Photonic research tackles timing and 74-femtosecond beam steering

A beam of light changes direction in 74 femtoseconds — quadrillionths of a second — after another beam pumps the material guiding it. That is the result reported by researchers at Caltech, who used light as a control signal rather than relying on electrons in an optical chip. In Seoul, a separate research team modeled a photonic circuit that could change how long light pulses take to cross the chip, bringing programmable timing to hardware that is usually fixed at manufacture.

The Caltech approach uses the optical Kerr effect: intense light changes a material's refractive index, which alters how a second, weaker beam travels through it. The team placed amorphous silicon on nanoscale pillars, with the pillars spaced to slow and recirculate the pump light. The resulting optical metasurface — an ultrathin, nanoengineered sheet — enabled beam deflection of up to 13 degrees. Earlier steering methods changed electronic properties in optical chips or liquid-crystal panels, with response times ranging from nanoseconds to picoseconds.

The Seoul architecture tackles a different weakness. Light is excellent for carrying large quantities of data, but separate signals may take different paths and need to meet at the right moment. Existing coupled-resonator-induced transparency, or CRIT, systems can delay selected frequencies, yet their resonator geometry fixes the delay and filtering behavior. The proposed design adds two tunable loop couplers: changing them alters how the resonators interact, allowing one circuit to adjust its transmission band, throughput and pulse delay while operating.

So what changes in practice? If these approaches become hardware systems, optical computers and high-speed communications equipment could gain faster beam routing and reusable delay, filtering and signal-processing functions instead of relying on separate fixed components. Caltech researchers say the steering speed could move below 74 femtoseconds, while the Seoul simulations indicate that their circuit could also manipulate transmitted-light frequency and remain functional despite modeled losses, backscattering, phase errors and thermal crosstalk.

The distance between a promising optical model and a manufactured device remains real. The Seoul result is based on theory, numerical modeling and three-dimensional electromagnetic simulations, not an experimental chip. Caltech's work is reported as a research result rather than a deployed computing product. The immediate achievement is narrower but useful: researchers are gaining new ways to make light change direction and arrive on cue, two conditions any practical photonic computer will need.