In this paper we show that the active interplay of nonlinear kinetics and transport phenomena in a chemical oscillator can be exploited to induce and control chaos. To this aim we use as a model system the ferroin-catalysed Belousov-Zhabotinsky (BZ) oscillating reaction, which is known to evolve to characteristic chaotic transient dynamics when carried out under batch and unstirred conditions. In particular, chemical chaos was found to appear and disappear by following a Ruelle-Takens-Newhouse (RTN) scenario. Here we use medium viscosity as a bifurcation parameter to tune the reaction-diffusion-convection (RDC) interplay and force the reaction in a specific sequence of dynamical regimes: either (i) periodic → quasi-periodic → chaotic or (ii) periodic → quasi-periodic or (iii) only periodic. The medium viscosity can be set by adding different amounts of surfactant (sodium dodecyl sulphate), known to have a little impact on the reaction mechanism, above its critical micelle concentration. Experimental results are supported by means of numerical simulations of a RDC model, which combines self-sustained oscillations to the related chemically-induced buoyancy convection.