Electric polarization, usually denoted by , is the electric dipole moment per unit volume of a material. It describes the separation of positive and negative electric charge within matter, whether that separation is induced by an applied electric field or exists spontaneously. Polarization connects microscopic charge distributions to the macroscopic behavior of insulators, capacitors, optical materials, sensors, and electrically switchable crystals.
In classical descriptions, polarization arises through the distortion of electron clouds, the relative displacement of ions, the orientation of polar molecules, or charge accumulation at internal interfaces. Its spatial variation produces bound charge, and its time variation contributes to electric current. In crystalline solids, quantum mechanics adds an important qualification: bulk polarization cannot generally be obtained simply by assigning a dipole moment to an arbitrarily chosen unit cell. Instead, measurable polarization changes are connected to charge transport and to the geometric properties of electronic wave functions.
Electric polarization is distinct from the polarization of electromagnetic waves, which describes the orientation and evolution of a wave’s electric-field vector. The two concepts interact in optics because an electromagnetic wave drives polarization in matter, and the resulting material response determines propagation, absorption, and many nonlinear optical effects.
Definition, units, and microscopic meaning
For a localized, electrically neutral collection of charges, the electric dipole moment is
where is a charge and its position. Because the total charge is zero, translating the coordinate origin does not change . For a continuous charge distribution, the corresponding expression is . The dipole vector points from negative toward positive charge. Two equal and opposite charges of magnitude , separated by a displacement directed from the negative charge to the positive one, therefore have .
In a material containing identifiable microscopic dipoles, the polarization can be expressed as a local average,
The averaging volume must contain enough microscopic constituents to smooth atomic fluctuations while remaining small relative to the scale over which the macroscopic fields vary. This separation of scales is essential: the microscopic electric field changes sharply near nuclei and electrons, whereas continuum electrodynamics describes fields averaged over many atomic spacings. For a uniformly polarized specimen of volume , its total dipole moment is , provided the specimen’s full charge distribution and surfaces are included consistently.
In the International System of Units, dipole moment is measured in coulomb metres and polarization in coulombs per square metre:
Although this is also the unit of surface charge density, polarization is a vector field defined throughout a material, not merely a charge density confined to its surface. Its dimensions anticipate an important physical relation: the component of polarization normal to a boundary determines a bound surface charge. A polarization of , for example, corresponds to , a unit commonly used for ferroelectric measurements.
The simple sum over permanent molecular dipoles is not a universal microscopic definition. An initially nonpolar atom can acquire an induced dipole because its negatively charged electron distribution shifts relative to its positively charged nucleus. In a solid, electronic charge may extend over several atoms, making any division into atomic dipoles partly conventional. The directly observable consequences—surface charge under specified boundary conditions, current during a change of state, and electrical response—are therefore more fundamental than a particular partition into microscopic dipoles.
Bound charge, displacement field, and electrical boundaries
The macroscopic charge associated with nonuniform polarization is called bound charge. In a continuum description, its volume density is
At a boundary between a polarized material and vacuum, the bound surface charge density is
where is the outward normal. Thus, a uniformly polarized block has no bound charge in its interior but can carry opposite bound charges on opposite faces. “Bound” does not mean immobile under all circumstances; it identifies charge represented through polarization rather than through a separately specified free-charge density.
This division leads to the electric displacement field,
where is the vacuum permittivity. Substituting the bound-charge expression into Gauss’s law gives
At an interface, with the normal directed from medium 1 to medium 2,
This formulation is particularly useful when the free charge on an electrode is known. It does not make an independently specified material response: predicting the fields still requires a relation between , , and the material’s state.
Bound surface charges themselves generate a field, often called the depolarization field, that opposes the polarization producing them. For an ideal uniformly polarized slab whose polarization is perpendicular to its faces, with no compensating free charge and no externally imposed field, the interior field is . For a uniformly polarized sphere under the corresponding ideal conditions, it is . These expressions describe the field produced by a specified polarization; a self-consistent material calculation must also determine how that field changes the polarization.
Electrical boundary conditions consequently affect the stable state of a specimen. Metallic electrodes can supply compensating charge, reducing the depolarization field. An unelectroded polar crystal may instead form domains, adsorb charged species, or redistribute mobile carriers. For a large parallel-plate capacitor with negligible fringing, the free electrode charge per area equals the normal component of . This distinction between electrode charge and polarization matters in precision measurements: the electrode also supplies the vacuum-field contribution . The division into free and bound charge must furthermore be chosen consistently, especially in materials containing mobile ions or poorly conducting interfaces.
Microscopic mechanisms of polarization
Electronic polarization occurs when the electronic charge distribution shifts or distorts relative to atomic nuclei. It exists in essentially all materials. In a weak local field, an isolated atom or molecule can often be described by
where is its polarizability tensor and is the field acting on it. Electronic response is generally fast because electrons are light, although its frequency dependence becomes pronounced near electronic excitations. The magnitude depends on how tightly the electrons are bound and on the spatial extent and symmetry of their distribution.
Ionic polarization results from relative displacement of oppositely charged sublattices or bonded ions. In sodium chloride, an electric field shifts the sodium and chloride sublattices in opposite directions. In oxide crystals such as barium titanate, relative motion of titanium, oxygen, and barium contributes to a much larger and more collective response. The relevant motions are linked to optical phonons, especially infrared-active modes. Ionic response is normally slower than purely electronic response because nuclei have much greater mass. Its strength is determined not only by nominal ionic charges but also by changes in chemical bonding as atoms move.
Orientational polarization occurs when molecules with permanent dipole moments preferentially align with an applied field. Thermal agitation opposes this ordering. For independent classical dipoles of magnitude , number density , and temperature , the idealized equilibrium result is
Here is the Boltzmann constant, and the field is the field acting on the dipoles. In the weak-field limit,
The inverse-temperature dependence illustrates the competition between alignment energy and disorder. Real polar liquids need more elaborate descriptions because local fields, molecular correlations, and interactions such as hydrogen bonding invalidate the independent-dipole approximation.
Interfacial polarization arises when charge moves over distances larger than an atom or molecule but accumulates at barriers. It is common in composites, polycrystalline ceramics, emulsions, and materials containing layers with different conductivities. The Maxwell–Wagner effect describes this type of charge accumulation in electrically heterogeneous media. Related low-frequency signals can originate at sample–electrode interfaces, particularly when electrodes block mobile ions. These responses can produce very large apparent permittivities without implying exceptionally large intrinsic atomic polarizability. The mechanisms are distinguished most effectively by their time scales, temperature dependence, microstructure dependence, and sensitivity to electrodes rather than by a single static measurement.
Linear response, susceptibility, and local fields
For a linear, homogeneous, isotropic material with no spontaneous polarization, the constitutive relation is
where is the dimensionless electric susceptibility. Combining this with the definition of gives
These equations are approximations valid over a specified range of field, frequency, temperature, and mechanical conditions. A material can be linear for a small alternating perturbation while being strongly nonlinear over a larger field excursion. In a material with spontaneous polarization, linear response is usually written for small changes about a chosen state, rather than by assuming that total polarization vanishes at zero field.
In anisotropic matter, polarization need not be parallel to the field. The relation becomes
The susceptibility tensor reflects crystal symmetry and experimental conditions. Its principal values may differ along different crystallographic axes, contributing to direction-dependent optical and electrical behavior. For reciprocal, nonmagnetic materials in electrostatic equilibrium, thermodynamic response generally yields a symmetric static susceptibility tensor. At finite frequency, components may be complex because the response includes both phase delay and dissipation. If polarization at one point depends on fields in neighboring regions, the response is also spatially nonlocal and cannot be represented fully by a position-independent tensor.
A central microscopic issue is that the local field acting on an atom is not generally equal to the macroscopic field. In the idealized Lorentz cavity construction for an isotropic medium, the local field is
provided additional near-neighbor contributions vanish or are incorporated appropriately. Combining this expression with produces the Clausius–Mossotti relation,
This links molecular or atomic polarizability to bulk permittivity, but it is not universal. Strong correlations, anisotropy, complex bonding, and proximity to an instability can make its assumptions inadequate.
The same reasoning connects polarization to refractive index. In a transparent, nonmagnetic medium, at the optical frequency concerned, leading to the Lorentz–Lorenz relation when combined with the local-field model. Static and optical permittivities must not be interchanged: slow molecular rotations and many ionic motions cannot follow visible-light oscillations. Liquid water, for example, has a low-frequency relative permittivity near 78 at room temperature, whereas its visible refractive index is about 1.33, giving an optical near 1.77. The difference primarily reflects response mechanisms available at low frequency but not at optical frequency.
Time dependence, dielectric relaxation, and absorption
A changing polarization contributes a polarization-current density,
Within Maxwell’s equations, this combines with the vacuum contribution to give
Charge carriers need not traverse the entire specimen for this current to be detected in an external circuit. Small internal displacements of bound charges can change the charge required on capacitor electrodes. This is why integrating a properly separated transient current can measure a change in polarization even in an insulating specimen.
A causal linear material responds to the history of the applied field. For a scalar isotropic response, one may write
In the frequency domain this becomes . With fields represented using the time dependence , write . For a passive material, positive represents absorption. The cycle-averaged dissipated power per volume is
The real and imaginary parts are constrained by causality through the Kramers–Kronig relations: dispersion and absorption are not independent material properties.
The Debye relaxation model describes a single exponential approach to equilibrium:
Here is a relaxation time, the static relative permittivity, and the background permittivity above this relaxation but below additional faster processes. The relaxation contribution to dielectric loss peaks at . Liquid water has a principal dielectric relaxation time of roughly 8 picoseconds near room temperature, placing the associated loss maximum near 20 gigahertz. Polymers, glasses, and biological materials commonly show broad distributions of relaxation times rather than one ideal Debye process.
Resonant polarization differs from relaxational polarization. A damped bound-charge oscillator gives a contribution of the form
where is the resonance frequency, the damping rate, and an oscillator-strength parameter with units of frequency squared. Ionic resonances occur mainly in the infrared and terahertz ranges, while many electronic resonances lie at higher frequencies. Finite conductivity adds an apparent loss contribution , so low-frequency loss cannot automatically be assigned to dipole relaxation. This separation is central to dielectric spectroscopy.
Spontaneous polarization, ferroelectricity, and nonlinear behavior
Some crystals possess a nonzero polarization in the absence of an applied electric field. Such spontaneous polarization is permitted by polar crystal symmetry. Ferroelectricity is the more restrictive property in which the spontaneous polarization can be reversed between stable orientation states by an electric field. A polar material is therefore not necessarily ferroelectric. Ferroelectricity was experimentally established in Rochelle salt by Joseph Valasek in 1921. Barium titanate subsequently became a central example of an oxide ferroelectric, while lead zirconate titanate and ferroelectric polymers became important technological material families.
In barium titanate, a structural transition near at ambient pressure separates the high-temperature cubic paraelectric phase from a tetragonal ferroelectric phase. Relative displacements of the titanium and oxygen sublattices contribute strongly to the polarization. Other ferroelectrics involve ordering among locally polar configurations, collective molecular motion, or combinations of these processes. “Displacive” and “order–disorder” descriptions identify useful limiting mechanisms, but real materials can exhibit features of both. Near a transition, a large susceptibility may reflect the low energetic cost of changing a collective polar distortion.
A phenomenological description uses polarization as an order parameter. For a simple uniaxial case, an appropriate free-energy density can be expanded as
The zero-field even powers express equivalence between opposite polarization states in the reference symmetry. With suitable coefficients, the free energy develops minima at nonzero . A positive quartic coefficient can support a continuous mean-field transition, whereas a negative quartic coefficient stabilized by a positive sixth-order term can produce a first-order transition. Elastic coupling, spatial gradients, fluctuations, and electrical boundary conditions must be included for more realistic predictions.
Macroscopic specimens often divide into domains with different polarization orientations. Domain formation can reduce electrostatic or elastic energy, and field-driven switching proceeds through nucleation and domain-wall motion as well as changes within each domain. The resulting polarization–field hysteresis loop is characterized by remanent polarization and coercive field. Both depend on measurement frequency, temperature, defects, geometry, and switching history; neither is simply an immutable microscopic constant. Leakage can also generate loop-like electrical signals, so hysteresis alone does not establish ferroelectric switching. Outside ferroelectrics, strong fields can produce nonlinear polarization through orientational saturation or higher-order electronic and ionic response without creating switchable spontaneous order.
Coupling to temperature, strain, and optical fields
Polarization is coupled to other thermodynamic and structural variables. In a reversible description with electric field as an independent variable, an appropriate electric free-energy density contains a term in its differential. This relates polarization to free-energy derivatives and generates reciprocal relations between electrical, mechanical, and thermal responses. The specification of controlled variables matters: susceptibility measured at fixed strain need not equal susceptibility measured at fixed stress because the latter allows the specimen to deform. Related distinctions apply between isothermal and adiabatic electrical measurements.
Piezoelectricity is the linear coupling between polarization and mechanical stress or strain. One common representation of the direct effect is
where is mechanical stress and is a piezoelectric coefficient. The reciprocal effect is field-induced strain. In ordinary bulk crystals, inversion symmetry forbids piezoelectricity because polarization changes sign under inversion whereas strain does not. Quartz is piezoelectric but not ferroelectric, demonstrating that electromechanical coupling does not require switchable spontaneous polarization. Electrostriction, by contrast, is a quadratic coupling between strain and polarization and is allowed even in centrosymmetric materials.
Pyroelectricity concerns the temperature dependence of spontaneous polarization. Its coefficient is
under specified electrical and mechanical constraints. For a uniformly heated specimen with a suitable electrode geometry, a temperature change produces a current proportional to electrode area, the relevant pyroelectric coefficient, and . Thermal expansion can also alter polarization through piezoelectric coupling, making mechanical constraints important. The reciprocal electrocaloric effect is a change in temperature under adiabatic variation of electric field. Both phenomena follow from the dependence of entropy and polarization on temperature and field, although large irreversible switching losses complicate simple equilibrium interpretations.
At optical frequencies, nonlinear polarization is often written as
These susceptibilities generally depend on the participating frequencies. The second-order term supports second-harmonic generation, sum-frequency generation, and optical rectification; third-order terms support effects such as the optical Kerr effect and four-wave mixing. In the electric-dipole approximation, bulk inversion symmetry forces to vanish, although surfaces, interfaces, and higher-multipole processes can still produce second-harmonic signals. Nonlinear optics thus uses polarization both as the material response driven by incident light and as the source of newly generated electromagnetic fields.
Polarization in periodic quantum solids
For a finite neutral molecule, the dipole moment can be calculated from nuclear positions and the electronic charge distribution. Applying the same recipe to an infinite periodic crystal is problematic. A periodic solid has no unique outer boundary, and choosing different but equivalent unit cells can assign different dipole moments to the cell without changing the physical bulk crystal. Moreover, electrons occupy extended quantum states rather than necessarily belonging to particular atoms. A cell-by-cell sum of position times charge therefore does not provide a unique bulk polarization from periodic charge density alone.
The modern theory of polarization resolves this difficulty by connecting polarization differences to the integrated current during a continuous insulating transformation. The electronic contribution is expressed through a Berry phase of occupied electronic states across reciprocal space. R. D. King-Smith and David Vanderbilt formulated an influential crystalline treatment in 1993, and Raffaele Resta developed closely related formulations and interpretations. In an equivalent localized description, polarization can be related to ionic positions and the centers of occupied Wannier functions. The essential information concerns how electronic states are arranged and evolve, not merely the static charge density.
Bulk polarization in a periodic crystal is multivalued. Equivalent branches differ by a polarization quantum,
where is the positive elementary charge, a lattice vector, and the unit-cell volume. This corresponds to transferring one electronic charge through a lattice translation in the periodic description. Measured changes are obtained by following a physically specified path and maintaining branch continuity. The ambiguity does not imply an arbitrary experimental result: surface termination, electrical boundary conditions, and the history of charge transfer supply information that a bulk periodic calculation intentionally omits.
The theory also clarifies how atomic displacement changes polarization. A Born effective charge tensor can be defined by
where is the displacement of atom . This dimensionless dynamical charge includes the electronic rearrangement accompanying atomic motion and need not equal an oxidation state. Large effective charges in some transition-metal oxides reflect strong changes in bonding during displacement. Berry-phase polarization is now integral to first-principles studies of ferroelectrics, lattice vibrations, piezoelectricity, and interfaces. Its ordinary equilibrium formulation applies to insulating systems; a homogeneous static field in a conducting bulk instead produces carrier redistribution and current, requiring a different treatment.
Experimental determination and interpretation
Polarization changes are commonly measured electrically. For a planar specimen with uniform normal displacement field, the external current associated with changing electrode charge satisfies
Integrating the current yields the change in . To obtain the polarization change, the vacuum-field term must be accounted for:
when conduction and other unwanted currents have been removed. If the field is held constant during the process, the final term vanishes. This approach is used for field-driven switching, temperature-induced polarization changes, and some mechanically driven measurements.
Small-signal capacitance provides a differential electrical response rather than, by itself, a spontaneous polarization. For an ideal parallel-plate specimen of thickness ,
In anisotropic or biased materials, the relevant permittivity is the component and differential response appropriate to the measurement geometry. Frequency-dependent electrical impedance reveals loss and relaxation as well as capacitance. Accurate interpretation requires attention to electrode contact, stray capacitance, sample thickness, field uniformity, and calibration. A large inferred permittivity can originate from thin interfacial layers or mobile charges rather than from the bulk dielectric response.
Ferroelectric switching measurements often integrate charge during a varying voltage. The Sawyer–Tower circuit, introduced in 1930, uses a reference capacitor to monitor transferred charge. Modern instruments use current amplifiers, charge integration, and pulse sequences. Positive-up–negative-down measurements compare nominally switching and nonswitching pulses to help separate switchable charge from other contributions. Their interpretation still depends on pulse width, relaxation, leakage, and the reproducibility of the specimen’s initial state. Coercive fields obtained at different frequencies or with different pulse protocols are therefore not automatically equivalent.
Structural and local probes provide complementary information. X-ray diffraction and neutron diffraction determine atomic displacements and symmetry; vibrational spectroscopy identifies polar modes and changes in their frequencies; optical second-harmonic generation can reveal broken inversion symmetry. Piezoresponse force microscopy maps local electromechanical response, but electrostatic forces, ionic redistribution, and contact mechanics can mimic parts of a ferroelectric signal. No single contrast mechanism establishes every aspect of polarization. Reliable determination rests on agreement among electrical response, symmetry, structure, and time dependence, with explicit measurement uncertainty where quantitative values are reported.
Technological uses and material constraints
In capacitors, polarization allows a material to store more electrode charge at a given voltage than vacuum would. For a reversible linear dielectric, the electrical energy density is
For nonlinear materials, recoverable energy must instead be obtained from the appropriate charging or discharge path, using the integral of . High permittivity alone does not ensure high energy storage: dielectric breakdown strength, hysteretic loss, leakage, thermal stability, and endurance all matter. Ferroelectric ceramics can provide large capacitance, whereas polymer films often offer high breakdown fields and low loss. Composite designs attempt to combine favorable features while controlling interfacial fields and defects.
Switchable polarization provides the state variable in ferroelectric memories. Conventional ferroelectric random-access memory detects switching charge, while ferroelectric field-effect devices use polarization to alter the electrical behavior of a semiconductor channel. Hafnium-oxide-based films became a major research and device platform after ferroelectricity was reported in suitably processed doped hafnium oxide in 2011. Their behavior depends strongly on phase stabilization, thickness, dopants, mechanical constraint, electrodes, and cycling history. Device limitations include incomplete switching, loss of retained states, fatigue, leakage, and changes in behavior during early electrical cycling.
Coupled polarization responses enable sensors and actuators. Piezoelectric ceramics convert voltage into mechanical displacement in positioning systems and ultrasonic transducers, and convert stress into electrical signals in force or vibration sensing. Pyroelectric detectors respond to changing incident radiation through heating and the resulting polarization change. Polar optical crystals are used for electro-optic modulation and frequency conversion. In each case, useful performance depends on a selected response coefficient together with bandwidth, loss, mechanical or thermal conditions, and stability; a large static polarization does not by itself imply a superior transducer.
Interfaces and reduced dimensions create additional constraints and opportunities. In ultrathin polar films, incomplete electrode screening can increase the depolarization field, while strain, surface chemistry, and atomic-scale bonding can stabilize or suppress particular polar structures. In layered electronic devices, polarization discontinuities can influence charge accumulation and band alignment, although the actual carrier distribution also depends on defects and charge reservoirs. Computational materials design increasingly combines electronic-structure calculations with models of domains, interfaces, and processing. The practical target is not polarization in isolation but a controlled, reproducible electrical response under the boundary conditions of a functioning device.