Statistical Models of Phase Separation

Introduction

Why do some liquids mix fully while others separate? Mixing creates entropy, favouring one phase, but unlike-molecule interactions can impose an energetic cost. A simple regular-solution model places these effects in one free-energy curve. Its shape reveals stable mixing, metastability, spinodal instability and a critical solution temperature.

Core explanation

For a symmetric binary mixture at fixed pressure, consider molar mixing Gibbs energy g mix(x) = RT[x ln x + (1−x)ln(1−x)] + w x(1−x), where x is mole fraction of B and w has energy-per-mole units. The first term is ideal configurational mixing entropy, always nonpositive for 0 < x < 1. The second is a simplified excess enthalpy: w > 0 penalises unlike contacts and favours demixing, while w < 0 favours mixing more strongly. The model assumes a composition-independent w for the interval considered.

The first derivative gives the local free-energy slope; the second derivative is g mix''(x) = RT[1/x + 1/(1−x)] − 2w. Positive curvature means a homogeneous composition resists infinitesimal fluctuations; negative curvature means small composition variations can lower the local free energy, the spinodal instability condition. Spinodal boundaries satisfy g mix'' = 0. At the symmetric critical composition x = 1/2, curvature is 4RT − 2w. If w is positive and roughly temperature independent, the model critical temperature is T c = w/(2R). Above that temperature, entropy wins throughout the symmetric composition range and the curve is convex.

Binodal compositions are not found merely by setting curvature to zero. They are the two compositions sharing a common tangent, which ensures equal component chemical potentials between phases. Between binodal and spinodal, a homogeneous mixture can be metastable: small fluctuations increase free energy, but a sufficiently large separated domain can lower it after overcoming interfacial cost. Inside the spinodal, infinitesimal fluctuations are favourable in the bulk model. At the critical point, binodal and spinodal endpoints merge for this symmetric idealisation.

The model is not a universal theory of all solutions. Real w can depend on T, pressure and composition; molecules can have unequal sizes, directional interactions or chemical association. These effects can shift critical compositions, create lower critical solution temperatures or yield multiple gaps. The simple formula is valuable because it exposes the competition clearly, not because every mixture follows its symmetric curve.

Step-by-step reasoning

Write g mix(x), identify the sign and units of w, and calculate first and second derivatives. At a chosen T inspect g'' at the composition of interest for local stability. For global two-phase equilibrium, use a common tangent or equal component μ rather than curvature alone. Check whether the symmetric constant-w assumptions apply before using T c = w/(2R).

Visual explanation

Plot g mix against x at high T as one convex bowl. At lower T with positive w, draw a double-well-like curve with two stable compositions joined by a common tangent. Mark outer binodal points and inner spinodal points. Above the curves sketch mixing entropy pushing downward and interaction energy pushing upward at intermediate x.

Real-world analogy

Mixing two groups creates many possible arrangements, an entropic advantage, but unfavourable cross-group contacts can make separation attractive. A warm, strongly shuffled situation may remain mixed, while colder conditions reveal the contact penalty. The analogy conveys competition; it does not supply the actual molecular interaction parameter.

Real-world example

Temperature-dependent liquid–liquid miscibility gaps can be interpreted by fitting an interaction term and comparing it with the RT mixing contribution. Such models help explain why a mixture is homogeneous above an upper critical solution temperature but separates into two layers below it. Experimental tie lines are needed to test the predicted compositions.

Why?

The logarithmic entropy term is most stabilising at intermediate composition and grows in importance with T. A positive w term is also largest near the middle but raises free energy there. At sufficiently low T, the energetic penalty can overcome entropy, causing nonconvexity and making a two-phase weighted average lower in G than a homogeneous middle composition.

Common misconception

Negative curvature is a local spinodal criterion, not the full condition for a miscibility gap. A mixture can already be globally unstable to two-phase separation while locally stable to tiny fluctuations. Also, w > 0 does not guarantee demixing at every T; the RT entropy term can dominate at sufficiently high temperature in this model.

Worked example

Let w = 10.0 kJ mol⁻¹ and R = 8.314 J mol⁻¹ K⁻¹. The symmetric model predicts T c = w/(2R) = 10000/(16.628) ≈ 601 K. At T = 500 K and x = 0.5, g'' = 4RT − 2w ≈ 4(8.314)(500) − 20000 = −3372 J mol⁻¹, where x is dimensionless. The negative curvature marks local instability at the central composition. At T = 700 K, the same curvature is positive, consistent with a convex central region above the model T c.

Quick check

1. What does g mix''(x) < 0 imply for small composition fluctuations in the homogeneous bulk model? Answer: They can lower Gibbs energy, so the homogeneous state is locally unstable and lies inside the spinodal region under the model assumptions.

Exam focus

Keep w and RT in the same energy-per-mole units. Differentiate both logarithmic terms correctly. Use curvature for local stability, common tangent for binodal coexistence and material balance for phase amounts. State that T c = w/(2R) belongs to the symmetric constant-w model, not all real mixtures.

Advanced insight

Adding an interfacial-gradient penalty to the bulk free-energy model gives a diffuse-interface theory of spinodal decomposition. Negative bulk curvature drives concentration fluctuations, while gradient energy penalises very short wavelengths, selecting characteristic early-stage patterns. This connects equilibrium free-energy shape to kinetic morphology.

Summary

The regular-solution model combines favourable ideal mixing entropy with an interaction term w x(1−x). Positive w can produce phase separation below a model critical temperature. Common tangents locate binodal coexistence; zero or negative curvature locates spinodal limits and local instability. Real mixtures may require more complex interactions.

Practice questions

1. If w = 0, is the symmetric model stable against demixing at any positive T for interior x? Answer: Yes in this model. g'' = RT[1/x+1/(1−x)] is positive for 0 < x < 1, so ideal mixing is convex and favours one homogeneous phase. 2. At x = 0.5, T = 300 K, what minimum w makes the curvature nonpositive? Answer: Set 4RT − 2w ≤ 0, giving w ≥ 2RT ≈ 2(8.314)(300) = 4.99 kJ mol⁻¹. 3. Why are binodal compositions generally outside the spinodal compositions on a miscibility-gap diagram? Answer: Global two-phase coexistence can lower G before the homogeneous state loses local stability. The intermediate metastable region lies between binodal and spinodal boundaries.