Nucleation and Metastability

Introduction

A liquid can remain supercooled and a solution can remain supersaturated even though another phase has lower Gibbs energy. The reason is that a tiny new-phase cluster creates an interface, which costs free energy. Only after the cluster grows enough does the bulk free-energy benefit outweigh that surface penalty. Classical nucleation theory turns this competition into a critical size and barrier.

Core explanation

Consider a spherical nucleus of a stable new phase inside a metastable parent phase. Let γ > 0 be interfacial free energy per area and Δg v > 0 the magnitude of the bulk Gibbs-energy advantage per volume of converting parent to new phase. A sharp-interface model gives ΔG(r) = 4πr²γ − (4π/3)r³Δg v. The positive area term dominates for very small radius; the negative volume term dominates for sufficiently large radius. The model assumes spherical shape, uniform γ and a size-independent bulk driving force.

Differentiate: dΔG/dr = 8πrγ − 4πr²Δg v. Apart from r = 0, the stationary point is r = 2γ/Δg v. It is a maximum, not a stable minimum. Substituting gives ΔG = 16πγ³/(3Δg v²). A nucleus smaller than r tends to shrink in this simplified free-energy picture; a larger one can grow while lowering ΔG. Thermal fluctuations must create a cluster near the barrier top before sustained growth becomes favourable.

Increasing supersaturation or supercooling generally increases the magnitude of the bulk driving force Δg v. The critical radius falls as 1/Δg v and barrier as 1/Δg v², so nucleation can become dramatically faster. A surface, dust particle or container wall may provide heterogeneous nucleation with a lower interfacial cost than a fully spherical nucleus in the bulk. This explains why a clean sample can remain metastable longer than a seeded one.

Classical nucleation formulas are approximations, especially when r approaches molecular dimensions. Real interfaces may be diffuse, γ may depend on curvature, elastic strain can matter in solids, and growth after nucleation requires transport. Thermodynamic favourability alone predicts the eventual lower-G phase, not the time needed to form it. Conversely, a system inside a spinodal region can separate through growing small fluctuations without the same classical nucleation barrier.

Step-by-step reasoning

Choose a sign convention: let Δg v be positive magnitude of bulk advantage. Write surface area and nucleus volume, then combine +γA and −Δg vV. Differentiate with r, solve for nonzero stationary radius and check it is a maximum. Calculate ΔG , then interpret how γ and Δg v alter both barrier and kinetic delay. Check whether continuum spherical assumptions are plausible at the resulting radius.

Visual explanation

Plot ΔG(r) starting at zero, rising to a maximum at r , then falling below zero at larger r. Decompose it into a positive r² surface curve and a negative r³ bulk curve. Add two sketches: a tiny cluster shrinking and a larger-than-critical cluster growing. A surface-supported cap illustrates heterogeneous nucleation.

Real-world analogy

Starting a new club requires an upfront organisational cost even if a large club would benefit everyone. Very small clubs dissolve; once enough members join, the benefits outweigh the setup cost and growth becomes easier. The analogy conveys a barrier between a metastable state and a lower-free-energy outcome, without implying molecules make deliberate choices.

Real-world example

Supercooled water can remain liquid below its equilibrium freezing point until an ice nucleus forms. Adding a seed crystal or contacting a suitable surface lowers the difficulty of starting the new phase and can trigger rapid freezing. The equilibrium phase diagram identifies ice as stable; nucleation kinetics explains the observed delay.

Why?

The new phase lowers bulk free energy in proportion to its volume, but creating a boundary costs energy in proportion to area. At small radii, area cost wins. At larger radii, volume gain grows faster and wins. Their competition creates a finite free-energy maximum that must be crossed before growth becomes downhill.

Common misconception

Being supersaturated does not guarantee immediate precipitation. A metastable state can persist until a suitable nucleus appears. Another mistake is to call r an equilibrium stable droplet size in this simple model; it is the barrier-top size. Also, seed particles lower a kinetic barrier without changing the bulk equilibrium chemical potentials substantially.

Worked example

Take γ = 0.010 J m⁻² and Δg v = 1.0 × 10⁷ J m⁻³. The critical radius is r = 2γ/Δg v = 0.020/(1.0 × 10⁷) = 2.0 × 10⁻⁹ m, or 2.0 nm. The barrier is ΔG = 16πγ³/(3Δg v²) ≈ 1.68 × 10⁻¹⁹ J. The nanoscale radius warns that the sharp-interface model may be approximate even though the algebra is correct.

Quick check

1. If Δg v doubles with γ unchanged, what happens to r and ΔG in the spherical model? Answer: r halves because it scales as 1/Δg v. The barrier becomes one quarter as large because it scales as 1/Δg v².

Exam focus

Define the sign of Δg v explicitly. Use area 4πr² and volume 4πr³/3, check units of γ and Δg v, and verify the nonzero stationary point is a maximum. Distinguish homogeneous from heterogeneous nucleation and thermodynamic phase stability from the kinetics of reaching it.

Advanced insight

The nucleation rate often has an exponential sensitivity to ΔG /kT multiplied by kinetic prefactors. Interfaces, strain and finite-size effects can change the barrier strongly. In a spinodally unstable composition region, infinitesimal fluctuations can grow, providing a different route to phase separation than activated classical nucleation.

Summary

Classical spherical nucleation balances positive surface cost 4πr²γ against negative bulk gain (4π/3)r³Δg v. It predicts r = 2γ/Δg v and ΔG = 16πγ³/(3Δg v²). These barriers explain metastable persistence and its sensitivity to supersaturation and surfaces, while real molecular-scale nuclei may depart from the simple model.

Practice questions

1. Why does a larger interfacial energy make homogeneous nucleation harder at fixed Δg v? Answer: It raises r linearly and ΔG as γ³, increasing the free-energy barrier that fluctuations must cross. 2. A cluster has r < r in the simple model. Is growth downhill in ΔG? Answer: No. On the rising side of the barrier, increasing r raises ΔG, so the small cluster tends to shrink rather than grow spontaneously. 3. Can a seed crystal change the equilibrium melting temperature without other changes? Answer: A seed mainly lowers the kinetic barrier to nucleation. It does not by itself change the bulk equality of phase chemical potentials or the equilibrium melting temperature under the same T,p conditions.