Classifying Phase Transitions
Ehrenfest classification, first-order and continuous transitions
Lesson 3096 of 4,500 · Chemical and Statistical Thermodynamics I
Learning objectives
- Distinguish first-order from continuous transitions
- Relate latent heat and discontinuous first derivatives to coexistence
- Explain the limits of the historical Ehrenfest classification
Introduction
The word transition covers boiling, freezing, magnetic ordering and many other changes, but their thermodynamic signatures differ. At ordinary boiling, two phases coexist and a finite latent heat is absorbed. At some critical or ordering transitions there is no latent-heat jump, yet response properties can become large or singular. Classifying transitions by Gibbs-energy derivatives links observable quantities to the shape of the equilibrium free energy.
Core explanation
At fixed composition, the differential of Gibbs energy is dG=−S dT+V dP. Hence S=−(∂G/∂T) P and V=(∂G/∂P) T. A first-order transition occurs when two equilibrium branches of G cross with different slopes, so entropy or volume (or another first derivative with respect to a field) changes discontinuously. The Gibbs energy itself remains equal for coexisting phases at the transition; otherwise one phase would not be in equilibrium with the other. Melting and ordinary liquid–vapour boiling below the critical point are typical examples.
If the entropy difference is ΔS, the reversible latent heat at the transition is L=TΔS per chosen material amount. Melting normally has positive ΔS and absorbs heat. Vaporisation has an even larger entropy increase in many conditions and absorbs latent heat. A volume jump often accompanies these changes. The Clapeyron relation dP/dT=ΔS/ΔV uses these first-derivative discontinuities to give the slope of a coexistence curve. At a critical endpoint, ΔS and ΔV between liquid and vapour tend toward zero, and simple finite-jump descriptions must be treated carefully.
The historical Ehrenfest scheme called a transition first order if first derivatives of G were discontinuous, second order if first derivatives were continuous but second derivatives showed discontinuities, and so on. This language is useful for linking derivatives to heat capacity, compressibility and expansion. However, modern critical phenomena may produce divergences or nonanalytic power laws rather than a simple finite jump in a particular higher derivative. Calling every continuous transition exactly “second order” can therefore conceal relevant physics.
For a continuous transition, the order parameter changes continuously from zero or one form to another at the transition in a typical description, and there is no latent heat associated with a finite entropy jump. Yet heat capacity or susceptibility may show a cusp, divergence or other singular behaviour. A ferromagnet approaching its Curie temperature in zero applied magnetic field is a classic ordering example. A supercritical liquid–vapour path above the critical point, by contrast, can change properties continuously without crossing a distinct phase boundary; not every smooth change is itself a continuous phase transition.
The phase rule describes coexistence dimensions but does not classify critical singularities. It says an ordinary two-phase pure-substance equilibrium lies on a curve in P–T space. At the liquid–vapour critical point the two phases become indistinguishable, so a naive two-phase label loses meaning at and beyond the endpoint. Statistical mechanics explains why fluctuations and correlation lengths become important there; finite systems round off ideal singularities.
In real calorimetry, a sharp heat-flow peak could reflect latent heat, a continuous-transition heat-capacity anomaly, or kinetic effects. One must test for hysteresis, phase coexistence and scaling as appropriate. The transition order is a property of the thermodynamic limit and equilibrium free energy, not merely the appearance of one instrument trace.
Step-by-step reasoning
Write G's natural-variable differential and identify its first derivatives S and V. Ask whether either has a finite jump at the transition. If so, it is first order and may carry latent heat. If first derivatives remain continuous, inspect response functions such as heat capacity and susceptibility, and avoid assuming every anomaly is a finite second-derivative jump.
Visual explanation
Sketch two Gibbs-energy curves crossing at a transition temperature. Their common value at the crossing but different slopes make the lower-envelope G continuous with a corner: a first-order transition. A continuous critical case instead has matching first derivatives, while the curve's curvature or higher behaviour changes sharply.
Real-world analogy
Joining two road segments at one height but different slopes creates a corner; joining them with the same slope but changing curvature gives a smoother yet detectable transition. Gibbs energy is analogous to the road height, while entropy and volume correspond to slopes with respect to thermodynamic coordinates. The analogy does not reproduce microscopic fluctuations.
Real-world example
When ice melts near ordinary pressure, temperature can remain nearly fixed while heat goes into changing phase amount; a finite fusion enthalpy is involved. At a liquid–vapour critical point, the separate liquid and vapour densities merge and the latent heat of vaporisation approaches zero. These observations show why the endpoint cannot be treated as an ordinary boiling event.
Why?
Why can G be continuous while entropy jumps? At coexistence, the two phases must have equal molar Gibbs energy, but the temperature derivatives of their G branches can differ. The stable lower branch switches at the crossing, producing a discontinuity in −∂G/∂T without a jump in G itself.
Common misconception
“First order” does not mean the phase change happens first in time or proceeds at a particular reaction rate. It refers to the lowest derivative of thermodynamic potential that is discontinuous. Also, a gradual change in colour or density along a path is not automatically a continuous phase transition; a true transition requires nonanalytic equilibrium behaviour in the thermodynamic limit.
Worked example
At a reversible melting transition at T=300 K, suppose the molar fusion entropy is 20 J mol⁻¹ K⁻¹ and molar volume increases by 2.0×10⁻⁶ m³ mol⁻¹. Latent heat is L=TΔS=6.0 kJ mol⁻¹. Clapeyron slope is ΔS/ΔV=20/(2.0×10⁻⁶)=1.0×10⁷ Pa K⁻¹, positive. The finite entropy and volume jumps identify an ordinary first-order boundary under the stated approximation.
Quick check
1. Which Gibbs-energy derivative directly gives entropy at fixed pressure? Answer: S=−(∂G/∂T) P. A finite entropy jump across coexistence means this first derivative is discontinuous.
Exam focus
State that G is equal across equilibrium phases even at a first-order transition. Connect latent heat to TΔS and phase-line slope to ΔS/ΔV. Use “continuous” carefully: no finite latent-heat jump does not mean no sharp heat-capacity or susceptibility anomaly.
Advanced insight
Landau theory uses an order parameter and free-energy expansion to describe many continuous and first-order transitions. Near a critical point, fluctuations can invalidate simple mean-field exponents; universality groups systems by dimensionality and symmetry. Ehrenfest's derivative labels remain a useful introduction but are not a complete classification of nonanalytic critical behaviour.
Summary
First-order transitions have a discontinuity in a first derivative of Gibbs energy, often producing latent heat and volume change. Continuous transitions lack a finite first-derivative jump but can have singular response functions. The historical Ehrenfest scheme organises derivative behaviour; modern critical phenomena require further statistical description.
Practice questions
1. If ΔS fus=10 J mol⁻¹ K⁻¹ at 350 K, what is molar latent heat? Answer: L=TΔS=3500 J mol⁻¹, or 3.5 kJ mol⁻¹. 2. Can two phases at equilibrium have different molar Gibbs energies? Answer: No. Their appropriate chemical potentials are equal at coexistence, though derivatives such as entropy may differ. 3. What happens to ordinary liquid–vapour latent heat approaching the critical point? Answer: It tends toward zero as the two phases become indistinguishable at the critical endpoint. 4. Does a smooth density change above the critical point necessarily mark a new phase transition? Answer: No. A supercritical fluid can vary continuously without crossing a coexistence boundary or another thermodynamic singularity.