4.3 Thermal expansion

Thermal expansion overview

Thermal expansion describes how materials change their dimensions when temperature changes. Because nearly every engineering application involves some temperature variation, understanding this property is essential for predicting material behavior, preventing mechanical failure, and designing devices that work reliably across different conditions.

Thermal expansion connects directly to other thermal properties you've studied in this unit, like specific heat capacity and thermal conductivity. The same lattice vibrations (phonons) that carry heat and store thermal energy also drive the dimensional changes we observe.

Thermal expansion definition

Thermal expansion is the tendency of a material to change its dimensions (length, area, or volume) in response to a temperature change. In most materials, dimensions increase with increasing temperature, though some materials exhibit negative thermal expansion (contraction upon heating).

The physical origin is straightforward: as temperature rises, atoms vibrate with greater amplitude. Because of the asymmetry in the interatomic potential (more on this below), this increased vibration shifts the average interatomic spacing to larger values. The cumulative effect across billions of atoms produces a measurable change in the material's size.

Thermal expansion coefficient

The thermal expansion coefficient quantifies how much a material expands or contracts per degree of temperature change.

• Linear thermal expansion coefficient ($\alpha$): the fractional change in length per unit temperature change.

$\alpha = \frac{1}{L_0}\frac{dL}{dT}$

Here $L_0$ is the initial length and $dL/dT$ is the rate of change of length with temperature. Typical values range from about $0.5 \times 10^{-6} \, \text{K}^{-1}$ for diamond to $\sim 25 \times 10^{-6} \, \text{K}^{-1}$ for aluminum.

• Volumetric thermal expansion coefficient ($\beta$): the fractional change in volume per unit temperature change.

$\beta = \frac{1}{V_0}\frac{dV}{dT}$

These coefficients are generally temperature-dependent, though they're often treated as constants over moderate temperature ranges.

Linear vs volumetric expansion

Linear expansion tracks the change along a single direction; volumetric expansion tracks the change in total volume.

• For isotropic materials (same properties in all directions), the relationship is:

$\beta \approx 3\alpha$

This follows from expanding $(1 + \alpha \Delta T)^3 \approx 1 + 3\alpha \Delta T$ for small $\alpha \Delta T$.

• Anisotropic materials (e.g., hexagonal or tetragonal crystals) can have different values of $\alpha$ along different crystallographic axes. In that case, the volumetric coefficient is the sum of the three principal linear coefficients: $\beta = \alpha_1 + \alpha_2 + \alpha_3$.

Microscopic origin of thermal expansion

Thermal expansion arises from the interplay between interatomic forces and thermal vibrations. To understand why materials expand, you need to look at the shape of the interatomic potential and the role of anharmonicity.

Asymmetric interatomic potential

The key to thermal expansion lies in the shape of the interatomic potential energy curve. A typical pair potential (like the Lennard-Jones or Morse potential) is asymmetric: the repulsive wall at short distances is steep, while the attractive tail at large distances falls off more gradually.

Because of this asymmetry, when atoms vibrate with greater amplitude at higher temperatures, they can move farther apart more easily than they can move closer together. The result is that the mean interatomic separation shifts outward as temperature increases. If the potential were perfectly symmetric (purely harmonic), the average position would stay fixed regardless of vibration amplitude, and there would be no thermal expansion.

Anharmonicity of lattice vibrations

In a purely harmonic potential, the restoring force is exactly proportional to displacement, and the time-averaged atomic position doesn't change with temperature. Real interatomic potentials are anharmonic: the potential energy contains cubic and higher-order terms beyond the quadratic.

You can write the potential near equilibrium as:

$U(x) = cx^2 - gx^3 - fx^4 + \cdots$

where $x$ is the displacement from equilibrium. The cubic term ($-gx^3$) is the leading source of asymmetry and is directly responsible for thermal expansion. It also governs phonon-phonon scattering, which is why thermal expansion and thermal resistance share the same microscopic origin.

Quasi-harmonic approximation

The quasi-harmonic approximation (QHA) provides a practical way to calculate thermal expansion from phonon properties. The idea is:

1. Treat the lattice vibrations as harmonic at any given volume, so you can compute phonon frequencies $\omega(\mathbf{q}, V)$ as a function of volume.
2. Calculate the Helmholtz free energy $F(T, V)$ using the harmonic phonon spectrum at each volume.
3. Minimize $F(T, V)$ with respect to volume at each temperature to find the equilibrium volume $V(T)$.

The QHA captures the essential physics of thermal expansion (volume-dependent phonon frequencies) without requiring a full treatment of anharmonic phonon interactions. It works well at moderate temperatures but breaks down near the melting point or in strongly anharmonic systems.

The connection to macroscopic thermodynamics comes through the Grüneisen parameter $\gamma$, which measures how phonon frequencies shift with volume:

$\gamma = -\frac{V}{\omega}\frac{d\omega}{dV}$

The volumetric expansion coefficient can then be expressed as:

$\beta = \frac{\gamma C_V}{B V}$

where $C_V$ is the heat capacity at constant volume and $B$ is the bulk modulus. This equation neatly links microscopic phonon behavior to the macroscopic expansion you measure in the lab.

Factors affecting thermal expansion

Several factors control how much a material expands with temperature. Knowing these helps you predict expansion behavior and select materials for temperature-sensitive applications.

Bonding strength and thermal expansion

Stronger interatomic bonds generally mean lower thermal expansion. The logic: stronger bonds create a deeper, narrower potential well, so atoms need more thermal energy to shift their average position.

• Covalent and ionic bonds (e.g., diamond, alumina) produce low $\alpha$ values.
• Metallic bonds give moderate $\alpha$ values.
• Van der Waals bonds (e.g., polymers, molecular solids) lead to high $\alpha$ values.

For example, diamond (strong covalent bonds) has $\alpha \approx 1 \times 10^{-6} \, \text{K}^{-1}$, while polyethylene (weak intermolecular forces) has $\alpha \approx 200 \times 10^{-6} \, \text{K}^{-1}$.

Lattice structure and thermal expansion

Crystal structure affects thermal expansion in two ways: through the coordination environment (how atoms are arranged around each other) and through the available space for vibration.

• Open structures like diamond cubic tend to have lower expansion coefficients because the extra space in the lattice allows atoms to vibrate without pushing neighbors apart as much.
• Close-packed structures (FCC, HCP) tend to have somewhat higher expansion because atomic vibrations more directly translate into volume changes.
• Crystal symmetry also determines whether expansion is isotropic or anisotropic. Cubic crystals expand equally in all directions; hexagonal crystals (e.g., zinc, graphite) can have very different expansion along the $c$-axis versus the basal plane.

Composition and thermal expansion

Chemical composition provides another lever for tuning thermal expansion.

• Alloying can modify bond strengths and lattice parameters. For instance, adding nickel to iron produces Invar (Fe-36%Ni), which has an anomalously low $\alpha \approx 1.2 \times 10^{-6} \, \text{K}^{-1}$ near room temperature due to a magnetic contribution that counteracts normal lattice expansion.
• Composite materials can be engineered with specific expansion properties by combining constituents with different $\alpha$ values. Carbon fiber-reinforced polymers, for example, can achieve near-zero expansion along the fiber direction.

Thermal expansion in crystals

Crystalline materials require a more detailed description of thermal expansion because their properties can vary with direction. This matters for single crystals, epitaxial thin films, and textured polycrystals.

Isotropic vs anisotropic expansion

• Isotropic expansion: Cubic crystals (e.g., NaCl, Cu, Si) have the same $\alpha$ in every direction. A single scalar value fully describes their expansion.
• Anisotropic expansion: Lower-symmetry crystals have direction-dependent expansion. Calcite ($\text{CaCO}_3$) actually expands along one axis and contracts along another when heated. Graphite expands much more along the $c$-axis (perpendicular to the layers) than within the basal plane.

Anisotropic expansion in polycrystalline materials can generate internal stresses between differently oriented grains, which is one reason some ceramics are prone to microcracking during thermal cycling.

Thermal expansion tensor

The full description of thermal expansion in a crystal uses a second-rank symmetric tensor $\alpha_{ij}$, relating thermal strain to temperature change:

$\varepsilon_{ij} = \alpha_{ij} \Delta T$

Because it's symmetric ($\alpha_{ij} = \alpha_{ji}$), the tensor has at most 6 independent components. Crystal symmetry reduces this further:

• Cubic: 1 independent component ($\alpha_{11} = \alpha_{22} = \alpha_{33}$, off-diagonals zero)
• Hexagonal/tetragonal: 2 independent components ($\alpha_{11} = \alpha_{22} \neq \alpha_{33}$)
• Orthorhombic: 3 independent components ($\alpha_{11} \neq \alpha_{22} \neq \alpha_{33}$)
• Monoclinic/triclinic: up to 4 or 6 independent components, including off-diagonal terms

Negative thermal expansion materials

Some materials contract upon heating along one or more directions. This negative thermal expansion (NTE) is unusual and technologically useful for creating composites with zero net expansion.

Mechanisms behind NTE include:

• Transverse vibrational modes: In framework structures, transverse vibrations of bridging atoms (e.g., oxygen in $\text{ZrW}_2\text{O}_8$) can pull the connected polyhedra closer together, shrinking the overall structure even as bond lengths increase.
• Rigid unit modes (RUMs): Cooperative rotations of rigid polyhedral units can reduce volume without compressing individual bonds.
• Magnetic or electronic transitions: Invar-type alloys show reduced expansion due to competing magnetic states.

Notable NTE materials:

• Zirconium tungstate ($\text{ZrW}_2\text{O}_8$): isotropic NTE over a wide temperature range (0.3 to 1050 K), with $\alpha \approx -9 \times 10^{-6} \, \text{K}^{-1}$
• Beta-eucryptite ($\text{LiAlSiO}_4$): strong NTE along the $c$-axis
• Certain metal-organic frameworks (MOFs): tunable NTE through framework design

Thermal expansion measurements

Accurate measurement of thermal expansion is essential for materials characterization and engineering design. The three main techniques each have distinct strengths.

Dilatometry techniques

Dilatometry is the most common method for measuring linear thermal expansion of bulk samples.

How it works:

1. Place the sample in a furnace or cryostat with controlled temperature.
2. A displacement sensor (capacitive, inductive, or optical) tracks the change in sample length as temperature varies.
3. Record $\Delta L$ vs. $T$ and compute $\alpha(T) = (1/L_0)(dL/dT)$.

Push-rod dilatometers are the standard workhorse, capable of measuring from cryogenic temperatures up to ~1600°C with resolution on the order of nanometers. The technique works for bulk solids, thin films (with appropriate fixtures), and even powders (in a container).

X-ray diffraction methods

X-ray diffraction (XRD) measures thermal expansion at the lattice level by tracking how lattice parameters change with temperature.

1. Collect diffraction patterns at a series of temperatures using a high-temperature or low-temperature stage.
2. Determine lattice parameters at each temperature from the peak positions using Bragg's law.
3. Fit lattice parameter vs. temperature to extract $\alpha$ along each crystallographic direction.

XRD is particularly valuable for anisotropic materials because it gives you the expansion along each crystal axis separately. It also works on powders and thin films. The main limitation is that it measures the intrinsic lattice expansion and doesn't capture contributions from grain boundaries or porosity.

Optical interferometry approaches

Optical interferometry (e.g., Fabry-Perot or Michelson configurations) offers the highest precision for thermal expansion measurements.

The sample sits between two reflective surfaces forming an optical cavity. As the sample expands or contracts with temperature, the cavity length changes, shifting the interference fringe pattern. By counting fringe shifts, you can determine length changes with sub-nanometer resolution.

This technique is the gold standard for calibration specimens and reference materials, but it requires optically flat, parallel sample surfaces and careful vibration isolation.

Thermal expansion applications

Thermal expansion is not just a property to account for; it can be actively exploited in device design or carefully managed to prevent failure.

Thermomechanical actuators and sensors

Thermal expansion drives several types of actuators and sensors:

• Bimetallic strips bond two metals with different $\alpha$ values (e.g., brass and steel). When heated, the strip bends because one side expands more than the other. These are used in thermostats, circuit breakers, and temperature indicators.
• Fiber Bragg grating sensors detect temperature through the thermal expansion of the fiber, which shifts the Bragg wavelength. They're widely used for structural health monitoring.
• MEMS thermal actuators use localized Joule heating to produce controlled displacements in microelectromechanical systems.

Thermal expansion in manufacturing

Thermal expansion mismatches are a major source of problems in manufacturing:

• Shrink fitting: A metal ring is heated so it expands, slipped over a shaft, then cooled. The contraction creates an extremely tight interference fit without adhesives or fasteners.
• Precision machining: Components machined at room temperature may not fit at operating temperature if expansion isn't accounted for. Tolerances must include thermal effects.
• Soldering and brazing: Joints between dissimilar materials develop residual stresses upon cooling due to $\alpha$ mismatch. This can cause warping or cracking if not managed through material selection or compliant interlayers.

Thermal expansion in materials design

Controlling thermal expansion is a core challenge in several fields:

• Electronic packaging: Silicon chips ($\alpha \approx 2.6 \times 10^{-6} \, \text{K}^{-1}$) are mounted on substrates and connected to circuit boards with much higher $\alpha$. Materials like Invar, Kovar, and glass-ceramics are chosen to bridge this mismatch and prevent solder joint fatigue.
• Aerospace composites: Carbon fiber-reinforced polymers can be designed with near-zero $\alpha$ along specific directions by controlling fiber orientation, which is critical for dimensionally stable space structures like telescope mirrors and antenna supports.
• Dental and biomedical: Porcelain coatings on metal dental crowns must have matched $\alpha$ to avoid cracking during the thermal cycling of eating and drinking.

Thermal expansion related phenomena

Thermal expansion doesn't act in isolation. It connects to several other phenomena that are important in solid state physics and engineering.

Thermal shock and thermal stress

Thermal shock occurs when a material experiences a rapid temperature change, creating a non-uniform temperature distribution and therefore non-uniform expansion. The resulting internal stresses can cause cracking or fracture.

A material's resistance to thermal shock depends on a combination of properties. A common figure of merit is:

$R = \frac{\sigma_f (1 - \nu)}{E \alpha}$

where $\sigma_f$ is the fracture strength, $\nu$ is Poisson's ratio, $E$ is Young's modulus, and $\alpha$ is the linear expansion coefficient. Materials with low $\alpha$, low $E$, and high strength resist thermal shock best. High thermal conductivity also helps by reducing temperature gradients.

This is why silicon nitride and silicon carbide are preferred over alumina for applications involving rapid heating or cooling.

Thermal expansion and phase transitions

Phase transitions can produce abrupt changes in thermal expansion behavior. During a first-order transition (e.g., a structural transformation), the volume typically changes discontinuously, producing a spike or discontinuity in the measured expansion curve.

Examples:

• The BCC-to-FCC transition in iron at 912°C involves a volume contraction of about 1%.
• Martensitic transformations in shape-memory alloys produce large, reversible dimensional changes.
• Ferroelectric transitions (e.g., in $\text{BaTiO}_3$) are accompanied by changes in lattice parameters and expansion coefficients.

These volume changes can generate significant internal stresses in constrained geometries, which is why thermal cycling through phase transitions can lead to fatigue and failure.

Thermal expansion and thermoelectricity

Thermoelectric devices convert temperature differences to voltage (Seebeck effect) or use voltage to pump heat (Peltier effect). Thermal expansion affects these devices in two ways:

• Performance: The thermoelectric figure of merit $ZT = \frac{S^2 \sigma T}{\kappa}$ (where $S$ is the Seebeck coefficient, $\sigma$ is electrical conductivity, and $\kappa$ is thermal conductivity) can be influenced by thermal expansion through changes in band structure and carrier concentration as the lattice expands.
• Reliability: Thermoelectric modules operate across large temperature gradients, so thermal expansion mismatches between the thermoelectric legs and the contact/substrate materials generate stresses that limit device lifetime. Matching $\alpha$ between components is a key design consideration.