The reaction of ozone with aqueous sodium bromide particles is investigated with a combination of aerosol chamber experiments, kinetics modeling, and molecular dynamics simulations. The molecular bromine production in the chamber experiments is approximately an order of magnitude greater than that predicted by known chemistry in the gas and bulk aqueous phases with use of a comprehensive computer kinetics model. Molecular dynamics simulations indicate that ozone has significant residence time at the air-solution interface, while making frequent contacts with bromide ions for as long as 50 ps in the surface layer of a 6.1 M NaBr solution. The formation of a complex between ozone and bromide ion, [(O3Br-)-Br-...], which can lead to production of Br-2 by reaction at the air-water interface, is therefore feasible. Experimentally observed Br-2 is well predicted by including an interface process with a reaction probability of [1.9 +/- 0.8] x 10(-6) (1 s) as the first step in a surface mechanism to produce additional gas-phase Br-2. An estimate of the impact of this interface reaction on bromine formation in the marine boundary layer shows that several ppt of bromine could potentially be produced during the night from this proposed surface chemistry.