Skip to Main Content
This paper gives an overview of recent demonstrations of optical 2R regeneration achieved by vertical microcavity mirror-based multiple-quantum-well saturable absorber (SA). The potential of the device to perform wavelength division multiplexing regeneration is first demonstrated through the first pigtailed SA chip implemented with eight independent fibers using a cost effective coupling technique. The cascadability and wavelength tunability assessment of this module associated with a power limiter fiber-based function has been experimentally demonstrated at 42.6 Gbit/s. Because this method of power limiting is not a suitable solution for all-optical multichannel 2R regeneration, a new SA structure allowing a power limiting function was proposed. We describe and characterize such a structure in this paper. This new SA opens the door to a complete passive all-optical 2R regeneration relying upon a single technology, as shown in this paper through the use of two SA: SA.0 for extinction ratio enhancement and SA.1 for high-power level equalization allowing receiver sensitivity (up to 3.5 dB) and Q-factor (up to 1.4 dB) improvement for a RZ signal at 42.6 Gbit/s. The limitation of SA.1 when the regenerator must be cascaded a large number of times is also described, leading to the observation that SA.1 should be more suitable for phase encoded formats that are more spectrally efficient than ON-OFF keying formats. A SA.1 used as a phase-preserving amplitude regenerator in a 42.6 Gbit/s RZ differential quadrature phase-shift keying transmission system is, therefore, assessed . A fiber launched power margin of 2 dB and a receiver sensitivity improvement of 5.5 dB are obtained. Finally, we use, for the first time an SA.1 as a phase-preserving amplitude regenerator of RZ differential quadrature phase-shift keying signals. The regenerator is assessed in a recirculating loop at 28 Gbaud. The system tolerance to nonlinear phase noise is enhanced by 3 dB and the distance improvem- nt factor was 1.3 for a bit error rate = 10-4.