplitude modulation

To sum up, we get the optimum operating point of the SOA. Under the circumstances where the To sum up, we get the optimum operating point of the SOA. Under the circumstances where the bias current is 200 mA and the input power is −25 dBm or −5 dBm of the SOA, the output optical bias current is 200 mA and the input power is 25 dBm or 5 dBm of the SOA, the output optical power can keep constant on the premise of ensu− ring low noise− relatively. The input power of −25 power can keep constant on the premise of ensuring low noise relatively. The input power of 25 dBm dBm is suitable for phase modulation, while −5 dBm is suitable for amplitude modulation. − is suitable for phase modulation, while 5 dBm is suitable for amplitude modulation. − 3. 10 10 Gb/s Gb/s Bidirectional Bidirectional Transmission Transmission for a DFB-EAM-SOA-Based ONU

3.1. Design Design and and Analysis Analysis InIn order to analyze the quality of the signal tr transmissionansmission using DFB-EAM-SOA, we first ﬁrst tested thethe unidirectional transmission transmission for for both both DS DS and US US;; then, we tested the bidirectional transmission under thethe optimal optimal conditions conditions we hadwe gotten;had gotten; ﬁnally, bothfinally, ONUs both were ONUs active were and testedactive bidirectionally. and tested bidirectionally.In Figure2, we proposeIn Figure a 2, network we propose schematic a network with schematic 50 GHz channel with 50 spacing GHz channel as in thespacing International as in the InternationalTelecommunication Telecommunication Union Telecommunication Union Telecommuni Standardizationcation Standardization Sector (ITU-T) Sector G.694.1 (ITU-T) standard G.694.1 for standardDWDM [for22, 23DWDM]. The [22,23]. optical The line optical terminal line termin (OLT)al is (OLT) composed is composed of several of several transceivers, transceivers, and eachand eachtransceiver transceiver corresponds corresponds to one to one ONU. ONU. The The main main parameters parameters list list of of the the simulation simulation settingsetting isis shown inin Table Table3 3..

Figure 2. The semiconductor optical ampliﬁer electro absorption modulator (SOA EAM)—based − − optical network unit (ONU) for 10 Gb/s bidirectional transmission (ODN: optical distribution network; Figure 2. The semiconductor optical amplifier electro−absorption modulator (SOA−EAM) − based C1, C2: circulator; Mux/Demux: multiplexer/demultiplexer; VOA: variable optical attenuator; PIN: optical network unit (ONU) for 10 Gb/s bidirectional transmission (ODN: optical distribution PIN diode; TIA: trans impedance ampliﬁer). network;− C1, C2: circulator;− Mux/Demux: multiplexer/demultiplexer; VOA: variable optical attenuator; PIN: PIN−diode; TIA: trans−impedance amplifier). Appl. Sci. 2020, 10, 8960 6 of 11

Table 3. Main parameters list of the simulation setting.

Parameters Value Wavelength of distributed feedback (DFB) laser in ONU1 1592.5 nm Wavelength of DFB laser in ONU2 1592.1 nm Wavelength of DFB laser in transceiver1 1550.1 nm Wavelength of DFB laser in transceiver2 1549.7 nm Bias voltage of electro-absorption modulator (EAM) 2 V − Length of single-mode fiber (SMF) 20 km Attenuation of SMF 0.2 dB/km Dispersion of SMF 16.75 ps/nm/km Non-linear coefficient of SMF 2.6 10 20 m2/w × − PMD coefficient of SMF 0.05 ps/sqrt(km) Extinction ratio of Mach–Zehnder modulator (MZM) 30 dB Symmetry factor of MZM 1 − Bandwidth of Mux/Demux 20 GHz Responsivity of PIN 1 A/W Dark current of PIN 10 nA Transimpedance of TIA 2000 Ohm

For DS transmission, the continuous wave light injected into the Mach–Zehnder modulator (MZM) employed at the OLT was modulated in non-return-to-zero (NRZ) format by a pseudo-random binary sequence (PRBS), which was generated by a distributed feedback (DFB) laser. The optical signals with different wavelengths were multiplexed and carried. After transmission over a 20 km standard single-mode fiber (SSMF), the signal was demultiplexed to different ONUs of corresponding wavelengths. The optical signal was converted into a low-noise electrical signal through a photoelectric detector and an equalizing filter. Finally, the bit error ratio (BER) was detected and analyzed. For US transmission, an SOA-EAM was selected to improve the transmission distance of data signals because of its wide modulation bandwidth and low chirp, as depicted in the schematic view of Figure2. There was leading edge distortion (as pre-distortion) to form an undershoot mode when data passed through the EAM, shown in the inset (I) of Figure2, while the carrier recombination limitation in the SOA saturation led to the formation of an overshoot mode, as the inset (II) of Figure2 shows; such mutually compensated optical processing can be adjusted by the bias point, and the whole system can emit an undistorted optical waveform (see the inset (III) of Figure2)[ 18]. Additionally, the chirp from the absorption processing of an EAM can be compensated for by adjusting the negative frequency chirp from the carrier depletion of the SOA, which can reduce the chirp generated by the SOA-EAM and the dispersion effect in the optical transmission system [17]. The signal was detected at the OLT after a 20 km transmission.

3.2. Results and Discussion Figure3 shows the BER against the received power in unidirectional US and DS transmission. Figure3a shows the BER characteristics with a signaling rate of 10 Gb /s in a unidirectional uplink. The horizontal axis shows the average received power. The amplitude shift keying (ASK) signal was generated with a peak-to-peak RF voltage of 2 V (VPP) for the EAM with a bias current of 80 mA (Ibias) for the SOA. At the FEC level of BER = 2.4 10 4 [20], the sensitivities reached 29.5 dBm, 27.5 dBm, × − − − and 21 dBm for input power (Pi) = 10 dBm, 5 dBm, and 20 dBm, respectively. There was a 2 dB − − − − power penalty between Pi = 10 dBm and 5 dBm, and an 8.5 dB power penalty between 10 dBm and − − − 20 dBm. The eye diagrams are also shown in the insets of Figure3a. The eye pattern at Pi = 20 dBm − − exhibits a thicker 1-level due to the OSNR degradation and at Pi = 10 dBm and 5 dBm shows the − − typical overshoot on account of the pattern eﬀect [24]. Appl. Sci. 2020, 10, 8960 7 of 11 Appl. Sci. 2020, 10, x FOR PEER REVIEW 7 of 11

(a)

(b)

(c)

Figure 3. Cont. Appl.Appl. Sci. 2020 Sci. 2020, 10, 10x FOR, 8960 PEER REVIEW 8 of 118 of 11

(d)

Figure 3. The BER against the received power in a 10 Gb/s unidirectional transmission: (a) the Figure 3. The BER against the received power in a 10 Gb/s unidirectional transmission: (a) the upstream upstream(US) BER (US) curves BER for curves input fo powerr input (Pi) power values; (Pi) (b) thevalues; US BER (b) curves the US for BER the peakcurvesto forpeak the RF peak voltage−to−peak − − PP bias RF (VvoltagePP) values; (V ()c )values; the US BER(c) the curves US for BER the biascurves current for (Ithebias )bias values; current (d) the (I US) and values; downstream (d) the (DS) US and downstreamBER curves (DS) for backBER curvesto back for (BtB) back and−to 20−back km ﬁbers. (BtB) and 20 km fibers. − −

At theWe FEC then modulatedlevel of BER the = EAM2.4 × 10 with−4 [20], VPP =the1 V,sensitivities 1.5 V, and 2reached V. The DFB −29.5 laser dBm, provides −27.5 dBm, an input and −21 power of 10 dBm, and there is I = 80 mA, as previous. The BER against the received optical power dBm for input− power (Pi) = −10 dBm,bias −5 dBm, and −20 dBm, respectively. There was a 2 dB power is plotted in Figure3b. At BER = 2.4 10 4, the received power sensitivities are 29.5 dBm, 28.4 dBm, penalty between Pi = −10 dBm and −×5 dBm,− and an 8.5 dB power penalty between− −10− dBm and −20 and 25 dBm for VPP = 2 V, 1.5 V, and 1 V, respectively. Figure3b also shows that the power penalties dBm. The− eye diagrams are also shown4 in the insets of Figure 3a. The eye pattern at Pi = −20 dBm are 1.1 dB and 3.4 dB at BER = 2.4 10− when VPP decreases from 2 V to 1.5 V and from 1.5 V to 1 V, exhibits a thicker 1-level due to the× OSNR degradation and at Pi = −10 dBm and −5 dBm shows the respectively. Figure3c shows the measured BER with I bias = 60 mA, 80 mA, and 100 mA at VPP = 2 V typicaland overshoot Pi = 10 dBm, on account in which of the the EAM pattern works effect under [24]. reverse bias and the SOA operates at forward − bias.We Thethen sensitivities modulated of the27 EAM dBm, with29.5 dBm,VPP = and1 V, 1.529 dBmV, and were 2 obtainedV. The DFB for I laser= 60provides mA, 80 mA,an input − − − bias powerand of 100 −10 mA dBm, at BER and of there 2.4 10is I4biasfor = 80 a 10 mA, Gb/ sas signal previous. with PRBS. The BER against the received optical power × − is plottedWe inshow Figure the 3b. BER At curvesBER = 2.4 of US × 10 and−4, the DS received unidirectional power transmissions sensitivities in are Figure −29.53 ddBm, under −28.4 the dBm, appropriate condition of Pi = 10 dBm, V = 2 V, and I = 80 mA. At BER = 2.4 10 4 for the and −25 dBm for VPP = 2 V, 1.5 V,− and 1 V, respectively.PP Figurebias 3b also shows that the× power− penalties unidirectional US transmission, the sensitivities−4 reached 29.8 dBm and 29.5 dBm for back-to-back are 1.1 dB and 3.4 dB at BER = 2.4 × 10 when VPP decreases− from 2 V −to 1.5 V and from 1.5 V to 1 V, (BtB) and 20 km fibers, respectively; in the DS direction, received sensitivities were 33.6 dBm and respectively. Figure 3c shows the measured BER with Ibias = 60 mA, 80 mA, and 100− mA at VPP = 2 V 33.3 dBm for BtB and 20 km fibers, respectively. and− Pi = −10 dBm, in which the EAM works under reverse bias and the SOA operates at forward bias. For bidirectional single-ONU and multi-ONU transmissions, the BER curves and eye diagrams The sensitivities of −27 dBm, −29.5 dBm, and −29 dBm were obtained for Ibias = 60 mA, 80 mA, and 100 with the optimal condition of Pi = 10 dBm, VPP = 2 V,and Ibias = 80 mA are shown in Figure4. Figure4a −4 − mApresents at BER theof 2.4 US × and 10 DS for received a 10 Gb/s sensitivities signal with of 29.5 PRBS. dBm and 33.5 dBm for BtB transmission, and the − − sensitivitiesWe show reachthe BER28.9 curves dBm and of US33.1 and dBm DS after unidirec the 20 kmtional fiber transmissions transmission at BERin Figure= 2.4 3d10 under4 for the − − × − appropriate10 Gb/s. These condition are 0.3 of dB Pi and = 0.6−10 dB dBm, larger V thanPP = the2 V, unidirectional and Ibias = US80 transmissionmA. At BER for = BtB 2.4 and × 10 20−4 km for the unidirectionalfibers, respectively, US transmission, and the penalty the sensitivities between BtB andreached 20 km − fibers29.8 dBm in bidirectional and −29.5 USdBm transmission for back-to-back is (BtB)0.6 and dB (0.320 km larger fibers, than therespectively; unidirectional in the US transmission)DS direction, due received to backscattering sensitivities and were reflection –33.6 eff dBmects and −33.3in dBm the bidirectional for BtB and system 20 km [25 fibers,]. respectively. ForAfter bidirectional accomplishing single-ONU the bidirectional and multi-ONU transmission tran ofsmissions, the single ONU,the BER a neighboring curves and ONU2 eye diagrams was added with 50 GHz spacing. The added ONU2 and transceiver2 used amplitude modulators, as shown with the optimal condition of Pi = −10 dBm, VPP = 2 V, and Ibias = 80 mA are shown in Figure 4. Figure in Figure2. Compared with the single-ONU transmission, the multi-ONU transmission needed to 4a presents the US and DS received sensitivities of −29.5 dBm and −33.5 dBm for BtB transmission, consider the negative effect of crosstalk and cross-phase modulation between signals [26]. As Figure4b and the sensitivities reach −4 28.9 dBm and −33.1 dBm after the 20 km fiber transmission at BER = 2.4 × shows, at BER = 2.4 10− for 10 Gb/s, for US, the sensitivities reach 28.8 dBm and 28.2 dBm for BtB −4 × − − 10 andfor 2010km Gb/s. fibers, These respectively; are 0.3 dB for DS,and the 0.6 received dB larger sensitivities than the are unidirectional33 dBm and US32.6 transmission dBm for BtB and for BtB − − and20 20 km km fibers, fibers, respectively. respectively, Compared and the with penalty the bidirectional between transmission BtB and 20 in km Figure fibers4a, therein bidirectional is a 0.7 dB US transmission is 0.6 dB (0.3 larger than the unidirectional US transmission) due to backscattering and reflection effects in the bidirectional system [25]. After accomplishing the bidirectional transmission of the single ONU, a neighboring ONU2 was added with 50 GHz spacing. The added ONU2 and transceiver2 used amplitude modulators, as shown in Figure 2. Compared with the single-ONU transmission, the multi-ONU transmission needed to consider the negative effect of crosstalk and cross-phase modulation between signals [26]. As Figure 4b shows, at BER = 2.4 × 10−4 for 10 Gb/s, for US, the sensitivities reach −28.8 dBm and −28.2 Appl. Sci. 2020, 10, x FOR PEER REVIEW 9 of 11 dBm for BtB and 20 km fibers, respectively; for DS, the received sensitivities are −33 dBm and −32.6 dBm for BtB and 20 km fibers, respectively. Compared with the bidirectional transmission in Figure Appl. Sci. 2020, 10, 8960 9 of 11 4a, there is a 0.7 dB and a 0.5 dB power penalty between the single-ONU and the multi-ONU in US and DS transmissions, respectively. It can be seen that the sensitivity penalty caused by the interferenceand a 0.5 between dB power multiple penalty between channels the was single-ONU less than and 1 dB, the so multi-ONU the influence in US of and the DS neighboring transmissions, ONU2 on ONU1respectively. was not It canserious. be seen that the sensitivity penalty caused by the interference between multiple channels was less than 1 dB, so the influence of the neighboring ONU2 on ONU1 was not serious.

(a)

(b)

FigureFigure 4. The 4. The BER BER against against the the received received power power in 1010 GbGb/s/s bidirectional bidirectional transmission: transmission: (a) single(a) singleONU−ONU − bidirectionalbidirectional transmission; transmission; (b ()b multi) multi−ONUONU bidirectional bidirectional transmission.transmission. − 4. Conclusions 4. Conclusions In this paper, we investigated an eﬀective optimization of an SOA and an SOA-EAM in an NGPON2-In this paper, ONU we and investigated analyzed their an impacts effective on optimization the system. The of mainan SOA characteristics and an SOA-EAM of the SOA, in an NGPON2-namely ONU the gain and and analyzed NF, were obtainedtheir impacts from theon relationship the system. between The main the bias characteristics current and theof inputthe SOA, namelypower. the Thegain proper and NF, conditions were obtained of the SOA from for the phase relationship and amplitude between modulation the bias were current a bias and current the input power.of 200 The mA proper with anconditions input power of the of SOA25 dBm for and phas5e dBm,and amplitude respectively. modulation We set up and were analyzed a bias thecurrent − − of 20010 mA Gb/s with single-ONU an input and power the multi-ONU of −25 dBm bidirectional and −5 dBm, DWDM-PON respectively. based We on an set SOA-EAM. up and analyzed The results the 10 show that the sensitivities in 20 km US and DS transmissions reach 28.9 dBm and 33.1 dBm for the Gb/s single-ONU and the multi-ONU bidirectional DWDM-PON −based on an SOA-EAM.− The results single-ONU and reach 28.2 dBm and 32.6 dBm for the multi-ONU at the BER of 2.4 10 4 with show that the sensitivities− in 20 km US and− DS transmissions reach −28.9 dBm and −33.1× dBm− for the VPP = 2 V (Pi = 10 dBm, Ibias = 80 mA), respectively. As evidenced by their measured performance, single-ONU and reach− −28.2 dBm and −32.6 dBm for the multi-ONU at the BER of 2.4 × 10−4 with VPP such integration devices as SOAs and SOA-EAMs have the potential for low-cost use in metro-area = 2 V (Pi = −10 dBm, Ibias = 80 mA), respectively. As evidenced by their measured performance, such network applications of NGPON2s. integration devices as SOAs and SOA-EAMs have the potential for low-cost use in metro-area network applications of NGPON2s. Appl. Sci. 2020, 10, 8960 10 of 11

Author Contributions: X.R.C. and G.Y.C. conceived and designed the experiments; X.R.C. performed the experiments; G.Y.C. supervised the project. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by [the Natural Science Foundation of Jiangsu Province] grant number [BK20170199, BK20180598], [the Fundamental Research Funds for the Central Universities] grant number [JUSRP11835], [Postgraduate Research & Practice Innovation Program of Jiangsu Province] grant number [SJCX20_0762] And The APC was funded by [the Natural Science Foundation of Jiangsu Province] grant number [BK20170199]. Acknowledgments: This work was supported in part by the Jiangsu Provincial Research Center of Light Industrial Optoelectronic Engineering and Technology. Conflicts of Interest: The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References

1. Luo, M.; Wu, D.; Li, W.; Zeng, T.; Zhou, L.; Meng, L.; Li, X.; Yu, S. 100 Gb/s (4 25 Gb/s) real-time coherent × UDWDM-PON with a large power budget. IEEE/OSA J. Opt. Commun. Netw. 2019, 12, A204–A213. [CrossRef] 2. Kim, K.O.; Doo, K.-H.; Lee, H.H.; Kim, S.H.; Park, H.; Oh, J.-Y.; Chung, H.S.; Doo, K.H. High Speed and Low Latency Passive Optical Network for 5G Wireless Systems. J. Light. Technol. 2019, 37, 2873–2882. [CrossRef] 3. Presi, M.; Ciaramella, E. Mutual Seeding of Directly Modulated R-SOAs for Full-Duplex and Single-Wavelength Short Reaches. IEEE Photon. Technol. Lett. 2018, 30, 2064–2067. [CrossRef] 4. Velásquez, J.C.; Tabares, J.; Prat, J. Differential 8-APSK monolithically integrated dual-EML transmitter for flexible coherent PONs. Opt. Lett. 2019, 44, 2760–2763. [CrossRef] 5. Chu, G.Y.; Cano, I.; Brenot, R.; Debrégeas, H.; Prat, J.; Maho, A.; Polo, V. First demonstration and field trial on multi-user UDWDM-PON full duplex PSK-PSK with single monolithic integrated dual-output-DFB-SOA based ONUs. Opt. Lett. 2016, 41, 4696. [CrossRef] 6. Chu, G.Y.; Cano, I.N.; Polo, V.; Kazmierski, C.; Brenot, R.; Prat, J. Monolithically Integrated Dual-Output DEML for Full Duplex DPSK-ASK and DPSK-SSB ONU in Ultra-Dense Channel Spaced Access Network. J. Light. Technol. 2016, 34, 2042–2048. [CrossRef] 7. Ngo, M.N.; Poingt, F.; Brenot, R.; Nguyen, H.T.; Gosset, C.; Erasme, D.; Deniel, Q.; Genay, N.; Guillamet, R.; Lagay, N.; et al. ElectroAbsorption Modulated Laser Integrated with a Semiconductor Optical Amplifier for 100-km 10.3 Gb/s Dispersion-Penalty-Free Transmission. J. Light. Technol. 2013, 31, 232–238. [CrossRef] 8. Reis, J.D.; Shukla, V.; Stauffer, D.R.; Gass, K. Technology Options for 400G Implementation. Optical Networking Forum (OIF) White Paper. Technical Report. July 2015. Available online: http://www.oiforum.com/wp- content/uploads/OIF-Tech-Options-400G-01.0.pdf (accessed on 1 April 2020). 9. ITU-T Recommendation. G.989.2: 40-Gigabit-Capable Passive Optical Networks (NG-PON2): Physical Media Dependent (PMD) Layer Specification; ITU: Geneva, Switzerland, 2019. 10. Niu, P.; Nablo, B.J.; Bhadriraju, K.; Reyes, D.R. Uncovering the Contribution of Microchannel Deformation to Impedance-Based Flow Rate Measurements. Anal. Chem. 2017, 89, 11372–11377. [CrossRef] 11. Wang, N.; Kim, I.; Vassilieva, O.; Ikeuchi, T.; Wen, H.; Antonio-Lopez, J.E.; Alvarado-Zacarias, J.C.; Liu, H.; Fan, S.; Habib, S.; et al. Low-crosstalk few-mode EDFAs using retro-reflection for single-mode fiber trunk lines and networks. Opt. Express 2019, 27, 35962–35970. [CrossRef] 12. Goki, P.N.; Imran, M.; Porzi, C.; Toccafondo, V.; Fresi, F.; Cavaliere, F.; Poti, L. Lossless WDM PON Photonic Integrated Receivers Including SOAs. Appl. Sci. 2019, 9, 2457. [CrossRef] 13. Zhou, H.; Shen, Y.; Chen, M.; Fei, C.; He, J. Cost-efficient, polarization-insensitive and widely-tunable AOWC of OFDM signal based on FWM in SOA. Opt. Express 2019, 27, 38553–38566. [CrossRef][PubMed] 14. Kotb, A.; Zoiros, K.E.; Guo, C. Ultrafast performance of all-optical AND and OR logic operations at 160 Gb/s using photonic crystal semiconductor optical amplifier. Opt. Laser Technol. 2019, 119, 105611. [CrossRef] 15. Melikyan, A.; Kaneda, N.; Kim, K.; Baeyens, Y.; Dong, P. Differential drive I/Q modulator based on silicon photonic electro-absorption modulators. J. Light. Technol. 2020, 38, 1. [CrossRef] 16. Moreno, V.; Connelly, M.J.; Romero-Vivas, J.; Krzczanowicz, L.; Mora, J.; Muriel, M.A.; Capmany, J.; Galue, V.M.Integrated 16-ps Pulse Generator Based on a Reflective SOA-EAM for UWB Schemes. IEEE Photon. Technol. Lett. 2016, 28, 2180–2182. [CrossRef] Appl. Sci. 2020, 10, 8960 11 of 11

17. Naughton, A.; Talli, G.; Porto, S.; Antony, C.; Ossieur, P.; Townsend, P.D. Design Optimization of R-EAM-SOA for Long-Reach Carrier-Distributed Passive Optical Networks. J. Light. Technol. 2014, 32, 4386–4392. [CrossRef] 18. Wu, J.-P.; Ding, W.-Z.; Chiu, Y.-J. Low-Pattern-Dependence Prechirp Optical Modulation by Using Saturation Behaviors of SOA-Integrated EAM. J. Light. Technol. 2013, 31, 3651–3657. [CrossRef] 19. Jeong, J.S.; Kim, H.-S.; Choi, B.-S.; Kim, D.C.; Kim, K.-S.; Park, M.-R.; Kwon, O.-K. Mitigation of Rayleigh crosstalk using noise suppression technique in 10-Gb/s REAM-SOA. Opt. Express 2012, 20, 26373–26378. [CrossRef] 20. Shindo, T.; Fujiwara, N.; Kanazawa, S.; Nada, M.; Nakanishi, Y.; Yoshimatsu, T.; Kanda, A.; Chen, M.; Ohiso, Y.; Sano, K.; et al. High Power and High Speed SOA Assisted Extended Reach EADFB Laser (AXEL) for 53-Gbaud PAM4 Fiber-Ampliﬁer-Less 60-km Optical Link. J. Light. Technol. 2020, 38, 2984–2991. [CrossRef] 21. Sancho, J.; Lloret, J.; Gasulla, I.; Sales, S.; Capmany, J. Figures of merit for microwave photonic phase shifters based on semiconductor optical ampliﬁers. Opt. Express 2012, 20, 10519–10525. [CrossRef] 22. Ruzbarsky, J.; Turan, J.; Ovsenik, L. Stimulated Brillouin scattering in DWDM all optical communication systems. In Proceedings of the 2016 26th International Conference Radioelektronika (RADIOELEKTRONIKA), Kosice, Slovakia, 19–20 April 2016; pp. 395–398. 23. ITU-T Recommendation. G.694.1: Spectral Grids for WDM Applications: DWDM Frequency Grid; ITU: Geneva, Switzerland, 2002. 24. Lee, D.-H.; Jeong, J.S.; Kim, K.-S.; Kim, H.-S.; Kim, D.C.; Park, M.-R.; Han, Y.-T.; Kwon, O.K.; Kwon, O.-K. Design and performance of 10-Gb/s L-band REAM-SOA for OLT Transmitter in next generation access networks. Opt. Express 2015, 23, 2339–2346. [CrossRef] 25. Yeh, C.-H.; Jhao-Ren, C.; You, W.-Y.; Chow, C.W. Hybrid WDM FSO Fiber Access Network With Rayleigh Backscattering Noise Mitigation. IEEE Access 2020, 8, 96449–96454. [CrossRef] 26. Ceballos-Herrera, D.E.; Lopez-Coyote, M.; Gutierrez-Castrejon, R. Crosstalk Characteristics of WDM Channels in Quasi-Homogeneous Multicore Fibers. IEEE Photon. Technol. Lett. 2020, 32, 759–762. [CrossRef]

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional aﬃliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).