In this paper we investigate the electronic structure of Zn 1-x Co x O by means of optical absorption measurements under pressure. Thin films of Zn 1-x Co x O with different Co content (from 1 to 30%) were prepared by pulsed laser deposition on mica substrates. Absorption spectra exhibit three main features that are clearly correlated to the Co content in the films: (i) absorption peaks in the infrared associated to crystal-field-split internal transitions in the Co 3d shell, with very small pressure coefficients due to their atomic character; (ii) a broad absorption band below the fundamental edge associated to charge transfer transitions, that exhibit relatively large pressure coefficients, indicating that the Co 3d final states must be strongly hybridized to the conduction band; and (iii) a blue-shifted fundamental absorption edge associated to band to band transitions with a pressure coefficient close to that of pure ZnO. In the up-stroke the transition pressure from wurtzite to rock-salt phase decrease almost linearly as the Co increases, from 9.5 GPa in pure ZnO to about 6.5 GPa for x = 30%. In the down-stroke pressure we observe a similar behavior, yielding a metastable rock-salt phase at room pressure, after a pressure cycle up to 15 GPa.

We report the direct proof of parametric amplification in a semiconductor waveguide. Thanks to low propagation losses and high nonlinear efficiency, we amplified an input signal at 2 μm in a selectively oxidized GaAs/AlAs multilayer waveguide through the down conversion of a pump at 1 μm. The 4.5% single-pass gain measured for a 30 mW cw pump is compatible with parametric oscillation, provided that distributed and concentrated waveguide losses are further reduced.

We report a quantitative study on the absorption of oxidized-AlAs (AlOx) integrated thin layers in the mid-infrared range (2–4μm). An AlOx absorption coefficient of 1600cm−1 is measured in a nonlinear waveguide, by investigating the phase-matching bandwidth of parametric fluorescence. Since mid-infrared optical losses of the idler mode are inferred from the corresponding signal features in the 1–2μm range, our technique takes advantage of common near-infrared sources and detectors.