## Abstract

We propose phase switching for either main-carrier or subcarriers of two consecutive signal blocks to achieve fading-free double-sideband direct-detection (DD). The proposed approach has twice of the electrical spectral efficiency (SE) of offset OFDM, and the same electrical SE as single-side band (SSB) OFDM. With this scheme, 40-Gb/s DD-OOFDM is successfully received over 80-km SSMF with single polarization and single detector.

©2013 Optical Society of America

## 1. Introduction

Although coherent detection has achieved dramatic success in the past few years due to adoption of powerful electronic digital signal processing [1–4], the optical gears for coherent detection are rather sophisticated, involving polarization multiplexing/de-multiplexing, double IQ modulators and balanced receivers. The high cost of the coherent systems constrains coherent communication for long-haul transport for the time being. As a consequence, direct modulation (DM) and direct detection (DD) has been developed for cost-sensitive short-reach high-speed communication, such as inter-cabinet communication [5–9]. However, the severe dispersion-induced fading becomes a fundamental limitation for DM/DD for 40 Gb/s and beyond communication when the electrical spectrum occupied is wider than 15 GHz and the reach is beyond 10 km.

There exists an interesting window of applications such as interconnect between data centers about 80 km apart that excludes either coherent detection due to high cost, or DM/DD due to serve dispersion-induced fading. This paper focuses on the prospect of using external modulation (EM) and DD for 40 Gb/s and beyond transmission for these medium reach applications. Although DD-OOFDM 120-Gb/s transmission has been reported by using offset OFDM, it has receiver complexity closed to that of coherent detection [10]. It is highly desirable that 40-Gb/s and beyond DD be achieved with single polarization and single detector, and preferably without a need for feedback path to perform pre-distortion. This proposition has placed a high premium on the high electrical spectral efficiency (SE) rather than optical SE for these medium reach systems. As the electrical bandwidth is the dictating factor for the transponder cost – high bandwidth means high modulator and receiver bandwidth, and faster signal processing elements.

Table 1 captures some representative DD experimental demonstrations, including offset single side band (SSB) and Virtual SSB. Although offset OFDM [8,10,13] can achieve fading free and high-speed transmission, it uses twice as much bandwidth as Virtual SSB [11,14], requiring expensive polarization-multiplexing to achieve the same high data rate.

In this work, we propose a novel alternative approach termed block-wise phase switching (BPS) where for two similar consecutive signal blocks, the phase of the main carrier or subcarriers are switched by 90 or 180 degrees, achieving complete separation of double sideband signals. We also propose modification of the conventional optical IQ modulation by including a path for supplying main carrier, enhancing modulation efficiency of IQ modulation. With such a scheme, we have achieved the first single-polarization and single-detector DD of 40 Gb/s over 80 km SSMF. BPS may offer an effective means to achieve DD-based 100G Ethernet with high electrical SE for medium reaches.

## 2. Principle of BPS

The main idea to achieve high electrical SE is to use double sideband modulation. The conventional double side band suffers from dispersion-induced fading at moderate distance of 80 km. The three schemes of BPS are depicted in Fig. 1, including (i) carrier-phase-switching where the main carrier phase is switched 90 degree for two consecutive blocks while the subcarriers remain the same, (ii) signal (subcarrier)-phase-switching where the signal subcarrier phase is switched by 90 degree while the main carrier phase remains the same, and (iii) (signal) set phase reversal (SPR) where the entire subcarriers are partitioned into two sets of C_{1} and C_{2}. C_{1} (or C_{2}) has no overlapping with their mirror image (over the main carrier). In the two consecutive blocks, one of the set, for instance, C_{1} is switched 180 degree. The simplest set partition is C_{1} for upper sideband and C_{2} for lower sideband as shown in Fig. 1(c). All the three schemes can be shown to achieve fading-free detection. However, in this work, only the scheme I is reported and the theory is provided next.

In scheme I, the photo currents for the first and second blocks, I_{1} and I_{2} are given by

The first term is DC that can be ignored. The second term is proportional to the fading-free complex signal, and the third terms represent the signal-to-signal beating noise (SSBN). Assuming we start with sufficient high carrier to signal ratio (CSR) so that the second term is dominant over the third term, an iterative detection can be used to remove the nonlinear term [11]. The preliminary symbol decision can be made and the estimated signal is reconstructed and subtracted from (3); the second iteration of the symbol decision can be made based on the new signal after removing nonlinear term. In this report, we find that we only need one iteration. For scheme I, we already have the time-domain signal as shown in (3) and the first iteration cancellation can be made without symbol decision, reducing computation complexity and error propagation. The principle of scheme II is the same as scheme I.

For scheme III, the operation of subtraction and summation between *I _{1}* and

*I*will recover lower and upper subbands. We define ${E}_{s}={E}_{u}+{E}_{l}$, where ${E}_{u}$ and ${E}_{l}$ respectively represent waveform of upper and lower side band of the signal. The photocurrents of the first and second blocks are given by Eq. (4) and Eq. (5)

_{2}Similar to schemes I and II, we assume a large CSR, and the second terms in (6) and (7) are dominate over the third. The${I}_{3}$and ${I}_{4}$will respectively upper and lower sidebands. The same iteration cancellation technique will be used to remove the nonlinearity terms.

The repetition of the same data over the two blocks in BPS reduces effective SE by half, but using the double sidebands recovers the overall SE. As such, BPS approach has twice of the electrical SE of offset OFDM [10] which requires half of the spectrum as the guard band. The signal block of BPS could comprise one OFDM symbol for multicarrier, or many symbols for single carrier, or narrow band signals for RF over fiber (RoF).

In medium reach systems, optical spectral efficiency is not as an important factor as long haul systems. Because it is relatively easy to introduce a pair of optical amplifiers (OAs) at the two sides to light a new fiber. The cost of OAs is shared by 10 or 20 WDM channels per fiber. The channel spacing can be 100 GHz or even 200 GHz such that the cross-channel nonlinearity is not a critical factor.

## 3. Experimental setup

Although the signal can be of single-carrier or multicarrier, for this proof-of-concept demonstration of BPS, the signal used is a multiband OFDM consisting 3 orthogonal bands. As shown in Fig. 2, an ECL laser is split into two branches, one fed into a multiband OFDM generator, the other to an optical IQ modulator for main carrier phase switching. The OFDM generator consists of an intensity modulator for tone generations, and an optical IQ modulator driven by an arbitrary waveform generator (AWG) at 20 GSa/s. FFT size is 2048 points; CP is 64-point. Only 1124 subcarriers out of 2048 are filled for each OFDM band; the tones spacing of 11.09375 GHz is multiple of subcarrier spacing, satisfying orthogonal band multiplexing (OBM) condition [14] and enabling direct detecting the three bands simultaneously. The main carrier phase is rotated between 0°, 90°, 180°, and 270° (inset (ii) of Fig. 2) aligned with OFDM frames by driving the IQ modulator with complex IQ values of (1,i, −1,i) supplied by an AWG on corresponding OFDM frames. This means the fiber length of the two paths (carrier path and signal path) are carefully matched, so that there is no time delay between the two paths. In practice, these two paths (and two modulators) should be implemented within a compact module, with fixed and matched optical fiber length. The main carrier only needs to switch the phase at frequency of ~10 MHz, and therefore its cost should be insignificant compared to with the signal optical IQ modulator requiring about 15~20 GHz bandwidth. In practice, the main carrier phase switching and OFDM generation can be achieved using a modified optical IQ modulator as shown in the inset (i) of Fig. 2, where an additional path is provided for the main carrier. The phase switching can be accomplished by a phase shifter (PS). For schemes II and III, the PS is not needed. The separation of the main carrier and signal generation is ideal for DD that provides a strong main carrier and a large linear dynamic range for the signals. The OFDM signal and the main carrier are combined (inset (iii) of Fig. 2), amplified to 6 dBm before launching into an 80-km SSMF fiber. At the receiving end, the signal is re-amplified and filtered with 100-GHz WDM filter, and fed into a photo-detector, the output of which is sampled with a real time oscilloscope at 50 GSa/s. The combined receiver bandwidth of the photodiode and the sampling scope is about 15 GHz. 16 training symbols with only odd subcarriers filled which are free from SSBN are used for channel estimation [11]. The training symbols in practice will be only used at acquisition stage and the subsequent channel estimation can be recovered via the data, and therefore should not be considered in overhead computation for this proof-of-concept demonstration. The OFDM signal processing involves (1) FFT window synchronization using Schmidl format to identify the start of the OFDM symbol, (2) channel estimation in terms of Jones Matrix *H*, (3) phase estimation for each OFDM symbol, and (4) constellation construction for each carrier and BER computation. About 4 millions bits are collected for BER computation.

## 4. Experimental results and discussion

The spectra of the photocurrents *I*_{1} and *I*_{2}, and ${I}_{1}+j\text{\hspace{0.05em}}{I}_{2}$(with 10 dB offset) after 80-km transmission are shown in Fig. 3 to illustrate the effectiveness of BPS when on. It can be seen for either *I*_{1} or *I*_{2}, severe fading took places at some frequencies, making conventional SSB signal extremely sensitive to dispersion. However, after combining *I*_{1} and *I*_{2,} the complex current $\tilde{I}$ shows fading free spectrum, signifying that the BPS format is immune to dispersion fading.

Figure 4 shows the Q^{2} factor performance versus OSNR after 80-km transmission for 4-QAM and 8-QAM with and without SSBN cancellation. The Q^{2} factors are calculated from BER. The CSR used is 11 dB, which is the optimal ratio for 80-km transmission. It can be seen that both of the modulation formats can achieve error-free performance with 20% FEC threshold for a raw data rate of 32.9 Gb/s for 4-QAM and 49.4 Gb/s 8-QAM. We also test the effectiveness of the iterative cancellation at 80-km for 1-band (not shown here) and 3-band OFDM signal. We find that the method is effective for narrow 1 band of 10.97 GHz that improves the Q by 3 dB, but is moderately successful for 33 GHz bandwidth. We deduce that for such a large bandwidth, the phase response of the photodetector can no longer be ignored as in this work. It is also noted the required OSNR for such 40 Gb/s system is relatively high. We attribute the system penalty to the relatively high CSR we used, as well as the residual SSBN as we mentioned above. Although the net effective data rate of 40 Gb/s at 8-QAM is modest for the first proof-of-concept experiment, it is mainly due to the limitation of the equipments used such as AWG that has 3 dB bandwidth of 4 GHz and the receiver that has 15 GHz bandwidth. With a better access to high performance DAC and ADC at 20 GHz bandwidth, we anticipate the fast progress of demonstration over 100 Gb/s using BPS method. Nevertheless, to the best of our knowledge, this is the first demonstration of 40-Gb/s signal over 80-km SSMF with single detector and single polarization without optical dispersion compensation.

## 5. Conclusion

In this paper, we have demonstrated the first 40-Gb/s transmission over 80-km SSMF with single polarization and single detector. Since the proposed BPS is an inherent fading free algorithm, the transmission distance and signal bandwidth will not be limited by chromatic dispersion. The repetition of the same data over the two blocks in BPS reduces effective SE by half, but using the double sidebands recovers the overall SE. As such, BPS approach has twice of the electrical SE of offset OFDM, and has the same electrical SE as single-side band modulated OFDM. Therefore, BPS could be an effective means to achieve DD-based 100 Gb/s Ethernet with high electrical SE for medium reaches.

## References and links

**1. **S. J. Savory, G. Gavioli, R. I. Killey, and P. Bayvel, “Electronic compensation of chromatic dispersion using a digital coherent receiver,” Opt. Express **15**(5), 2120–2126 (2007). [CrossRef] [PubMed]

**2. **C. R. S. Fludger, T. Duthel, D. V. den Borne, C. Schulien, E. D. Schmidt, T. Wuth, J. Geyer, E. D. Man, G.-D. Khoe, and H. D. Waardt, “Coherent equalization and POLMUX-RZ-DQPSK for robust 100-GE transmission,” J. Lightwave Technol. **26**(1), 64–72 (2008).

**3. **W. Shieh and C. Athaudage, “Coherent optical orthogonal frequency division multiplexing,” Electron. Lett. **42**(10), 587–589 (2006). [CrossRef]

**4. **X. Liu, F. Buchali, and R. W. Tkach, “Improving the nonlinear tolerance of polarization-division-multiplexed CO-OFDM in long-haul fiber transmission,” J. Lightwave Technol. **27**(16), 3632–3640 (2009). [CrossRef]

**5. **S. C. J. Lee, F. Breyer, S. Randel, M. Schuster, J. Zeng, F. Huiskens, H. P. A. van den Boom, A. M. J. Koonen, and N. Hanik, “24-Gb/s transmission over 730 m of multimode fiber by direct modulation of 850-nm VCSEL using discrete multi-tone modulation,” in Opt. Fiber Commun. Conf. (OFC), Anaheim, CA, 2007, PDP6.

**6. **J. L. Wei, D. G. Cunningham, R. V. Penty, and I. H. White, “Feasibility of 100G ethernet enabled by carrierless amplitude/phase modulation and optical OFDM.” European Conference and Exhibition on Optical Communication (ECOC), 2012, P6.05. [CrossRef]

**7. **D. Qian, N. Cvijetic, J. Hu, and T. Wang, “108 Gb/s OFDMA-PON with polarization multiplexing and direct detection,” J. Lightwave Technol. **28**(4), 484–493 (2010). [CrossRef]

**8. **B. Schmidt, A. J. Lowery, and J. Armstrong, “Experimental demonstrations of electronic dispersion compensation for long-haul transmission using direct-detection optical OFDM,” J. Lightwave Technol. **26**(1), 196–203 (2008). [CrossRef]

**9. **M. Schuster, S. Randel, C. A. Bunge, S. C. J. Lee, F. Breyer, B. Spinnler, and K. Petermann, “Spectrally efficient compatible single-sideband modulation for OFDM transmission with direct detection,” IEEE Photon. Technol. Lett. **20**(9), 670–672 (2008). [CrossRef]

**10. **B. Schmidt, Z. Zan, L. B. Du, and A. J. Lowery, “120 Gbit/s Over 500-km using single-band polarization-multiplexed self-coherent optical OFDM,” J. Lightwave Technol. **28**(4), 328–335 (2010). [CrossRef]

**11. **W. R. Peng, X. X. Wu, K. M. Feng, V. R. Arbab, B. Shamee, J. Y. Yang, L. C. Christen, A. E. Willner, and S. E. Chi, “Spectrally efficient direct-detected OFDM transmission employing an iterative estimation and cancellation technique,” Opt. Express **17**(11), 9099–9111 (2009). [CrossRef] [PubMed]

**12. **N. Cvijetic, M. Cvijetic, M. Huang, E. IP, Y. Huang, and T. Wang, “Terabit optical access networks based on WDM-OFDMA-PON,” J. Lightwave Technol. **30**, 493–503 (2012). [CrossRef]

**13. **W. R. Peng, B. Zhang, K. M. Feng, X. X. Wu, A. E. Willner, and S. Chi, “Spectrally efficient direct-detected OFDM transmission incorporating a tunable frequency gap and an iterative detection techniques,” J. Lightwave Technol. **27**(24), 5723–5735 (2009). [CrossRef]

**14. **Q. Yang, Y. Tang, Y. R. Ma, and W. Shieh, “Experimental demonstration and numerical simulation of 107-Gb/s high spectral efficiency coherent optical OFDM,” J. Lightwave Technol. **27**(3), 168–176 (2009). [CrossRef]