4.6 Structural phase transitions

Types of phase transitions

Structural phase transitions occur when a material's crystal structure reorganizes, often changing its electronic, magnetic, or mechanical properties in the process. These transitions sit at the heart of condensed matter physics because they connect microscopic atomic arrangements to macroscopic material behavior.

First vs second order

The Ehrenfest classification sorts phase transitions by which derivative of the free energy is discontinuous at the transition point.

• First-order transitions have a discontinuous first derivative of the free energy (entropy, volume). They involve latent heat, and the two phases coexist at the transition temperature. Familiar examples: melting of ice, boiling of water, and many solid-solid structural transitions.
• Second-order transitions have a continuous first derivative but a discontinuous (or divergent) second derivative (heat capacity, susceptibility). There's no latent heat and no phase coexistence. Instead, critical fluctuations dominate near the transition.

In practice, many structural transitions fall somewhere between these idealized categories, which is why the continuous/discontinuous language is often more useful.

Continuous vs discontinuous transitions

This is the modern way to classify transitions, based on the behavior of the order parameter:

• Continuous (second-order) transitions: the order parameter goes smoothly to zero at the critical temperature. Critical phenomena, diverging correlation lengths, and universality emerge here. Ferromagnetic and superconducting transitions are classic examples.
• Discontinuous (first-order) transitions: the order parameter jumps abruptly at the transition. These often show hysteresis and metastable states. Martensitic transformations in shape memory alloys are a good structural example.

Landau theory overview

Landau theory is a phenomenological framework for describing phase transitions near the critical point. The core idea: expand the free energy $F$ as a power series in the order parameter $\phi$:

$F = F_0 + a(T)\phi^2 + b\phi^4 + \cdots$

where the coefficient $a(T)$ changes sign at the critical temperature $T_c$, driving the transition. For a second-order transition, $a(T) \propto (T - T_c)$, and minimizing $F$ gives the order parameter behavior.

Landau theory predicts mean-field critical exponents and provides a starting point for classifying transitions by symmetry. Its main limitation is that it neglects fluctuations, which become dominant near $T_c$ in low-dimensional systems, causing mean-field predictions to break down.

Order parameters

An order parameter is a quantity that distinguishes the ordered phase from the disordered one. It's zero above the transition and nonzero below it, making it the central variable in any theoretical description of a phase transition.

Definition and significance

• Quantifies the degree of broken symmetry in the ordered phase
• Vanishes in the high-symmetry (disordered) phase
• Allows you to write down a free energy functional and derive thermodynamic predictions
• Choosing the right order parameter is often the first step in analyzing any transition

Examples in different systems

Different transitions call for different order parameters:

• Ferromagnetic transitions: magnetization $\mathbf{M}$
• Ferroelectric transitions: electric polarization $\mathbf{P}$
• Liquid-gas transitions: density difference $\Delta \rho = \rho_l - \rho_g$
• Superconducting transitions: the complex gap function $\Delta$ (or equivalently, the macroscopic wavefunction amplitude)
• Structural transitions: atomic displacement from the high-symmetry position, or the amplitude of a specific phonon mode

Temperature dependence

The order parameter typically decreases as temperature rises toward $T_c$. Near the transition, this behavior follows a power law:

$\phi \propto (T_c - T)^{\beta}$

Mean-field (Landau) theory predicts $\beta = 1/2$, giving $\phi \propto (T_c - T)^{1/2}$. In real systems, fluctuations modify this exponent. For example, the 3D Ising model gives $\beta \approx 0.326$, which matches experiments on uniaxial ferromagnets much better than the mean-field value.

Symmetry considerations

Symmetry is the organizing principle behind structural phase transitions. Almost every structural transition involves a reduction in symmetry, and the mathematical language of group theory tells you what transitions are allowed and what their consequences will be.

Symmetry breaking in transitions

When a system undergoes a structural transition, its ground state typically has lower symmetry than the governing Hamiltonian. This is spontaneous symmetry breaking.

• A cubic perovskite distorting to a tetragonal structure breaks some rotational symmetries while preserving others.
• Ferromagnetic ordering picks out a preferred direction, breaking the rotational symmetry of the paramagnetic phase.
• The reduction in symmetry often enables new physical properties (piezoelectricity, ferroelectricity) that are forbidden by the higher symmetry.

Group theory applications

Group theory provides the mathematical backbone for analyzing symmetry changes:

• The high-symmetry and low-symmetry phases each have a space group, and the low-symmetry group is a subgroup of the high-symmetry one.
• Group theory predicts which order parameter components are allowed and how they couple in the Landau free energy expansion.
• It also determines selection rules for spectroscopic probes and constrains the form of elastic and dielectric tensors.

Goldstone modes

When a continuous symmetry is spontaneously broken, the Goldstone theorem guarantees the existence of gapless (massless) excitations. The number of these modes equals the number of broken symmetry generators.

• Spin waves (magnons) in ferromagnets arise from broken rotational symmetry of spins.
• Acoustic phonons in crystals arise from broken translational symmetry.

These Goldstone modes dominate the low-energy physics of the ordered phase and control thermodynamic properties at low temperatures (e.g., the $T^{3/2}$ Bloch law for magnetization in ferromagnets).

If only a discrete symmetry is broken (as in an Ising-like structural distortion), no Goldstone mode appears.

Thermodynamic aspects

Thermodynamics connects the microscopic physics of a transition to measurable quantities like heat capacity, susceptibility, and compressibility.

Free energy landscapes

Visualizing the free energy $F(\phi)$ as a function of the order parameter is a powerful way to understand transitions:

• Second-order transitions: the free energy has a single minimum that smoothly shifts from $\phi = 0$ (above $T_c$) to $\phi \neq 0$ (below $T_c$).
• First-order transitions: multiple minima coexist near the transition. The system can get trapped in a metastable minimum, which is why you see hysteresis and superheating/supercooling.

Critical exponents

Near a continuous transition, physical quantities diverge or vanish as power laws of $|T - T_c|$. The key exponents are:

• $\alpha$: specific heat, $C \propto |T - T_c|^{-\alpha}$
• $\beta$: order parameter, $\phi \propto (T_c - T)^{\beta}$
• $\gamma$: susceptibility, $\chi \propto |T - T_c|^{-\gamma}$
• $\nu$: correlation length, $\xi \propto |T - T_c|^{-\nu}$

These exponents are not independent; they satisfy scaling relations like $\alpha + 2\beta + \gamma = 2$ (Rushbrooke's equality).

Universality classes

Systems with the same dimensionality, order parameter symmetry, and range of interactions share the same critical exponents, regardless of microscopic details. This is universality.

• Ising class ($d=3$, scalar order parameter): uniaxial ferromagnets, liquid-gas transitions, order-disorder alloys
• XY class ($d=3$, 2-component order parameter): superfluid helium, planar magnets
• Heisenberg class ($d=3$, 3-component order parameter): isotropic ferromagnets

Universality is one of the deepest results in the theory of critical phenomena and is why studying simple models like the Ising model teaches you about real materials.

Microscopic mechanisms

The macroscopic signatures of a structural transition originate from specific atomic-scale processes. Understanding these mechanisms is essential for predicting and engineering transitions.

Atomic displacements

Structural transitions generally fall into two microscopic categories:

• Displacive transitions: atoms shift by small amounts from their high-symmetry positions. The transition is driven by a lattice instability. The perovskite family is the textbook example: in $\text{BaTiO}_3$, the Ti ion displaces relative to the oxygen octahedron, producing ferroelectricity.
• Order-disorder transitions: atoms or molecular groups that were randomly distributed among several equivalent sites become ordered onto a subset of those sites. $\text{KH}_2\text{PO}_4$ (KDP) is a classic case, where proton ordering drives the ferroelectric transition.

Many real transitions have mixed displacive and order-disorder character.

Electronic structure changes

Phase transitions can reshape the electronic landscape:

• Metal-insulator transitions (e.g., $\text{VO}_2$) involve dramatic band structure changes at the transition, often coupled to a structural distortion.
• Spin-state transitions in transition metal compounds involve changes in orbital occupancy driven by competition between crystal field splitting and Hund's coupling.
• Charge-transfer transitions in organic conductors redistribute electrons between molecular sites.

Phonon softening

A phonon mode whose frequency drops toward zero signals an impending structural instability. This is the soft mode concept:

1. As temperature approaches $T_c$ from above, a specific phonon mode decreases in frequency: $\omega^2 \propto (T - T_c)$.

2. At $T_c$, the mode "freezes out" ($\omega \to 0$), and the corresponding atomic displacement pattern becomes the static distortion of the new phase.

3. Below $T_c$, the mode hardens again as the new structure stabilizes.

Soft modes are directly observable through inelastic neutron scattering or Raman spectroscopy. The ferroelectric transition in $\text{SrTiO}_3$ is a well-studied example (though it's complicated by quantum fluctuations that suppress the transition near 0 K).

Experimental techniques

Characterizing structural transitions requires probes sensitive to changes in atomic positions, lattice dynamics, and thermodynamic quantities.

X-ray diffraction methods

X-ray diffraction is the primary tool for detecting changes in crystal symmetry:

• Powder diffraction identifies phases and tracks lattice parameter changes through a transition. Rietveld refinement extracts detailed structural information.
• Single-crystal diffraction provides the most complete structural picture, including subtle distortions and superlattice reflections that signal symmetry lowering.
• Synchrotron sources offer high brilliance for time-resolved studies (watching a transition unfold) and high-pressure experiments using diamond anvil cells.

Neutron scattering

Neutrons complement X-rays in important ways:

• They scatter strongly from light elements (hydrogen, oxygen, lithium) that X-rays miss.
• They carry a magnetic moment, making them sensitive to magnetic order.
• Elastic neutron scattering determines crystal and magnetic structures.
• Inelastic neutron scattering measures phonon and magnon dispersions directly, which is how soft modes are mapped across the Brillouin zone.

Calorimetry measurements

Calorimetry directly measures the thermodynamic signatures of transitions:

• Differential scanning calorimetry (DSC) is the workhorse technique: it measures heat flow vs. temperature, giving you transition temperatures and latent heats (for first-order transitions) or heat capacity anomalies (for second-order transitions).
• Adiabatic calorimetry provides the most precise absolute heat capacity values.
• AC calorimetry probes the frequency-dependent thermal response, useful for studying dynamic and glassy aspects of transitions.

Computational approaches

Simulations and first-principles calculations let you explore transitions at a level of detail that experiments alone cannot reach.

Density functional theory

DFT calculates electronic structure from first principles by solving the Kohn-Sham equations. For structural transitions, DFT can:

• Compare total energies of competing crystal structures to predict which is stable
• Calculate phonon spectra (via density functional perturbation theory) and identify soft modes
• Determine elastic constants and Born effective charges

DFT works well for weakly correlated systems but struggles with strongly correlated materials (e.g., Mott insulators), where methods like DFT+U or dynamical mean-field theory (DMFT) are needed.

Molecular dynamics simulations

MD simulations integrate Newton's equations for atoms interacting through a potential (either empirical or DFT-derived):

• They capture the real-time dynamics of a transition, including nucleation events and domain wall motion.
• They're useful for studying kinetics and non-equilibrium processes.
• The main limitations are system size (typically thousands to millions of atoms) and simulation time (nanoseconds to microseconds), which can miss slow processes like diffusion-controlled transitions.

Monte Carlo methods

MC methods sample configuration space stochastically rather than following real-time dynamics:

• The Metropolis algorithm generates configurations weighted by the Boltzmann distribution, giving equilibrium thermodynamic properties.
• MC is particularly effective for lattice models (Ising, Potts, etc.) near critical points.
• Cluster algorithms (Wolff, Swendsen-Wang) dramatically reduce critical slowing down, making it feasible to extract accurate critical exponents.

Examples in materials

Ferroelectric transitions

Ferroelectrics develop a spontaneous electric polarization $\mathbf{P}$ below a Curie temperature $T_c$:

• $\text{BaTiO}_3$ is the prototypical displacive ferroelectric. It undergoes a sequence of transitions: cubic $\to$ tetragonal $\to$ orthorhombic $\to$ rhombohedral as temperature decreases, each involving Ti displacement along a different crystallographic direction.
• $\text{KH}_2\text{PO}_4$ (KDP) is an order-disorder ferroelectric where proton ordering along hydrogen bonds drives the polarization.
• Applications span capacitors, piezoelectric sensors and actuators, infrared detectors, and non-volatile ferroelectric RAM (FeRAM).

Martensitic transformations

These are diffusionless, first-order structural transitions where atoms move cooperatively by less than one interatomic distance:

• The parent phase (austenite) transforms to a lower-symmetry product phase (martensite) via a shear-like mechanism.
• In NiTi (Nitinol), the transformation is reversible, producing the shape memory effect: deform the martensite, heat above the transition, and it snaps back to the austenite shape.
• In steels, rapid quenching traps the martensitic phase, producing a hard but brittle microstructure.
• These transitions show large transformation strains (up to ~10%) and significant thermal hysteresis.

Charge density waves

A charge density wave (CDW) is a periodic modulation of the conduction electron density coupled to a periodic lattice distortion:

• CDWs are driven by electron-phonon coupling combined with Fermi surface nesting: if large parallel segments of the Fermi surface are connected by a single wavevector $\mathbf{q}$, the electronic susceptibility diverges at that wavevector, and a lattice distortion with periodicity $2\pi/|\mathbf{q}|$ becomes favorable.
• Classic CDW materials include $\text{NbSe}_2$, $\text{TaS}_2$, and quasi-1D conductors like $\text{K}_{0.3}\text{MoO}_3$ (blue bronze).
• CDW order competes with superconductivity in many of these materials, making their phase diagrams rich and actively studied.

Kinetics of transitions

Thermodynamics tells you which phase is stable; kinetics tells you how fast the system gets there and what path it takes.

Nucleation and growth

This is the dominant mechanism for first-order transitions:

1. Nucleation: thermal fluctuations create small clusters of the new phase. A nucleus must exceed a critical size $r^*$ to be stable, because below that size the surface energy cost outweighs the bulk free energy gain.
2. Growth: once a supercritical nucleus forms, it grows by atoms attaching at the interface.
3. The nucleation rate depends exponentially on the free energy barrier $\Delta G^*$, which itself depends on the degree of undercooling or superheating.

Classical nucleation theory gives the critical radius as $r^* = \frac{2\gamma}{\Delta g}$, where $\gamma$ is the surface energy per unit area and $\Delta g$ is the bulk free energy difference per unit volume between the two phases.

Spinodal decomposition

Inside the spinodal region of a phase diagram (where $\frac{\partial^2 F}{\partial \phi^2} < 0$), there is no nucleation barrier. Any small fluctuation in composition grows spontaneously:

• The system separates into two phases through a continuous, wavelength-dependent process described by the Cahn-Hilliard equation.
• The resulting microstructure is an interconnected, labyrinthine pattern with a characteristic length scale that coarsens over time.
• Spinodal decomposition is important in alloy design, polymer blends, and glass-ceramics.

Time-dependent Ginzburg-Landau theory

TDGL extends Landau theory to dynamics by writing an equation of motion for the order parameter:

$\frac{\partial \phi}{\partial t} = -\Gamma \frac{\delta F}{\delta \phi} + \eta(t)$

where $\Gamma$ is a kinetic coefficient and $\eta(t)$ is a stochastic noise term representing thermal fluctuations. TDGL predicts relaxation times, dynamic critical exponents, and scaling behavior. It's widely applied to superconductors (where $\phi$ is the superconducting order parameter) and liquid crystals.

Applications and implications

Functional materials design

Structural transitions are the working principle behind many functional materials:

• Shape memory alloys (NiTi, Cu-Al-Ni): exploiting the martensitic transformation for actuators, stents, and orthodontic wires
• Piezoelectrics (PZT, $\text{BaTiO}_3$): converting mechanical stress to voltage (and vice versa) for sensors, ultrasound transducers, and energy harvesting
• Magnetocaloric materials ($\text{Gd}_5\text{Si}_2\text{Ge}_2$): the large entropy change at a coupled magnetic-structural transition enables solid-state refrigeration
• Phase change materials (paraffins, salt hydrates): storing and releasing thermal energy through solid-liquid transitions for building temperature regulation

Phase change memory devices

Phase change memory (PCM) exploits the rapid, reversible transition between amorphous and crystalline states in chalcogenide glasses (e.g., $\text{Ge}_2\text{Sb}_2\text{Te}_5$, known as GST):

• A short, intense electrical pulse melts and quenches the material into the amorphous (high-resistance) state.
• A longer, lower-power pulse anneals it into the crystalline (low-resistance) state.
• Switching times are on the order of nanoseconds, and the resistance contrast between states is large (typically $10^2$ to $10^3$).
• PCM is a leading candidate for next-generation non-volatile memory and has potential in neuromorphic computing, where the gradual crystallization process mimics synaptic plasticity.

Structural transitions in geology

Phase transitions in minerals shape the interior structure of Earth and other planets:

• The olivine-to-wadsleyite transition at ~410 km depth and the ringwoodite-to-bridgmanite + ferropericlase transition at ~660 km depth define the boundaries of the mantle transition zone. These are detected as seismic velocity discontinuities.
• Ice polymorphism is relevant to the interiors of icy moons (Europa, Enceladus) and giant planets, where pressures stabilize exotic ice phases (ice VII, ice X) with very different properties from ordinary ice.
• Understanding these transitions constrains models of mantle convection, plate tectonics, and planetary evolution.