How does light stay inside a glass fibre?

In 1966, Charles Kao and George Hockham were investigating an unpromising material for long-distance communication: glass. Light could already be guided along fibres, but too much of it disappeared on the journey. A strand that demonstrated an optical effect was not necessarily a useful telephone line.

Kao’s decisive argument was that much of the loss came from impurities that manufacturing could remove. The challenge was to make exceptionally transparent glass, not abandon glass altogether. The Nobel Committee’s account describes the work at Standard Telecommunication Laboratories.

That history reveals two separate questions hidden inside an optical cable. Why does the light follow the glass? And, once guided, how much useful signal survives? The first is a problem of boundaries. The second made fibre communication a demanding achievement in materials science.

Two transparent materials can make a guide

A conventional communications fibre has a central core, surrounded by cladding. Both are transparent materials. The crucial difference is their refractive index: the core’s is slightly higher. Refractive index tells us how light travels in a material. A higher index means a lower phase speed, the speed at which a wave’s peaks move. The index compares that speed with light’s speed in a vacuum, where there is no matter.

The cladding is therefore part of the optical system, not simply packaging. Protective coatings and the surrounding cable structure do a different job: they help the fragile fibre withstand handling and its environment. Light is not travelling through a hollow glass tube, and the cladding does not need a metallic mirror coating. The Fiber Optic Association introduces this core–cladding arrangement.

The difference between the two glasses creates conditions under which light can remain guided. To see why, begin with a single ray meeting their boundary.

The angle that changes what happens

When light reaches a boundary between materials, some can reflect and some can pass into the second material, changing direction through refraction. Consider a ray travelling from the higher-index core towards the lower-index cladding. To measure the angle, draw an imaginary line at right angles to the boundary. This is the normal. The angle between the incoming ray and that line is the incidence angle. As it increases, the ray passing into the cladding bends farther away from the normal.

At the critical angle, the transmitted ray would run along the boundary. Beyond that angle, an ideal lossless interface produces total internal reflection: the ray reflects back into the higher-index material. Two conditions matter together: light approaches from the higher-index side, and its angle from the normal exceeds the critical angle. OpenStax explains the geometry and its derivation from Snell’s law.

This angle convention often causes confusion. A ray almost parallel to the boundary has a large angle from the normal. “Large incidence angle” does not mean that the ray strikes the surface nearly head-on.

In the diagram, the illustrative indices are 1.48 and 1.46. Their critical angle is about 80.6°. The drawn ray arrives at about 83.5°, so it satisfies the reflection condition. These numbers describe the teaching example, not every commercial fibre.

What the zigzag picture leaves out

The familiar drawing of a ray bouncing along a fibre is useful, but light is an electromagnetic wave. A more complete description uses guided modes: patterns of the light wave that can travel along the fibre. Which patterns can travel depends on the core’s size, the light’s wavelength and how refractive index changes across the fibre. Wavelength is the distance between neighbouring wave peaks.

The light wave does not stop suddenly at the edge of the core. Part of its field extends into the cladding and becomes weaker farther from the core. That extension is compatible with guidance; it does not automatically mean the signal is escaping. Rüdiger Paschotta’s fibre-mode tutorial shows calculated field patterns and explains this distinction.

This also clarifies the name single-mode fibre. It does not mean that one photon travels alone, or that the fibre carries only one bit at a time. The word means that the fibre supports one basic pattern across its width. The ray picture is an introduction to the mechanism, not a literal photograph of the light inside a modern cable.

Keeping light inside is not the same as keeping it intact

Guided light still weakens. Absorption turns some light energy into other forms, such as heat. Scattering sends some light away from the useful guided path. Engineers call the reduction in optical power attenuation. Tight bends can introduce additional loss, so an optical path cannot be treated as an arbitrarily flexible line.

There is also dispersion, which spreads a light pulse in time. In multimode fibre, different modes can have different travel times. Single-mode fibre avoids that particular source of spreading. But different wavelengths within a pulse can still take different times to arrive, spreading the pulse out. These arrival-time differences are called group delays. A signal may therefore become both weaker and less sharply defined. The FOA’s optical-fibre guide distinguishes attenuation, bending loss and dispersion.

A receiver needs to recover information from what arrives. Merely showing that a ray remains guided does not establish how far, or at what data rate, a complete system will work.

From a boundary effect to a useful material

At Corning, Robert Maurer, Donald Keck and Peter Schultz pursued the manufacturing problem. Their work involved glass composition, purification and the difficulty of making the core and cladding into one usable fibre. In 1970, their team achieved the low-loss breakthrough that helped turn Kao’s proposal into a practical technology. Corning’s historical account records the team and its experimental progress.

The achievement joined two kinds of precision: controlling how the materials differed at their boundary, and controlling what unwanted material remained within the glass.

Try the interactive model below. At its initial 84°, the ray reflects internally. Lower the angle below about 80.6° and the model shows a transmitted branch entering the cladding. Some reflection also occurs below the threshold; the simplified display draws only the transmitted branch there.

Then return to how undersea cables carry the internet. A cable’s strength keeps its fibres physically safe; the optical design gives light a guided path. And as the companion story on bandwidth and delay explains, even an excellent path does not make the journey instantaneous. The remarkable thing is that a carefully made boundary can turn transparent glass into a route for information.