= Entropy concavity and positive heat capacity
For an equilibrium <entropy> $S(E,V,N)$ that is concave in <energy> at fixed <volume (thermodynamics)> and <particle number>, the thermodynamic relation $(\partial S/\partial E)_{V,N}=1/T$ gives, wherever the derivatives are finite and invertible,
$$
\left(\frac{\partial^2S}{\partial E^2}\right)_{V,N}
=-\frac1{T^2}\left(\frac{\partial T}{\partial E}\right)_{V,N}
=-\frac1{T^2C_V}.
$$
Strict concavity therefore implies positive <heat capacity at constant volume>. A straight coexistence segment can instead have constant <temperature> over an <energy> interval; finite strictly positive heat-capacity conclusions do not apply there.
There is also a direct <canonical ensemble> statement for a temperature-independent <Hamiltonian operator>:
$$
C_V=\frac{\langle H^2\rangle-\langle H\rangle^2}{k_BT^2}\ge0.
$$
The numerator is the <canonical energy variance>. Positivity follows from a <variance>, without assuming an <ideal gas> or noninteracting particles.
The concavity hypothesis is essential. Isolated systems with long-range gravitational interactions can have <negative heat capacity> in a <microcanonical ensemble>; the familiar virialized-gas example does not belong to a stable canonical branch in contact with an infinite <heat bath>. A claim that every stable equilibrium has strictly positive <heat capacity> would erase this ensemble distinction.
Back to article page