Magnets placed beside a common type of optical fibre can change the light inside it, engineers at Wits University have found. The effect went unnoticed for decades because of a shortcut in fibre theory that dates to the 1970s.
The work, by MSc student Mohammed Raiyan Sharif and Mitchell Cox, who heads the Wits Optical Communication Lab, has been published in the journal Optics Letters. It could lead to magnetic sensors that are easier to install and, in the longer term, new ways of computing with light.
Optical fibre is best known for carrying internet traffic, but it is also used to measure magnetic fields, such as those around power lines and substations. These sensors rely on an effect Michael Faraday discovered in 1845: a magnetic field running along a fibre slightly twists the direction in which the light wave vibrates. Measure the twist and you know how strong the current is.
The catch is that the fibre usually has to be wrapped around the conductor, which can be awkward to install. Because the “transverse Faraday effect” responds to a field coming from the side, a future sensor could simply lie next to the source, such as a fibre alongside an electrical cable in a trench.
It wasn’t hard to 3D print a rig to test our intuition. Working out the theory for what we were seeing was the challenging part.
MSc student Mohammed Raiyan Sharif
Wits cautions that the magnetic fields near most cables are much weaker than those used in the lab, so it is not yet clear how sensitive a practical sensor could be.
In the multimode fibre the team tested, light travels along many paths, bouncing off the mirror-like edge of the glass core. At the far end, it forms a random-looking pattern of bright and dark spots, rather like the view through a kaleidoscope. Sharif and Cox filmed this pattern while removing magnets, pair by pair, from a rail beside the fibre. The pattern changed.
“It wasn’t hard to 3D print a rig to test our intuition. Working out the theory for what we were seeing was the challenging part,” says Sharif. “The key was to return to the underlying equations and question assumptions that are usually taken for granted in fibre optics.”
The effect was missed because of the simplified model engineers use to describe light in fibre. Light in a fibre vibrates almost entirely across the strand; a tiny part vibrates along it, and the standard model leaves that part out. But where the glass of the core differs sharply from the glass around it, that part matters near the core’s edge – and it is exactly what a sideways magnetic field acts on.
The experiments matched the team’s theory. In the strongest result, the before and after patterns were only about 78% alike. When the light was set to vibrate in line with the magnetic field, there was no measurable change, as the theory predicted.
“Often these things are right under our noses, hidden in the assumptions we have made for decades,” says Cox.
Computing with light
The bigger prize may be control. If an electrical coil replaced the permanent magnets, the way light travels through the fibre could be adjusted electrically – like reshaping a kaleidoscope’s mirrors while looking through it.
“Besides sensing, I think that being able to control different modes [patterns of light] inside the fibre electrically could unlock a new way to do optical computing,” says Cox. “We can model the way light gets distorted when it propagates through fibres, frosted glass windows or even the atmosphere. If we change our perspective on those models, we can actually use light to do mathematical computations for us, for free.”
The lab already uses light-based computing to extend the range of wireless optical links, which beam data through the air on lasers. The aim is affordable “fibre” without the fibre.
Elsewhere at Wits, the Structured Light Laboratory last month unveiled a neural network that runs on light.
The research was supported by the CSIR. — (c) 2026 NewsCentral Media
