{
  "id": "thermal-expansion",
  "title": "Thermal Expansion",
  "lang": "en",
  "source_lang": "en",
  "summary": "Thermal expansion is the temperature-dependent change in a material’s dimensions, governed by atomic interactions, structure, and mechanical constraints.",
  "category": "science",
  "subcategory": "physics",
  "type": "concept",
  "aliases": [
    "Thermic expansion"
  ],
  "generated_at": "2026-11-03",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/thermal-expansion/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/thermal-expansion/",
      "json": "https://aiwiki.page/zh/thermal-expansion.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "Thermal expansion is the change in the length, area, or volume of a material caused by a change in temperature. Most substances expand when heated and contract when cooled, although some contract over particular temperature ranges or expand in one direction while contracting in another. The effect occurs in solids, liquids, and gases and reflects the way temperature changes their microscopic motion and equilibrium structure.\n\nThermal expansion connects atomic-scale interactions with observable behavior: the movement of a bridge bearing, the rise of liquid in a thermometer, changes in crystal dimensions, and part of the rise in global sea level. Its quantitative description requires both a material property—the expansion coefficient—and specification of conditions such as pressure, temperature range, direction, and mechanical restraint. Ordinary expansion within a single phase must also be distinguished from abrupt volume changes during melting, freezing, or other transformations."
    },
    {
      "heading": "Expansion coefficients and dimensional change",
      "text": "For a freely expanding specimen under specified loading conditions, the instantaneous linear expansion coefficient is\n\n\\[\\alpha_L=\\frac{1}{L}\\left(\\frac{\\partial L}{\\partial T}\\right)_p,\\]\n\nwhere \\(L\\) is a measured length and \\(T\\) is temperature. The subscript indicates constant pressure; for solids, a fuller specification may require the applied stresses to remain fixed. The coefficient has units of inverse kelvin, \\(\\mathrm{K}^{-1}\\). Its numerical value is the same when a temperature interval is expressed in degrees Celsius, because a Celsius degree and a kelvin have equal size. Absolute temperatures, however, must be expressed on an absolute scale in equations that contain temperature itself rather than a temperature difference.\n\nIf the coefficient is approximately constant and the dimensional change is small, expansion is described by\n\n\\[\\Delta L \\simeq \\alpha_L L_0\\Delta T.\\]\n\nThis familiar engineering expression uses an initial length \\(L_0\\) and a representative coefficient for the interval. For example, a steel member 10 m long, with \\(\\alpha_L=12\\times10^{-6}\\ \\mathrm{K}^{-1}\\), lengthens by about 6 mm during a 50 K temperature increase if it is free to move. Actual steel coefficients depend on composition and temperature. Tabulated values should therefore be understood as properties of particular materials under stated conditions, not universal constants attached to broad material names.\n\nThe volume expansion coefficient is\n\n\\[\\alpha_V=\\frac{1}{V}\\left(\\frac{\\partial V}{\\partial T}\\right)_p.\\]\n\nFor an isotropic material undergoing unconstrained expansion, \\(\\alpha_V=3\\alpha_L\\); correspondingly, the area coefficient is \\(2\\alpha_L\\). These are differential relations. Finite changes calculated from constant-coefficient linear formulas agree only to first order, because products of small strains have been omitted. For temperature-dependent coefficients, the exact length ratio is\n\n\\[\\frac{L(T_2)}{L(T_1)} =\\exp\\!\\left[\\int_{T_1}^{T_2}\\alpha_L(T)\\,dT\\right].\\]\n\nThis distinction matters over wide temperature ranges, especially in cryogenic systems and high-temperature processing.\n\nExpansion also changes density. For a fixed mass in a single phase,\n\n\\[\\alpha_V=-\\frac{1}{\\rho}\\left(\\frac{\\partial\\rho}{\\partial T}\\right)_p.\\]\n\nConsequently, density measurements can determine volume expansion without directly measuring dimensions. Neither linear nor volume coefficients should be extrapolated indiscriminately across a phase transition, where dimensions may change discontinuously or the derivative may become very large. In addition, heating can cause moisture loss, chemical change, permanent deformation, or structural relaxation. Such changes may accompany thermal expansion but are not identical to the reversible dimensional response conventionally represented by an expansion coefficient."
    },
    {
      "heading": "Atomic origins and lattice vibrations",
      "text": "In a solid, neighboring atoms occupy configurations determined by their chemical bonds and surrounding structure. Heating usually increases the energy stored in their motion. Expansion does not simply occur because atoms “vibrate more”: a perfectly harmonic model can produce increasing vibration amplitudes without changing the mean separation. The essential ingredient in ordinary lattice expansion is anharmonicity—the departure of the interaction energy from a purely quadratic dependence on displacement.\n\nA schematic interaction potential near an equilibrium separation can be written\n\n\\[U(x)=U_0+\\frac{1}{2}kx^2+\\frac{1}{3}ax^3+\\frac{1}{4}bx^4+\\cdots.\\]\n\nThe quadratic term gives the harmonic-oscillator approximation. Higher-order terms make the energy landscape asymmetric. For many common interatomic interactions, compression encounters a steeper energy increase than an equivalent extension, so thermal sampling shifts the average separation outward. This is a useful picture, although real solids involve collective motions and many-body interactions rather than independent pairs of atoms connected by identical springs.\n\nIn crystalline solids, collective vibrations are described by phonons. Their frequencies generally depend on crystal volume. This dependence is expressed for a vibrational mode \\(i\\) by its Grüneisen parameter,\n\n\\[\\gamma_i=-\\frac{\\partial\\ln\\omega_i}{\\partial\\ln V},\\]\n\nwhere \\(\\omega_i\\) is the mode frequency. Many modes soften as the crystal expands and therefore have positive \\(\\gamma_i\\). As such modes become thermally populated, their contribution to the free energy favors a larger equilibrium volume. Modes with negative parameters can instead favor contraction. The net expansion depends on the balance among these contributions and on which modes are excited at a given temperature.\n\nA commonly used relation is\n\n\\[\\alpha_V=\\frac{\\bar{\\gamma}C_V}{K_TV},\\]\n\nwhere \\(C_V\\) is the constant-volume heat capacity of the specimen, \\(K_T\\) its isothermal bulk modulus, and \\(\\bar{\\gamma}\\) an appropriate heat-capacity-weighted average. Within the quasiharmonic description, this connects expansion with vibrational thermodynamics and resistance to compression. Strong bonding and a large bulk modulus often accompany small expansion, although structural and electronic effects prevent this from being a universal rule.\n\nQuantum mechanics becomes especially important at low temperature. As progressively fewer vibrational modes are thermally excited, their contribution to expansion generally diminishes. In a simple insulating crystal in the low-temperature Debye regime, the phonon contribution often varies approximately as \\(T^3\\). Metals may additionally have contributions from conduction electrons, while magnetic ordering can alter dimensions through coupling between magnetism and the lattice. Near structural or electronic instabilities, these additional effects can dominate the ordinary vibrational response."
    },
    {
      "heading": "Thermodynamic relations and fluid expansion",
      "text": "Thermal expansion is a response property of an equation of state, which relates pressure, volume, and temperature. It is linked to other measurable properties rather than being independent of them. Defining the isothermal compressibility as\n\n\\[\\kappa_T=-\\frac{1}{V}\\left(\\frac{\\partial V}{\\partial p}\\right)_T,\\]\n\ndifferentiation of the equation of state gives\n\n\\[\\alpha_V=\\kappa_T\\left(\\frac{\\partial p}{\\partial T}\\right)_V.\\]\n\nThus, a substance that expands strongly when heated at constant pressure generally develops a substantial pressure increase when heating occurs at fixed volume, subject to its compressibility.\n\nFor an ideal gas, \\(pV=nRT\\), where \\(n\\) is the amount of substance and \\(R\\) is the gas constant. At fixed pressure and fixed \\(n\\),\n\n\\[\\alpha_V=\\frac{1}{T}.\\]\n\nAt 300 K, this is about \\(3.33\\times10^{-3}\\ \\mathrm{K}^{-1}\\), much larger than typical solid expansion coefficients. The inverse-temperature dependence shows why a single constant coefficient is inappropriate over a broad gas-temperature interval. Real gases approach this relation at sufficiently low density but depart from it when molecular interactions become important, particularly near condensation and the critical point.\n\nExpansion is also connected to entropy. A Maxwell relation derived from the Gibbs free energy gives\n\n\\[\\left(\\frac{\\partial S}{\\partial p}\\right)_T =-\\left(\\frac{\\partial V}{\\partial T}\\right)_p =-\\alpha_VV.\\]\n\nFor a material with positive expansion, increasing pressure at fixed temperature lowers entropy. Another identity is\n\n\\[C_p-C_V=\\frac{TV\\alpha_V^2}{\\kappa_T},\\]\n\nwhere the two heat capacities refer to the same quantity of material. This relation applies to a simple compressible system in an equilibrium, differentiable single-phase region. It explains why constant-pressure and constant-volume heat capacities differ even without a phase change: allowing expansion changes the thermodynamic response to heating.\n\nLiquids generally have larger volume expansion coefficients than dense crystalline solids, although their values vary widely. Their response reflects changes in local molecular arrangement and intermolecular interactions as well as thermal motion. In an open vessel, the observed liquid level depends on both liquid expansion and vessel expansion. For small changes, the apparent volume coefficient is approximately the liquid’s true coefficient minus the vessel’s volume coefficient. A liquid-in-glass thermometer therefore measures differential expansion, not the liquid’s expansion in isolation.\n\nPressure can strongly modify liquid expansivity, and coefficients may change sign. Near a phase boundary, the distinction between continuous expansion and phase conversion is essential: a large volume increase produced by boiling is mainly the replacement of liquid by vapor, not an unusually large ordinary liquid expansion. Measurements and thermodynamic models must identify which phase or mixture of phases is present."
    },
    {
      "heading": "Directional expansion and materials that contract on heating",
      "text": "Thermal expansion is not necessarily the same in every direction. In a crystal, a small unconstrained thermal strain can be represented by a second-rank tensor,\n\n\\[d\\varepsilon_{ij}^{\\mathrm{th}}=\\alpha_{ij}\\,dT.\\]\n\nCrystal symmetry restricts the independent components of \\(\\alpha_{ij}\\). Cubic crystals have isotropic thermal expansion in their ideal, unstressed state, whereas lower-symmetry crystals can have different coefficients along different directions. The volume coefficient is the tensor trace, \\(\\alpha_V=\\alpha_{11}+\\alpha_{22}+\\alpha_{33}\\). A crystal can therefore contract along one direction while still increasing in total volume.\n\nGraphite provides a prominent example. Strong bonding within its carbon sheets and much weaker interactions between sheets produce sharply different expansion parallel and perpendicular to the layers. Sheet vibrations can generate negative in-plane expansion over substantial temperature intervals, while the separation between layers increases. In polycrystalline materials, the macroscopic response depends on how crystallites are oriented, how they constrain one another, and whether pores or microcracks accommodate their differential movement.\n\nNegative thermal expansion can also occur in all directions. Cubic zirconium tungstate, \\(\\mathrm{ZrW_2O_8}\\), is a widely studied example, exhibiting contraction from very low temperatures to roughly 1,050 K, with a structural transition within that range. Its open framework permits collective movements of connected coordination polyhedra that shorten some average separations as thermal motion increases. More generally, transverse vibrations can pull connected structural units closer together. Such behavior demonstrates that thermal expansion depends on geometry and collective motion, not merely on whether individual bonds lengthen.\n\nWater has a different, environmentally important anomaly. At approximately atmospheric pressure, liquid water reaches its maximum density near \\(3.98\\,^\\circ\\mathrm{C}\\). Between that temperature and freezing, cooling makes it expand; equivalently, heating water within this interval makes it contract. The behavior reflects competition between ordinary thermal effects and the organization of its hydrogen-bond network. Freezing produces a further, discontinuous volume increase of about 9% under ordinary conditions. That jump belongs to the liquid–solid transition and should not be confused with the negative expansion coefficient of the liquid below its density maximum.\n\nLow expansion can also arise from competing mechanisms. Invar, conventionally an iron–nickel alloy containing about 36% nickel by mass, has unusually small expansion near room temperature because magnetic and lattice effects partly compensate. Its coefficient depends on composition, treatment, and temperature; “low expansion” does not mean dimensionally invariant under all conditions. Likewise, composites combining positive- and negative-expansion constituents can achieve a small bulk response, but constituent stiffness, interfaces, and internal stresses determine the result rather than a simple arithmetic average alone."
    },
    {
      "heading": "Measurement, historical development, and uncertainty",
      "text": "The expansion of heated substances was central to the development of thermometry. Early thermoscopes showed changes in gas or liquid volume without providing a fully standardized temperature scale. Eighteenth-century liquid-in-glass thermometers turned reproducible expansion into a practical measurement method. Daniel Gabriel Fahrenheit introduced his temperature scale in 1724, and Anders Celsius proposed a centigrade scale in 1742. Modern temperature measurement is not defined by the expansion of a particular substance, but these instruments established the practical importance of relating dimensional change to temperature.\n\nDirect measurement of expansion is called dilatometry. In a push-rod dilatometer, a specimen’s length change is transmitted to a displacement sensor through a rod. The apparatus must distinguish specimen expansion from expansion of the rod, holder, and measuring frame. Reference materials and calibration procedures are therefore essential. Contact forces must be controlled, especially for soft specimens or materials that creep at elevated temperature. Heating and cooling measurements can reveal whether the response is reversible or accompanied by permanent changes.\n\nOptical methods reduce some mechanical-contact effects. Interferometry determines displacement through changes in optical phase or fringe position, relating a length change to a known light wavelength. It can resolve changes far smaller than a micrometre under suitable conditions. Capacitive dilatometers instead infer displacement from a change in capacitance and are particularly valuable in low-temperature and high-sensitivity studies. Neither method is automatically free of systematic error: alignment, support deformation, electric fields, thermal gradients, and the measuring environment can all influence results.\n\nX-ray diffraction and neutron diffraction determine expansion through changes in crystal lattice parameters. For X-rays, changes in diffraction angles are interpreted using Bragg’s law. These methods distinguish crystallographic directions and can follow several phases within one specimen. Lattice expansion need not exactly match the expansion of an entire porous or microcracked body, because bulk dimensions also reflect microstructure. Comparing diffraction with macroscopic dilatometry can therefore reveal internal accommodation mechanisms.\n\nThe measurement uncertainty of an expansion coefficient includes uncertainties in temperature, temperature uniformity, initial dimensions, displacement, and instrument correction. Numerical differentiation of measured length data can amplify noise, so investigators often fit a smooth function or report an average coefficient,\n\n\\[\\bar{\\alpha}_L=\\frac{L(T_2)-L(T_1)}{L(T_1)(T_2-T_1)}.\\]\n\nThis interval coefficient is not necessarily the instantaneous value at the midpoint. High-quality reporting specifies the temperature interval, specimen composition and orientation, thermal history, applied load, atmosphere, and whether measurements were taken during heating or cooling. Precision dimensional metrology commonly references dimensions to \\(20\\,^\\circ\\mathrm{C}\\), making reliable expansion corrections important whenever measurements occur at another temperature."
    },
    {
      "heading": "Restraint, internal stress, and mechanical failure",
      "text": "A freely heated object may change dimensions without developing substantial stress. Restraint changes the outcome. If a uniform bar is prevented from lengthening, its tendency to expand is opposed by compressive elastic strain. Under ideal uniaxial conditions, the resulting thermal stress is approximately\n\n\\[\\sigma=-E\\alpha_L\\Delta T,\\]\n\nwhere \\(E\\) is Young’s modulus and tensile stress is taken as positive. Cooling under the same restraint produces tension. This formula assumes uniform temperature, constant properties, small strains, and linear elasticity; different constraints require different equations.\n\nThe magnitude can be substantial. Taking \\(E=200\\ \\mathrm{GPa}\\), \\(\\alpha_L=12\\times10^{-6}\\ \\mathrm{K}^{-1}\\), and \\(\\Delta T=50\\ \\mathrm{K}\\) gives a compressive stress of about 120 MPa for a fully restrained steel bar. In a real structure, supports may move, joints may slip, and yielding or buckling may limit the stress. Nevertheless, the estimate explains why a seemingly modest temperature change can have major mechanical consequences. On a railway, continuously welded rails are managed through specified stress-free temperatures, fastening systems, and track resistance rather than by placing an expansion gap at every rail joint.\n\nNonuniform temperature can generate stress even without external restraint. When the surface of a hot body cools rapidly, it attempts to contract while the warmer interior resists. The resulting tensile surface stresses can initiate cracks in brittle materials. This is a principal mechanism of thermal shock in glass and ceramics. Resistance depends on expansion, stiffness, strength, fracture toughness, geometry, and thermal conductivity, which controls the development and relaxation of temperature gradients. A low expansion coefficient is helpful but is not by itself a complete measure of thermal-shock resistance.\n\nBonded materials experience mismatch stress when their expansion coefficients differ. In the simplest approximation, their unconstrained differential strain is\n\n\\[\\Delta\\varepsilon=(\\alpha_1-\\alpha_2)\\Delta T.\\]\n\nBonding forces the components toward compatible deformations, producing internal stresses and often curvature. Thin films on thick substrates are commonly constrained in two in-plane directions, so their stress depends on a biaxial modulus rather than simply on Young’s modulus. The temperature at which a bond or film becomes mechanically fixed also establishes the reference state for subsequent mismatch.\n\nRepeated temperature cycles can produce fatigue, interface separation, and crack growth. At high temperatures, creep may relax stresses during a dwell period and redistribute them during later cooling. Polymers and glasses may also undergo time-dependent relaxation, particularly near a glass transition. Mechanical analysis therefore often requires temperature-dependent material properties and loading history, not just a single expansion coefficient and the largest temperature difference."
    },
    {
      "heading": "Devices, manufacturing, and dimensional stability",
      "text": "Thermal expansion can be used to create controlled motion. A bimetallic strip consists of two bonded metals with different expansion coefficients. Heating makes one layer seek a greater length than the other, so the strip bends. Its curvature depends on the coefficient difference, temperature change, layer thicknesses, and elastic properties. Bimetallic elements have been used in thermostats, dial thermometers, and thermal circuit breakers. Their operation illustrates how differential expansion can amplify small material strains into a useful displacement.\n\nManufacturing processes exploit the same effect. In a shrink fit, an outer component is heated, an inner component is cooled, or both, to create temporary clearance. As the components return toward a common temperature, interference develops and produces contact pressure. Bearings, collars, and other assemblies can be joined in this way. Conversely, thermal expansion complicates precision machining: a warm workpiece may be cut to a dimension that becomes incorrect after cooling, while heat generated in the machine can shift the tool position. Temperature control and expansion compensation address different parts of this dimensional problem.\n\nMaterial selection is especially important in optics and electronics. Borosilicate glass has lower expansion than common soda–lime glass, and fused silica is lower still, with a room-temperature linear coefficient of roughly \\(0.5\\times10^{-6}\\ \\mathrm{K}^{-1}\\). Certain glass-ceramics achieve near-zero expansion over selected ranges through compensating phase contributions. Such materials are used for telescope mirrors and precision reference structures. Low bulk expansion reduces dimensional change, but spatial temperature gradients can still alter shape and optical performance.\n\nElectronic assemblies contain materials with markedly different coefficients. Near room temperature, silicon has a linear coefficient of about \\(2.6\\times10^{-6}\\ \\mathrm{K}^{-1}\\), compared with roughly \\(17\\times10^{-6}\\ \\mathrm{K}^{-1}\\) for copper and \\(23\\times10^{-6}\\ \\mathrm{K}^{-1}\\) for aluminum. Repeated heating during operation therefore strains solder joints and interfaces. Package geometry, compliant connections, substrate selection, and heat-flow paths influence reliability. The relevant material property is not always the smallest possible coefficient, but a compatible combination of expansion, stiffness, conductivity, and processing temperature.\n\nExpansion also affects resonators, sensors, and devices built for astronomy. A resonator’s frequency may change because its dimensions, elastic constants, and internal stress all vary with temperature. An optical instrument can drift through both geometric expansion and changes in refractive index. Athermal design arranges compensating effects so that a selected output remains comparatively stable. Such compensation is necessarily tied to a temperature range and operating condition; cancellation at one temperature does not guarantee cancellation over an arbitrarily wide interval."
    },
    {
      "heading": "Oceans, buoyancy, and geological processes",
      "text": "Thermal expansion has major consequences for the ocean. Warming seawater generally increases its specific volume and lowers its density, producing thermosteric sea-level rise even if ocean mass does not change. This is distinct from the addition of water through melting land ice. The expansion depends on the initial temperature, salinity, and pressure, so equal amounts of added heat do not produce equal sea-level changes everywhere. Warm surface water generally has a larger expansion coefficient than cold surface water, while pressure modifies the response at depth.\n\nFor a simplified vertical column, the thermosteric contribution can be approximated by\n\n\\[\\Delta h \\simeq \\int_{-H}^{0}\\alpha_V(z)\\,\\Delta T(z)\\,dz,\\]\n\nwhere \\(H\\) is the column depth and the coefficient is evaluated for the relevant seawater state. This expression isolates temperature-driven volume change and assumes a suitable reference configuration. Full calculations use seawater thermodynamics and must account for salinity and pressure. Ocean observations show that thermal expansion is a major component of contemporary global mean sea-level rise, alongside changes in land-ice and terrestrial-water storage.\n\nDensity differences caused by temperature also contribute to buoyancy. In many fluid-flow models, they are represented by\n\n\\[\\rho\\simeq\\rho_0[1-\\alpha_V(T-T_0)].\\]\n\nCombined with gravity, this relation produces the familiar tendency for warmer fluid to rise. It underlies a common approximation in which density variations are retained in buoyancy forces but neglected elsewhere. Its validity is limited when density changes are large, compressibility is important, or the expansion coefficient varies strongly. Salinity can reinforce or oppose thermal buoyancy in seawater, while the density maximum of freshwater changes the seasonal circulation of lakes.\n\nIn Earth’s interior, thermal expansion contributes to mantle buoyancy and the development of convection associated with plate tectonics. Hot mantle material is generally less dense than compositionally similar cold material at the same pressure. However, density also depends on mineral composition, pressure, and phase transformations. Cooling oceanic lithosphere becomes denser and tends to subside as it moves away from spreading ridges. Models of seafloor depth and lithospheric cooling therefore incorporate thermal contraction together with heat transport and mechanical structure.\n\nAt the surface, repeated heating and cooling can generate differential stresses in rocks because adjacent minerals expand differently and exposed surfaces develop temperature gradients. These stresses can contribute to cracking and physical weathering, particularly when they act alongside moisture, salt growth, or pre-existing fractures. At greater depths, expansion of heated pore fluids can increase pressure where drainage is restricted. Whether a heated rock mainly expands, fractures, or develops elevated fluid pressure depends on its permeability, confinement, mineral structure, and the relative rates of heating and fluid escape."
    }
  ],
  "links": [
    "temperature",
    "pressure",
    "kelvin",
    "density",
    "mass",
    "phase-transition",
    "chemical-bond",
    "quantum-harmonic-oscillator",
    "phonon",
    "quantum-mechanics",
    "electron",
    "equation-of-state",
    "ideal-gas",
    "gas-constant",
    "critical-point",
    "entropy",
    "gibbs-free-energy",
    "heat-capacity",
    "graphite",
    "water",
    "hydrogen-bond",
    "calibration",
    "interferometry",
    "capacitance",
    "braggs-law",
    "measurement-uncertainty",
    "metrology",
    "railway",
    "thermal-conductivity",
    "glass-transition",
    "silicon",
    "astronomy",
    "refractive-index",
    "ocean",
    "sea-level-rise",
    "heat",
    "gravity",
    "plate-tectonics"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}