Researchers at the University of Witwatersrand, South Africa, encoded data using a parameter that remains unchanged when light passes through turbulence, showing that their scheme can both increase bit rate and reduce crosstalk compared with encoding based directly on orbital angular momentum. [Image: Wits University]
Technology that exploits the spatial profile of light beams to encode data has the potential to boost optical bandwidth by many orders of magnitude. But such schemes are vulnerable to noise, with imperfections in optical fibers or atmospheric turbulence limiting their effective range—in the case of free space, to as little as a few hundred meters.
Now researchers in South Africa have shown, paradoxically, how communicating parties could transmit such structured light at high bit rates with minimal distortion by avoiding having to detect the structures themselves (Laser & Photonics Rev., doi: 10.1002/lpor.202370027). The trick is to exploit a beam’s “vectorness”—a kind of relative rather than absolute measure of light’s spatial properties whose encoding fidelity is limited only by the sensitivity of receiver detectors.
Susceptibility to distortions
Space- and mode-division multiplexing enhance rates of data transmission by encoding data in light’s spatial properties, such as multiple spatial modes. One promising approach to such multiplexing exploits orbital angular momentum (OAM). Whereas spin angular momentum is manifested as a light beam’s polarization, OAM instead is seen as a spiraling of the beam’s wavefront as it propagates—with distinct modes, in this case, corresponding to different degrees of twistedness.
Unfortunately, this and other types of spatial multiplexing are susceptible to transmission distortions. In optical fibers, stress and other imperfections introduced during manufacturing or bending of the fiber while it is being deployed can merge what should be distinct modes. While in the atmosphere, changes in air density caused by pressure and temperature fluctuations prompt variations in the refractive index that a light beam experiences along its path.
Vectorness from 0 to 1
To get around these problems, Andrew Forbes, Keshaan Singh and colleagues at the University of Witwatersrand have shown how to exploit a parameter known in quantum information science as concurrence, but which they term vectorness. This is a measure of how hard it is to separate a light beam’s polarization pattern from its intensity pattern.
The parameter can take on a spectrum of values between 0 and 1, with the latter extremity designating a perfect vector beam in which two orthogonal polarizations, such as horizontal and vertical, each have a unique spatial mode. A vectorness of 0 instead corresponds to a scalar field, in which the spatial modes are identical and thus can be factorized out from the two polarizations.
Forbes and colleagues used vectorness because it remains unchanged even as light crosses a turbulent medium.
Forbes and colleagues used this quantity because it remains unchanged even as light crosses a turbulent medium—the orthogonal spatial modes are combined in such a way that the function remains invariant under the unitary transformation characterizing atmospheric distortion. Since vectorness is tuned by adjusting a beam’s amplitude, the invariance means that even small tuning steps should yield distinct values of the parameter when the beam is measured by a receiver.
And smaller steps, in turn, mean more modes—enough modes that each consecutive beam (in the form of a laser pulse) can encode a string of data comprising multiple bits. This contrasts with conventional amplitude modulation, which is restricted to just a single bit—either on or off—thanks to the distorting effects of transmission.
Almost complete fidelity
To put this encoding scheme to the test, Forbes and coworkers used an interferometer to independently modulate the horizontal and vertical components of a diagonally polarized plane wave via a pair of holograms. By varying the incoming beam’s amplitude, they were able to adjust the efficiency of one hologram compared with the other, thereby tuning the vectorness. After directing the beam through a short channel of hot air to induce turbulence, they used diffraction gratings to separate the different polarizations and then imaged each of them with a CCD to measure their Stokes intensities. Combining those intensities yielded a value for the transmitted vectorness.
By sticking to 3 bits, they could transmit a black-and-white image (of a statue head) with almost complete fidelity (99.73%) despite the distortion.
The researchers first studied the effect of turbulence on beams with different tuning steps. For just eight modes—corresponding to 3 bits—they found no distortion as the beams traveled through the hot air. As they increased the bit number, some crosstalk did creep in—reaching 17% for 5 bits—but they also showed that, by sticking to 3 bits, they could transmit a black-and-white image (of a statue head) with almost complete fidelity (99.73%) despite the distortion.
Then the researchers used the same experimental setup to compare the performance of their scheme with that of OAM encoding. As they point out, higher OAM modes require larger apertures. This restricted the number of modes in that case to just 11. In contrast, they used as many as 50 modes for tests of vectorness. Even then, they recorded higher fidelities under the effects of turbulence—about 85%, compared with the roughly 60% with OAM.
What’s more, the researchers showed that the reduced fidelity in their scheme was almost entirely due to detector limitations—it dropping a mere 0.2 percentage points when they directed the beam through the hot air, compared with a turbulence-free passage. This, they say, opens up the possibility of high-speed optical communication without adaptive optics or computer-based correction procedures.
Swapping CCD cameras for photodiodes?
Forbes and coworkers add that the scheme provides a number of other benefits, including its suitability for stressed optical fiber and other media that scramble polarization (the redundancy of information at the output meaning that sender and receiver can use different specific polarizations). But they acknowledge the difficulty of scaling—the need for an exponential increase in the resolution of both emitter and receiver as the number of bits rises. They reckon, however, that this will not prove an insurmountable hurdle, arguing that swapping out their CCD camera for fast, sensitive photodiodes is “all that is required for a real-world implementation.”
He also says that he and his colleagues have started experiments in more realistic settings—using both spooled fiber and a free-space link 260 m long.
Forbes points out that these photodiodes are “superfast and supercheap”—being used routinely in telecommunications today. The catch is that they need to be employed in conjunction with similarly fast optical modulators at the input, which he says are commonplace in communication labs but are not present in his own lab.
But, he explains, his group has collaborators “now testing exactly this.” He also says that he and his colleagues have started experiments in more realistic settings—using both spooled fiber and a free-space link 260 m long. He adds that vectorness could also be exploited in other kinds of noisy channel, such as underwater communication between submarines—light, in this case, offering a line-of-sight link with radio silence.
