Horacio Soto

Here, a technique for implementing an all-optical wavelength converter based on the quantum-confined Stark effect (QCSE) influenced by piezoelectric fields in a semiconductor optical amplifier with compressively strained zinc-blende multi-quantum well grown along the [111]A direction and embedded in the intrinsic layer of a p-i-n diode is presented. The originality and crucial aspect of the technique used is that the piezoelectric fields, induced by the compressive strain within the quantum wells (QWs) of the amplifier with an orientation parallel to that of the built-in field of the structure, make it possible to accelerate the absorption recovery and to perform a fast wavelength conversion over a wide range of the continuum. Specifically, the built-in p-i-n electric field and the piezoelectric fields induce a QCSE and unevenly tilt the potential energy profile of the QWs. This dramatically reduces the energy between the effective height of barriers and the quantized energy of carriers, remarkably due to the piezoelectric fields, thereby decreasing the escape time of carriers from wells and accelerating absorption recovery. Consequently, a strong negative chirp is induced into the converted signal pulses, allowing their compression after passing through a blue-shifted optical filter. Up- and down-conversions at 150 and 1300 Gb/s, respectively, were theoretically predicted in an ideal case, and experimentally, both were error-free demonstrated at 40 Gb/s in a total range of 29 nm, employing a straightforward scheme, with the possibility of operating at 100 Gb/s.

Here, we present a method, based on Stark shifts induced by the quantum-confined Stark effect, to directly determine the piezoelectric field in strained zinc-blende quantum wells (QWs) grown along the 111 direction and embedded in the p–i–n diode structure (p-iMQW-n) of a semiconductor optical amplifier. Under short-circuit and open-circuit conditions and resonant optical excitation, we experimentally determine the energy of the 1s e-hh excitonic resonance and the potential difference across the p-iMQW-n junction. Using these parameters in an analytical expression derived in this work, we directly determined a piezoelectric field of −108.71 ± 8.51 kV/cm acting on each of the eight 12-Å-wide strained [111]-oriented In_0.687Ga_0.313As /In_0.807Ga_0.193As_0.304P_0.696 QWs of a semiconductor optical amplifier (SOA). The theoretical prediction of −112.24 kV/cm was in very good agreement, within the experimental error, with the value of the piezoelectric field extracted through the proposed procedure, which was also indirectly validated by comparing the experimental value of the 1s e-hh excitonic resonance Stark shift with the extracted one. Even though the implemented technique was specially designed to experimentally determine the piezoelectric field acting on each of the QWs of a multiple quantum well SOA, it can be applied to any p–i–n structure with electrodes and strained [111]-oriented zinc-blende QWs embedded in its intrinsic region. Remarkably, the method proposed here allows direct experimental determination of the piezoelectric field through an analytical expression, with which it is also possible to estimate, with high reliability, the precision of the result and how it is affected by the accuracy of each measuring instrument intervening in the procedure.

This communication addresses, theoretically and experimentally, the effect of the cross polarization modulation (XPolM) phenomenon on the performance of a wavelength converter based on a four wave mixing process (FWM) developed in a semiconductor optical amplifier (SOA). The main studied parameter for evaluating the performance of the wavelength converter is the conversion efficiency, whose degradation, provoked by the XPolM, is analyzed when the four wave mixing process is developed with input signals linearly co-polarized at 0, 45, 90, and -45 degrees. The analysis is based on the theoretical and experimental determinations of the polarization mismatch of the pump and probe signals. Finally, it is demonstrated that classic models lacking the inclusion of the XPolM effect can reflect a significant error in the estimation of the conversion efficiency of a wavelength converter based on FWM.