DISCUSIÓN DE RESULTADOS
5.1 Contrastación teórica y práctica de los resultados
5.1.2 Contrastación teórica y práctica de las teorías consideradas en las bases teóricas
2.1 Introduction
As part of the evaluation of the potential applicability of forward-biased semiconductor optical ampMers as ultrafast all-optical switches, it is desirable to obtain information on the nature and magnitude of the nonlinear absorption mechanisms which occur in the amplifiers. Pump-probe experiments utilising ultra-short pump and probe pulses allow these nonlinear absorption mechanisms to be isolated and measured in the time domain. In this chapter, results fi'om pump-probe measurements of absorption-related nonlinearities in amplifiers operating at the transparency point are presented and discussed. Four amplifiers were used for these experiments. The structure of the active regions of each of the amplifiers was different in that both bulk and multiple quantum well devices (MQW) were investigated. One amplifier contained a bulk active region, and three others had four quantum wells, eight quantum wells and sixteen quantum wells respectively. The relative magnitudes of the absorptive nonlinear effects in the amplifiers were measured.
The chapter begins with a description of the pump-probe technique. Previous work on the investigation of gain nonlinearities in optical amplifiers using this method is reviewed, and this is followed by a discussion of the type and size of the nonlinearities that have been observed. The results of measurements performed during the course of this project are presented and the relative magnitudes of the nonfinearities which have been resolved are determined by using a suitable fitting algorithm.
2.2 Bacltgroumd
2,2,1
Pump-Probe Measiaremmts
Pump-probe experiments are a practical and accurate means of providing time- domain measurements of ultrafast dynamics in semiconductors. The principle of a pump-probe experiment is to induce nonlinear effects in a medium using a high intensity pump pulse, and then monitor the transmission characteristics of a low intensity probe pulse in the medium. The time delay between the low intensity probe pulse and the high intensity pump pulse is varied in a predetermined manner, and the corresponding changes in the transmission of the probe pulse are recorded.
In general, a beam of optical pulses is split into a high intensity pump beam and a low intensity probe beam. The probe beam travels a longer optical path than the pump beam before both are incident on a sample. Upon propagation through (or reflection from) a semiconductor, the pump pulse alters the gain and refractive index of the semiconductor, and the transmission of the probe pulse is affected accordingly. The probe pulse is separated from the pump pulse using a polarising beam-splitter or some other method of discrimination and its intensity is measured independently of the pump pulse. The transmission of the probe pulse is determined as a frinction of its delay relative to the pump pulse, thus giving information on both the size and temporal nature of gain nonlinearities in the semiconductor
Measuring the transmission of an amplifier as a function of the delay between pump and probe pulses allows measurements of ultrafast effects to be transformed into measurements of relative displacement. The resolution thus obtained is well in excess of the response time of ultrafast detectors. Typically, the position of the mirror determining the delay between the pump and probe can be determined to the order of one micron. This corresponds to a time resolution of ~3 fs, and the resolution of the pump-probe measurement is therefore limited only by the duration of the detected pump and probe pulses.
Pump-probe investigations of optical amplifiers described in the literature have tended to be influenced directly by the availability of ultrashort pulse sources in particular spectral bands, in parallel with the development of optical amplifiers operating at those wavelengths. Research in the late 1980’s involved dye lasers, used
in the near infra-red to investigate gain dynamics in AlGaAs amplifiers biased close to the transparency point\ It was observed that the propagation of high intensity pump pulses through the AlGaAs amplifiers heated the distribution of carriers in the amplifiers via the creation of hot carriers, or by the removal of cool carriers through stimulated emission. The heated carrier distribution returned to the temperature of the semiconductor lattice through the emission of optical phonons with a time constant of 0.9 ps.
With the development of techniques to reduce the duration of pulses generated by dye lasers, the increased temporal resolution of gain dynamics in AlGaAs optical amplifiers revealed evidence of a second nonlinearity with a recovery time of between
150 and 300
The development of InGaAsP amplifiers with bandgap transitions around 1.5pm, combined with the availability of coupled-cavity mode-locked KC1:T1^(1) lasers, opened up the possibility of pump-probe investigations of gain dynamics in a new material'*’®. Experiments with pump and probe pulses of around 150 fs duration revealed gain dynamics that were more complex than those seen in AlGaAs. Carrier heating was still an important mechanismi in amplifiers operating at gain, but amplifiers biased to transparency or loss also exhibited a positive transient with a recovery time of ~ 200 fs. Mark and Mork^'^, and Hall et a/.® have concluded that this transient is a result of spectral hole burning in combination vrith a delay in the heating of the carrier distribution.