1. Introduction

To accommodate the explosive growth of Internet and telecommunication traffic in networks, wavelength-division multiplexing (WDM) has emerged as a promising technique in which multiple channels are operated along a single fiber simultaneously, each on a different wavelength [1]. WDM networks can rely upon a data stream’s wavelength to help determine the routing and use wavelength conversion to enable reconfigurable paths and rapid resolution of contention [2]. Moreover, many network switching applications can beneficially use the exchange of two different data streams located on two separate wavelength channels in order to increase the efficiency and flexibility of networks [3,4]. Desirable features of such data exchange would be ultrafast response, transparency to the modulation format, and capability of achieving the exchange function in a single optical element (e.g., a single fiber).

Optical nonlinearities (e.g., χ⁽²⁾ and χ⁽³⁾) are potential candidates for achieving data exchange with the aforementioned features. Previous reports of data exchange between different wavelengths include the use of degenerate FWM in an optical parametric loop mirror (PALM) [5,6], non-degenerate FWM in a highly nonlinear fiber (HNLF) [7–10], and cascaded sum- and difference-frequency generation (cSFG/DFG) in a periodically poled lithium niobate (PPLN) waveguide [11]. For example, PALM-based wavelength interchange for on-off keying (OOK) [5] and 10-Gbit/s differential phase-shift keying (DPSK) signals [6], non-degenerate FWM based wavelength exchange for 10-Gbit/s non-return-to-zero (NRZ) signals [7], byte-level exchange for 10-Gbit/s return-to-zero (RZ) signals [8], pulsed pump exchange for tributary demultiplexing [9], phase-transparent data exchange for 40-Gbit/s DPSK signals [10], and PPLN-based time- and channel-selective data exchange between 40-Gbit/s WDM channels [11] have been demonstrated. The reported results of optical data exchange were for binary formats, i.e., OOK or DPSK [5–10,11]. Advanced multi-level modulation format is of great importance for enabling high-spectral efficiency transmission of high-bit-rate services. In particular, differential quadrature phase-shift keying (DQPSK) which doubles the spectral efficiency compared with OOK/DPSK, has attracted significant attention recently [12]. A laudable goal would be to achieve data exchange for spectrally-efficient DQPSK signals.

In this paper, we investigate the optical data exchange of DQPSK signals using the signal depletion effect of non-degenerate FWM in an HNLF [13]. 100-Gbit/s return-to-zero DQPSK (RZ-DQPSK) data exchange between two different wavelengths is demonstrated with a power penalty of less than 5 dB at a bit-error rate (BER) of 10⁻⁹. In addition, we discuss and characterize the tolerance of the temporal pump phase misalignment (+/−2 ps) and the dynamic range of the input signal power (~20 dB) for the 100-Gbit/s DQPSK data exchange.

2. Concept and principle of operation

Figure 1 depicts the concept and operation principle for DQPSK data exchange between two different wavelengths (S1: ${\lambda }_{S1}$, S2: ${\lambda }_{S2}$). The four-level phase information carried by two DQPSK signals at different wavelengths is swapped after the data exchange. To perform the data exchange of DQPSK signals, phase-transparent data exchange is highly desirable which can be implemented by exploiting the signal depletion effect of non-degenerate FWM in a single HNLF. As shown in Fig. 1(a), when signal 1 (S1: ${\lambda }_{S1}$) and two continuous-wave (CW) pumps (P1: ${\lambda }_{P1}$, P2: ${\lambda }_{P2}$) are sent through the HNLF with S1 and P1 set symmetrically with respect to the zero-dispersion wavelength (ZDM) of the HNLF, S1 and P1 photons are consumed to produce photons of S2 and P2 during the non-degenerate FWM process. Thus the depletion of S1 is expected with its data information transparently copied onto a newly generated S2. Similarly, as shown in Fig. 1(b), the depletion of S2 accompanied by the generation of S1 can be achieved as S2 and two pumps are launched into the HNLF. As shown in Fig. 1(c), in the presence of two signals and two pumps at the input of HNLF with S1(S2) and P1(P2) symmetric relative to the ZDW of the HNLF, S1(S2) can be extinguished and converted to S2(S1), resulting in the implementation of data exchange between S1 and S2.

 

Fig. 1 Concept and principle of non-degenerate FWM based signal depletion and data exchange. (a) Signal 1 (S1) depletion. (b) Signal 2 (S2) depletion. (c) S1, S2 data exchange.

Download Full Size | PDF

We can derive simple linear relationships ($A_{S1}^{\text{'}}\propto A_{S2}\cdot A_{P2}\cdot A_{P1}^{*}$,$A_{S2}^{\text{'}}\propto A_{S1}\cdot A_{P1}\cdot A_{P2}^{*}$) of the complex amplitudes between the output signals ($A_{S1}^{\text{'}}$,$A_{S2}^{\text{'}}$) and input signals and pumps ($A_{S1}$, $A_{S2}$, $A_{P1}$, $A_{P2}$) under the non-depletion approximation and proper control of pump powers [10]. The linear relationships of complex amplitudes imply that non-degenerate FWM based data exchange has the characteristic of transparency to the modulation format, including the phase transparency. We can further obtain the corresponding phase relationships of ${\varphi }_{S1}{}^{\text{'}}={\varphi }_{S2}+{\varphi }_{P2}-{\varphi }_{P1}$ and ${\varphi }_{S2}{}^{\text{'}}={\varphi }_{S1}+{\varphi }_{P1}-{\varphi }_{P2}$. It is worth noting that phase modulation applied to the two pumps (${\phi }_{P1}$, ${\phi }_{P2}$) helps effectively suppress the stimulated Brillouin scattering (SBS) effect. As a result, the pump power is efficiently utilized in the non-degenerate FWM process, which benefits the effective signal depletion and data exchange. Remarkably, the pump phase transfer (${\phi }_{P1}-{\phi }_{P2}\ne 0$) to the exchanged signals does not impact on the OOK data exchange but could cause severe degradation on the DQPSK data exchange. Fortunately, according to the deduced phase relationships, it is possible to cancel the pump phase transfer by applying the precisely identical phase modulation to the two pumps (i.e., ${\phi }_{P1}$ = ${\phi }_{P2}$) [14], which makes it applicable to implement the data exchange of two DQPSK signals.

3. Experimental setup

Figure 2 shows the experimental setup for data exchange of 100-Gbit/s DQPSK signals. Two CW lasers are sent to a 100-Gbit/s (50-GSymbol/s) RZ-DQPSK transmitter (Tx) to produce two single-polarized 100-Gbit/s 2⁷-1 pseudo-random binary sequence (PRBS) RZ-DQPSK signals (S1, S2) at different wavelengths. The 100-Gbit/s RZ-DQPSK transmitter is a thermally stable Lithium Niobate Mach-Zehnder modulator with a nested Mach-Zehnder interferometer structure (SHF 46214A) followed by a pulse curve. Two CW pumps (P1, P2) together with two 100-Gbit/s RZ-DQPSK signals (S1, S2) are coupled into a 1-km piece of HNLF with a nonlinear coefficient of 9.1 W⁻¹·km⁻¹, a ZDW of ~1552 nm, a dispersion slope of ~0.04 ps/nm²/km, and a fiber loss of 0.45 dB/km. The DQPSK data exchange is realized in the HNLF based on the signal depletion effect of the non-degenerate FWM process. Note that the two pumps are phase-modulated for the SBS suppression using the same phase modulator (PM) driven by a 10-Gbit/s 2⁷-1 PRBS. These two pumps are separated using band-pass filters (BPFs), polarization adjusted by polarization controllers (PCs), amplified via high-power erbium-doped fiber amplifiers (EDFAs), and recombined with an optical coupler (OC). The BPFs after EDFAs are employed to suppress the amplified spontaneous emission (ASE) noise. It is highly desired to precisely match the two pump paths using tunable optical delay lines (ODLs) for the pump phase cancellation in the exchanged signals. The temporal pump phase misalignment could cause performance degradation on the data exchange. The exchange performance is optimized by the proper adjustment of pump powers and polarization states of pumps and signals. A 50-GHz delay line interferometer (DLI) is used to demodulate the in-phase (Ch. I) and quadrature (Ch. Q) components of 100-Gbit/s DQPSK signals.

 

Fig. 2 Experimental setup for 100-Gbit/s DQPSK data exchange. Rx: direct detection receiver.

Download Full Size | PDF

4. Experimental results and discussions

Figure 3 depicts the measured spectra for 100-Gbit/s DQPSK optical data exchange. Two 100-Gbit/s RZ-DQPSK signals (S1: 1539.5 nm, S2: 1545.3 nm) and two pumps (P1: 1564.4 nm, P2: 1558.6 nm) are sent to the HNLF. The input power of each signal and pump fed into the HNLF is about 0 and 23.8 dBm, respectively. Figure 3(a) and (b) depict spectra of the signal depletion (S1: ~19.8 dB, S2: ~19.1 dB) due to non-degenerate FWM when only S1 or S2 and two pumps are present at the input of HNLF. Shown in Fig. 3(c) is the spectrum of 100-Gbit/s RZ-DQPSK data exchange in the presence of both two signals and two pumps.

 

Fig. 3 Measured spectra for 100-Gbit/s DQPSK data exchange. (a) Depletion of S1 (S1: ON, S2: OFF, P1: ON, P2: ON); (b) Depletion of S2 (S1: OFF, S2: ON, P1: ON, P2: ON). (c) Data exchange between S1 and S2 (S1: ON, S2: ON, P1: ON, P2: ON).

Download Full Size | PDF

Figure 4 displays the measured temporal waveforms of the demodulated in-phase (Ch. I) and quadrature (Ch. Q) components for the 100-Gbit/s DQPSK data exchange. It can be clearly observed that the data information carried by two 100-Gbit/s RZ-DQPSK signals is successfully swapped after the non-degenerate FWM based data exchange. In addition, by comparing the waveforms after wavelength conversion (i.e. only S1 or S2 is present) and after data exchange (i.e. both S1 and S2 are present), we can observe the degradation of waveforms after data exchange with added noise, which can be ascribed to the beating effect of in-band interference between the newly converted signal and the original residual signal. Figure 5 shows the measured constellation diagrams of different signals for the 100-Gbit/s DQPSK data exchange. Four phase levels of all 100-Gbit/s RZ-DQPSK signals are clearly observed.

 

Fig. 4 Demodulated waveforms (Ch. I and Ch. Q) for 100-Gbit/s DQPSK data exchange. (a1)(a2) S1 before data exchange (P1: OFF, P2: OFF). (b1)(b2) S2 before data exchange (P1: OFF, P2: OFF). (c1)(c2) S1 after wavelength conversion (WC: S2 to S1) (S1: OFF, S2: ON, P1: ON, P2: ON). (d1)(d2) S1 after data exchange (Ex.: S2 to S1) (S1: ON, S2: ON, P1: ON, P2: ON). (e1)(e2) S2 after wavelength conversion (WC: S1 to S2) (S1: ON, S2: OFF, P1: ON, P2: ON). (f1)(f2) S2 after data exchange (Ex.: S1 to S2) (S1: ON, S2: ON, P1: ON, P2: ON). (a1)-(f1) Ch. I. (a2)-(f2) Ch. Q.

Download Full Size | PDF

 

Fig. 5 Constellation diagrams for 100-Gbit/s DQPSK data exchange. (a) S1: back-to-back. (b) S2: back-to-back. (c) S1 after wavelength conversion (WC: S2 to S1) (S1: OFF, S2: ON). (d) S2 after wavelength conversion (WC: S1 to S2) (S1: ON, S2: OFF). (e) S1 after data exchange (Ex.: S2 to S1) (S1: ON, S2: ON). (f) S2 after data exchange (Ex.: S1 to S2) (S1: ON, S2: ON).

Download Full Size | PDF

Figure 6 plots the BER curves and balanced eye diagrams for the 100-Gbit/s DQPSK data exchange. Less than 1.2-dB power penalty at a BER of 10⁻⁹ is obtained for the 100-Gbit/s DQPSK wavelength conversion with only one signal (S1 or S2) present. Less than 5-dB power penalty at a BER of 10⁻⁹ is observed for the 100-Gbit/s DQPSK data exchange. The extra power penalty of data exchange compared to wavelength conversion could be ascribed to the beating effect between the newly converted signal and the original residual signal.

 

Fig. 6 BER curves and balanced eye diagrams for the 100-Gbit/s DQPSK data exchange.

Download Full Size | PDF

We further investigate the tolerance of the temporal pump phase misalignment and the dynamic range of the input signal power for the 100-Gbit/s RZ-DQPSK data exchange. Figure 7 depicts the impact of time misalignment between two phase-modulated pumps on the performance of wavelength conversion and data exchange (i.e., relative power penalty compared to the case of perfect pump phase alignment). It is found that the performance of the wavelength conversion and data exchange suffers rapid degradation as the pump phase misalignment goes beyond +/−2 ps. Such phenomena can be briefly explained as follows. Large temporal pump phase misalignment causes incomplete pump phase cancellation, and thus the resultant residual pump phase transfer to the phase noise degrades the performance of the wavelength conversion and data exchange. Figure 7 also displays typical balanced eye diagrams of the exchanged signals under different values of temporal pump phase misalignment. The reduced eye opening and increased noise are observed with pump phase misalignment of 3 and 4 ps as compared to Fig. 6 with perfect pump phase alignment. In particular, under an even larger pump phase misalignment of −10 ps, we can see nearly complete eye closure of the exchanged signals. As a consequence, for the phase-transparent optical data exchange of 100-Gbit/s DQPSK signals, the optimal pump phase cancellation in the exchanged signals is of great importance to achieve the optimized exchange performance.

 

Fig. 7 Impact of pump phase misalignment on the exchange performance.

Download Full Size | PDF

Figure 8 shows the received power at a BER of 10⁻⁹ as a function of the input signal power. As the input signal power changes from −12.0 to 8.1 dBm, the variation of the received power at a BER of 10⁻⁹ is estimated to be less than 3.5 dB. An approximate 20-dB dynamic range of the input signal power is achieved for the non-degenerated FWM based optical data exchange of 100-Gbit/s RZ-DQPSK signals.

 

Fig. 8 Dynamic range of the input signal power for 100-Gbit/s DQPSK data exchange.

Download Full Size | PDF