In 1976, Tom Kibble asked what the early universe would leave behind as it cooled through a series of phase transitions. Water freezing into ice does not crystallise uniformly. Separate regions choose their own orientation, and where those choices disagree, the lattice is forced into defects: seams and dislocations that persist long after the transition ends. Kibble realised that the vacuum of quantum field theory might behave the same way.

Suppose a symmetry of the fundamental forces broke as the cosmos expanded and cooled. Causally disconnected regions would have had no way to coordinate their choice of vacuum state. Where incompatible choices met, energy could become trapped in thin, stable filaments of the unbroken phase. These are cosmic strings. Each is far narrower than a proton, yet a single string could stretch across the observable universe. At grand unification scales, each would carry roughly a billion billion tonnes per metre of length.

For a time, cosmic strings were serious candidates for seeding the galaxies themselves. That role has been ruled out, yet the strings have not disappeared from cosmology. They have become something subtler: a probe. Whether or not they exist, the search for them translates directly into constraints on energy scales no particle accelerator will ever reach. In this sense, the sky becomes a laboratory for the first fraction of a second.

At high temperatures, many field theories possess symmetries that are hidden at low energies. The canonical example is a complex scalar field whose potential is shaped like the bottom of a wine bottle. Above a critical temperature, the field sits at the central peak, symmetric under rotations of its phase. As the universe cools, the minimum shifts to a circle of equivalent states, and the field must settle somewhere on that ring. The choice is arbitrary. Nothing in the physics prefers one point over another.

The crucial ingredient is causality. During the transition, the field's phase can only be correlated over distances smaller than the horizon. In practice, the correlation length is much shorter, set by the dynamics of the transition itself. Separate domains therefore select phases independently. Now trace a closed loop through several domains. If the phase winds through a full 2π as you travel around, continuity demands that somewhere inside the loop the field cannot sit on the vacuum ring at all. It is trapped at the symmetric peak, and that trapped region extends as a line.

This is a topological statement, not a dynamical accident. Strings form whenever the vacuum manifold contains non-contractible loops, which mathematically means its first homotopy group is non-trivial. Analogous classifications predict domain walls when the vacuum is disconnected and monopoles when it contains non-contractible spheres. Because the winding cannot be undone by smooth deformation, the string cannot simply dissolve. It can only close on itself as a loop or stretch to infinity. Its tension, μ, scales roughly as the square of the symmetry-breaking energy η.

None of this is purely speculative mathematics. Abrikosov vortices thread type-II superconductors, carrying quantised magnetic flux through precisely this kind of winding. They are the condensed-matter cousins of the Nielsen-Olesen string solution in particle physics. Wojciech Zurek extended Kibble's argument to laboratory transitions, predicting how defect density should depend on cooling rate. Experiments in superfluid helium, liquid crystals, and cold atomic gases have since observed defects forming in ways broadly consistent with Kibble-Zurek scaling.

The cosmological picture carries an important caveat. Grand unified theories that produce strings often also produce monopoles, which would overclose the universe. Inflation is the standard remedy, but it also dilutes any defects formed before it. Strings therefore survive only if they form during or after inflation's end, as happens naturally in hybrid inflation models. String theory adds another possibility. Brane inflation can produce fundamental strings and D-strings stretched to cosmic scales. These so-called cosmic superstrings would interact in distinctive ways.

Takeaway

Cosmic strings are what happens when geometry meets causality. When distant regions cannot coordinate, the topology of their independent choices can lock energy into the fabric of space permanently.

A straight cosmic string does not attract nearby matter in the familiar Newtonian sense. Its tension equals its energy density, and in general relativity the two contributions cancel, leaving spacetime locally flat. Globally, however, the geometry is conical. A wedge of angle δ = 8πGμ is removed from the space around the string, and its edges are joined together. The relevant dimensionless parameter is Gμ/c². For grand unification scales, this sits near one part in a million.

This missing wedge produces a remarkably clean lensing signature. A galaxy behind a string appears twice, with the two images separated by an angle comparable to the deficit, a few arcseconds for GUT-scale strings. Because the space on either side remains flat, neither image is magnified or distorted. A string crossing the sky would leave a chain of such identical pairs along its path. One candidate, CSL-1, generated considerable excitement in 2003 until Hubble imaging revealed two distinct galaxies. Even the cleanest predictions still demand careful observation.

Moving strings also imprint themselves on the cosmic microwave background through the Kaiser-Stebbins effect. As a string sweeps across the line of sight, photons passing on either side receive relative Doppler boosts. This creates a sharp step in temperature of order 8πGμ multiplied by the string's transverse velocity. The result is a network of line discontinuities: non-Gaussian, edge-like features unlike the smooth Gaussian fluctuations produced by inflation. Strings also generate B-mode polarisation through vector and tensor perturbations, offering a second CMB channel.

The richest signal may come from gravitational waves. Simulations show that a string network evolves toward a scaling solution, keeping a roughly constant number of long strings per Hubble volume. It does this by chopping itself into loops whenever strings intersect and reconnect. Those loops oscillate relativistically and radiate gravitational energy until they decay. Cusps, where a segment briefly approaches light speed, and kinks, where the string is sharply bent, produce characteristic bursts. Countless decaying loops across cosmic history combine into a stochastic background spanning many decades in frequency.

That broad spectrum is what makes strings such a valuable target. A single parameter, Gμ, sets the amplitude across every band. Pulsar timing arrays probe nanohertz frequencies, ground interferometers listen at hundreds of hertz, and future space missions will cover the range in between. Stephen Hawking noted in 1989 that a small fraction of loops might collapse entirely within their Schwarzschild radius, forming black holes. That would be yet another fingerprint these relics could leave. Each signature probes different physics, and agreement between them would be extraordinarily compelling.

Takeaway

A cosmic string is nearly invisible up close yet reshapes spacetime on large scales. Some of the most consequential structures reveal themselves only through the geometry they leave behind.

The first major constraint came not from searching for strings directly but from mapping what they could not explain. In the 1980s, strings were a leading alternative to inflation for seeding galaxies. Defects act as active, incoherent sources, continuously perturbing matter throughout cosmic history. Inflation instead sets up perturbations once, with coherent phases, producing a sequence of acoustic peaks in the CMB power spectrum. When BOOMERanG, MAXIMA, and then WMAP measured those peaks clearly, strings were decisively excluded as the primary seeds of structure.

What remains is the question of a smaller, subdominant contribution. Planck's temperature and polarisation data limit cosmic strings to at most a few percent of the observed anisotropy power. That translates into bounds of roughly Gμ below 10⁻⁷. The exact value depends on whether strings are modelled as idealised Nambu-Goto lines or as field-theoretic Abelian-Higgs solutions. Searches for Kaiser-Stebbins edges and non-Gaussian features in CMB maps yield comparable limits. Classic grand unification strings, with Gμ near 10⁻⁶, are therefore ruled out.

Gravitational waves have pushed the limits further. Pulsar timing arrays treat millisecond pulsars as a galaxy-sized detector. Their limits on the stochastic background constrain Gμ to roughly 10⁻¹⁰ or 10⁻¹¹, depending on assumptions about loop sizes and network dynamics. In 2023, NANOGrav and partner collaborations reported evidence for a nanohertz background, adding an intriguing twist. Supermassive black hole binaries remain the conventional explanation. However, metastable strings and cosmic superstrings with low reconnection probabilities can also fit the spectrum, and the question remains open.

At higher frequencies, LIGO, Virgo, and KAGRA have searched both for the stochastic background and for individual cusp and kink bursts. Their limits depend heavily on the model. Some loop distributions are constrained only to around 10⁻⁸, while others are pushed below 10⁻¹⁴. This spread is itself instructive. Progress now depends as much on understanding loop production in simulations as on improving detectors. Future instruments such as LISA, the Einstein Telescope, and Cosmic Explorer could reach Gμ as low as 10⁻¹⁷.

What do these numbers mean for fundamental physics? Because tension scales with the square of the symmetry-breaking scale, Gμ below 10⁻¹¹ pushes any string-forming transition below about 10¹⁴ GeV, well beneath grand unification. Several explanations remain. Perhaps such transitions never produced strings, or inflation erased them. Perhaps the strings are metastable, or the relevant physics lies at lower energies. Each possibility reshapes model building. A null result is not an empty result. It removes regions of theory space that no collider could ever reach.

Takeaway

In cosmology, a carefully measured absence can be as informative as a detection. Every tighter limit on cosmic strings tells us how the universe behaved at energies we will never recreate.

Cosmic strings occupy an unusual place in physics. They follow almost inevitably from symmetry breaking and topology, they have close analogues in laboratory systems, and they make predictions sharp enough to test. Yet after five decades, no confirmed detection exists, and the allowed tension has fallen by more than four orders of magnitude.

That history is not a story of failure. Each tighter bound has refined our picture of the early universe: which symmetries broke, when inflation ended, and how the vacuum settled into the state we now inhabit. Detected or not, strings give us a window onto physics at energies far beyond any experiment on Earth.

Perhaps the nanohertz background hides a string contribution, or perhaps LISA will hear the faint chorus of decaying loops. Or perhaps the universe cooled smoothly, leaving no scars at all. Either answer would tell us something profound about the conditions from which everything we see emerged.