Two-Qubit Gate

Two-qubit gates are quantum operations that act on a pair of qubits and can create entanglement between them. They are the essential ingredient that distinguishes quantum computation from classical — without two-qubit gates, a quantum computer with only single-qubit operations could be efficiently simulated classically. The most common two-qubit gates are the CNOT (controlled-NOT), CZ (controlled-Z), and various parameterized entangling gates.

The quality of two-qubit gates is typically the performance bottleneck in quantum processors. Two-qubit gates are slower, noisier, and more susceptible to crosstalk than single-qubit gates. On superconducting processors, two-qubit gates take 100-600 nanoseconds with fidelities of 99-99.9%. On trapped-ion systems, they take 10-200 microseconds with fidelities up to 99.9%. These fidelities may seem high, but in a circuit with hundreds of two-qubit gates, even 99.5% fidelity per gate leads to cumulative errors that overwhelm the computation.

Different qubit technologies implement two-qubit gates through different physical mechanisms. Superconducting processors use cross-resonance interactions (IBM), tunable coupler-mediated CZ gates (Google), or parametric gates (Rigetti). Trapped-ion processors use laser-driven Molmer-Sorensen gates that entangle ions through their shared motional modes. Neutral atom systems use Rydberg blockade interactions. The choice of native two-qubit gate affects circuit compilation, as different gate types have different connectivity constraints and decomposition properties.