Ridge microstructures were prepared by etching through samples consisting of a series of stacked InAs x P 1-x quantum wells (QWs) with step graded composition grown on InP by molecular beam epitaxy. Different etching techniques were used: wet etching with HCl/H 2 O and reactive ion etching (RIE) with CH 4 /H 2 . These microstructures were characterized by low temperature micro-photoluminescence. The photoluminescence (PL) emission associated with each QW was clearly identified. The PL was measured in detail across etched ridge stripes of various widths. Variations of the integrated PL intensities across the etched stripes were observed. The PL intensities for all QWs increase gradually from the edge to the center of the ridge microstructures. The PL intensity measured at the ridge center is systematically reduced for ridges which are 10 or 20 µm wide as compared to ridges which are 30 µm wide or larger. On the other hand, the spectral peak position of the PL lines remained constant with high accuracy (0.2 to 0.4 meV) across the microstructures. These observations are discussed in terms of the different possible mechanisms which determine the PL intensity variations, namely non-radiative recombination at the etched walls and effects of stray electric fields which result from the etching process. Based on this discussion, we compare quantitatively the different etching processes which we have used. Altogether, this study illustrates the contribution that specially designed test structures, coupled with advanced spectroscopic characterization, can provide to the development of semiconductor photonic devices (e.g. lasers or waveguides) involving etching processes.