A photonic qubit architecture based on single-photon encoding (typically dual-rail or polarization) with state preparation, gates, and measurement implemented using linear optical elements, single-photon sources, and photodetectors.
Physics
A common encoding is dual-rail:
[ |0_L\rangle = |1\rangle_a|0\rangle_b, \qquad |1_L\rangle = |0\rangle_a|1\rangle_b ]
Beam splitters and phase shifters implement single-qubit unitaries. Entangling gates are induced through measurement and feed-forward (KLM-style), trading deterministic interactions for optical network + ancilla overhead.
Hamiltonian
In integrated linear optics, passive transformations are generated by a quadratic bosonic Hamiltonian:
which induces a unitary mode transformation acting on creation operators as:
Single-qubit operations in dual-rail encoding are realized with beam splitters and phase shifters, while effective entangling operations are implemented by measurement-induced nonlinearities (ancilla + feed-forward in KLM-style protocols).
Figure

Why it matters
- Photons have low decoherence during transmission, making this architecture naturally network-compatible.
- The KLM result established that scalable QC is possible with linear optics + measurement alone.
- Provides a central route to distributed and communication-first quantum architectures.
Key Metrics
| Metric | Value | Notes | Fidelity reference |
|---|---|---|---|
| Native two-qubit interaction | Measurement-induced (probabilistic) | No direct photon-photon interaction required | knill-2000-efficient-linear-optics-quantum, adami-1998-quantum-computation-with-linear |
| Deterministic universal QC path | Yes (with ancilla + feed-forward overhead) | Core KLM result | knill-2000-efficient-linear-optics-quantum |
| Hardware bottleneck | Loss + source/detector efficiency | Dominant practical scaling limit in reviews | kok-2005-review-article-linear-optical |