The liquid-phase transport phenomena which occur at the surface of iron-base alloys during corrosion have been analysed. These mechanisms determine either the maintenance of bare metal or the precipitation of solid corrosion products, the build-up of a corrosion deposit and the control of its thickness, and finally, the kinetics of the electrochemical reactions under the deposit. Although it is shown that pure ''precipitation-redissolution'' or ''direct formation'' reactions are impossible, the only conceivable mechanisms are nevertheless closely related, because the transport of iron between the metal and the external corrosive medium occurs chiefly either via the solid phase of the deposit (for ''soluble'' deposits), or via the liquid phase permeating its porosities (for ''insoluble'' deposits). It is also shown that, depending on the precipitation conditions, any given solid compound Fe(n)X2 can lead to three types of deposit with quite different properties. (i) ''Soluble'' deposits: moderately protective, steady-state corrosion insensitive to potential, but highly sensitive to turbulence; (ii) ''Insoluble cationic'' deposits (controlled by the removal of Fe2+ cations by liquid-phase diffusion): highly protective, corrosion rate slightly sensitive to potential, and insensitive to turbulence. (iii) ''Insoluble anionic'' deposits (controlled by the diffusional supply of the precipitatable anion X(n-): slightly or unprotective, corrosion slight or insensitive to the presence of the deposit; possibly profuse deposit if steady state corrosion is not attained. This theoretical analysis can retrospectively explain numerous experimental observations reported in the literature, such as the incubation time before the drop in corrosion rates, the multiple forms of CO2 and H2S corrosion, the role of Ca2+ ions, erosion-corrosion and bacterial corrosion. This analysis also paves the way for the reliable laboratory prediction of real corrosion rates under deposits.
