The growth of barite {O 0 I } and {2 I O} faces from aqueous solutions moderately supersaturated with respect to barite ({Jbarite ;::: :; 12 for experiments on {O 01} surfaces and {Jbarile ;::: :; 7 for experiments on {21 O} surfaces) and bearing different concentrations of carbonate has been studied in situ using an atomic force microscope (AFM). Nanoscopic observations show that, above a certain carbonate concen tration threshold in the aqueous solution, the advancement of mono layers (�3.5A in height) on barite {0 01} and (210} surfaces is strongly inhibited. However, inhibition never affects the growth of the first monolayer, whose growth rate increases in the presence of carbonate. In contrast, the second monolayer growth rate decreases as the concentration of carbonate in the solution increases. For high carbonate concentrations in the solution, growth stops after the formation of the first monolayer. While on barite {O 0 I } faces, the for mation of a second monolayer does not occur for carbonate concentrations higher than 0. 2 mM, on barite (21 O} faces the complete inhibition of the second monolayer is observed for carbonate concentrations higher than 0.05 mM. Once growth on {O 0 I} or {21 O} faces is completely inhibited, i. e. such surfaces are in the "dead zone", growth can be recovered by increasing supersaturation. In order to study the recovery behaviour of barite {O 0 I } and (2 I O} faces from the "dead zone", an additional series of AFM experiments have been conducted. In these experiments, carbonate-free aqueous solutions with increasing supersaturations with respect to barite were passed over (00 I} and (21 O} surfaces previously "poisoned" with carbonate. Our experimental results show that the recovery of growth on barite {O O I} faces requires an important increase of the solution supersaturation. In contrast, the recovery of barite {21 O} surface growth does not require any supersaturation increase, but spontaneously occurs in a few minutes. Our observations of inhibition and growth recovery on barite surfaces at a nano-scale are discussed and compared with the descriptions given by the classical crystal growth inhibition models.