The cosmonuclide Cl-36 is relevant for hydrological applications due to its conservative chloride form. Cl-36 measurements are commonly performed by Accelerator Mass Spectrometry (AMS) while stable chlorine concentration measurements are usually performed by ion chromatography (IC). They can also be achieved by Isotope Dilution-Accelerator Mass Spectrometry (ID-AMS) which offers the advantage of the simultaneous determination of both Cl and Cl-36 contents. The range of applicability of this method has been assessed on rock samples but, to our knowledge, no experimental tests were conducted on water samples characterized by low chloride contents. This study aims at investigating the accuracy and precision of the ID-AMS method for such low chlorine water samples for which chloride and Cl-36 determinations remain highly challenging. First, internal and external errors were evaluated on the measured ratios and uncertainties on Cl and Cl-36 concentrations were assessed at +/- 5% and +/- 7% respectively using standard error propagation calculations. Focused on the 0.1 to 10 mg center dot L-1 concentration range, the Cl concentrations of thirty-five gravimetric standards were determined by ID-AMS leading to a 1:1 line (R-2 = 0.99) and a 1-sigma uncertainty on [Cl-] of +/- 5%, consistently with the propagated uncertainty. Finally, the optimal sample volume to reach accurate Cl-36 and Cl measurements with ID-AMS for water sample concentrations within the 0.1 to 10 mg center dot L-1 range has been theoretically estimated. (C) 2015 Elsevier B.V. All rights reserved.