Cosmic String

What kind of object is this

The universe is full of zero-dimensional objects: particles. We are made of them. Astronomy is mostly about three-dimensional objects: clouds, stars, galaxies. Cosmic strings are something different — they are one-dimensional defects: a thread of frozen-in symmetry mismatch, threading through the present-day vacuum like a runner in stocking that nothing can repair.

The closest everyday analogue is a vortex line in liquid helium-4, or a flux tube in a type-II superconductor: a singular line around which a phase wraps by 2π. In both cases the line itself is a stable, energetically expensive structure that cannot be smoothly contracted away because of topology, not dynamics. Cosmic strings are the cosmological generalisation. They live in the spacetime continuum rather than a lab dewar, but the mathematics is the same.

The discriminator from any "ordinary" astrophysical object is that the energy per length is set by very high-energy physics — typically the energy scale η of the broken symmetry that produced it. For a Grand Unified Theory (GUT) string at 10¹⁶ GeV, μ ~ η² ~ 10²² g/cm in natural units. That's about the mass of an asteroid per centimetre of length. A galactic-scale string would weigh as much as a galaxy.

The Kibble mechanism — how strings get made

In 1976 Tom Kibble proposed that any phase transition in the early universe involving a spontaneously broken symmetry should leave behind topological defects. The argument is irresistible. Before the transition, the order parameter (a complex Higgs-like scalar field, say) sits at the unstable maximum of its potential. As the universe cools below a critical temperature, it must roll down into the trough — but the trough is a circle (in the case of a broken U(1) symmetry), and the choice of which point in the trough to settle at is made independently in each causally disconnected patch.

By causality, no patch larger than the horizon at the time of transition can agree on a single value. So the field freezes in with random phase angles, varying spatially. Where the field winds around the trough an integer number of times around a closed loop in space, you cannot interpolate continuously to a single value of the order parameter — the field must, somewhere inside that loop, return to the unstable maximum. That's the string: a line along which the symmetric, false-vacuum state is trapped.

For other symmetry groups, defects of different dimensionalities are predicted. Z₂ symmetry breaking produces domain walls (two-dimensional); broken U(1) produces strings (one-dimensional); broken SU(2) produces monopoles (zero-dimensional); broken SU(3) and higher can produce textures (no defects but topological field configurations). Cosmologically, monopoles are very dangerous (they would dominate the universe's mass density unless inflation dilutes them away), domain walls are similarly catastrophic, and only strings sit in the "Goldilocks" zone of being potentially observationally allowed.

The single number that matters: Gμ/c²

Everything observationally relevant about a cosmic string depends on a single dimensionless number,

Gμ/c² = G η² / c⁴   (order of magnitude)

where μ is the linear-mass density and η is the symmetry-breaking scale. Concrete benchmarks:

Scale	η (GeV)	Gμ/c²	Status (2026)
Grand Unified Theories	~10¹⁶	~10⁻⁶	Ruled out by CMB and pulsar timing
Intermediate-scale strings	~10¹³	~10⁻⁹	Borderline / disfavoured by Planck
NANOGrav-allowed strings	~10¹¹–10¹²	~10⁻¹¹	Compatible with 2023 PTA hint
Electroweak strings	~10²	~10⁻³³	Topologically unstable; never form a network

The trend is unambiguous: every new observational window — Planck CMB (2015, 2018), Planck non-Gaussianity, NANOGrav 15-year data (2023), and direct lensing searches — has tightened the upper bound on Gμ/c². Strings at the GUT scale, popular in the 1980s, are now firmly excluded. Strings at intermediate scales are alive but cornered. Below 10⁻¹¹ the network is still allowed by everything we can measure.

Why the geometry is conical

Around a thin, straight, static string, general relativity gives a metric that locally looks like flat Minkowski but has a global identification: the spatial slice is a flat cone with deficit angle

δ = 8πGμ/c²

Picture a flat sheet of paper, cut out a thin wedge, and glue the two cut edges together. The resulting cone is flat everywhere except at the tip — and yet the geometry is non-trivial. Walk in a closed loop around the tip and you come back to your start having "turned" through 2π − δ instead of 2π. There is no gravitational force pulling you toward the tip; you measure no curvature anywhere except at the singular line. But the cone's circumference is shorter than 2πr.

This is the central, beautiful surprise of cosmic-string gravity: strings do not attract. A particle at rest beside a string stays at rest. Two parallel strings do not pull on each other. The metric is locally flat. What strings do instead is steal a wedge of space.

Gravitational lensing — the double-galaxy signature

If the deficit wedge of a string passing between us and a distant galaxy happens to cross our line of sight, the observer sees two identical images of that galaxy. The pair is separated by an angle

θ = 8πGμ/c² × D_LS / D_S

where D_LS is the angular-diameter distance from lens to source and D_S is the distance from observer to source. For Gμ/c² ~ 10⁻⁷, θ is a few arcseconds — comparable to but distinct from cluster-scale lensing.

The signature is unmistakable in principle. Unlike lensing by an extended mass distribution, string lensing produces:

• Two identical images. No distortion, no magnification, no shear. The two images are perfect copies.
• Zero time delay. Both paths have the same length (just routed around different sides of the deficit wedge).
• Straight separation lines. Pairs caused by the same string lie along a line on the sky tracing the string's projection — many candidate pairs in a row.
• Sharp boundary. The lensing turns on and off across the deficit wedge — a small spatial offset moves a source completely out of the lensed region.

This unique combination drove a flurry of excitement when CSL-1, an apparent double galaxy in the CAPODIMONTE Deep Field, was claimed in 2003 as a cosmic-string candidate by Sazhin and collaborators. Follow-up Hubble imaging in 2008 resolved CSL-1 into two genuinely different elliptical galaxies that happened to look similar at lower resolution. No confirmed string-lensing pair has been found since.

Kaiser-Stebbins — strings as step discontinuities on the CMB

A moving cosmic string drags spacetime with it. If a string is translating at velocity v transverse to the line of sight, and CMB photons stream past it, photons on the leading side of the deficit wedge are Doppler-blueshifted relative to those on the trailing side. The result is a step discontinuity in CMB temperature across the string's projected location:

ΔT / T = 8πGμβγv/c    (Kaiser-Stebbins 1984)

where β = v/c and γ is the Lorentz factor. For typical string velocities β ~ 0.3 and Gμ/c² ~ 10⁻⁸, the step is ΔT/T ~ 10⁻⁸ — at the edge of what Planck and ground-based experiments can resolve. The line-like geometry is the key discriminator: a step function in temperature, straight on the sky, would be unmistakable. Searches in Planck and CMB-S4 forecast data so far show no convincing detection. The non-detection translates to a bound Gμ/c² ≲ 10⁻⁷ to 10⁻⁹ depending on the assumed network properties.

Stochastic gravitational waves — the strongest current constraint

A cosmological network of cosmic strings inevitably produces gravitational waves. Long strings (longer than the horizon) intercommute and chop themselves into loops; loops oscillate at speeds approaching c; oscillating loops radiate gravitationally with power roughly

P_GW = Γ G μ² c     Γ ≈ 50

Cusps — points where the string momentarily moves at the speed of light — radiate especially efficiently in sharp bursts, and kinks — sharp corners where intercommutation has joined two segments — contribute a smoother background. Integrated over cosmic history, the result is a stochastic gravitational-wave background with a roughly flat (or slowly rising) spectrum extending across many decades in frequency.

This is observable. Pulsar timing arrays (NANOGrav in the US, EPTA in Europe, PPTA in Australia, and the joint International PTA) monitor millisecond pulsars at nanohertz frequencies; LIGO/Virgo/KAGRA cover the hectohertz band; LISA, once flying, will cover millihertz. The 2023 NANOGrav 15-year data set yielded a 3-4σ detection of a stochastic background at frequencies of a few nanohertz. The conventional astrophysical interpretation is a supermassive black hole binary population — but the data are also fit by cosmic-string network models with Gμ/c² ~ 10⁻¹¹ to 10⁻¹⁰. The string interpretation is not statistically preferred, but it is not excluded by NANOGrav alone; future data and complementary observations (lensing, CMB) will discriminate.

Why strings are not the seeds of structure

In the 1980s and early 1990s cosmic strings were a leading alternative to inflation as the source of the primordial density perturbations that eventually grew into galaxies and clusters. A string network seeds density perturbations on the same horizon scale at every epoch — a so-called "active seed" model. The arithmetic of how those perturbations grow into the cosmic web seemed promising.

What killed strings as the dominant structure seed was the CMB. The angular power spectrum of CMB temperature anisotropies has a series of acoustic peaks at multipoles ℓ ≈ 220, 540, 800, ... predicted precisely by inflation: a coherent, super-horizon set of primordial perturbations enters the horizon and starts oscillating in the photon-baryon fluid, producing standing-wave peaks. Strings, by contrast, are causally generated at the horizon — perturbations from strings are incoherent and produce a smooth, peakless contribution to the angular power spectrum. COBE in 1992 ruled out strings as 100% of the seeds; WMAP in 2003 confirmed the picture; Planck 2015 and 2018 limit the string contribution to at most a few percent.

So strings are no longer the engine of structure formation. But they may still contribute. The current constraint on the string fraction of the CMB anisotropy power is ≲ 1–3% depending on model details. A subdominant string network at Gμ/c² ~ 10⁻¹¹ is consistent with everything we observe and would produce wakes — sheet-like density enhancements trailing fast-moving strings — that could leave imprints in 21 cm surveys and dwarf-galaxy populations.

How to tell strings from inflation

Observable	Inflation prediction	Cosmic-string prediction
CMB temperature power spectrum	Sharp acoustic peaks (Doppler/Sachs-Wolfe)	Smooth, peakless contribution
Spatial features in CMB	Gaussian random field	Line-like step discontinuities (Kaiser-Stebbins)
Non-Gaussianity	Very small f_NL ~ 1	Large at sharp features; non-trivial bispectrum
B-mode polarisation	Tensor modes from inflation; r ≲ 0.06 (2025 limits)	Vector modes; distinct angular spectrum
GW spectrum	Nearly scale-invariant, low amplitude	Roughly flat or rising across many decades; cusp bursts
Lensing morphology	Smooth distortion; magnification + shear	Double, undistorted images along a line

No single observation has decided the question definitively, because the inflation hypothesis explains the bulk of the data and strings are sub-dominant. But every observation tightens the noose: Planck, BICEP/Keck, NANOGrav, and future LiteBIRD, CMB-S4, SKA, and LISA experiments will between them either detect a string network or push the bound to Gμ/c² ≲ 10⁻¹².

A note on terminology — cosmic strings vs string theory

"Cosmic strings" and the "fundamental strings" of string theory are different objects historically, but the boundary has blurred. The cosmic strings discussed here are solitonic defects: classical solutions of a field theory, macroscopic in extent, with energy scale μ. They are well-defined in any field theory with the right symmetry-breaking pattern, with no reference to extra dimensions or strings of quantum gravity. The fundamental strings of perturbative string theory are very different — they are quantum objects at the Planck scale and do not normally appear as cosmologically extended structures.

Where things get interesting is in flux-compactified string theory and in models of cosmic superstrings. In such theories, the fundamental F-strings and D-brane wrapping modes can produce stable, macroscopic, cosmologically stretched strings — "cosmic superstrings". These have all the observational signatures of solitonic cosmic strings, plus a couple of extras: F- and D-strings have different tensions and can form bound states with quantised junctions where three strings meet. A non-detection of stochastic GW or Kaiser-Stebbins steps with the right junction statistics would constrain whole classes of string-theory cosmologies.

Real-world searches, by year

• 1976 — Kibble. The original paper proposing topological defect formation in cosmological phase transitions.
• 1980s — Vilenkin, Turok, Albrecht. Cosmic strings developed as the leading alternative to inflation. Simulations of string-network evolution converge on a scaling solution.
• 1984 — Kaiser & Stebbins. The CMB step-temperature signature.
• 1992 — COBE. First CMB anisotropy detection; large-scale power consistent with inflation; strings disfavoured but not excluded.
• 2003 — CSL-1 candidate. Apparent double galaxy discovered in CAPODIMONTE deep field; later resolved as two distinct galaxies by Hubble in 2008.
• 2013, 2015, 2018 — Planck. Successive papers tightening the string contribution to the CMB to ≲ 1–3%, bounding Gμ/c² ≲ 10⁻⁷.
• 2017 — LIGO O1/O2. Direct gravitational-wave non-detection of string bursts at hectohertz frequencies sets Gμ/c² ≲ 10⁻⁸ for some loop distributions.
• 2023 — NANOGrav 15-year. First evidence of a nanohertz stochastic gravitational-wave background. Fitted equally well by SMBH binaries and by cosmic-string networks with Gμ/c² ~ 10⁻¹¹.
• 2024–present — SKA, LISA, CMB-S4 forecasts. Future experiments designed to either detect or push Gμ/c² below 10⁻¹².

Common pitfalls

• Confusing cosmic strings with string-theory strings. Cosmic strings are macroscopic field-theory solitons; string-theory strings are microscopic quantum objects. Only in special cosmological models do the two pictures merge ("cosmic superstrings").
• Thinking strings exert a gravitational pull. They don't, in the Newtonian sense. The geometry around a static string is locally flat — there is only the global deficit. Mass falls toward strings only via the conical wake effect when relative motion is involved.
• Conflating the deficit angle with a curvature. The deficit is concentrated at the string itself; outside the string the spacetime is flat. It's a defect, not a smooth curvature field.
• Forgetting strings are subdominant. Inflation is the established account of structure formation. A discussion of strings always means "in addition to" the inflationary picture, not "instead of."
• Reading too much into the NANOGrav hint. The 2023 detection is statistically robust as a stochastic GW background, but the source identification is unresolved. SMBH binaries are the conservative interpretation; cosmic strings are an allowed alternative; new physics is one of several other options.