Sputter ion sources are generally used in Accelerator Mass Spectrometry (AMS) because they produce high currents and very little memory effect. However, gas sources are promising for AMS because of their potential for isobar attenuation. In this study a gas ion source was designed based on electron transfer from a primary negative ion beam (20 keV energy) to an iodine monochloride (ICl) gas target.  ICl is promising for ³⁶Cl work, because its sulphur analogue, disulphur diiodide (S₂I₂) is thermodynamically unstable at room temperature. In a preliminary phase of this work, the total electron loss cross-sections of various ions in ICl were investigated.  These cross-sections for Cu^- and Sn^- on ICl at 20 keV were determined to be 8 x 10^-15 cm² and 1.2 x 10^-14 cm² respectively, consistent with values expected for resonant electron transfer, allowing for the Franck-Condon principle. Atoms with higher and lower electron affinities had lower cross section as expected. During the design phase, longitudinal ion extraction geometry was selected for the ion source. Although this simplified the design, it also led to a significant energy spread in the extracted ions.  Test runs of the ion source indicated that relatively high ion extraction efficiencies could be achieved. With a primary beam of Cu^-, and a quadratic axial potential for extraction, the secondary beam of Cl^- was about 20% as large as the Cu^- beam.  Mass spectra showed that I^-, Cl^-, ICl^-, and I₂^- were all extracted from the cell.  The Cu^- beam produced the most favourable ratio (2.6) of Cl^-/ICl^- whereas the value for a S^- beam was lowest at 0.76. Overall this study has validated the basic concept of a resonant charge transfer gas ion source.  Future studies will concentrate on improvements to the ion extraction geometry using, for example, a radio-frequency mulipole with a small longitudinal electric field gradient.