Search

Silicon photonics can be used to increase the versatility of wavelength division multiplexing (WDM). Ultra-dense wavelength division multiplexing (uDWDM) shrinks channel spacing between WDM channels to decrease guard bands and increase spectral efficiency. Microring modulators (MRMs) provide a simple solution for uDWDM transmitters - they require no multiplexer/demultiplexer. When used with comb lasers, MRMs provide extremely compact and low power solutions. The spectral efficiency is limited by interchannel interference (ICI) from incomplete isolation of channels. We report an MRM-uDWDM transmitter achieving channel spacing comparable to the symbol rate in each WDM channel, reaching a spectrum occupation of 80%. By choosing the correct operating point, we demonstrate experimentally our fabricated MRM chip design causes no discernable ICI power penalty. To the best of our knowledge, this is the first demonstration of MRM-uDWDM to support such dense channel spacing. [ABSTRACT FROM AUTHOR]

A frequency conversion mixer or single side band modulator using two cascaded MZM is proven experimentally. The operation of the circuit is modelled by a transfer matrix approach and verified by simulation in support of the experiment. A 10 GHz shift of the optical carrier in both left and right direction is demonstrated. The residual sideband suppression relative to the enhanced sideband is 22 dB for the best cases. Numerical analysis shows that the circuit has 3-dB optical and 3-dB electrical intrinsic advantage over the functionally equivalent DP-MZM. [ABSTRACT FROM AUTHOR]

Silicon photonics has enormous potential for ultrahigh-capacity coherent optical transceivers. We demonstrate an in-phase and quadrature (IQ) modulator using silicon photonic traveling-wave modulators optimized for higher order quadrature amplitude modulation (QAM). Its optical and RF characteristics are studied thoroughly in simulation and experiment. We propose a system-orientated approach to optimization of the silicon photonic IQ modulator, which minimizes modulator-induced power penalty in a QAM transmission link. We examine the tradeoff between modulation efficiency and bandwidth for the optimal combination of modulator length and bias voltage to maximize the clear distance between adjacent constellation points. This optimum depends on baud rate and modulation format, as well as achievable driving voltage swing. Measured results confirm our prediction using the proposed methodology. Without precompensating bandwidth limitation of the modulator, net data rates up to 232 Gb/s (70 Gbaud 16-QAM) on single polarization are captured, indicating great potential for 400+ Gb/s dual-polarization transmission. [ABSTRACT FROM AUTHOR]

Silicon photonic modulators have strong nonlinear behavior in phase modulation and frequency response, which needs to be carefully addressed when they are used in high-capacity transmission systems. We demonstrate a comprehensive model for depletion-mode Mach–Zehnder modulators (MZMs) on silicon-on-insulator, which provides a bridge between device design and system performance optimization. Our methodology involves physical models of p–n–junction phase-shifters and traveling-wave electrodes, as well as circuit models for the dynamic microwave-light interactions and time-domain analysis. Critical aspects in the transmission line design for high-frequency operation are numerically studied for a case of p–n–junction loaded coplanar-strip electrode. The dynamic interaction between light and microwave is simulated using a distributed circuit model solved by the finite-difference time-domain method, allowing for accurate prediction of both small-signal and large-signal responses. The validity of the model is confirmed by the comparison with experimental results for a series push–pull MZM with a 6 mm phase shifter. The simulation shows excellent agreement with experiment for high-speed operation up to 46 Gb/s. We show that this time-domain model can well predict the impact of the nonlinear behavior on the large-signal response, in contrast to the poor prediction from linear models in the frequency domain. [ABSTRACT FROM PUBLISHER]

We present an experimental demonstration of multilevel coded modulation (MLCM) for 16-ary constellations. Performance of MLCM coding is compared for two 16 QAM constellations: 1) a phase noise optimized and 2) a square constellation. We quantify experimentally the phase noise regimes where the optimized constellation outperforms square 16 QAM. While MLCM coding is designed for one linewidth at a given symbol time, for the same overhead we show it outperforms uniform rate coding over a wide range of linewidths. The MLCM coding strategy minimizes block error rate under some simplifying assumptions. We show that under realistic conditions, the MLCM also gives good post forward error correction and quantify the advantage over uniform rate coding experimentally. [ABSTRACT FROM PUBLISHER]

We analyze the bit error rate (BER) performance of optical coherent communication systems employing multi-section monolithic tunable lasers with large excess phase noise. The BER of the coherent system utilizing a second-order decision-directed digital phase-locked loop (DD-PLL) for phase tracking was computed numerically. Experimental results have also been demonstrated with a 16-quadratic-amplitude modulation coherent system deploying a sampled-grating distributed Bragg reflector laser at 16 Gbaud that confirms the performance of the second-order DD-PLL in tracking the excess phase noise. [ABSTRACT FROM PUBLISHER]

Application of multi-level coded modulation (MLCM) for 16-ary constellations in coherent systems is studied. An MLCM system with Reed-Solomon component codes and multi-stage decoder is considered. A systematic numerical method for finding set-partitioning and optimal code rates is presented. The method only requires the probability density function of received samples and can be applied to any constellation regardless of irregularity or lack of symmetry. Performance of the designed MLCM system is verified in presence of nonlinear phase noise and normal phase noise. For nonlinear phase noise limited system, it is shown that the block error rate (BLER) of the system can be improved using our approach for set-partitioning. For phase noise limited system, both BLER and bit error rate (BER) are studied for a phase noise optimized and square 16-quadrature amplitude modulation (16QAM). Post forward-error correction (FEC) BER performance of the optimized constellation over square 16QAM is studied for different levels of phase noise. It is shown that the optimized constellation along with MLCM system decreases the required signal-to-noise ratio by several dB at high phase noise regime and low post FEC BER. [ABSTRACT FROM PUBLISHER]