Two-qubit states are quantum states of two qubits, normalized vectors in a four-dimensional Hilbert space. Two-qubit states are fundamental to understanding entanglement and form the basis for many quantum algorithms and protocols.

Two-qubit states span a Hilbert space of dimension $2 \times 2 = 4$. They are classified into two families:

States that factor into independent single-qubit states: $$|\psi\rangle = |\psi\rangle_A \otimes |\phi\rangle_B$$

These exhibit no entanglement; measuring one qubit does not constrain the other. The default product states are the computational basis states $|00\rangle$, $|01\rangle$, $|10\rangle$, $|11\rangle$.

States that cannot be factored into independent single-qubit states. The most important entangled two-qubit states are:

• Bell states: maximally entangled basis states
  • Bell 00 (|Φ⁺⟩)
  • Bell 01 (|Ψ⁺⟩)
  • Bell 10 (|Ψ⁻⟩)
  • Bell 11 (|Φ⁻⟩)

The four computational (Z) basis states form the default measurement basis for two-qubit systems:

• Computational 00 (|00⟩): both qubits in ground state
• Computational 01 (|01⟩): first qubit ground, second excited
• Computational 10 (|10⟩): first qubit excited, second ground
• Computational 11 (|11⟩): both qubits in excited state

These product states are separable (unentangled) and are the eigenstates of the Z operator on both qubits. They form a complete orthonormal basis for two-qubit Hilbert space.

The four Bell states form a complete orthonormal basis for two-qubit Hilbert space. Any two-qubit state can be decomposed in the Bell basis, enabling “Bell measurement” to distinguish all four states.

A two-qubit state $\rho$ is separable if it can be written as: $$\rho = \sum_i p_i \rho_i^A \otimes \rho_i^B$$

where $p_i \geq 0$ and $\sum_i p_i = 1$. Separable states (pure or mixed) exhibit no entanglement and can be prepared classically.

A state is entangled if it is not separable. Bell states are maximally entangled; general two-qubit entangled states have varying degrees of entanglement.

For two-qubit states:

• Concurrence: quantifies bipartite entanglement; ranges from 0 (separable) to 1 (maximally entangled)
• Entanglement of formation: minimum number of Bell pairs needed to create the state
• Bell inequality violation: entangled states violate CHSH inequality; maximally entangled states violate it maximally

Two-qubit gates couple qubits and create entanglement:

• CNOT (controlled-NOT): Pauli X on target qubit, controlled by source qubit state
• CZ (controlled-Z): Pauli Z on target, controlled by source
• SWAP: exchange qubit states
• iSWAP: SWAP with $i$ phase on exchange
• Parametric gates (XX, YY, ZZ, CPHASE): angle-tunable entangling gates

See two-qubit gates for details on each gate.

The most common Bell state, $|\Phi^+\rangle$, is created via:

1. Prepare initial state $|00\rangle$
2. Apply Hadamard to first qubit: $\frac{1}{\sqrt{2}}(|0\rangle + |1\rangle) \otimes |0\rangle$
3. Apply CNOT with first as control: yields $|\Phi^+\rangle$

Other Bell states are obtained by applying single-qubit rotations (X or Z gates) before or after CNOT.

• Quantum teleportation: Bell pair enables sending unknown qubit state via two classical bits
• Superdense coding: Bell pair allows encoding two classical bits by manipulating one qubit
• Quantum key distribution: E91 protocol uses Bell states to establish shared secret key

• Entanglement swapping: connecting Bell pairs extends entanglement across networks
• Quantum repeaters: swap Bell pairs to extend range beyond direct gate distance
• Distributed quantum computing: Bell pairs form links between distant quantum processors

• Deutsch-Jozsa algorithm: uses Bell state phase patterns
• QAOA: two-qubit entanglement creates search space for optimization
• VQE: Bell state ansätze parameterize trial states

• Surface codes: local two-qubit interactions detect errors via Bell-like measurements
• Stabilizer codes: measure two-qubit Pauli products to extract syndrome information

Pure states: two-qubit density matrix $\rho = |\psi\rangle\langle\psi|$ with rank 1. Maximum purity.

Mixed states: density matrix with rank > 1. Represent statistical mixtures or decoherence-affected states. Purity $\text{Tr}(\rho^2) < 1$.

Common two-qubit decoherence sources:

• Dephasing: loss of phase coherence in superposition; affects X and Y basis overlaps
• Amplitude damping: decay to lower energy state (usually $|00\rangle$)
• Bit flip: random flips $|0\rangle \leftrightarrow |1\rangle$
• Depolarization: random projection to random state

Bell states are particularly fragile under decoherence due to maximal entanglement.

Two-qubit entanglement generalizes to multi-qubit systems:

• Three qubits: Bell states embed as subspaces; GHZ and W states have different three-qubit entanglement structures
• $n$ qubits: entanglement complexity grows; two-qubit correlations are building blocks for global correlations
• Graph states: two-qubit CZ gates create larger graph-structured entanglement

Characterizing a two-qubit state requires:

1. Measuring in multiple bases (Z, X, Y on each qubit independently)
2. Collecting statistics from many trials
3. Reconstructing density matrix via classical processing

Full state tomography of a two-qubit state requires measurements in all $3 \times 3 = 9$ two-qubit bases (e.g., ZZ, ZX, ZY, XZ, XX, XY, YZ, YX, YY).

Two-qubit systems are the simplest non-trivial quantum computers. Scaling to many qubits requires:

• Precise two-qubit gate control (errors accumulate with circuit depth)
• Minimal crosstalk between gate operations
• Long coherence times for many-qubit entanglement maintenance
• Classical processing for measurement feedback and error correction