Linear Optical Quantum Computing
The KLM scheme shows universal quantum computing is possible with single photons, linear optics, and measurement, without direct photon-photon interaction.
The surprising result
Knill, Laflamme, and Milburn showed in 2001 that photons passing through beam splitters and phase shifters, together with single-photon sources and detectors, suffice for universal quantum computation. No nonlinear medium is required. The trick is that measurement itself induces an effective nonlinearity between photons.
Measurement-induced gates
Send photons through an interferometer that includes auxiliary photons, then detect the auxiliaries. Conditioned on a particular detection outcome, the remaining photons emerge in an entangled state that could not arise from linear optics alone. The gate works only when the heralding detection succeeds, so it is probabilistic.
Making it deterministic
- Teleportation-based gates that use pre-prepared entangled resources
- Feed-forward: quickly reconfiguring optics based on earlier measurement outcomes
- Heavy multiplexing of probabilistic sources and gates to raise the success rate toward one
Resource cost
Because success is probabilistic, useful computation demands many trials, ancillas, and fast switching, translating into large photon and component overheads. Modern proposals lean on fusion-based or cluster-state approaches that build a large entangled resource from small, easy-to-make pieces, tolerating loss with built-in error correction.
Linear optics reframed what a quantum computer needs: not strong interactions, but the ability to prepare, interfere, and measure photons with high efficiency. The practical barrier moved from physics to engineering, chiefly source purity, detector efficiency, and low-loss integrated photonics.