Abstract
Quantitative modeling shows that artificial acetate buffering of test solutions should be implemented only below pH 4.6. In addition, this would allow a major simplification of the initial pH adjustment, which could then be made before saturation in the test gas, and subsequently an extension to autoclave testing. Above pH 4.6, not only artificial buffering is useless but it may even become detrimental when the constraining pH adjustment protocol of EFC 16 is not strictly followed (e.g. with some traditional test cells or in autoclave testing). Acetic acid addition or reduced buffer content can also introduce detrimental artefacts and should be avoided. These unforeseen pH artefacts may explain the unexpected behaviors often observed when testing martensitic stainless steels. Modifications of EFC 16, EFC 17 and ISO 15156 are therefore suggested.
Also the true value of the in-situ pH in corrosive production wells in the presence of CO2 and free acetic acid is partially unknown. This undoubtedly contributes to the present failures of the prediction tools which rely only on laboratory corrosion rate measurements, whatever the quality of their respective data base. Conversely, the efficiency of indirect prediction tools based on a field data base (such as CORMED) is not altered.
