Corrosion occurring in subsurface nuclear waste repositories yields the production of dihydrogen (H2), the possible accumulation of which being a major security concern. The development of accurate and theoretically assessed mathematical and numerical models is therefore a priority to quantify the in situ production of H2 along large time scales. Despite important efforts of the mathematical community during the last 15 years, there is so far no satisfactory mathematical framework for the so-called Diffusion Poisson Coupled Model (DPCM) proposed by Bataillon and collaborators (Bataillon et al., 2012, 2010). This model describes the evolution of the oxide layer covering the metal by taking into account the oxidation of the metal, the transfer within the oxide layer of the charge carriers driven by some self-consistent electric potential, and the dissolution of the oxide at the interface with some aqueous solution. The main reason for the aforementioned gap in the theory is that no thermodynamic potential, serving as a Lyapunov functional, has been shown to be dissipated along time, in accordance to the second principle of thermodynamics. Assuming pressure and temperature to be constant, Gibbs free energy is indeed expected to decay, up to some exchange with the surrounding metal and solution. We propose an update of the DPCM model which fulfils some variant of Onsager's reciprocal relation, ensuring therefore the compatibility of the model with the second principle of thermodynamics. The main differences with the original DPCM model are the following: (i) the transport of ferric cations and oxygen vacancies in the oxide are driven by a vacancy diffusion process with a nonlinear mobility including a saturation effect; (ii) the motion of the interfaces is driven by the difference of grand potential (or Landau free energy) density rather than by a difference of chemical potential as in the original DPCM; (iii) a correction in the charge carrier fluxes is incorporated to take the volume expansion stemming from the iron oxidation into account; (iv) an update of the boundary condition for the electrons between the oxide and the metal.