Keywords: SIMS; matrix effects; ion yield; quantification; silicon; oxygen Since the beginning of the development of SIMS techniques, oxygen is well known to enhance the ion yield of electropositive species.[1] This 'oxygen matrix effect' is widely used, either as primary beam ions or using an oxygen leak ('oxygen flooding') near the surface of the sample, or even by combining both; oxygen modifies the near surface concentration by means of a partial or complete oxidation leading to the increase of the ionization probability, useful yield and ion yield of matrix and trace elements. The case of silicon has been extensively studied in order to measure experimentally the variations of the ion yields with oxygen surface concentration, [2,3] and to propose a theoretical model for the mechanisms of atomic secondary ions formation. [4] Experimentally, two main problems are to be solved: the accurate measurement of the oxygen atomic concentration at the surface whatever the oxygen level used, and the reliable measurements of the (useful) ion yields enhanced by the oxygen induced matrix effects.In some remarkable previous studies under Ar + beam, oxygen atoms are introduced either by flooding oxygen near the receding surface (external source of oxygen) [2,5] or by implantation of one isotope of oxygen atoms [6] (internal source of oxygen). Williams' ' 18 O method' [2] consists in implanting an 18 O marker (for quantitative oxygen measurements of the following flooded oxygen) and then in introducing 16 O (matrix effects 'enhancer') during the analysis with a constant superficial concentration, determined by the oxygen pressure. This method has been discussed and commented in e.g. [7,8,9] Wittmack's ' 16 O method' does not use any preliminary isotopic marker but in situ implantation of 16 O by the primary column of the SIMS instrument itself. The implantation profile, or at least the concentration of the maximum of the implantation peak, must be known for quantitative oxygen concentration measurements. According to how the implantation has been performed, this method is limited to oxygen concentrations lower than 35 at.%. [6] In this work, we propose a new 18 O Isotopic Comparative Method ' 18 O-ICM' to investigate the oxygen concentration dependence of secondary ion yield enhancement in the 0-10 at.% range under Ar + primary beam. This method has been previously developed to measure the variations of the ion yields of highly concentrated boron in silicon.[10] For this, a silicon sample is multiimplanted with 18 O (ICM isotopic reference) to obtain a near-flat profile with a sufficiently low concentration of oxygen to remain in dilute regime. A high dose of 16 O implantation is subsequently (or previously) performed, so that the whole profile of 16 O is contained in the near-flat 18 O region. Moreover, silicon is uniformly boron doped for investigating the dopant ion yield. The ICM allows to deduce from the knowledge of the 18 O concentration the real profile of 16 O and total oxygen concentration at the surface,