4.6 Metastability and two-state systems

The nucleation barrier of the previous lesson is one face of a general phenomenon: a free-energy landscape with two wells. A system can be trapped in the higher well — metastable — and the rate at which it escapes, or the equilibrium split between the two wells, is governed by the barrier and the temperature. This lesson treats both limits: the slow escape over a large barrier, and the fast equilibration of a two-state system.

Metastability and the limit of stability

A phase that is not the global free-energy minimum but is locally stable — a positive curvature, no downhill direction nearby — is metastable. Supercooled water below $0^\circ$ C and superheated water above $100^\circ$ C are metastable: the more stable phase is available, but the barrier to nucleating it (the previous lesson) blocks the way. A metastable state persists for a time set by the Boltzmann escape rate $e^{-\Delta G^*/k_B T}$ , which can be microseconds or longer than the age of the universe depending on the barrier.

Metastability has a limit. As the driving force grows, the barrier shrinks, and at the spinodal it vanishes entirely: the curvature of the trapping well reaches zero and the state becomes unstable to infinitesimal perturbations. Beyond the spinodal there is no waiting for a rare fluctuation — the transformation is immediate and spontaneous, the same loss of stability that drives spontaneous phase separation in 4.2.

Two-state systems

When the barrier is low — comparable to $k_B T$ — the system does not stay trapped; it hops freely between the wells, and what matters is the equilibrium population of each. Take two states with free-energy difference $\Delta G$ that depends linearly on an external bias $F$ ,

\Delta G(F) \;=\; \Delta G_0 - \alpha F,

where $\alpha$ measures how strongly the bias couples to the switch. Boltzmann-weighting the two states gives the probability of finding the system in the upper one,

P(F) \;=\; \frac{1}{1 + e^{\Delta G(F)/k_B T}} \;=\; \frac{1}{1 + e^{(\Delta G_0 - \alpha F)/k_B T}}.

This is a sigmoid in the bias $F$ — the Fermi function — centred where $\Delta G = 0$ , at $F = \Delta G_0/\alpha$ , with a transition width $k_B T/\alpha$ .

_openF_1/2

F (bias)0.00

P_open(F)26.9%

F_1/2 = ΔG₀/α0.50

width k_BT/α0.50

Bias F = 0.00 Zero-bias offset ΔG₀ = 0.50 Coupling α = 1.00 Temperature k_BT = 0.50

The Fermi-function sigmoid is the equilibrium occupancy of a two-state system with energy gap ΔG(F) = ΔG₀ − αF. The midpoint sits at F_1/2 = ΔG₀/α, and the width (in F) is set by k_BT/α. Cool the bath (lower T) and the transition sharpens toward a step. This is the operative model for ion-channel gating, MET-channel mechanotransduction, and any other two-level system in thermal contact with a bath.

Slide the bias $F$ : the system switches from one state to the other as $F$ crosses the midpoint. Lower the temperature and the transition sharpens — in the $T\to 0$ limit it becomes a step, since thermal fluctuations no longer blur the crossover. Raise the coupling $\alpha$ and the transition steepens at fixed $T$ . As always, it is the dimensionless ratio $(\Delta G_0 - \alpha F)/k_B T$ that decides the split, not any of the energies alone. The same two-state law describes a molecular conformational switch, a magnetic moment flipping in a field, and any thermally activated system biased between two configurations.

⏳ The history — Helmholtz, Gibbs, and the invention of free energy

Hermann von Helmholtz introduced freie Energie in 1882, showing that $U - TS$ is the maximum work extractable from a system in contact with a heat bath at fixed temperature. Josiah Willard Gibbs, in his 1873–1878 monograph On the Equilibrium of Heterogeneous Substances, independently built the same machinery for the constant- $T,p$ case, introducing the potential $U + pV - TS$ that now bears his name.

Gibbs’s monograph — published in three instalments in the obscure Transactions of the Connecticut Academy of Arts and Sciences — laid out essentially the entire modern thermodynamics of phase equilibria: the chemical potential, the phase rule, and the analysis of multiphase systems. It was so far ahead of its time that it went largely unread for two decades, until Wilhelm Ostwald translated it into German in 1892.

The Clausius–Clapeyron relation predates both: Émile Clapeyron wrote it in 1834, and Clausius gave it a clean derivation in 1850 — the historical bridge between the empirical observation of latent heat and the modern thermodynamic potentials.

Read the original: On the Equilibrium of Heterogeneous Substances (J. Willard Gibbs, 1876)