{
  "id": "ferroelectricity",
  "title": "Ferroelectricity",
  "lang": "en",
  "source_lang": "en",
  "summary": "Ferroelectricity is the existence of spontaneous electric polarization that can be reversed by an applied electric field, enabling distinctive dielectric, electromechanical, and memory functions.",
  "category": "science",
  "subcategory": "physics",
  "type": "concept",
  "aliases": [
    "Ferroelectric behavior",
    "Ferroelectric order"
  ],
  "generated_at": "2026-11-05",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/ferroelectricity/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/ferroelectricity/",
      "json": "https://aiwiki.page/zh/ferroelectricity.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "Ferroelectricity is a property of certain materials that possess a spontaneous electric polarization whose direction can be reversed by an applied electric field. Polarization describes the electric dipole moment per unit volume. In a ferroelectric phase, the arrangement of atoms, ions, or molecular units supports a polar state even without an external field. Reversibility distinguishes ferroelectricity from polarity alone: a material may have a permanently polar crystal structure yet not permit practical electrical switching before it breaks down.\n\nThe name reflects an analogy with ferromagnetism, not the presence of iron. Ferroelectric materials can form domains, exhibit hysteresis, and undergo transitions between ordered and disordered phases. Their coupled electrical, mechanical, thermal, and optical responses underpin capacitors, ultrasound transducers, actuators, infrared detectors, optical modulators, and nonvolatile memories. Understanding these properties requires connecting atomic-scale distortions to collective ordering, domain motion, interfaces, and the conditions under which polarization is measured."
    },
    {
      "heading": "Polar symmetry and the meaning of switchable polarization",
      "text": "In a simple microscopic picture, polarization arises when the centers of positive and negative electric charge do not coincide. A crystal can therefore carry an electric dipole moment in each repeating unit cell. The bulk polarization is a vector, conventionally expressed in coulombs per square metre. A value of \\(0.1\\ \\mathrm{C\\,m^{-2}}\\), for example, corresponds to \\(10\\ \\mathrm{\\mu C\\,cm^{-2}}\\), a unit frequently used in experimental reports.\n\nCrystal symmetry constrains whether a spontaneous polarization is possible. Of the 32 crystallographic point groups, ten are polar: they permit a vector that remains unchanged by every symmetry operation of the group. A ferroelectric phase belongs to one of these polar classes, but membership does not establish ferroelectricity. Electrical switching between distinct polar states must also be possible. Wurtzite zinc oxide, for example, is polar and piezoelectric, but ordinary undoped material is not conventionally classified as a readily switchable ferroelectric.\n\nFerroelectricity is related to, but narrower than, pyroelectricity and piezoelectricity. A pyroelectric has a spontaneous polarization that varies with temperature, producing a measurable electrical response when its temperature changes. A piezoelectric couples mechanical stress and electrical polarization. Ferroelectric phases are pyroelectric and piezoelectric by symmetry, although measured responses can cancel in a specimen containing differently oriented domains. Conversely, most piezoelectrics are not ferroelectric; quartz is a familiar example.\n\nA specimen need not display a net dipole moment to be ferroelectric. Oppositely polarized regions can cancel macroscopically while retaining local order. Applying a sufficiently strong field can favor one orientation and produce a poled state. The polarization remaining after the field is removed is called the remanent polarization, \\(P_r\\). It is not necessarily identical to the spontaneous polarization, \\(P_s\\), because domain populations, incomplete switching, and interfacial effects influence the measured remanence.\n\nFor an infinite periodic solid, absolute polarization cannot be defined simply by assigning charges to a particular unit cell. The modern quantum theory expresses electronic polarization through a Berry phase of occupied electronic states and defines it modulo a polarization quantum. Observable changes in polarization along a specified structural or switching path remain well defined. This framework, developed in the early 1990s, made polarization calculable from electronic structure without treating a crystal as a collection of isolated molecular dipoles."
    },
    {
      "heading": "Discovery and development of the field",
      "text": "The first generally recognized demonstration of ferroelectricity was Joseph Valasek’s work on Rochelle salt, potassium sodium tartrate tetrahydrate, reported in 1921. Valasek observed a hysteretic relation between electrical polarization and applied field, drawing an analogy with magnetic hysteresis. Rochelle salt has an unusual temperature interval of ferroelectric order, approximately between \\(-18\\ ^\\circ\\mathrm{C}\\) and \\(24\\ ^\\circ\\mathrm{C}\\). Its sensitivity to temperature, humidity, and mechanical conditions initially made the phenomenon appear characteristic of a rather exceptional material.\n\nThe identification of ferroelectricity in potassium dihydrogen phosphate, commonly called KDP, in 1935 provided a chemically different example. KDP becomes ferroelectric near 123 K and helped establish the importance of hydrogen bonding and configurational ordering. Replacing hydrogen with deuterium shifts its transition temperature substantially, demonstrating that isotope substitution and nuclear motion can strongly influence collective electrical order. Such findings showed that ferroelectricity was not restricted to Rochelle salt or to one structural mechanism.\n\nBarium titanate transformed the field during the 1940s. Investigations in several countries established its exceptionally large dielectric response and ferroelectric behavior. Unlike the early water-soluble crystals, it could be fabricated as a robust oxide ceramic. Its discovery linked ferroelectricity to technologically useful dielectrics and opened a large family of related compounds with the perovskite structure. Ceramic processing also made it possible to obtain useful macroscopic properties from polycrystalline materials rather than relying exclusively on large single crystals.\n\nTheoretical developments accompanied this expansion. In 1949, William Cochran and Philip Anderson independently articulated ideas connecting ferroelectric transitions with unstable lattice vibrations; the soft-mode picture was developed further during the following decade. Alexander Devonshire applied a phenomenological free-energy expansion to barium titanate, accounting for its sequence of structural transitions. Studies of domains, elastic coupling, and switching subsequently connected these equilibrium descriptions to the hysteretic behavior of real samples.\n\nLater discoveries broadened the range of material platforms. Ferroelectricity was reported in polyvinylidene fluoride in 1971, following the discovery of strong piezoelectricity in the polymer in 1969. Thin-film oxide research expanded alongside semiconductor fabrication and scanning-probe methods. In 2011, ferroelectric behavior was reported in silicon-doped hafnium oxide films, establishing a major route toward integration with established microelectronics. The historical progression thus moved from specialized crystals to ceramics, polymers, epitaxial heterostructures, and ultrathin films, each with different mechanisms and processing constraints."
    },
    {
      "heading": "Microscopic mechanisms and material families",
      "text": "Many widely used ferroelectrics are oxide perovskites with approximate composition \\(ABO_3\\). In barium titanate, BaTiO\\(_3\\), the high-temperature average structure is cubic, with titanium surrounded by an oxygen octahedron. On entering a polar phase, relative displacements of titanium, oxygen, and barium remove inversion symmetry. At room temperature, bulk barium titanate is normally tetragonal, with spontaneous polarization along one of the equivalent parent-cubic axes. Its transition to the cubic paraelectric phase occurs near \\(120\\ ^\\circ\\mathrm{C}\\), with the precise temperature depending on composition, defects, and specimen conditions.\n\nThese displacements are not adequately understood as rigid charged spheres moving independently. Chemical bonding, particularly hybridization between oxygen and transition-metal electronic states, helps stabilize off-centered configurations. Effective charges describing polarization changes can differ substantially from nominal ionic charges. Lead titanate, lead zirconate titanate, and potassium sodium niobate illustrate how related structures support different polarization magnitudes, transition temperatures, and electromechanical responses.\n\nTwo traditional limiting descriptions are displacive and order–disorder ferroelectricity. In a displacive transition, a collective lattice distortion develops from a comparatively symmetric reference structure. In an order–disorder transition, locally polar units exist above the transition but acquire long-range orientational order on cooling. The distinction is useful but not absolute: a material can exhibit local off-centering together with a soft collective vibration. Experimental observations therefore need not fit exclusively into one category.\n\nHydrogen-bonded crystals provide an important alternative mechanism. In KDP, proton configurations along hydrogen bonds couple to distortions of the phosphate groups and the surrounding lattice. Organic and molecular ferroelectrics can instead involve molecular reorientation, proton transfer, or collective charge redistribution. In ferroelectric phases of polyvinylidene fluoride and its copolymers, aligned chain conformations organize molecular dipoles into switchable polar regions. Their flexibility and processing routes differ markedly from those of dense oxide ceramics.\n\nNot all ferroelectric polarization is the primary structural instability. In an improper ferroelectric, another ordered distortion induces polarization through coupling in the free energy. Hexagonal manganites such as YMnO\\(_3\\) are examples in which a nonpolar structural distortion is central. In some magnetic compounds, a particular spin arrangement induces a small polarization, connecting ferroelectricity with multiferroics. Such materials are valuable for studying coupled orders, although coexistence of magnetic and polar order does not automatically imply a large electrically controllable magnetic response.\n\nFluorite-derived hafnium oxide represents another structural family. Ferroelectricity is commonly associated with a metastable polar orthorhombic phase, often assigned space group \\(Pca2_1\\). Dopants, zirconium alloying, film thickness, grain structure, mechanical constraint, and electrode processing affect the competition among polar and nonpolar phases. Its behavior cannot simply be transferred from the familiar perovskite model."
    },
    {
      "heading": "Phase transitions and thermodynamic descriptions",
      "text": "Ferroelectric order often appears through a phase transition from a higher-symmetry paraelectric phase. Polarization serves as an order parameter: it distinguishes symmetry-related ordered states and commonly becomes zero in the high-temperature phase. Some compounds have several ferroelectric phases rather than a single polar state. On cooling, bulk barium titanate passes from cubic to tetragonal, then orthorhombic, and finally rhombohedral structures, with corresponding changes in the allowed polarization directions.\n\nA common phenomenological treatment expands an appropriate free-energy density in powers of a scalar polarization component:\n\n\\[f(P,T,E)=f_0+\\frac{1}{2}\\alpha(T)P^2 +\\frac{1}{4}\\beta P^4+\\frac{1}{6}\\gamma P^6-EP.\\]\n\nThis expression describes a simplified uniaxial case under specified mechanical and electrical conditions. The absence of odd powers at zero field reflects the equivalence of \\(P\\) and \\(-P\\) in a centrosymmetric reference phase. The term \\(-EP\\) favors polarization aligned with the applied field. Equilibrium states satisfy \\(\\partial f/\\partial P=0\\), with local stability requiring positive curvature.\n\nIf \\(\\alpha=a(T-T_0)\\), \\(a>0\\), and \\(\\beta>0\\), the simplest mean-field model produces a continuous transition. Below \\(T_0\\), two symmetry-related minima appear. If \\(\\beta<0\\) and \\(\\gamma>0\\), the model can instead produce a first-order transition, with discontinuous polarization and a range of metastability. In actual crystals, polarization has several components, and elastic strain introduces additional coupling terms. Mechanical clamping can therefore change both the transition and the dielectric response.\n\nAbove many ferroelectric transitions, the relative permittivity approximately follows a Curie–Weiss relation,\n\n\\[\\varepsilon_r \\simeq \\varepsilon_b+\\frac{C}{T-T_0},\\]\n\nover a suitable temperature interval. Here \\(C\\) is a Curie constant and \\(\\varepsilon_b\\) represents a background contribution. The extrapolated \\(T_0\\) need not equal the observed transition temperature, especially for a first-order transition. A large permittivity reflects a strong response to an applied field, not necessarily a large spontaneous polarization.\n\nMicroscopically, a displacive instability can appear as a soft phonon whose frequency decreases as the transition is approached. Because its restoring force becomes small, a modest field can induce a large distortion. Disorder, damping, and coupling to other modes complicate this idealized picture. In relaxor ferroelectrics, the dielectric maximum is broad and depends strongly on measurement frequency; chemical disorder and correlated local polar structures prevent a simple description by a uniform order parameter.\n\nAt sufficiently low temperatures, quantum fluctuations can suppress ordering that a classical treatment would predict. Strontium titanate is a standard quantum-paraelectric example under ordinary bulk conditions: its dielectric response grows strongly on cooling, but it does not develop the expected conventional ferroelectric phase. Isotope substitution or strain can change this balance, illustrating the role of zero-point motion in structural order."
    },
    {
      "heading": "Domains, switching, and hysteresis",
      "text": "A uniformly polarized crystal creates bound charge wherever polarization has a component normal to a surface. Unless compensating charges are available, the resulting depolarizing field opposes the polarization. Forming domains can reduce this electrostatic cost, although creating domain walls consumes energy. Domain patterns consequently reflect a competition among electrostatic energy, wall energy, elastic compatibility, defects, and boundary conditions.\n\nA 180-degree domain wall separates regions whose polarizations are antiparallel. Non-180-degree walls, such as 90-degree walls in tetragonal perovskites, also separate differently oriented spontaneous strains. These are ferroelastic as well as ferroelectric variants. Reorienting them can produce substantial dimensional changes, explaining why domain engineering is important in actuators and piezoelectric ceramics. In a simplified stripe-domain model, balancing wall and stray-field energies often produces a domain width proportional to the square root of specimen thickness, although real structures can depart substantially from this scaling.\n\nElectrical reversal usually proceeds through nucleation and domain-wall motion rather than simultaneous reversal of every unit cell. A reversed nucleus forms at an electrode, defect, grain boundary, or other favorable site, then grows under the field. The measured coercive field, \\(E_c\\), is therefore a kinetic and microstructural quantity as well as a material characteristic. It depends on pulse duration, temperature, film thickness, defect population, and the history of the specimen.\n\nThe polarization–field hysteresis loop records this path dependence. Starting from a poled state, an opposing field reduces the net polarization, drives it through zero, and eventually establishes the opposite state. The field at the zero crossing is commonly called the coercive field, while the zero-field intercept gives the remanent polarization. The loop area,\n\n\\[w_{\\mathrm{loss}}=\\oint E\\,dP,\\]\n\nrepresents electrical energy dissipated per unit volume over a cycle when the loop describes the relevant polarization response. Leakage and other irreversible electrical processes can add to measured losses.\n\nDefects alter switching in several ways. Charged vacancies and defect complexes can pin walls, generate internal bias, or redistribute under repeated cycling. These processes contribute to aging, imprint, and fatigue. Imprint favors one polarization state; fatigue reduces switchable polarization with cycling. Some hafnium-oxide films instead show an initial increase in switchable polarization, called wake-up, associated with evolving defect distributions and phase or domain configurations. Its microscopic origin depends on the film and processing route.\n\nDomain walls are themselves distinct nanoscale environments. Their local symmetry, strain, and charge can differ from those of the domains they separate. Certain walls show enhanced conductivity or other electronic responses. Such behavior is not universal: wall orientation, compensation of bound charge, chemical defects, and the host material determine whether a wall is conducting, insulating, mobile, or strongly pinned."
    },
    {
      "heading": "Experimental identification and measurement",
      "text": "A central experimental task is to demonstrate reversible spontaneous polarization rather than merely a nonlinear electrical response. In a capacitor geometry, the measured current contains several contributions. A simplified expression is\n\n\\[I=A\\frac{dP}{dt}+C_{\\mathrm{lin}}\\frac{dV}{dt}+I_{\\mathrm{leak}},\\]\n\nwhere \\(A\\) is electrode area, \\(C_{\\mathrm{lin}}\\) represents a chosen linear background contribution, and \\(I_{\\mathrm{leak}}\\) is leakage current. Interfacial charging, ionic motion, and instrumentation effects may add further terms. Integrating the current without separating these contributions can produce a misleading apparent polarization.\n\nThe Sawyer–Tower circuit, introduced in 1930, measures charge by placing a reference capacitor in series with the specimen. Modern ferroelectric testers generally use controlled voltage waveforms and electronic charge integration. Frequency-dependent measurements, switching-current peaks, and carefully chosen pulse sequences help distinguish genuine reversal from conduction. In positive-up-negative-down measurements, usually abbreviated PUND, switching and nonswitching pulses are compared to estimate the reversible charge. This method is useful but remains sensitive to relaxation and to differences in the sample state between pulses.\n\nStructural methods provide complementary evidence. X-ray diffraction and neutron diffraction identify lattice symmetry and atomic displacements, while temperature-dependent measurements can reveal structural transitions. Electron microscopy can resolve polar distortions and domain configurations at small length scales. A polar structure alone is insufficient proof of ferroelectricity, however: structural evidence must be connected to switching between states rather than simply to the absence of inversion symmetry.\n\nPiezoresponse force microscopy uses an electrically driven scanning-probe tip to detect local electromechanical motion. It can image domain contrast and examine nanoscale switching patterns. Yet local hysteresis is not conclusive by itself. Electrostatic forces, surface charging, electrochemical strain, and ionic migration can mimic aspects of ferroelectric behavior. Strong evidence combines reproducible writing and reversal, appropriate controls, and independent structural or electrical measurements.\n\nThermal and spectroscopic probes illuminate different parts of the phenomenon. calorimetry can detect anomalies or latent heat at a transition. Dielectric spectroscopy measures how response and loss vary with frequency and temperature. Raman spectroscopy and infrared measurements track lattice vibrations and symmetry changes, including soft modes. Pyroelectric-current measurements reveal changes in polarization during heating or cooling, but currents from trapped-charge release can complicate interpretation.\n\nReported properties must therefore specify measurement conditions. Coercive field without pulse duration, dielectric constant without frequency, or remanent polarization without a waveform and leakage assessment can be difficult to compare. Electrode area, thickness, crystallographic orientation, mechanical constraint, and temperature all affect the result. These dependencies are intrinsic to the connection between local order and macroscopic measurement, rather than merely experimental inconveniences."
    },
    {
      "heading": "Thin films, interfaces, and reduced dimensions",
      "text": "Reducing a ferroelectric to a thin film changes the balance of energies that stabilizes its polarization. Bound surface charge produces a depolarizing field unless compensated by electrode charge, mobile carriers, adsorbates, or a domain pattern. Real electrodes screen imperfectly, and interfaces can introduce an additional voltage drop. A polarization state stable in a bulk crystal may therefore become unfavorable in an otherwise similar ultrathin film.\n\nThere is no universal minimum thickness for ferroelectricity. Stability depends on composition, orientation, screening, interfaces, temperature, and the polarization pattern permitted by the geometry. Some perovskite heterostructures sustain polar states at thicknesses of only a few unit cells under suitable conditions. Other films lose a measurable switchable state at substantially larger thicknesses because of depolarization, interfacial layers, defects, or leakage. A critical thickness quoted for one device stack is not a general constant of the material.\n\nEpitaxial strain provides a powerful means of controlling order. A film grown coherently on a substrate can be forced into an in-plane lattice spacing different from its unconstrained value. Through polarization–strain coupling, this constraint alters phase stability, transition temperature, and preferred polarization direction. Strain can stabilize polar states in materials that are nonferroelectric in ordinary bulk form. Relaxation through dislocations, however, changes both the mechanical constraint and the defect landscape.\n\nInterfaces also influence band alignment and carrier transport. Polarization can raise or lower an interfacial barrier, accumulate or deplete carriers in a neighboring semiconductor, and alter tunneling through an ultrathin insulating layer. In ferroelectric tunnel junctions, reversal changes the electrical resistance through its effect on the tunnel barrier. The measured resistance contrast can additionally depend on defects, interface chemistry, and ionic redistribution, so separating mechanisms remains important.\n\nArtificial superlattices extend these controls by alternating layers of different oxides. Electrostatic continuity, elastic coupling, and interfacial symmetry can create polar phases unlike those of either constituent alone. Competing energies can stabilize continuously rotating polarization patterns, including vortices and related textures. These structures show that a ferroelectric need not consist only of uniformly polarized domains separated by sharp walls.\n\nLayered materials provide another route to reduced-dimensional order. Ferroelectric switching has been investigated in systems such as copper indium thiophosphate and ultrathin indium selenide, while relative sliding between layers can generate switchable interfacial polarization in suitable van der Waals stacks. In these systems, environmental screening, contact geometry, thickness, and structural stacking are central to interpreting measurements. Their mechanisms differ from those of a conventional three-dimensional oxide film bounded by metal electrodes."
    },
    {
      "heading": "Devices, functional responses, and engineering constraints",
      "text": "Ferroelectric materials are widely used where a large dielectric response is valuable, even if polarization switching is not the intended operation. Barium-titanate-based multilayer ceramic capacitors combine thin dielectric layers with internal electrodes to achieve large capacitance in a small volume. Composition and grain structure are engineered to control temperature dependence and reliability. The useful capacitance can vary with applied direct-current bias, temperature, and aging, reflecting the nonlinear response of the dielectric.\n\nElectromechanical devices exploit coupling between polarization, strain, and domain configuration. Lead zirconate titanate, commonly abbreviated PZT, is used in ultrasound transducers, precision actuators, and sensors. Compositions near its morphotropic phase boundary can show strong responses because competing structural states and domain reorientation allow substantial deformation. Lead-free alternatives include potassium sodium niobate and bismuth sodium titanate systems, but replacing a material requires matching processing, temperature stability, loss, and reliability as well as a single piezoelectric coefficient.\n\nOptical and thermal functions use different aspects of polar order. Lithium niobate supports electro-optic modulation and nonlinear optical frequency conversion; periodically reversed domains enable quasi-phase matching of interacting optical waves. Pyroelectric detectors convert temperature changes into electrical signals, including those produced by absorbed infrared radiation. The electrocaloric effect uses field-induced changes in entropy and temperature, particularly near transitions. Large effects in a material do not by themselves establish the efficiency of a complete cooling device, which also depends on heat transfer and electrical losses.\n\nFerroelectric memories encode information in polarization states. In a capacitor-based memory cell, a read pulse distinguishes switching from nonswitching charge; the read operation is commonly destructive and requires restoration of the stored state. A ferroelectric field-effect transistor instead uses polarization to shift a transistor’s threshold voltage, allowing the state to be sensed through channel conduction. Hafnium-oxide-based ferroelectrics are particularly relevant because related oxides and deposition processes are already established in semiconductor manufacturing.\n\nDevice reliability involves several distinct limitations. Retention concerns preservation of a state over time; endurance concerns survival through repeated switching; imprint concerns a bias toward one state. Variability arises from grain structure, phase distribution, defects, and interfaces, especially as device area decreases. A film with a large polarization may still be unsuitable if it requires excessive voltage, leaks strongly, breaks down readily, or loses its state under operating conditions.\n\nOther applications seek tunable capacitance, multilevel conductance, or reduced switching energy. Negative-capacitance transistor concepts exploit the free-energy landscape associated with a ferroelectric instability, with behavior depending critically on stabilization, dynamics, and the surrounding circuit. Analog and neuromorphic devices use partial domain switching to produce intermediate electrical states, but reproducibility and separation from charge trapping or ion migration remain central engineering issues. For pulse-power capacitors, ordinary broad ferroelectric hysteresis is often a disadvantage: slim-loop relaxor or antiferroelectric compositions can recover more of the stored energy with lower dissipation."
    }
  ],
  "links": [
    "polarization",
    "electric-field",
    "ferromagnetism",
    "electric-charge",
    "electric-dipole",
    "unit-cell",
    "piezoelectricity",
    "temperature",
    "electronic-band-structure",
    "dielectric",
    "chemical-bond",
    "hydrogen-bond",
    "phase-transition",
    "order-parameter",
    "permittivity",
    "phonon",
    "zero-point-energy",
    "electric-current",
    "calorimetry",
    "raman-spectroscopy",
    "semiconductor",
    "capacitance"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}