Partial deterministic non-demolition Bell measurement

We propose the scheme implementing partial deterministic non-demolition Bell measurement. When it is used in quantum teleportation the information about an unknown input state is optimally distributed among three outgoing qubits. The optimality means that the output fidelities saturate the asymmetric cloning inequality. The flow of information among the qubits is controlled by the preparation of a pair of ancillary qubits used in the Bell measurement. It is also demonstrated that the measurement is optimal two-qubit operation in the sense of the trade-off between the state disturbance and the information gain.