Scientists in China have shown how to switch between two types of light vortex known as skyrmions by directing differently shaped laser beams at a metasurface to generate terahertz pulses (Optica, doi: 10.1364/OPTICA.578501). They reckon that their scheme could be used for robust encoding of wireless data streams in the wide but under-exploited terahertz region of the electromagnetic spectrum.

Skyrmions in free space

Skyrmions are stable vortex-like structures whose topology remains intact even when subject to significant environmental noise. They were originally discovered as persistent disturbances in magnetic domains, suggesting a new way to store digital data. But more recently they have also been observed in electromagnetic phenomena, manifesting themselves as soliton-like pulses of light with a definite, unchanging topology.

To date, however, researchers have concentrated on generating these optical skyrmions in nonpropagating systems such as evanescent waves. The latest work instead looks to free space and the possibility of generating and switching between different topological states of electromagnetic beams. Its aim in particular was to show how the stability of such structures could be used to encode light pulses in terahertz waves, which have frequencies hundreds of times higher than the radio waves used in 5G wireless networks but which are more susceptible to interference from atmospheric turbulence.

Looking to nonlinear metasurfaces

The work has been carried out by Xueqian Zhang and Jiaguang Han of Tianjin University, Yijie Shen of Nanyang Technological University and colleagues in China, Singapore and the United States, who have exploited what is known as a nonlinear metasurface. Metasurfaces are artificial two-dimensional materials containing arrays of sub-wavelength structures to guide and shape beams of light. In this case, the metasurface consisted of more than 5 million horseshoe-shaped pieces of gold, each measuring about 200 nanometers across and arranged in two distinct but interwoven patterns on an 8-nm-thick layer of indium tin oxide.

The ring-shaped beam produced magnetic doughnuts threaded with electric current, yielding what are known as electric skyrmions.

The researchers used the metasurface to generate two types of toroidal light pulse—single-cycle electromagnetic pulses in the shape of a doughnut. They did so by directing an infrared laser beam at the metasurface, having structured the beam either in a ring or radial formation. The ring-shaped beam produced magnetic doughnuts threaded with electric current, yielding what are known as electric skyrmions. The radial pulses instead produced doughnut-shaped electric currents threaded with magnetic fields, resulting in magnetic skyrmions.

Zhang and colleagues switched between the two different beams—and associated skyrmions—by placing a half-wave plate and a vortex half-wave retarder between the laser and metasurface (rotating the plate to change the beam's polarization, which resulted in two different structures from the retarder). They then scanned the terahertz pulses at different points in time and space to reconstruct the evolving pulse fields, which they say allowed them to clearly identify the two different types of skyrmion and demonstrate that they could switch easily and reliably between the two.

Toward practical use

The researchers envision in future switching between the two types of skyrmion at high rates to encode the 0s and 1s that make up digital data. Other characteristics that could be varied to enhance bit rate, they suggest, are the pulse intensity and purity of the skyrmion topology.

However, Zhang and colleagues point out that the technology is not yet practical. For one thing, they say, they need to improve the efficiency with which they convert the input infrared laser beam into terahertz pulses. In the current experiment, they were only able to produce about one terahertz photon for every 100 million infrared photons. They might be able to increase this ratio, they say, by making a metasurface from a different material—one that could withstand higher beam intensities before getting damaged (to increase the input photon count) and which might also have better nonlinear characteristics (to boost efficiency). Potential candidates for such a material, they add, include spintronic emitters and photoconductive antennas.

They add that they also need to make the system more stable, with a larger detection sensitivity and signal-to-noise ratio, as well as smaller and more robust. They are confident that they can overcome these varied challenges but are reluctant to speculate as to when the technology could be commercialized; “That is really hard to say currently,” notes Zhang.