A system made of an atmosphere, a liquid metal and an oxide phase at a given temperature and total pressure is zero-variant and allows only a particular set of partial pressures, compatible with thermodynamic equilibrium. For any different gas composition the system will show a tendency to evolution. Therefore, classical thermodynamics cannot give the answers to a number of problems of major interest, such as those concerning the conditions for interface saturation and oxide formation. Strictly speaking these are kinetic problems, but they can still be treated as pseudo-thermodynamic questions. The key to the problem is in considering the characteristic times of evolution, which suggest stationary state approximation for the condensed phases. In many instances, the evaluation of gas-atmosphere mass exchanges under stationary conditions makes it possible to determine the effective oxygen pressure at which the oxidation of the metal becomes evident. Surface tension measurements allow this condition to be detected with a good accuracy. According to experimental evidence, the effective oxygen pressure can be many orders of magnitude greater than the equilibrium value. The problem needs different theoretical approaches according to the molecular mechanisms involved. Moreover, from the experimental point of view, there are particularly delicate questions regarding the accuracy and the significance of the oxygen control and measurement. (C) 2005 Springer Science + Business Media, Inc.