Post-selection-based continuous variable quantum information processing
Jie Zhao · ANU Open Research (Australian National University) · 2019
Quantum communication and computation harness the intriguing and bewildering nature of quantum mechanics to realize information processing tasks that have no classical analog. Nonetheless, this supremacy comes with fundamental limits that, in some scenarios, pose undesirable bounds on the performance of these quantum technologies. One such example is the well-known quantum no-cloning theorem imposed by the Heisenberg uncertainty principle. It states that an unknown quantum state cannot be duplicated with arbitrarily high accuracy. Very recently, however, post-selection was proposed as a way out: it was demonstrated that in various quantum information tasks, deterministic bounds can be overcome by forgoing determinism. In this thesis, we investigate post-selection as a novel approach to enhance the performance of versatile continuous-variable (CV) quantum information processing and envisage it to become a useful component of the general Gaussian toolbox. The first part of this thesis examines applications of post-selection in purely linear systems. In particular, two implementations of the noiseless linear amplifier (NLA), the measurement-based NLA and the physical NLA, are investigated and compared in terms to their abilities to preserve the state Gaussianity and their success probability. We show that the inevitable signal-to-noise ratio (SNR) degradation accompanying a linear quantum amplifier can be circumvented by resorting to a probabilistic scheme. Amplification with a signal transfer coefficient of Ts>1 is realised by combining a measurement-based NLA with a deterministic linear amplifier. We also construct a quantum cloning machine based on this hybrid amplifier for arbitrary coherent input states. We demonstrate a production of multiple clones (up to five) with fidelity of each clone exceeding the corresponding no-cloning limit. We then consider employing the post-selection algorithm in information protocols involving nonlinearity. First, we develop two squeezers as optical parametric amplifiers, each producing fairly pure squeezed output field up to 11.2dB (after correcting the detection loss). The squeezers are served as the nonlinear source in the remaining part of this thesis. We demonstrate a high fidelity quantum squeezing gate which is one indispensible building block for constructing a universal CV quantum computer. An inverse-Gaussian filter is incorporated into the feedforward line, leading to an enhancement in precision of the inline dual-homodyne measurement and therefore combats efficiently the correlation degradation due to loss and noise introduced during feedforward. As one example, we show that a fidelity of 98.49% for a target squeezing of -2.3dB is obtained with only -6dB ancilla squeezing, which would otherwise require -20.5dB initial squeezing using a conventional deterministic setup. Additionally, we introduce a CV quantum teleportation scheme using post-selection, allowing for a significantly improved fidelity against the conventional deterministic CV teleporter. The intuition behind this improvement is that post-selection effectively distilled the accessible entanglement and therefore a high fidelity only originally achievable with a higher amount of initial squeezing is now obtainable with only modest amount of squeezing, coming at an expense of finite success probability.