Phase-Flip Errors

Phase-Flip Errors

The phase-flip error has no classical counterpart at all. It is the second half of the Pauli error group, and understanding it is the real entry point to quantum error correction. Classical computers need only protect against bit flips (0 becomes 1). Quantum computers need to protect against bit flips and phase flips (a sign change on the |1⟩ amplitude), because both are equally likely to occur on real hardware.

The Gate Matrix and What It Does

A phase-flip error applies the Pauli Z gate:

Z = [[1,  0],
     [0, -1]]

On the computational basis states:

So on a general state α|0⟩ + β|1⟩ the effect is α|0⟩ − β|1⟩. Crucially, the probabilities in the computational basis are unchanged: |α|² and |β|² are the same before and after. Measured directly in the Z basis, a phase-flip looks like nothing happened, the error is silent.

On superpositions, however, the effect is dramatic. Z|+⟩ = |−⟩, and Z|−⟩ = |+⟩. The state has completely flipped into its orthogonal partner. The information is still there; it is just hiding in the phase.

The Invisible Error, Made Visible

The trick for detecting phase flips is to change basis. A phase flip in the Z basis is a bit flip in the X basis. Formally:

H · Z · H = X

That single identity is the whole strategy. If you wrap any Z error in a pair of Hadamards, it becomes an X error, which you already know how to detect. The workflow for a single-qubit phase-flip detector is: