Skip to content

Poynting Vector

The local electromagnetic energy-flux density whose surface-normal integral gives electromagnetic power crossing a boundary and whose divergence closes the field-energy balance.

Version
v2 · 2026-08-30 · History
Domain-specific #
2519
Origin domain
electromagnetism
Subdomain
electromagnetic energy conservation
Aliases
Electromagnetic Energy Flux Density

Core Idea

The Poynting vector is the local flux density assigned to electromagnetic field energy. In the standard instantaneous macroscopic SI convention,

\[ \mathbf S(\mathbf r,t)=\mathbf E(\mathbf r,t)\times\mathbf H(\mathbf r,t), \]

where \(\mathbf E\) is electric-field strength and \(\mathbf H\) is magnetic-field strength. In vacuum, where \(\mathbf B=\mu_0\mathbf H\), this becomes

\[ \mathbf S=\frac{1}{\mu_0}\mathbf E\times\mathbf B. \]

Its SI unit is watt per square metre. The component \(\mathbf S\cdot\hat{\mathbf n}\) is electromagnetic power per area crossing an oriented surface, and

\[ P_A(t)=\int_A \mathbf S\cdot d\mathbf a \]

is the instantaneous electromagnetic power crossing that surface in the chosen outward-normal direction.[1][n1]

The vector earns its physical meaning through the electromagnetic energy balance, not from the cross product in isolation. In a common microscopic/vacuum form,

\[ \frac{\partial u_{\rm em}}{\partial t}+\nabla\cdot\mathbf S=-\mathbf J\cdot\mathbf E, \]

where \(u_{\rm em}\) is field-energy density and \(\mathbf J\cdot\mathbf E\) is the rate per volume at which the field does work on charge. Thus a region's field energy decreases through outward flux, work on matter, or both. J. H. Poynting's 1884 paper made this field-centered transfer explicit: energy supplied and dissipated in an electric circuit is conveyed through the surrounding electromagnetic field, not merely imagined as material carried inside the conductor.[2]

The default identity must be stated with its convention. Instantaneous \(\mathbf E\times\mathbf H\), vacuum \(\mathbf E\times\mathbf B/\mu_0\), complex harmonic power density, and alternative partitions of field and material energy are related but are not silently interchangeable. The retained abstraction is the convention-declared local electromagnetic energy-flux vector that closes the corresponding conservation equation.

Structural Signature

Electric and magnetic fields + declared microscopic/macroscopic and time-domain convention + oriented local cross product + compatible field-energy density and work term -> an electromagnetic energy-flux vector whose divergence and surface integral close local and global power balance.

The mandatory roles are:

  • the electromagnetic state: electric and magnetic fields at a position and time, or their declared harmonic phasors;
  • the field convention: \((\mathbf E,\mathbf H)\) for the usual macroscopic form or \((\mathbf E,\mathbf B)\) with \(1/\mu_0\) for vacuum/microscopic form;
  • the directed cross product: orientation fixed by the right-hand rule and the order electric cross magnetic;
  • the flux-density output: \(\mathbf S\), with units W m\(^{-2}\), resolved normal to any test surface;
  • the compatible energy density: the \(u_{\rm em}\) used in the same energy partition;
  • the exchange term: normally \(\mathbf J\cdot\mathbf E\), or a declared free/bound-current counterpart, representing field-to-matter work;
  • the continuity relation: \(\partial_t u+\nabla\cdot\mathbf S=\text{exchange/sources}\); and
  • the boundary readout: surface integration converts local flux density to net electromagnetic power.

The invariant is not that the arrow always equals a ray or wavevector. It is that the declared \(\mathbf S\), energy density, and exchange terms form one consistent local balance. In a uniform traveling plane wave, \(\mathbf S\) points along energy propagation. In standing, reactive, static-crossed-field, near-field, anisotropic, or material situations, instantaneous or local arrows require the full balance and may circulate, oscillate, or differ from a simple phase-propagation direction.[1][3]

What It Is Not

  • Not electromagnetic energy density. \(u_{\rm em}\) is energy per volume in J m\(^{-3}\); \(\mathbf S\) is energy crossing area per time in W m\(^{-2}\).
  • Not total power. Power is a scalar surface integral of the vector's normal component. A point value cannot by itself give the power through a finite aperture, cable cross-section, or enclosure.
  • Not Poynting's theorem. The vector is one term in the electromagnetic conservation law; the theorem relates its divergence to storage and work.
  • Not field momentum. In vacuum the electromagnetic momentum density satisfies \(\mathbf g=\mathbf S/c^2\), but energy flux and momentum density have different units and roles. In matter, momentum partition introduces the Abraham–Minkowski problem and requires explicit convention.[n2]
  • Not necessarily radiation intensity. For a traveling plane wave, time-averaged \(\langle\mathbf S\rangle\) gives intensity and propagation direction. Near fields, standing waves, evanescent fields, and reactive systems do not reduce so simply.
  • Not a phasor without a convention. Peak-amplitude phasors yield \(\langle\mathbf S\rangle=\tfrac12\operatorname{Re}(\mathbf E\times\mathbf H^*)\); RMS phasors omit the factor \(1/2\). Mixing these conventions doubles or halves power.
  • Not a generic cross product. Orthogonality and right-hand orientation are mathematical ingredients; the energy-flux meaning comes from Maxwell's equations and the compatible balance law.

Scope of Application

Poynting Vector applies throughout classical electromagnetism wherever local or boundary-resolved electromagnetic power matters. In radiation and optics it relates fields to irradiance, beam power, antenna patterns, absorption, reflection, transmission, and scattering. In microwave and radio-frequency engineering it calculates power carried by waveguide and transmission-line modes. In antennas, the far-field surface integral gives radiated power while the complex theorem separates real radiated/dissipated power from reactive storage.[n3][n4]

In circuits and power delivery, it exposes the field route hidden by lumped \(VI\) bookkeeping. Around a coaxial cable or pair of conductors, the fields in the surrounding dielectric carry longitudinal power; near a resistive conductor, flux tilts into the surface and accounts for Joule heating. Contemporary calculations for parallel DC wires confirm that integrated field power equals battery/load power in the lossless case and decreases by exactly the power entering resistive conductors.[n5]

In materials, the vector remains essential but the partition must be declared. The familiar macroscopic \(\mathbf E\times\mathbf H\) form pairs with a balance that distinguishes free and bound responses according to constitutive assumptions. A microscopic/vacuum-style \(\mathbf E\times\mathbf B/\mu_0\) form moves some terms between field storage and work on matter. Literature on multiple Poynting theorems emphasizes that several algebraically valid flux-density packages can be written; their meanings depend on which energy density and residual material terms accompany them.[3][4]

The node does not cover non-electromagnetic heat flux, acoustic intensity, fluid energy flux, or Umov's more general historical energy-flow construction. Those share the parent Flow but retain different fields and conservation equations.

Clarity

A reliable use requires five declarations:

  1. Are the fields instantaneous real fields or harmonic phasors?
  2. Are phasor magnitudes peak or RMS?
  3. Is the system vacuum/microscopic, or a macroscopic material model using \(\mathbf H\)?
  4. Is the desired result a local flux density, a surface-integrated power, or a cycle average?
  5. Which sign convention and surface normal define positive outward flow?

With those declarations, apparent formula conflicts resolve. For a closed volume \(V\) with outward normal,

\[ \frac{d}{dt}\int_V u_{\rm em}\,dV =-\oint_{\partial V}\mathbf S\cdot d\mathbf a -\int_V\mathbf J\cdot\mathbf E\,dV. \]

A positive outward surface integral removes field energy from the volume; positive \(\mathbf J\cdot\mathbf E\) transfers field energy to matter. Reversing the normal reverses the reported flux but not the physics.

The dimensional check is decisive. \((\mathrm{V/m})(\mathrm{A/m})=\mathrm{W/m^2}\). Integrating across square metres produces watts. If a calculation reports joules, amperes, or watts per metre without an additional geometric factor, it has not yet produced the intended quantity.

Manages Complexity

The vector converts six coupled field components into a local power-flow ledger. Rather than infer where energy travels from conductor current, voltage, ray sketches, or field amplitude alone, the analyst computes one oriented quantity and integrates it over the boundary relevant to the question. This collapses complicated spatial structure into a conserved readout while preserving enough locality to distinguish where power enters, exits, circulates, or is absorbed.

It also bridges field and circuit descriptions. A transmission line may be summarized by \(P=VI\), but Poynting Vector reconstructs the spatial distribution that produces the same scalar power. The agreement is a conservation check; the field picture adds where the power lies and how geometry, dielectric material, conductor loss, or mode shape redistributes it.

Finally, it separates propagation from storage. The real part of the complex surface flux tracks cycle-averaged delivered or radiated power, while the imaginary part of the complex Poynting theorem relates imbalance of electric and magnetic stored energy and reactive exchange.[n4] This prevents large oscillating near fields from being mistaken for equally large net exported power.

Abstract Reasoning

Several deductions follow directly from the structure.

If \(\nabla\cdot\mathbf S<0\) at a point and no work term offsets it, electromagnetic field energy is accumulating locally. If the outward surface integral of \(\mathbf S\) is positive, the enclosed region is losing electromagnetic field energy unless a source replenishes it. If \(\mathbf J\cdot\mathbf E>0\), the field is doing work on charges; in an ohmic conductor this becomes heating. A negative value represents matter or a source delivering energy to the field.

For a uniform plane wave in vacuum, \(\mathbf E\perp\mathbf B\perp\mathbf k\), \(B=E/c\), and

\[ S=\frac{E^2}{\mu_0c}=\epsilon_0cE^2. \]

For \(E=E_0\cos(kz-\omega t)\), time averaging gives

\[ I=\langle S\rangle=\frac12\epsilon_0cE_0^2, \]

where \(E_0\) is the peak amplitude.[n6][1] Thus intensity scales quadratically with field amplitude, and doubling amplitude quadruples average flux.

But the converse inference is unsafe: a nonzero instantaneous \(\mathbf S\) does not automatically prove net radiation. Compute a cycle average and an enclosing-surface integral, then examine falloff and reactive contributions. Standing waves can carry opposing instantaneous flows with zero net average, and crossed static fields can produce circulating flux consistent with stored momentum and angular-momentum bookkeeping.

Knowledge Transfer

Within electromagnetism, the same method transfers literally from plane waves to cables, waveguides, cavities, antennas, optical beams, scattering, absorbing media, and numerical field solvers. The field variables and geometry change, but the procedure remains: choose the compatible convention, construct \(\mathbf S\), project on a boundary normal, integrate, and close the storage/work ledger.

The transfer to computational practice is especially direct. Finite-element and finite-difference solvers output complex \(\mathbf E\) and \(\mathbf H\); post-processing forms the real time-averaged normal flux and integrates it over ports, absorbers, or far-field surfaces. A negative or unbalanced result is not merely a plotting anomaly: it is a diagnostic for normal orientation, phasor normalization, mesh truncation, missing material loss, or an inconsistent field convention.

Outside electromagnetism, the portable skeleton is local conserved transport: density, flux, source, sink, boundary, and continuity. That structure belongs to Flow and Conservation Laws. Heat flux, acoustic intensity, mass flux, and probability current use parallel reasoning, but calling them Poynting vectors would import the wrong fields, units, and exchange terms. Cross-domain transfer should map roles, not rename the quantity.

Examples

Vacuum plane wave. Let \(\mathbf E=E_0\cos(kz-\omega t)\hat{\mathbf x}\) and \(\mathbf B=(E_0/c)\cos(kz-\omega t)\hat{\mathbf y}\). Then

\[ \mathbf S=\epsilon_0cE_0^2\cos^2(kz-\omega t)\hat{\mathbf z}. \]

The cross-product order fixes \(+\hat{\mathbf z}\), the instantaneous magnitude oscillates at twice the field phase frequency, and the cycle average is \(\tfrac12\epsilon_0cE_0^2\hat{\mathbf z}\). The fields are the electromagnetic state, the vacuum formula declares the convention, the vector is local flux density, and integrating over a transverse area yields beam power.

Coaxial transmission line. For inner radius \(a\), outer radius \(b\), voltage \(V\), and current \(I\) in an ideal coaxial line,

\[ E_r=\frac{V}{r\ln(b/a)},\qquad H_\phi=\frac{I}{2\pi r}. \]

Their cross product points along the cable:

\[ S_z=\frac{VI}{2\pi r^2\ln(b/a)}. \]

Integrating over the dielectric cross-section gives

\[ \int_a^b S_z\,2\pi r\,dr=VI. \]

The field ledger reproduces the circuit ledger while showing that ideal-line power occupies the dielectric region between conductors rather than being a material stream confined inside the metal.

Resistive conductor. Tangential electric field outside the wire and circumferential magnetic field produce a Poynting vector with an inward normal component. Surface integration equals the local rate of Joule heating. Along a resistive pair the longitudinal power declines by exactly the amount flowing into conductor surfaces.[n5] This is a practical example in which the vector's direction is not merely the direction of a free-space wave.

Structural Tensions

Local arrow versus global power. \(\mathbf S\) is pointwise, while delivered or radiated power is an oriented surface integral. A striking vector plot can overemphasize local circulation that contributes nothing to net export. The diagnostic is to integrate over a physically justified boundary.

Instantaneous transfer versus cycle average. Real fields yield instantaneous flow, including double-frequency oscillation and reversal. Engineering power is often the cycle-average real part of a complex product. The diagnostic is to state the time representation before comparing values.

Peak versus RMS phasors. Both conventions are legitimate; the factor \(1/2\) moves with the amplitude definition. The diagnostic is to reconstruct the real field from the phasor before using a power formula.

Field energy versus field–matter partition. In material media, \(\mathbf E\times\mathbf H\) and \(\mathbf E\times\mathbf B/\mu_0\) accompany different choices about bound response, energy density, and work terms. The diagnostic is conservation of the complete declared package, not allegiance to a cross product by itself.[3][4]

Energy flux versus momentum. Vacuum connects them by \(\mathbf g=\mathbf S/c^2\), tempting identification. Their dimensions and balance laws remain distinct, and material momentum adds further convention dependence.[n2]

Simple formula versus localization ambiguity. A continuity equation determines divergence and net boundary flow but permits divergence-free rearrangements unless additional physical criteria fix localization. Feynman explicitly discusses this ambiguity while retaining the simple standard expression.[1] The diagnostic is whether the question concerns measurable net transfer or a claimed unique microscopic path.

Structural–Framed Character

Poynting Vector is strongly structural within a sharply bounded physical domain. It is defined by fields, cross product, units, orientation, and a continuity equation; no human norm or institutional frame enters. The choice among microscopic, macroscopic, material, and phasor conventions is technical rather than evaluative.

It is nevertheless domain-specific rather than prime because its literal variables are electromagnetic fields and currents governed by Maxwell's equations. The fully portable structure—local flux closing a conservation law—already belongs to Flow and Conservation Laws.

Structural Core vs. Domain Accent

The structural core is a vector flux density whose divergence measures local depletion and whose boundary-normal integral measures net transport. That core recurs in heat, mass, probability, charge, and fluid transport and is owned by Flow, Conservation Laws, Boundary, and Reservoir-Flux Network.

The domain accent supplies \(\mathbf E\), \(\mathbf H\), \(\mathbf B\), \(\mu_0\), \(\mathbf J\cdot\mathbf E\), Maxwell's equations, W m\(^{-2}\), electromagnetic material partitions, complex power, radiation, waveguides, and antennas. Those elements make the node autonomous inside electromagnetism but block prime classification.

Poynting Vector strictly specializes Flow: it is structured directional transport of electromagnetic field energy with a local rate, medium, boundary, and continuity relation. Conservation Laws explains Poynting's theorem, but the vector is a flux term rather than itself a conservation law. Boundary explains the conversion from local flux to net surface power. Momentum is quantitatively related in vacuum but remains a distinct physical quantity.

Only Flow is proposed as the minimal DAG parent. Conservation Laws, Boundary, and Momentum are explanatory relations rather than parallel taxonomic genera.

Relationships to Other Abstractions

Local relationship map for Poynting VectorParents appear above the current abstraction, mutual partners to the right, and children below. Node labels state whether each abstraction is prime or domain-specific; colors identify relation types.Poynting VectorDOMAINPrime abstraction: Flow — is a kind ofFlowPRIME

Current abstraction Poynting Vector Domain-specific

Parents (1) — more general patterns this builds on

  • Poynting Vector is a kind of Flow Prime

    Poynting Vector strictly specializes Flow: it is structured directional transport of electromagnetic field energy with a local rate, medium, boundary, and continuity relation.

Hierarchy path (1) — routes to 1 parentless root

  • Poynting VectorFlow

Neighborhood in Abstraction Space

Poynting Vector sits in a sparse region of the domain-specific corpus (91st percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Unclustered & Miscellaneous (1565 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-09-08

Not to Be Confused With

  • Poynting's theorem: the complete electromagnetic energy-balance equation.
  • Electromagnetic energy density: stored energy per volume rather than transported power per area.
  • Intensity or irradiance: commonly the time-averaged normal Poynting flux for a propagating wave or incident surface, not every instantaneous vector field.
  • Complex Poynting vector: a harmonic-domain construction whose real part gives average active power under a declared amplitude convention and whose imaginary part describes reactive exchange.
  • Power: a scalar obtained after surface integration.
  • Momentum density: \(\mathbf S/c^2\) in vacuum, not \(\mathbf S\) itself.
  • Wavevector \(\mathbf k\): encodes phase propagation; it need not be parallel to energy flux in anisotropic or more complex media.
  • Radiation pressure: force per area related to momentum transfer, not energy flux itself.
  • Umov vector: a historically broader energy-flow construction for continuous media; “Umov–Poynting vector” should remain context-qualified.

Notes

[n1] Zangwill gives the advanced electrodynamics treatment and convention-aware energy balance.

[n2] Pfeifer et al. review electromagnetic momentum in dielectric media and the convention-dependent material partition. ↩a ↩b

[n3] Haus and Melcher derive real and complex Poynting theorems and connect field flux with circuit and antenna power.

[n4] MIT 6.013 section 12.5 distinguishes the real time-average complex flux from the imaginary reactive-energy relation. ↩a ↩b

[n5] Boulé verifies by explicit integration that DC field power between conductors matches battery/load power and enters resistive wires as loss. ↩a ↩b

[n6] OpenStax verifies vacuum plane-wave flux, direction, units, time average, and intensity-amplitude relations.

References

[1] Feynman, Leighton, and Sands derive energy density and \(\mathbf S=\epsilon_0c^2\mathbf E\times\mathbf B\), interpret surface-normal flow, work plane-wave and conductor examples, relate field momentum, and discuss localization ambiguity. withdrawn registry ↩a ↩b ↩c ↩d

[2] Poynting 1884 is the primary derivation and field-transfer argument. registry

[3] Kinsler, Favaro, and McCall compare multiple flux/energy/residual packages in macroscopic materials. withdrawn registry ↩a ↩b ↩c

[4] Richter, Florian, and Henneberger analyze energy conservation and field/material exchange in bounded media. withdrawn registry ↩a ↩b