The basic idea that parity measurements could be used to create entanglement is theoretically established by more than 10 years. Moreover, parity measurements can be realized in different solid state implementations, e.g. for superconducting qubits in resonant superconducting cavities and for quantum dots coupled to an electronic Mach-Zehnder interferometer or to a quantum point contact. A major problem in all these implementations is that the unavoidable back-action of the measurement hinder the realization of certain maximally entangled states. Importantly, such a decoherence has been recently reported in experiments with superconducting qubits in the groups of Dicarlo in Delft and Siddiqi in Berkeley.

This talk will put forward a scheme to overcome this dangerous decoherence effect: in fact we propose to use two different quantum feedback protocols to stabilize the steady state of the system to any maximally entangled states defined a priori. I will present a microscopic derivation of the detector outcome and its back-action, which allows to show that a direct feedback based on parity measurement allows to overcome the measurement-induced dephasing within each parity subspace, leading to a larger and longer-lasting entanglement. While this scheme is inefficient for external noise sources (e.g. gate voltage fluctuations), we show that a more elaborate scheme requiring two simultaneous measurements and a single feedback channel leads to a steady state entanglement, whose amount depends on the feedback efficiency.

By combining these continuous feedback schemes with parity measurements, this work provides a further step towards the key goal of entanglement generation and qubits stabilization for quantum information processing.