How tiny holes in silicon could spot single particles of light
A new germanium-silicon chip design traps light and triggers electrical signals in separate layers, modelling single-photon detection without massive cooling equipment.
Reading level
The full story with proper science words explained.
Catching light one packet at a time
Detecting a single particle of light usually requires bulky hardware cooled down near absolute zero. That is a headache if you want to build compact quantum computers or secure communication networks. Standard computer chips are made of silicon, but silicon is practically blind to the infrared light sent through telecom fibres.
Germanium can absorb these telecom signals, but it is riddled with microscopic defects. If you run a high electric field through germanium to amplify a tiny signal, those flaws trigger constant false alarms. A new computer-modelling study suggests a clever way around this problem by redesigning a single-photon detector from the inside out.
Separating the capture from the spark
The team designed a device using a SAM architecture, which stands for separate absorption and multiplication. Instead of forcing one material to do everything, the layout splits the work between two distinct zones. Germanium catches the photon, and silicon handles the amplification.
Think of it like a bank with two rooms. The incoming photon arrives in a quiet reception lobby made of a 140-nanometre-thin slice of germanium. Here, the light hands over its energy to create a free electric charge. That charge then glides into an underlying silicon vault where a strong electric field triggers an avalanche of electrons, producing a measurable signal. Because the high electric field stays confined to the silicon, the defect-prone germanium never gets stressed enough to sound false alarms.
Squeezing light with photonic crystals
Germanium absorbs light relatively slowly, which usually requires a chunky slab that would slow down the detector. To solve this, the researchers modelled a hexagonal pattern of tiny holes in the adjacent silicon, spaced 610 nanometres apart. Each hole is 530 nanometres wide—hundreds of times thinner than a human hair.
This pattern creates a photonic bandgap, a structure that blocks certain frequencies of light from escaping. The trapped light bounces back and forth, forcing it into the thin germanium layer without etching or damaging the germanium itself. The simulations showed that even a compact germanium footprint between 4 and 16 micrometres long achieved an internal quantum efficiency of over 90% to 95%, meaning nearly every absorbed photon converts into a useful charge. The design also dropped the needed breakdown voltage from about 28 volts down to roughly 18 volts.
What we still don't know
These findings come from rigorous numerical simulations, not physical devices built in a cleanroom. The team modelled electric fields, photon paths, and carrier movements under idealized conditions where charges slip smoothly between layers without getting trapped.
The study could not fully measure the true dark count rate—the rate of phantom clicks caused by heat and real-world crystal defects. We do not yet know how real devices will handle high-speed Geiger mode, where voltages run high enough to spot individual photons instantly. The next test belongs to the engineers who must etch these microscopic hole patterns into actual silicon chips.
Science words
- SAM architecture
- A layout that separates the light-absorbing region from the signal-amplifying region to reduce noise.
- Photonic bandgap
- A nanoscale barrier pattern that prevents specific wavelengths of light from passing through, trapping them.
- Internal quantum efficiency
- The percentage of absorbed light particles successfully converted into mobile electric charges.
- Dark count rate
- The frequency of false detection signals triggered by random heat or material flaws instead of light.
- Geiger mode
- A high-voltage operating state sensitive enough for a single charge to trigger an avalanche signal.
Check it yourself
This story is based on a real research paper in Nanophotonics by Singh, Devine, Islam et al. We write with AI help and check it against the paper, but the original is the final word.