integrated quantum photonics

Integrated quantum photonics builds a quantum processor out of light instead of electrical circuits. Where a superconducting chip moves microwave signals along metal wires, a photonic chip guides single particles of light through tiny channels etched into glass-like material. Picture a maze of hair-thin transparent pipes printed onto a wafer: photons enter, run through splitters and bends, interfere with each other, and come out the far side carrying the answer. Because the information rides on light, the same telecom wavelengths and optical fibers that already carry the internet can plug straight into the chip, and much of it runs at or near room temperature rather than a hair above absolute zero.

The building blocks are waveguides, beamsplitters, phase shifters, and detectors, all patterned into silicon or silicon nitride. A waveguide is a buried strip of higher-index material that traps light and steers it around the chip. A beamsplitter is just two waveguides brought close enough that a photon can hop between them, which is how two light paths interfere. Phase shifters tune the path length to set the interference, and at the end, detectors — often superconducting ones that must still be cooled — click when a photon arrives. Encode a qubit in which path a photon takes, or in its polarization, and a careful mesh of these elements can run quantum logic and quantum links on a single piece of silicon.

The honest catch is that photons barely interact with each other, which is exactly why they are clean carriers but also why two-qubit gates are hard. The standard trick makes those gates probabilistic: the operation only succeeds when detectors happen to click the right way, perhaps a small fraction of the time, so you run many attempts and keep the good ones. Generating exactly one photon on demand, getting it into the waveguide without losing it, and detecting it efficiently are all still imperfect. Photonics is a serious contender precisely because it is telecom-friendly and partly warm, but like every modality today it is small, lossy, and unproven at scale — no approach has won.