A note on the relation between the hypothesis of local equilibrium and the Clausius-Duhem inequality (Q2640401)

The paper is devoted to different interpretations of Clausius-Duhem inequality (DCI) with regard to the domain on which constitutive equations are defined: The ``physical space'' \((x\in {\mathbb{R}}^ 3)\) is rigorously distinguished from the ``state space'' which is spanned by all equilibrium states of the system. By this distinction (a fiber bundle \({\mathbb{R}}^ 3\) consisting of the nonequilibrium states of the system is not introduced) the possibility of two interpretations of the CDI arises: On the ``physical space'' we get the one containing nonequilibrium contact quantities [\textit{W. Muschik}, ibid. 8, No.3, 219-228 (1983)], on the ``state space'' we get the usual CDI of irreversible thermodynamics including a nonnegative entropy production, an unexpected fact because the ``state space'' was defined only to include equilibrium states. The solution is that the ``state space'' includes the internal variables as independent variables, and therefore it can not be the equilibrium subspace [\textit{W. Muschik}, ibid. 15, No.2, 127-137 (1990)]. Consequently the accompanying processes introduced by a projection of the ``physical space'' onto the ``state space'' are not reversible, and the existence of the assumed equation \(dS=dQ^ 0/T\) on the ``state space'' is doubtful. These difficulties are avoidable by introducing a comprehensive nonequilibrium state space and an embedding axiom [\textit{J. Bataille} and the author, ibid. 4, No.4, 228-229 (1979)]. Finally a distance between equilibrium and nonequilibrium states is introduced by the Deborah number.