The influence of oxygen on the cerium-catalyzed Belousov-Zhabotinsky reaction is investigated in a system that is closed except for transport of oxygen from the atmosphere across a free surface. Experiments show that the introduction of oxygen increases the complexity of oscillations and shortens their duration. We determine the dependence of the oscillatory pattern on the concentration of oxygen in the reaction mixture, as measured with an oxygen electrode and controlled through the content of oxygen above the free surface. We show that the influence of the stirring rate under aerobic conditions is mainly an oxygen effect and report oscillations of the oxygen concentration. Calculations based on a modified Oregonator model show that the influence of oxygen can be qualitatively accounted for through the effect of two rate constants.

We describe a model of the Briggs−Rauscher reaction closely related to those due to Furrow and Noyes and to DeKepper and Epstein. The model has been fitted to 16 experimental basic quenching parameters at a Hopf bifurcation point. The fit is quite good. Predictions of the model are compared with experiments close to the Hopf bifurcation as well as at several operating points away from it and with experiments under batch conditions. The angular frequency of the finite limit cycle oscillations near the Hopf bifurcation is half the value observed in the experiments. The relative phases of the concentration oscillations of O2 and I- agree reasonably well with the experiments, but the average values are 1 or 2 orders of magnitude smaller in the simulations. The model demonstrates that it is possible to account for a paradoxical behavior of perturbation by OH- as compared to that of H+ and that quenching from a finite limit cycle may be impossible to realize. It qualitatively explains four distinct quenching experiments by dilution. At three separate operation points away from the Hopf bifurcation the model reproduces two observed patterns, namely, one of bistability between a stationary state and a limit cycle and one of mixed mode oscillations, but it fails to reproduce observed burst oscillations at the appropriate point. The model can reproduce transient batch mode oscillations and other transient behavior. The model of Furrow and Noyes with a more recent set of rate constants and a reduced version of it, derived by sensitivity analysis, are also compared with experiments.