Stability implies concavity/convexity of thermodynamic potentials

Thermodynamic stability forces entropy to be concave in extensive variables (equivalently, energy is convex), yielding positivity of response functions.
Stability implies concavity/convexity of thermodynamic potentials

Statement

Consider an equilibrium with a differentiable entropy representation S=S(U,V,N)S=S(U,V,N) for the . Assume standard thermodynamic stability in the entropy-maximum sense: for fixed additive constraints, the equilibrium state maximizes total entropy under allowed exchanges (see ).

Then S(U,V,N)S(U,V,N) is a concave function of the extensive variables: for 0λ10\le \lambda\le 1 and any two admissible states,

S(λx+(1λ)y)  λS(x)+(1λ)S(y),x=(U,V,N), y=(U,V,N). S(\lambda x + (1-\lambda) y)\ \ge\ \lambda S(x) + (1-\lambda) S(y), \qquad x=(U,V,N),\ y=(U',V',N').

Equivalently, in the energy representation U=U(S,V,N)U=U(S,V,N), the internal energy is convex in (S,V,N)(S,V,N).

Key hypotheses

  • Existence of equilibrium states and differentiable fundamental relations for S(U,V,N)S(U,V,N) or U(S,V,N)U(S,V,N).
  • 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

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 ) 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.