A charge-density wave (CDW) is an ordered state of matter in which the electronic charge density varies periodically in space, introducing a modulation beyond the ordinary periodicity of a crystal. In most established examples, this electronic modulation is coupled to a periodic displacement of the atomic lattice. The resulting state is therefore a coordinated reorganization of electrons and atoms rather than simply a standing wave of independent particles. Charge-density waves occur in chain compounds, layered metals, and several families of strongly interacting electronic materials.
Their importance in condensed-matter physics comes from the different processes they bring together: electronic interactions, lattice vibrations, broken translational symmetry, and collective motion. A CDW can reconstruct a metal’s electronic states, open an energy gap over part or all of its Fermi surface, and support unusual electrical conduction through motion of the modulation itself. Although the classic explanation involves an instability of a one-dimensional metal, many real charge-density waves require mechanisms beyond that idealized model.
The subject extends from equilibrium phase transitions and crystal structure to nonlinear transport, ultrafast dynamics, and competition with superconductivity. The term encompasses several related forms of order, so identifying a CDW requires attention to its wavevector, symmetry, electronic character, lattice coupling, and spatial and temporal coherence.
Spatial order and broken symmetry
The defining quantity is the spatially varying electronic density. For a simple modulation with one wavevector, it can be written as
where is the mean electron number density, is the ordering wavevector, is the modulation amplitude, and specifies the modulation’s position relative to the crystal. The complex quantity is an order parameter. The corresponding electronic charge density is , where is the positive elementary charge. Real structures can contain several ordering wavevectors and higher harmonics, particularly when the modulation is strongly nonsinusoidal.
A crystal already has a nonuniform electron density because electrons occupy atomic orbitals and chemical bonds. That ordinary periodicity is not itself a CDW. The additional modulation must introduce new translational order, generally visible at wavevectors distinct from those of the parent reciprocal lattice. A CDW can enlarge the crystallographic unit cell or, if its period is incommensurate with the lattice, produce a structure that cannot be described by any finite three-dimensional supercell.
The accompanying atomic displacement is often represented schematically by
where labels atoms within the parent unit cell. Different atoms can move in different directions and with different relative phases. Electronic density, bond lengths, orbital occupation, and local bonding can therefore acquire distinct but coupled modulations. In some materials, the most prominent microscopic signature is a modulation of bonds rather than a large transfer of charge between atomic sites. Describing such a state solely as alternating “charge-rich” and “charge-poor” atoms can be misleading.
The ordered state breaks the translational symmetry of the parent crystal. A unidirectional CDW can also break rotational symmetry by choosing one crystallographic direction, whereas a multidirectional state may preserve some of the parent rotational symmetry. These are examples of spontaneous symmetry breaking: symmetry-related patterns are possible, and different regions can select different domains. A “single-” state contains one independent modulation direction and its complex conjugate; a “triple-” state contains three symmetry-related directions.
Charge-density order is distinct from spin-density order, in which spin polarization varies periodically. It is also not automatically equivalent to macroscopic separation into regions of different composition or charge. A CDW has an identifiable spatial modulation and coherence, even when disorder limits that coherence to finite domains.
The Peierls instability and the one-dimensional model
The classical starting point is the Peierls instability, associated with Rudolf Peierls’s analysis of one-dimensional metals. Consider a chain with lattice spacing , one relevant orbital per site, and a half-filled electronic band. Its Fermi points lie at , with in the simplest spin-degenerate model. A lattice distortion with wavevector doubles the repeat distance and couples electronic states near the two Fermi points. The resulting mixing opens a gap at the Fermi energy.
The physical balance involves an electronic energy gain and a lattice energy cost. Opening the gap pushes occupied states downward in energy and unoccupied states upward. For a small distortion amplitude , an ideal one-dimensional metal at zero temperature has a singular electronic energy lowering with a form proportional to , whereas the elastic cost begins as . Within the idealized adiabatic treatment, the electronic term can favor a distortion even for weak coupling. The distorted lattice and modulated electronic density stabilize each other.
This instability can also be expressed through the electronic response. The Lindhard response function describes the density response of noninteracting electrons to a weak external perturbation. In one dimension, its static form has a logarithmic singularity at . Coupling the electrons to a phonon can then strongly reduce the frequency of a lattice vibration at that wavevector. The related anomaly in phonon dispersion is called a Kohn anomaly. If the relevant mode becomes unstable, the equilibrium structure acquires a periodic distortion.
The weak-coupling mean-field treatment has mathematical similarities to the Bardeen–Cooper–Schrieffer theory of superconductivity. In its simplest form, it gives
where is the zero-temperature single-particle gap, is the Boltzmann constant, and is the mean-field transition temperature. This is a model result, not a universal experimental ratio. Strong coupling, fluctuations, incomplete Fermi-surface gapping, and multiple bands can all cause substantial departures.
Strictly one-dimensional systems also have strong fluctuations. For short-range interactions, the ideal mean-field picture does not generally yield conventional long-range order at a nonzero temperature. Real quasi-one-dimensional CDW crystals consist of many coupled chains, and interactions between chains can establish three-dimensional phase coherence. A temperature range with substantial local correlations can therefore precede the actual ordering transition.
Peierls’s argument explains why low dimensionality makes metals susceptible to density modulation. It does not imply that every chain material must undergo the simplest dimerization, nor that every observed CDW is a weak-coupling Peierls state.
Microscopic mechanisms beyond perfect nesting
In a metal, the Fermi surface separates occupied from unoccupied electronic states at zero temperature. “Nesting” occurs when translation by a particular wavevector brings extended portions of that surface into approximate coincidence. Many low-energy electron–hole excitations can then share a similar momentum transfer, enhancing the electronic response at that wavevector. This generalizes the special role of in one dimension and provides a useful route to CDW formation in higher-dimensional materials.
Nesting is neither a sufficient nor a necessary explanation for all charge-density waves. The response depends not only on the geometry of the Fermi surface but also on orbital character, electronic velocities, matrix elements, and states away from the Fermi energy. A plausible nesting vector inferred from a band-structure plot does not establish the microscopic mechanism. Conversely, a material can develop a CDW even when its calculated electronic susceptibility has no sharp maximum at the observed ordering wavevector.
One important alternative emphasis is momentum-dependent electron–phonon coupling. The strength with which an electronic state interacts with a lattice vibration can vary substantially with wavevector and phonon polarization. A particular phonon branch may soften because its coupling is unusually strong, rather than because the uncoupled electronic response is sharply peaked. The instability then reflects the combined electronic and lattice response. Layered -NbSe is a prominent example in which a simple nesting-only description is inadequate.
Electron–electron interactions can also favor charge modulation. Repulsion between electrons may stabilize spatial arrangements that reduce unfavorable local configurations, especially in narrow bands or at particular fillings. Multiorbital systems add interactions involving orbital occupation and bond formation. Depending on the material, the ordered state can involve site-centered charge modulation, bond-centered modulation, or a combination. The experimentally measured density pattern need not resemble a simple sinusoidal redistribution of otherwise unchanged conduction electrons.
Another proposed route involves excitons, bound or correlated electron–hole pairs. If electron-like and hole-like states become coherently mixed at a finite wavevector, the resulting order can generate a density modulation and couple to the lattice. The transition near 200 K in -TiSe, which produces a superstructure, has long been studied in relation to excitonic correlations and lattice-driven mechanisms. The relative importance of those contributions remains a subject of research; calling the material a CDW system does not by itself settle that question.
These mechanisms are not mutually exclusive. A lattice displacement modifies the electronic band structure, while electronic redistribution changes interatomic forces. Determining what drives a transition therefore requires comparing temperature-dependent phonons, electronic spectra, structural changes, and quantitative calculations rather than assigning causation from one isolated observation.
Commensurability, phase transitions, and defects
A CDW is commensurate when its wavevector has rational components relative to the reciprocal lattice. In one dimension, the condition can be written as , where is an integer and is a reciprocal-lattice vector. The combined structure then repeats after an integer number of parent unit cells. An incommensurate modulation has no such finite repeat distance. A material can pass through several modulation periods as temperature, pressure, or composition changes.
The distinction has consequences for the phase of the order parameter. An ideal incommensurate modulation can be translated continuously relative to the lattice without selecting a new discrete registry. In a commensurate state, the lattice can favor particular phases. A phenomenological contribution to the free energy takes the form
where measures the commensurability-locking strength. This term creates discrete preferred positions for the modulation and can oppose its motion. Higher-order commensurabilities often have weaker locking, although the actual strength depends on microscopic structure.
A basic Landau description of the phase transition starts with
using a convention in which represents the preferred modulation wavevector relative to the chosen reference. The coefficients summarize electronic and elastic contributions, and symmetry determines which additional terms are allowed. Several coupled order parameters are required for multidirectional states. Their interactions can favor stripes, multiple-wavevector patterns, or transitions between different structures.
Close to a commensurate period, a system can accommodate mismatch by forming domains separated by narrow phase-changing boundaries, often called discommensurations. The phase is nearly locked inside each domain but changes across a boundary. A nearly commensurate state is consequently not always a uniformly distorted version of a commensurate crystal. Its spatially averaged wavevector can instead reflect the density and arrangement of domain walls. Such distinctions are important in interpreting diffraction patterns and real-space images.
Topological defects include dislocations of the modulation. Around a CDW dislocation, the phase can wind by , and the modulation amplitude is typically reduced near the core. Defects influence domain growth, transport, and recovery after a rapid temperature change. Their motion can relax strain in the CDW even when the underlying atomic lattice remains intact.
Impurities introduce additional complications because they locally favor particular phases and distort the modulation. Random pinning can limit correlation lengths and broaden an otherwise sharp transition. Thus, the temperature at which a probe detects growing CDW correlations need not coincide with the establishment of long-range order, and hysteresis may reflect domain rearrangement or a first-order structural transition rather than a simple change in amplitude alone.
Collective excitations and sliding conduction
A charge-density wave has collective degrees of freedom that differ from ordinary single-electron excitations. Small changes in its order parameter separate approximately into amplitude fluctuations and phase fluctuations. The amplitude mode, often called an amplitudon, changes the strength of the modulation. The phase mode, or phason, shifts the modulation’s position. These modes are generally mixtures of electronic and lattice motion because the equilibrium CDW couples both subsystems.
For an ideal incommensurate state without pinning, a uniform phase shift costs no energy, giving a low-energy phason. Real crystals modify this picture. Impurities and commensurability produce restoring forces, while long-range Coulomb interactions strongly affect charged longitudinal motion. Screening by residual mobile carriers, damping, and coupling between chains further determine the measured dispersion. A low-frequency resonance therefore cannot be interpreted without considering both the CDW and its electromagnetic environment.
A particularly striking phenomenon is collective electrical conduction in some quasi-one-dimensional compounds. At small applied electric fields, the CDW is pinned and ordinary carriers dominate the current. Above a threshold field, the modulation can move through the lattice, producing an additional nonlinear contribution to the electric current. This is called sliding CDW conduction. It does not mean that the entire ionic crystal translates: the phase of the coupled density and displacement pattern advances.
For a simple displacement of a modulation along its wavevector, the phase satisfies , up to the convention chosen for the modulation. The collective current is proportional to the phase velocity, with a coefficient determined by the participating electronic states. A characteristic modulation frequency is
In appropriate sliding states, this frequency scales with the collective current. Periodic voltage oscillations, known as narrow-band noise, and synchronization to an externally applied alternating field provide important evidence of coherent motion. Synchronization can produce steps or other structures in current–voltage curves.
Pinning and depinning are spatially complex. Impurities can distort the phase over distances much larger than an atomic spacing, and different regions may begin moving at different fields. Near electrical contacts, ordinary current must convert into collective current and back again. Phase slips, in which the amplitude locally decreases enough to change the number of modulation periods, help accommodate this conversion and the associated CDW strain.
Sliding is not a universal property of every material described as having charge order. Strongly commensurate states, short-range modulations, and many layered systems do not show the classic transport signatures. Even in sliding compounds, the low-field state generally retains dissipation and pinning; the phenomenon is not equivalent to zero-resistance superconductivity or an ordinary equilibrium persistent current.
Established material families and historical development
The modern subject grew from attempts to understand low-dimensional conductors. Herbert Fröhlich proposed in 1954 that a coupled electronic and lattice modulation could support collective electrical motion. Peierls’s instability supplied a complementary explanation for why a one-dimensional metal might distort in the first place. In the 1970s, improved synthesis and measurements of chain and layered compounds connected these ideas to observed superstructures, electronic gaps, and nonlinear transport.
The organic conductor TTF–TCNQ became an important early system. It contains stacks of donor and acceptor molecules and undergoes several low-temperature transitions associated with charge-density modulation, beginning near 54 K. Its behavior demonstrated both the relevance of quasi-one-dimensional electronic structure and the limitations of treating real compounds as isolated identical chains. Different stacks, interchain interactions, and fluctuations all matter.
Inorganic transition-metal trichalcogenides provided especially clear examples of collective transport. NbSe undergoes CDW transitions near 145 K and 59 K. Both remove electronic states from parts of the Fermi surface, but enough ungapped states remain for the low-temperature material to be metallic. Blue bronze, commonly written KMoO, develops CDW order near 180 K and is another major system for studying pinning, dielectric response, phase dynamics, and nonlinear conduction. These compounds established that a CDW need not make a crystal a complete insulator.
Layered transition-metal dichalcogenides exhibit a wider range of ordering geometries. In -NbSe, CDW order appears near 33 K and superconductivity near 7.2 K. The modulation has a period close to three in-plane lattice spacings. In -TaS, cooling produces a sequence conventionally described as incommensurate, nearly commensurate, and commensurate CDW phases. The commensurate low-temperature structure contains clusters of 13 tantalum atoms arranged in a “Star of David” pattern. Its electronic properties require consideration of correlations and stacking as well as the in-plane distortion.
Rare-earth tritellurides, written Te, contain conducting tellurium sheets and show strong unidirectional CDWs. Some members develop a second modulation along the perpendicular in-plane direction at a lower temperature. Changes across the rare-earth series alter lattice dimensions and electronic structure, making this family useful for studying the relation between dimensionality, competing modulation directions, and lattice effects.
These examples also show why chemical composition alone does not specify a CDW state. The and labels refer to different layer structures and coordination environments, which can support very different ordering patterns. Polytype, sample stoichiometry, strain, defects, and thermal history can all alter transition temperatures and domain arrangements. Numerical transition temperatures are consequently characteristic values rather than exact constants shared by every specimen.
Experimental identification and measurement
X-ray diffraction and electron diffraction are central tools for establishing the ordering wavevector. A periodic modulation produces satellite reflections near parent Bragg peaks, typically at positions such as . Additional satellites and harmonics reveal more complicated structures. Their widths give information about correlation lengths, although instrumental resolution, strain, finite domains, and disorder must be separated. Ordinary diffraction is often strongly sensitive to atomic displacement, so a satellite peak does not directly measure the electronic charge amplitude.
Resonant X-ray scattering adds chemical and electronic selectivity by tuning the incident photon energy near an absorption edge. It can enhance sensitivity to particular atomic species and valence states. This has been especially important for weak charge order in copper-oxide superconductors. The measured intensity nevertheless depends on the resonant form factor, polarization, geometry, and details of the modulation. Extracting an absolute charge difference requires a model; intensity is not simply interchangeable with electrons transferred per atom.
Scanning tunneling microscopy provides real-space images with atomic resolution. Its differential conductance is related, under suitable conditions, to the local electronic density of states at the selected energy. Fourier analysis can reveal CDW wavevectors, domain orientations, defects, and relative phases. However, a conductance modulation is not identical to a modulation of the total charge density. Surface reconstruction, impurity-induced standing waves, and quasiparticle interference can also create periodic patterns. Energy dependence, temperature evolution, and comparison with diffraction help distinguish them.
Angle-resolved photoemission spectroscopy measures occupied electronic states as a function of energy and momentum. CDW formation can cause band folding, avoided crossings, spectral-weight redistribution, and partial energy gaps. These observations test how the ordering wavevector connects states in the parent electronic structure. A gap may vary strongly around the Fermi surface, and a metallic CDW can retain ungapped pockets. Photoemission interpretation also requires accounting for experimental matrix elements and surface sensitivity.
Lattice and collective dynamics are examined with inelastic X-ray or neutron scattering, Raman spectroscopy, infrared measurements, and time-resolved optical techniques. Phonon softening and linewidth changes constrain electron–phonon interactions. Raman-active amplitude modes provide information about the ordered state, while low-frequency optical and dielectric measurements can reveal pinned collective modes. Electrical transport supplies complementary evidence through threshold fields, nonlinear conductivity, and narrow-band noise.
No single signature is definitive in every material. A resistance anomaly can accompany many kinds of transition, and a structural superlattice need not have the same electronic origin as a classic Peierls CDW. The strongest identifications combine wavevector-resolved structural evidence with electronic and dynamical measurements, while distinguishing static long-range order from short-range or fluctuating correlations.
Competition and coexistence with superconductivity
Charge-density waves and superconductivity often appear in neighboring regions of a material’s phase diagram. Both reorganize low-energy electronic states, but they break different symmetries. A conventional CDW involves coherent mixing of electron and hole states separated by an ordering wavevector and breaks translation symmetry. Superconductivity involves coherent electron pairing, commonly into Cooper pairs, and is associated with phase coherence of a charged pair field. The presence of one order can therefore alter the conditions for the other without making them equivalent.
A simple phenomenological description introduces a superconducting order parameter alongside the CDW order parameter . A coupling term is
A positive represents competition in this approximation: developing one order raises the energetic cost of the other. A negative coefficient favors coexistence. Real materials may need additional terms for multiple wavevectors, spatial variation, and multiband effects. Even with a competitive coupling, the two orders can coexist if they primarily involve different electronic states or if neither completely suppresses the other.
Layered -NbSe provides a well-established example of coexistence. Its CDW transition occurs substantially above its superconducting transition, and both orders are present at low temperature. This is possible partly because the CDW does not gap the entire Fermi surface. Altering pressure, thickness, or disorder can change the balance, but changes in the two transition temperatures do not alone establish a universal causal relation. The same tuning parameter can independently modify phonons, electronic bands, scattering, and interactions.
Copper-oxide superconductors introduced a different setting, involving strongly interacting electrons and doping-dependent phases. Experiments reported in 2012 established short-range charge correlations in underdoped YBaCuO using resonant X-ray scattering. In these systems, characteristic in-plane wavevectors are often near 0.3 reciprocal-lattice units, although they vary with composition. Magnetic fields that weaken superconductivity can enhance charge order, and under suitable conditions a more three-dimensionally coherent ordered state develops. These observations provide evidence for competition in parts of the phase diagram.
Charge order in cuprates must also be distinguished from the broader pseudogap phenomenon, which involves a suppression of low-energy electronic spectral weight and has its own temperature and doping dependence. They are related research topics, not interchangeable names. Stripe states in some cuprates combine charge and spin modulations, while a pair-density wave denotes a spatially modulated superconducting pair field. Such orders can couple by symmetry, but detecting charge modulation does not by itself establish a modulated superconducting condensate.
A major experimental task is therefore to separate coexistence within the same microscopic region from spatial separation into different domains, and to distinguish direct competition from responses to a common underlying change in electronic structure.
Dimensional control and nonequilibrium states
Reducing a CDW material to a few layers or a single layer changes more than its geometric thickness. It alters dielectric screening, interlayer bonding, phonon spectra, and sometimes carrier concentration. Substrates can introduce strain, charge transfer, and additional screening. Consequently, a monolayer need not preserve either the transition temperature or the modulation pattern of the bulk crystal. Layered dichalcogenides provide experimentally accessible systems in which these effects can be studied through exfoliation, molecular-beam growth, and controlled interfaces.
Dimensionality also changes the role of fluctuations. A weakly pinned incommensurate phase has an approximately continuous phase degree of freedom, whose long-wavelength fluctuations are especially important in two dimensions. Commensurability, substrate coupling, finite sample size, and residual interlayer interactions can stabilize order or change the nature of the transition. Statements that “two-dimensional CDW order is forbidden” are therefore too broad: the outcome depends on symmetry, interaction range, and the physical environment.
Ultrafast experiments investigate CDW dynamics by exciting a material with a short laser pulse and following the response with optical, photoemission, or diffraction probes. Electronic occupations can change before the atoms substantially move. A CDW-related electronic gap may weaken rapidly, while the periodic lattice distortion evolves on vibrational timescales. Comparing the two responses helps distinguish electronic and structural contributions, although a driven state cannot generally be represented by a single equilibrium temperature.
A coherent amplitude-mode oscillation can be launched when excitation suddenly changes the equilibrium value of the distortion. Its frequency, damping, and phase provide information about the effective energy landscape. Time-resolved diffraction can measure the accompanying atomic motion directly, while time-resolved photoemission follows band reconstruction. The two signals may recover at different rates because restoring a local modulation amplitude is not the same as rebuilding phase coherence across a macroscopic domain.
A notable nonequilibrium result was reported in 2014 for -TaS: a short optical pulse could produce a long-lived metastable state distinct from the ordinary low-temperature equilibrium state. Such behavior has motivated research on switching through domain-wall configurations and other collective rearrangements. The microscopic state reached depends on excitation conditions, temperature, sample geometry, and thermal history; a persistent resistance change alone does not uniquely identify its structure.
Electrical and optical switching have suggested possible applications in memory elements, oscillators, and nonlinear devices. Practical implementation faces heating, variability, contact effects, endurance, and the difficulty of controlling domains reproducibly. Current fundamental questions include how defects nucleate during rapid ordering, how electronic and lattice coherence recover on different timescales, and whether a driven material can sustain an ordered configuration unavailable in equilibrium.