Raman depth-profiling microspectroscopy is currently emerging as a fast and non-invasive method for the local content measurement of water diffusing across the perfluorosulfonic acid polymer used as electrolyte in low-temperature fuel cells. However, water depth profiles obtained thereby are affected by the gradual degradation of the Raman signal as the exciting radiation is focused deeper into the sample and, thus, usually exhibit artificial gradients. The appropriate way to rectify raw data, in order to measure reliable inner water contents, is discussed in the case of perfluorosulfonic acid membranes soaked in water. The method is tested on a tri-layer material composed by stacked ionomers with different chemical composition, ionic exchange capacity and swelling. Comparison of Raman spectra of the different ionomers, obtained under strictly the same hydration and optical conditions, allows critical discussion of previously reported band assignments. It is shown that Raman bands arising from the polymer backbone and from the side chain can be readily discriminated. Reliable water concentration profiles can then be obtained from the ratio between the Raman signal of sorbed water and of the polymer phase. Comparison with depth profiles obtained by using the pure water signal as internal reference shows that quantitative measurement requires the careful choice of the adequate Raman bands for representing the ionomer occupation of the volume probed by the exciting radiation. Different to what usually performed, the ionomer Raman signal to be integrated for compensating optical losses should include both bands arising from the polymer backbone and side chains. Last, obtaining accurate water concentration profiles also needs the knowledge of the so-called optical factor β, i.e. the ratio between the scattering cross sections of ionomer chemical groups and sorbed water, which has to be measured by independent Raman experiments. Copyright © 2012 John Wiley & Sons, Ltd.