Hydrogen sulfide breakthrough in producing wells occurred after the breakthrough of CO2 in the Long Coulee Glauconite F reservoir in southern Alberta, where acid gas (98% CO2, 2% H2S) has been injected since 2002. It was hypothesized that the preferential solubility of H2S in formation brine is responsible for the delay in H2S breakthrough. To study the chromatographic separation of H2S and CO2, a series of experiments were conducted to measure the solubility of CO2 and H2S in formation brine at in-situ conditions. Immiscible displacement experiments were performed in a slim tube packed with silica sand to study the breakthrough behaviour of different gas components. The experiments were then modelled using a compositional simulator, and the effect of different factors on the delayed breakthrough of H2S was examined using a series of sensitivity studies. It was confirmed that the preferential solubility of H2S over CO2 leads to it being stripped off at the leading edge of the gas displacement front, resulting in its delayed breakthrough. A similar delay in H2S breakthrough occurs even at higher H2S concentrations (e.g. 30%) in the injected gas. Through the simulation studies it was shown that the delay in H2S breakthrough becomes more pronounced if the gas front is more diffusive. For example, it was shown that, when gravity forces or mobility ratio favour stable displacement, CO2 and H2S breakthroughs occur closer to each other. This is of significance, particularly for monitoring of impure CO2 storage in deep saline aquifers, where the impurity may consist of H2S. Detection of CO2 at a monitoring well would indicate that the more noxious H2S is likely to show up after some time lag. This paper describes the experiments and the simulation studies and presents the implications of the chromatographic partitioning of H2S and CO2 for geological storage of acid gas or impure CO2