An all-optical frequency upconversion technique using a quasi optical single sideband (q-OSSB) signal in a nonlinear semiconductor optical amplifier (NSOA) for radio-over-fiber applications is proposed and experimentally demonstrated. An optical radio frequency signal (fRF = 37.5 GHz) in the form of a q-OSSB signal is generated by mixing an optical intermediate frequency (IF) signal (fIF = 2.5 GHz) with an optical local oscillator signal (fLO = 35 GHz) utilizing coherent population oscillation and cross gain modulation effects in an NSOA. The phase noise, conversion efficiency, spurious free dynamic range (SFDR), and transmission characteristics of the q-OSSB signal are investigated.
©2011 Optical Society of America
The demands for broadband wireless multimedia services require advanced networks that can support high data rates for end users [1–3]. A radio-over-fiber (RoF) system is an attractive candidate for providing such broadband wireless services since it utilizes an optical fiber having ultra-low transmission loss and ultra-wide bandwidth characteristics for transmission of optical signals modulated with radio frequency (RF) signals. The RoF systems that transmit optical signals in the form of an optical double sideband (ODSB) suffer from the power fading problem associated with the optical fiber chromatic dispersion and inefficiency in bandwidth, resulting in system performance degradation . The use of an optical single sideband (OSSB) signal can solve the power fading problem and increase efficiency in bandwidth, and various methods for the generation of OSSB signals using a Mach-Zehnder modulator (MZM) , an electroabsorption modulated laser (EML) , and narrow optical filters  have been reported. Recently, the generation of quasi-OSSB (q-OSSB) signals using coherent population oscillation (CPO) effects in a nonlinear semiconductor optical amplifier (NSOA) has been reported [7–13]. The q-OSSB signals generated using an NSOA were effective in reducing the power fading problem.
In this paper, the all-optical frequency upconversion and generation of q-OSSB signals using cross gain modulation (XGM) and CPO effects, respectively, in a nonlinear semiconductor optical amplifier (NSOA) are proposed and experimentally demonstrated.
2. Concept of q-OSSB signal upconversion
Figure 1 shows the concept of q-OSSB upconversion utilizing the CPO and XGM effects in an NSOA. An optical intermediate frequency (IF) signal (λIF) and an optical local oscillation (LO) signal (λLO), the spectra of which are shown in Fig. 1(a), are generated in the central station (CS). These optical signals are directed to an NSOA at the CS. When optical signals with sufficient power are directed in an NSOA biased in a gain saturation region, the CPO effects take place. The CPO effects change the third order susceptibility of the waveguiding medium in the NSOA and modify phase and power of the sidebands of the optical signals, producing a q-OSSB signal [11,13]. Unlike common OSSB signals quasi-OSSB signals have a remnant of one of the sidebands that has to be completely removed. At the same time, mixing the optical IF signal and the optical LO signal due to the XGM effect in the NSOA takes place, producing an optical RF signal having the spectra shown in Fig. 1(b). Dynamics associated with CPO and XGM effects are governed by the carrier lifetime of the NSOA [12–14]. The carrier lifetime of a typical NSOA is in the order of several tens of pico-seconds, and thus, the bandwidth of the q-OSSB upconversion utilizing the CPO and XGM effects is approximately several tens of GHz. The optical RF signal is then transmitted through a single mode fiber (SMF) to the remote antenna station (RAS), where it is converted into an electrical RF signal by a photodetector. The spectra of the electrical RF signal are shown in Fig. 1(c). Unlike other all-optical OSSB upconversion schemes [14–16], the proposed q-OSSB upconversion scheme utilizes a single NSOA for OSSB signal generation and upconversion, which can make the RoF system simple and compact.
3. Experiment and results
The characteristics of the all-optical frequency upconverter for a q-OSSB signal utilizing an NSOA including the conversion efficiency, noise performance, and linearity, and dispersion of the upconverted q-OSSB signal transmitted over an SMF are investigated.
3.1. Experimental setup
The experimental setup for upconverted q-OSSB signal generation and transmission for an RoF system is shown in Fig. 2 . An optical IF signal in the form of an optical double sideband (ODSB) was generated by modulating a laser diode (LD1) having a wavelength (λIF) of 1540.56 nm using an electroabsorption modulator (EAM1). The frequency of the electrical IF signal (fIF) was 2.5 GHz. For two-tone measurements, the frequencies of the electrical IF signals were 2.5 GHz ± 2.5 MHz. The power of the optical IF signal was varied between −25 and −5 dBm by a variable optical attenuator (VOA1). An optical LO signal in the form of an ODSB was also generated by modulating a laser diode (LD2) having a wavelength (λLO) of 1551.72 nm using an electroabsorption modulator (EAM2). The frequency of the electrical LO signal (fLO) was 35 GHz. An erbium-doped amplifier (EDFA) and the VOA2 were used to maintain the optical LO signal power between −25 and −2 dBm. After being combined by an optical coupler (OC), the optical LO and IF signals were directed to the NSOA (SOA-NL-OEC-1550). The NSOA was biased at a current of 350 mA so that it was operated in a gain saturation region. The EAM1 and EAM2 used for the intensity modulation of the optical IF and LO signals, respectively, were biased at the zero chirp point, where the best conversion efficiency was observed. The upconverted q-OSSB signal was transmitted through a 39 km-long single mode optical fiber and opto-electronically converted to an electrical RF signal by a photodetector (PD) and then amplified by a power amplifier (PA). The relative frequency responses of the PD and PA are shown in Fig. 3 .
3.2. Spectra of the upconverted q-OSSB signal
Figure 4(a) shows the spectrum of the optical LO signal at the input of the NSOA and that of the upconverted q-OSSB signal at the output of the NSOA. In this experiment the electrical IF and LO powers were set to 0 dBm, the optical IF and LO powers were set to −10 and −5 dBm, and the electrical IF and LO frequencies were set to 2.5 and 35 GHz, respectively. The NSOA was operated at the gain saturation region (350mA). Due to the CPO effects, the magnitudes of the two sidebands of the upconverted q-OSSB signal were modified. As shown in Fig. 4(a), the power of the red-shifted sideband is larger than that of the blue-shifted one. The ratio of the red-shifted sideband power to the blue-shifted sideband power as a function of the optical LO power for different optical IF powers is shown in Fig. 4(b). The ratio increased as the optical LO power was increased since the CPO effects become more effective as the optical LO signal power is increased . However, these effects decreased as the optical IF power was increased. The effect of the optical IF power on the ratio of the sideband power is under investigation. Figure 4(c) shows the electrical spectrum of the q-OSSB signal after it was converted by a photodetector and then amplified by a power amplifier with a power gain of 20 dB. The optical IF and LO signal powers were identical to those used for Fig. 4(a). The rather poor ratio of the RF to LO signal power can be improved by increasing the ratio of the electrical IF signal power (i.e. input to EAM1) to the electrical LO power (i.e. input to EAM2). The poor signal-to-noise ratio observed in Fig. 4(c) is attributed to a large resolution bandwidth of the ESA used for measuring the data.
3.3. Phase noise characteristics of the upconverted q-OSSB signal
Figure 5 shows phase noise of the upper sideband of the electrical RF signal along with those of the electrical IF and LO signals for comparison. The electrical RF signal was obtained by opto-electronic conversion of the q-OSSB RF signal (at the output of the NSOA) using a photodetector and the electrical IF and LO signals were obtained by opto-electronic conversion of the optical IF and LO signals (at the input of the NSOA) using a photodetector, respectively. The phase noise of the upconverted electrical RF signal was limited by that of the electrical LO signal source. The phase noise of the upconverted RF signal was −74.7 dBc/Hz at the offset frequency of 1 kHz.
3.4. Conversion efficiency
Figure 6 shows the conversion efficiency, which is defined as the ratio of the power of the electrical RF signal (the USB signal shown in Fig. 4(c)) to the power of the electrical IF signal (measured at the input of the NSOA), as a function of the optical LO power for different optical IF powers. The effect of the optical IF and LO signal powers on the conversion efficiency was investigated since the NSOA is an all-optical frequency upconverter. It was observed that the conversion efficiency increased as the optical LO power was increased and decreased as the optical IF power was increased. The tendency of the conversion efficiency is similar to that of the optical sideband of the q-OSSB signal shown in Fig. 4(b).
3.5. Linearity of the q-OSSB upconverter
Characterization of the linearity of the q-OSSB upconverter was done by a two tone test with fIF = 2.5 GHz ± 2.5 MHz. Figure 7 shows the powers of the fundamental and third inter-modulation components of the electrical RF signal (USB) as a function of the electrical IF power after q-OSSB up-conversion. It also shows the powers of the fundamental and third inter-modulation components of the electrical IF signal detected at the output of VOA1 as a function of the electrical IF power before q-OSSB up-conversion. In this two tone test experiment, the optical IF and LO powers were fixed at −10 dBm and −6 dBm, respectively, and the electrical LO power was set to 4 dBm. The minimum detectable signal (MDS) levels estimated from the noises of various components were −130.9 dBm before q-OSSB up-conversion and −121.5 dBm after q-OSSB up-conversion . The MDS levels were limited by the relative intensity noise (RIN) from the lasers (LD1 and LD2). The RINs of LD1 and LD2 were −140 and −145 dB/Hz. The estimated SFDRs of the link before and after q-OSSB up-conversion were 60.1 dB·Hz2/3 and 36.4 dB·Hz2/3, respectively.
Figure 8 shows the estimated SFDRs of the q-OSSB upconversion as a function of the electrical LO power for different optical LO and IF powers. In the electrical LO power range from 0 to 4 dBm, the SFDRs of the q-OSSB upconversion increased linearly as the electrical LO power was increased. The SFDRs also increased as the optical IF and LO powers were increased. The SFDR was larger than 38 dB-Hz2/3 for the optical LO and IF powers between −10 dBm and −2 dBm.
3.6. Transmission characteristics of the upconverted q-OSSB signal
The effect of the fiber chromatic dispersion on the transmission characteristics of the upconverted q-OSSB signal was investigated by transmitting the signal over a 39 km-long SMF. For this experiment the electrical IF and LO powers were set to 0 dBm and the optical IF and LO powers were set to −20 dBm and −3 dBm, respectively. The ratio of the two optical sidebands was approximately 15 dB. Figure 9 shows the electrical powers of the upconverted q-OSSB signal and ODSB signal detected by a photodetector measured with the IF frequency set to 2.5 GHz and the electrical LO frequency scanned from 7.5 to 40 GHz. As can be seen in the figure, while conspicuous periodic dips occurred in the electrical power of the ODSB signal, they were suppressed in the electrical power of the q-OSSB signal. The results indicate that the upconverted q-OSSB signal is much less susceptible to optical fiber dispersion.
A novel q-OSSB up-conversion scheme using CPO and XGM effects in an NSOA has been proposed and experimentally demonstrated. The phase noise of the up-converted RF signal is −74.7 dBc/Hz at the RF frequency of 37.5 GHz and with the offset frequency of 1 kHz. The SFDRs of the q-OSSB up-conversion are larger than 38 dB·Hz2/3 for the optical IF and LO powers investigated. The upconverted q-OSSB signal showed reduced power fading problem associated with the single mode fiber dispersion. The results indicate that a q-OSSB up-converter using CPO and XGM effects in an NSOA is simple and compact and can be used for RoF systems requiring high spectral efficiency.
This work was supported in part by a grant from the NRF (2011-0000341) and Bio-imaging Research Center and WCU (R31-2008-000-10026-0) programs at GIST.
References and links
1. H.-S. Kim, T. T. Pham, Y.-Y. Won, and S.-K. Han, “Simultaneous wired and wireless 1.25 Gb/s bidirectional WDM-RoF transmission using multiple optical carrier suppression in FP-LD,” J. Lightwave Technol. 27(14), 2744–2750 (2009). [CrossRef]
2. C. Lim, A. Nirmalathas, M. Bakaul, P. Gamage, K.-L. Lee, Y. Yang, D. Novak, and R. Waterhouse, “Fiber-wireless networks and subsystem technologies,” J. Lightwave Technol. 28(4), 390–405 (2010). [CrossRef]
3. P. Bonenfant, “The evolution of SONET/SDH over WDM,” Opt. Photon. News 14(3), 32–37 (2003). [CrossRef]
4. G. H. Smith, D. Novak, and Z. Ahmed, “Overcoming chromatic-dispersion effects in fiber-wireless systems incorporating external modulators,” IEEE Trans. Microw. Theory Tech. 45(8), 1410–1415 (1997). [CrossRef]
5. H. Kim, “EML-based optical single sideband transmitter,” IEEE Photon. Technol. Lett. 20(4), 243–245 (2008). [CrossRef]
6. J. Park, W. V. Sorin, and K. Y. Lau, “Elimination of the fibre chromatic dispersion penalty on 1550nm millimeter-wave optical transmission,” Electron. Lett. 33(6), 512–513 (1997). [CrossRef]
7. S. E. Schwarz and T. Y. Tan, “Wave interactions in saturable absorbers,” Appl. Phys. Lett. 10(1), 4–6 (1967). [CrossRef]
8. M. Sargent III, “Spectroscopic techniques based on Lamb’s laser theory,” Phys. Rep. 43(5), 223–265 (1978). [CrossRef]
9. R. W. Boyd, M. G. Raymer, P. Marum, and D. J. Harter, “Four-wave parametric interactions in a strongly driven two-level system,” Phys. Rev. A 24(1), 411–423 (1981). [CrossRef]
11. M. Park, K.-C. Kim, and J.-I. Song, “Generation and transmission of a quasi-optical single sideband signal for radio-over-fiber systems,” IEEE Photon. Technol. Lett. 23(6), 383–385 (2011). [CrossRef]
13. Y. Chen, W. Xue, F. Ohman, and J. Mork, “Theory of optical-filtering enhanced slow and fast light effects in semiconductor optical waveguides,” J. Lightwave Technol. 26(23), 3734–3743 (2008). [CrossRef]
14. T. Durhuus, B. Mikkelsen, C. Joergensen, S. Lykke Danielsen, and K. E. Stubkjaer, “All-optical wavelength conversion by semiconductor optical amplifiers,” J. Lightwave Technol. 14(6), 942–954 (1996). [CrossRef]
15. H.-J. Kim and J.-I. Song, “All-optical single-sideband upconversion with an optical interleaver and a semiconductor optical amplifier for radio-over-fiber applications,” Opt. Express 17(12), 9810–9817 (2009). [CrossRef] [PubMed]
16. H.-J. Kim and J.-I. Song, “Full-duplex WDM-based RoF system using all-optical SSB frequency upconversion and wavelength re-use techniques,” IEEE Trans. Microw. Theory Tech. 58(11), 3175–3180 (2010). [CrossRef]
17. V. J. Urick, M. S. Rogge, P. F. Knapp, L. Swingen, and F. Bucholtz, “Wide-band predistortion linearization for externally modulated long-haul analog fiber-optic links,” IEEE Trans. Microw. Theory Tech. 54(4), 1458–1463 (2006). [CrossRef]