The combination of sample injection into an air carrier stream (i.e., air segmentation) with a low sample consumption system has been evaluated for the analysis of microsamples through ICP-AES. A PFA micronebulizer has been coupled to: (i), a double pass spray chamber; (ii), a Cinnabar cyclonic spray chamber; and (iii), a torch integrated sample introduction system, TISIS. Three matrices have been studied: in addition to water two concentrated acid solutions (2 mol l⁻¹ nitric acid and 1.7 mol l⁻¹ acetic acid) and Na 5,000 µg ml⁻¹. A simulation of the evolution of the drop size distributions of the aerosols with time was carried out in order to evaluate the extent of solvent evaporation inside the chamber. The total mass of solvent evaporated inside the chamber was estimated and it was concluded that, at 25 °C, about 4–6 s residence time were required to promote the maximum evaporation of the solvent. In order to ensure this, discrete sample introduction into an air carrier stream (i.e., air segmentation) was used. Narrow peaks (i.e., with a full width at half maximum, FWHM, as short as 10 s) were obtained for a 10 µl injected sample. The peaks found for the Cinnabar and TISIS were narrower than those for the double pass spray chamber. More importantly, the interferences caused by inorganic as well as organic matrices were less severe in discrete than in continuous mode. The theoretical simulations allowed explanation of these results in terms of the enhancement of the solvent evaporation both for water and matrices in this operating mode. The enhanced solvent evaporation with respect to the situation in continuous mode minimized differences in analyte transport towards the plasma induced by these compounds. Despite this, in discrete mode a residual matrix effect was found that was attributed to the aerosol transport process. Internal standardization (IS) was applied to transient signals and the interferences were compensated for in virtually all the cases. Good results were obtained for the four emission lines taken as internal standards (i.e., Mg 280.270, Co 228.616, Cr 205.552 and Cu 324.754). However, for acetic acid and a few lines, IS was not efficient for removing interferences. The methodology was validated by analyzing two reference solid samples of foods (i.e., bovine liver and mussel tissue). By using Cd 214.438 as internal standard and under discrete mode 100% recoveries were found.