Key Points
- Scientists have developed a new, more efficient sensor for light-based computers.
- The sensor solves a key problem in the field of “programmable photonics.” It is a tiny light detector that is enhanced with germanium ions.
- The new device is highly sensitive and doesn’t degrade the light signal it’s monitoring.
- This breakthrough represents a significant step toward developing practical, large-scale photonic computers.
Scientists have developed a new type of sensor that addresses a significant challenge in the pursuit of building powerful, light-based computers. The breakthrough published in Advanced Photonics by researchers at the Hong Kong University of Science and Technology is a new kind of on-chip power monitor that is both highly sensitive and extremely efficient.
The technology in question is “programmable photonics,” which uses particles of light (photons) instead of electrons to perform complex calculations. This approach promises to be significantly faster and more energy-efficient than traditional electronics, making it ideal for demanding tasks such as AI and deep learning.
A key challenge has been finding a good way to monitor the light signals on the chip. Existing sensors either had to absorb too much light to obtain a good reading, which degraded the signal, or they weren’t sensitive enough to work at the low power levels required.
The new device solves this problem. It’s a tiny light detector that is integrated directly into the chip’s “waveguides,” which are the pathways that guide the light. The researchers enhanced the detector’s performance by “implanting” it with germanium ions. This allows the sensor to detect light more effectively without introducing the unwanted side effects that have plagued previous attempts.
The new germanium-implanted sensor is highly sensitive, exhibits very low signal loss, and generates minimal unwanted “dark current” (noise). This combination makes it a perfect solution for monitoring the delicate light signals inside a photonic chip without disturbing them.
The researchers say this is a major step toward making large-scale, light-based computing a practical reality.
