The computational density increases by orders of magnitude! The world's first photonic computing processor using polarization of light

Polarization is one of the fundamental properties of light in which information can be encoded and (de)multiplexed. In quantum information technology, polarization is the basis of quantum key distribution (QKD). In a paper published in Science Advances on June 15 [1], Oxford University researchers have developed the world's first photonic computing processor that uses the polarization of light to maximize Information storage density and computational performance using nanowires.

Photonic Computing Processor

Light has an exploitable property -- different wavelengths of light don't interact -- that's what fiber optics use to carry parallel streams of data. Likewise, light of different polarizations will not interact. Each polarization can be used as an independent channel, enabling more information to be stored in multiple channels, greatly improving information density.

June Sang Lee, lead author and PhD student at Oxford University's Department of Materials, said: "We all know that the advantage of photonics over electronics is that light is faster and more functional over large bandwidths. Therefore, our goal It's about taking full advantage of the combination of photonics and tunable materials for faster, denser information processing."

​

The research team, in collaboration with Professor C. David Wright from the University of Exeter, developed a HAD (Hybrid Active Dielectric) nanowire using a hybrid glassy material that was shown to be switchable when illuminated by light pulses material properties. Each nanowire exhibits a selective response to a specific polarization direction, so information can be processed simultaneously using multiple polarizations in different directions.

Using this concept, the researchers developed the first photonic computing processor that exploits the polarization of light.

Photonic computations take place through multiple polarization channels, increasing the computational density by orders of magnitude compared to conventional electronic chips. Computation is faster because the nanowires are modulated by nanosecond light pulses.

Hybrid nanowires can selectively switch devices based on polarization

Since the invention of the world's first integrated circuit in 1958, packing more transistors into an electronic chip of a given size has been a common method of maximizing computing density, known as "Moore's Law." However, as the need for specialized hardware for artificial intelligence and machine learning begins to push the boundaries of existing computing, the main question in electrical engineering has become "How do we pack more functionality into a single transistor?" For over a decade, Oxford Researchers in the laboratory of Professor Harish Bhaskaran in the University's Department of Materials have been investigating the use of light as a means of computing.

Professor Bhaskaran, who led the work, said: "This is just the beginning of everything we hope to see in the future, exploiting the various degrees of freedom offered by light, including polarization, to significantly parallelize information processing. This is certainly early work. , but with exciting ideas that combine electronics, nonlinear materials, and computation."

Polarization-selective switching of hybrid active dielectric nanowires for parallel photonic computing

Reference link:

[1] https://www.science.org/doi/10.1126/sciadv.abn9459

[2] https://phys.org/news/2022-06-world-ultra-fast-photonic-processor-polarization.html