Group Velocity¶

Read the leading propagation velocity of a narrow-band wave packet's envelope from the local slope of its dispersion relation, while treating bandwidth, curvature, loss, and causality as explicit limits on that interpretation.

Version: v3 · 2026-09-19 · History
Domain-specific #: 1966
Domain group: Natural Sciences
Origin domain: Physics
Subdomains: Wave Physics, Dispersion and Wave Packets → Physics
Aliases: Group speed, Wave-packet group velocity, Group-velocity vector

Core Idea¶

Group velocity is the first-order propagation velocity associated with a local slope of a wave mode's dispersion relation. In one dimension, for angular frequency (omega(k)) and a narrow spectrum centered at (k_0),

\[ v_g(k_0)=\left.\frac{d\omega}{dk}\right|_{k_0}. \]

In several dimensions it is the vector

\[ \mathbf v_g(\mathbf k_0)=\nabla_{\mathbf k}\omega(\mathbf k_0). \]

The physical interpretation comes from a nearly monochromatic packet. Superpose plane waves \(e^{i(kx-\omega(k)t)}\) with spectral amplitudes concentrated near (k_0). Linearizing the dispersion relation,

\[ \omega(k)\approx\omega_0+(k-k_0)\omega'_0, \]

factorizes the field into a carrier \(e^{i(k_0x-\omega_0t)}\) and an envelope depending on (x-v_gt). To this order the carrier phase travels at \(v_p=\omega_0/k_0\), while the envelope translates at (v_g). MIT wave and transport notes, Brillouin's monograph, and standard texts derive the same distinction.^[1]^[2]^[3]

The derivative is exact as a property of a differentiable dispersion branch. “Velocity of the envelope” is an approximation whose fidelity depends on bandwidth, dispersion curvature, propagation distance, loss, gain, and mode structure. Group velocity often agrees with energy transport in lossless systems and with the motion of a slowly distorted packet peak. It is not universally the speed of information, energy, a sharp front, a particle, or every point of a changing envelope.^[4]^[5]

How would you explain it like I'm…

Speed of the Ripple Bunch

Sometimes waves travel in a little bunch, like a clump of ripples on a pond. The tiny ripples inside the bunch can move at one speed while the whole bunch moves at another. Group velocity is, roughly, how fast the bunch's outline moves along.

Wave Bundle Speed

Waves often travel in a bundle, like a short burst of ripples. Inside the bundle are little crests, and around them is the bundle's overall shape. In many materials, waves of different sizes travel at different speeds. Because of this, the crests and the bundle's shape can move at different speeds. The speed of the crests is called phase velocity. The speed of the bundle's shape is called group velocity. The bundle can slowly change shape as it goes, so group velocity describes its motion best when the bundle is made of very similar waves.

Envelope Speed from Dispersion

Every kind of wave has a dispersion relation linking its frequency ω to its wavenumber k (how tightly packed the crests are). Group velocity is the slope of that relation, dω/dk, at the packet's central wavenumber. Phase velocity, ω/k, is the speed of individual crests. When you add up waves with nearly the same k, the result is a carrier wave moving at the phase velocity inside an envelope moving at the group velocity. That envelope picture is an approximation: packets with a wide spread of wavenumbers, strongly curved dispersion, long distances or losses can distort. Group velocity is not always the speed of energy, information or a sharp front.

Group velocity is defined from a wave mode's dispersion relation ω(k) as v_g = dω/dk evaluated at a central wavenumber k0; in several dimensions it is the gradient ∇_k ω. Its physical meaning comes from a nearly monochromatic packet: superpose plane waves e^{i(kx − ωt)} with amplitudes concentrated near k0 and linearize ω(k) ≈ ω0 + (k − k0)ω'0. The field then factors into a carrier e^{i(k0x − ω0t)} and an envelope that depends on x − v_g t. To first order, carrier phase moves at v_p = ω0/k0 and the envelope translates at v_g. The derivative is an exact property of a differentiable dispersion branch, but 'envelope velocity' is an approximation whose accuracy depends on bandwidth, curvature of ω(k), propagation distance, loss or gain, and mode structure. In lossless systems v_g often matches energy transport and the motion of a slowly distorting packet peak. It is not universally the velocity of information, energy, a sharp front, a particle, or every point of a changing envelope.

Structural Signature¶

A group-velocity claim has these roles:

Wave mode: a branch of allowed oscillatory solutions in a declared medium or field.
Dispersion relation: \(\omega=\omega(\mathbf k)\), including branch and parameter choices.
Evaluation point: a central wavevector \(\mathbf k_0\) or central frequency.
Local spectral slope: \(d\omega/dk\) in one dimension or \(\nabla_{\mathbf k}\omega\) in several dimensions.
Packet regime: a spectral distribution sufficiently concentrated that first-order expansion is informative.
Envelope observable: a peak, centroid, modulation, or related packet feature whose evolution is being approximated.
Validity statement: bandwidth, curvature, attenuation, gain, anisotropy, and distance over which the interpretation holds.

The next Taylor term organizes the principal failure mode:

\[ \omega(k)=\omega_0+\omega'_0q+\tfrac12\omega''_0q^2+\cdots, \qquad q=k-k_0. \]

The first derivative translates the leading-order envelope. The second derivative produces group-velocity dispersion: different nearby components have different slopes and the packet spreads or chirps. Higher derivatives matter for broad spectra, long propagation, or operation near resonances and band edges.

If \(\omega=ck\), then (v_g=v_p=c), and every Fourier component shares a common speed. If \(\omega=ck+\omega_c\), the group velocity remains © and a packet may translate without first-order distortion, while \(v_p=c+\omega_c/k\) differs. Equality of group and phase velocities therefore requires proportional—not merely affine—dispersion.

What It Is Not¶

Group velocity is not phase velocity. Phase velocity tracks a surface of constant carrier phase and equals \(\omega/k\) in one dimension. The slope and secant of a nonlinear dispersion curve generally differ.

It is not the wave packet. A wave packet is the localized spectral superposition; group velocity is a dispersion-derived velocity used to describe part of its evolution.

It is not group delay. For a propagation length (L), group delay is commonly \(\tau_g=d\phi/d\omega\) for transfer phase \(\phi(\omega)\), and equals (L/v_g) only under the relevant uniform-propagation assumptions.

It is not group-velocity dispersion. That concerns variation of group delay or velocity with frequency and is governed by second-order dispersion, not the first derivative alone.

It is not automatically energy velocity. Equality holds in important lossless settings, but absorption, gain, strong spatial dispersion, or non-Hermitian mode structure can separate them.^[6]^[4]

It is not automatically signal or front velocity. Pulse peaks can be reshaped, and anomalous dispersion can yield negative or superluminal group velocities without moving a causal front faster than light.^[5]

It is not particle velocity. In quantum wave packets it can coincide with the classical velocity for particular dispersions, but that is a derived correspondence, not the general definition.

Scope of Application¶

Group velocity recurs anywhere linear or linearized waves admit dispersion branches. In optics it describes narrow-band pulse propagation through transparent media, fibers, waveguides, photonic crystals, and resonators. Written using refractive index \(n(\omega)\) for an isotropic transparent medium,

\[ v_g=\frac{c}{n+\omega\,dn/d\omega}, \]

subject to the assumptions behind a real dispersion relation.

In water-wave theory, group speed organizes propagation of wave energy and packets. In solid-state physics, gradients of electronic and phonon band dispersions give semiclassical carrier and wave-packet velocities. In quantum mechanics, a free nonrelativistic particle has \(\omega=\hbar k^2/(2m)\), hence \(v_g=\hbar k/m=p/m\). In atmospheric and geophysical fluids, group-velocity vectors indicate how wave packets and influence propagate even when phase crests travel differently.

The common abstraction is local geometry of a dispersion surface. The physical meaning of the derivative must be re-established for the chosen system; it should not be imported uncritically from transparent optics to dissipative or amplifying media.

Clarity¶

Group velocity makes three separations explicit. First, carrier motion and modulation motion are different observables. Second, the derivative belongs to a particular dispersion branch at a particular point, not to “the medium” without frequency and mode qualification. Third, the derivative's exact mathematical value and the envelope-translation interpretation have different logical status.

These separations resolve familiar paradoxes. A crest can enter at the rear of a water-wave group and disappear at the front because phase and envelope travel at different speeds. In an anisotropic crystal, (mathbf k) is normal to phase fronts while \(\nabla_{\mathbf k}\omega\) can point elsewhere, so phase normal and energy-ray direction need not coincide. A negative group velocity can describe motion of a reshaped peak rather than backward causal transmission.

Manages Complexity¶

A full packet evolution is a Fourier integral containing a continuum of components. Group velocity compresses its leading translational behavior into one local derivative. The central value \(\omega'_0\) predicts where the packet center should move; \(\omega''_0\) predicts leading spreading; higher derivatives refine the model only when needed. This derivative hierarchy turns an unwieldy integral into a controlled approximation.

Dispersion diagrams then become operational maps. Their slopes immediately show direction and speed, flat bands indicate small group velocity, sign changes indicate counter-propagating envelope motion, and curvature warns of distortion. The reduction is powerful because it makes its own limits visible: a steep or rapidly varying slope demands narrower bandwidth or a higher-order calculation.

Abstract Reasoning¶

Group-velocity reasoning supports these moves:

Differentiate the correct branch. Identify the physical mode and hold the proper medium parameters fixed before taking \(\nabla_{\mathbf k}\omega\).
Distinguish slope from ratio. Compare \(d\omega/dk\) with \(\omega/k\) rather than assuming packet and crest motion coincide.
Audit bandwidth. Check whether the spectrum samples a region over which the linear approximation is adequate.
Estimate distortion. Use (omega''), bandwidth, and propagation time to decide whether a single translated envelope is meaningful.
Trace direction. In several dimensions, follow the gradient normal to the constant-frequency surface, not necessarily the wavevector direction.
Separate observables. Declare whether the measurement tracks a peak, centroid, energy flux, information onset, or phase.
Protect causality. Do not infer superluminal information transmission from a superluminal or negative group velocity alone.

Knowledge Transfer¶

The exact derivative framework transfers across mechanical, electromagnetic, quantum, elastic, acoustic, oceanic, and plasma waves. What travels is the relation between spectral geometry and packet kinematics: local slope drives leading translation, curvature drives spreading, and branch topology constrains direction.

The interpretation transfers conditionally. Energy-velocity equality may hold in a conservative lossless problem but fail in a lossy periodic material. A quantum packet's group velocity may match (p/m), whereas a photonic-crystal packet follows a band gradient. The abstraction encourages reuse of the derivation while requiring the system-specific conservation, mode, and causality hypotheses to be rechecked.

Examples¶

Two nearby components¶

For two waves with \((k_1,\omega_1)\) and \((k_2,\omega_2)\), trigonometric addition produces a rapidly oscillating carrier multiplied by a slowly varying beat envelope. The beat velocity is

\[ \frac{\omega_2-\omega_1}{k_2-k_1}, \]

which tends to \(d\omega/dk\) as the separation shrinks. This finite-difference construction explains the name “group” velocity.

Deep-water gravity waves¶

For \(\omega=\sqrt{gk}\),

\[ v_p=\sqrt{g/k},\qquad v_g=\frac12\sqrt{g/k}=\frac12v_p. \]

Wave crests therefore move through the group, appearing at its rear and disappearing at its front. University of Texas notes and the Feynman Lectures give this standard example.^[7]^[8]

Free quantum particle¶

For \(\omega=\hbar k^2/(2m)\), group velocity is \(\hbar k/m=p/m\). The packet center follows the classical free-particle speed, while nonzero curvature (hbar/m) causes spreading. The example shows both the usefulness and incompleteness of the first derivative.

Affine dispersion¶

If \(\omega=ck+\omega_c\), every nearby component has group velocity ©, but the phase velocity is \(c+\omega_c/k\). A rigidly translating modulation therefore does not by itself imply equal phase and group speeds.

Anomalous dispersion¶

Near a resonance, rapid dispersion and attenuation can make a calculated group velocity exceed ©, diverge, or become negative. The packet peak is reshaped by interference and filtering; the result does not authorize faster-than-light signaling. Front and signal analyses are the correct causality tests.^[5]

Structural Tensions¶

Derivative exactness versus packet approximation. The slope of a dispersion curve is exact; identifying it with one observable envelope velocity is asymptotic and regime-bound.

Localization versus narrow bandwidth. A spatially tight packet requires a broad spectrum, while a clean single group velocity requires a narrow spectrum. Improving one condition can degrade the other.

Translation versus distortion. First-order dispersion moves the envelope; second and higher orders reshape it. Over long enough distances, a velocity alone no longer describes the packet.

Phase direction versus energy direction. In isotropic media the vectors often align. In anisotropic media the dispersion gradient and wavevector can point differently.

Operational convenience versus causal meaning. Group delay and peak motion are easy to measure, but pulse reshaping can detach them from information onset or energy transport.

Conservative equivalence versus dissipative ambiguity. Lossless systems often equate group and energy velocities. Loss and gain require complex dispersion and competing velocity definitions.

Structural–Framed Character¶

Group Velocity is moderately framed by wave physics.

Vocabulary travels: 0.45 framed. Dispersion slope, envelope, and group velocity recur across many physical wave domains but remain technical wave concepts.
Evaluative weight: 0.00 framed. The definition is descriptive and mathematical.
Institutional origin: 0.00 framed. No authority or convention fixes the mechanism beyond ordinary notation.
Human-practice bound: 0.00 framed. The abstraction applies to nonhuman physical systems.
Import versus recognize: 0.75 framed. Other fields recognize slope-driven transport, but the name group velocity arrives with wave-mode vocabulary.

Aggregate: 0.24 framed. The structure is highly general within wave physics, yet its literal identity presupposes dispersion relations, spectral modes, and packet interpretation, so it remains domain-specific rather than prime.

Structural Core vs. Domain Accent¶

The portable core is local derivative control: the slope of a spectral relation at an operating point determines leading movement, while curvature determines departure from rigid translation. This pattern resembles sensitivity analysis and linearization.

The domain accent is decisive. The independent variables are frequency and wavevector; the object is a superposition of wave modes; the observable is phase, envelope, energy, or signal propagation; and Fourier bandwidth controls validity. Removing those commitments leaves generic local linearization already represented elsewhere. Retaining them yields group velocity.

This entry presupposes Wave.

Group Velocity most directly presupposes Wave. A dispersion relation and its phase/envelope decomposition exist only after a wave mode, field or medium, and superposition regime are declared. The relation is compositional rather than taxonomic, because a velocity is not itself a propagating disturbance.

It relates closely to Derivative and Local Linearization through \(\nabla_{\mathbf k}\omega\), and to Approximation through the narrow-band expansion. These generic operations do not provide the wave-specific observable or its physical boundaries.

Wave Packet is the closest domain-specific neighbor. It is the localized object whose motion group velocity helps describe, and its entry mentions the formula and spreading. That embedded role does not exhaust group velocity's independent branch geometry, multidimensional direction, energy-velocity conditions, lossy-media ambiguity, or causality boundary, so neither Wave Packet nor any combination of these neighbors covers the subject exactly.

Relationships to Other Abstractions¶

Current abstraction Group Velocity Domain-specific

Parents (1) — outgoing typed relations

Group Velocity presupposes Wave Prime

Group Velocity most directly presupposes Wave.

Children (1) — abstractions linked here by typed relations

Group-Velocity Dispersion Domain-specific presupposes Group Velocity

The local frequency derivative of inverse group velocity presupposes a selected wave mode and its group-delay relation; it is not a velocity.

Hierarchy path (1) — routes to 1 parentless root

Group Velocity → presupposes Wave

Neighborhood in Abstraction Space¶

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

Family — Unclustered & Miscellaneous (2769 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-25

Not to Be Confused With¶

Phase velocity: \(\omega/k\), the motion of constant phase.
Wave packet: the localized superposition whose envelope may move at group velocity.
Dispersion relation: the full function \(\omega(\mathbf k)\), not its local gradient.
Group delay: derivative of transfer phase with respect to frequency.
Group-velocity dispersion: frequency dependence of group delay or second-order dispersion.
Energy velocity: energy flux divided by energy density; equal only under stated conditions.
Signal velocity: speed of a usable signal feature under a chosen definition.
Front velocity: speed of the earliest nonanalytic onset, central to causality.
Particle velocity: material or quantum-particle motion, which may coincide only in special cases.
Phase delay: carrier-phase shift divided by angular frequency.
Wave equation: governing differential equation; it may generate a dispersion relation but does not equal its slope.
Velocity addition: relativistic composition rule for object or frame velocities, not spectral differentiation.

References¶

[1] MIT OpenCourseWare, “Lecture 13: Dispersive Medium, Phase Velocity, Group Velocity,” 8.03SC Physics III: Vibrations and Waves (2016), https://ocw.mit.edu/courses/8-03sc-physics-iii-vibrations-and-waves-fall-2016/resources/mit8_03scf16_lec13/. registry ↩

[2] Gang Chen, Nano-to-Macro Transport Processes, MIT OpenCourseWare lecture notes, discussion of wave packets, group velocity, and energy velocity, https://ocw.mit.edu/courses/2-57-nano-to-macro-transport-processes-spring-2012/resources/2e4ecaa5cf55f03bcefbc8ccce79aed6_mit2_57s12_lec_notes_2004/. registry ↩

[3] Léon Brillouin, Wave Propagation and Group Velocity (Academic Press, 1960), ISBN 9780121349680.not authoritative registry ↩

[4] P. Y. Chen et al., “Group Velocity in Lossy Periodic Structured Media,” Physical Review A 82 (2010): 053825, https://doi.org/10.1103/PhysRevA.82.053825. registry ↩a ↩b

[5] Robert Fox, C. G. Kuper, and S. G. Lipson, “Faster-than-Light Group Velocities and Causality Violation,” Proceedings of the Royal Society A 316 (1970): 515–524, https://doi.org/10.1098/rspa.1970.0093. registry ↩a ↩b ↩c Show verification details

[6] Steven G. Johnson, “Group Velocity and Energy Velocity,” MIT 18.369 Nanophotonics course materials (2016), https://math.mit.edu/~stevenj/18.369/spring16/. registry ↩ Show verification details

[7] Richard Fitzpatrick, “Gravity Waves in Deep Water,” University of Texas at Austin, https://farside.ph.utexas.edu/teaching/336L/Fluidhtml/node148.html. registry ↩

[8] Richard P. Feynman, Robert B. Leighton, and Matthew Sands, “Waves,” The Feynman Lectures on Physics, vol. I, ch. 51, https://www.feynmanlectures.caltech.edu/I_51.html. registry ↩ Show verification details