Eve's computational-basis intercept-resend attack applies a projective measurement of to Bob's photon and discards her outcome from Alice and Bob's description. Their resulting density matrix is
This separable state keeps but removes . Hence
Using the same signed Bell-test statistic gives
The ideal value is halved, and the attack no longer violates the CHSH inequality.
The key data show why several settings matter. For , Eve's measurement leaves perfect anticorrelation, so this subset has zero bit errors after Bob flips his assignment. For , , so correct anticorrelation has probability : its quantum bit error rate is . If both key-setting pairs occur equally often, their combined raw-key error rate is . The Bell-test change detects this attack even though the vertical/horizontal key subset alone would not.

Articles by others on the same topic (0)

There are currently no matching articles.