Stability implies concavity/convexity of thermodynamic potentials
Statement
Consider an equilibrium thermodynamic system with a differentiable entropy representation for the entropy . Assume standard thermodynamic stability in the entropy-maximum sense: for fixed additive constraints, the equilibrium state maximizes total entropy under allowed exchanges (see thermodynamic stability ).
Then is a concave function of the extensive variables: for and any two admissible states,
Equivalently, in the energy representation , the internal energy is convex in .
Key hypotheses
- Existence of equilibrium states and differentiable fundamental relations for or .
- Additivity/ability to form composite systems and exchange conserved quantities.
- Stability as an entropy maximum (or, dually, minimum of a suitable free energy at fixed intensive parameters).
Conclusions
- Concavity of implies the Hessian of (where it exists) is negative semidefinite; convexity of implies the Hessian of is positive semidefinite.
- Standard response-function positivities follow (under appropriate regularity and choice of independent variables), e.g.
- Heat capacity at constant volume satisfies in stable regions (see also C_V positivity from stability ).
- Isothermal compressibility satisfies (see compressibility positivity ).
- Stability restricts curvature of thermodynamic potentials obtained by Legendre transforms (e.g., convexity/concavity properties of Helmholtz free energy and Gibbs free energy ).
Proof idea / significance
The core idea is the “mixing” argument: take two macrostates, form a composite system that can redistribute the conserved quantities, and use the entropy-maximum principle (a formulation of the second law
) to show that the entropy of the combined equilibrium state dominates the weighted average of entropies. That inequality is precisely concavity.
Concavity/convexity is the mathematical backbone of stability, ensuring uniqueness of equilibrium in single-phase regions and governing fluctuations/response via curvature.