Subsystems for all-optical coherent quantum-noise cancellation

Abstract

Quantum mechanics dictates that a measurement always disturb the measured system. In weak continuous measurements, the trade-off between measurement precision and back-action onto the system yields an optimal measurement sensitivity, which is known as the Standard Quantum Limit (SQL) in opto-mechanical measurements, such as gravitational-wave detection. It corresponds to finding the optimal optical power in a compromise between quantum shot noise and quantum radiation-pressure noise. Coherent quantum-noise cancellation (CQNC) aims at overcoming the SQL and reducing back-action noise via the introduction of an effective negative-mass oscillator. In an alloptical set-up, this oscillator is realised by a detuned optical resonator coupled to incoming light with a beam-splitter and a down-conversion interaction and needs to be matched to the measured system in resonance frequency, damping and coupling strengths. This thesis explores the nature of CQNC and a potential all-optical realisation in theory and experiment, with a particular emphasis on the beam-splitter and the down-conversion interaction. Two possible set-ups are compared theoretically and critical parameters determined. Available opto-mechanical devices were characterised and confirmed to be suitable for CQNC. The down-conversion coupling strength gDC is linked to experimentally obtainable parameters. More than 2.3 dB reduction in uncertainty of two-mode squeezed light were observed. The squeezing measurements yielded gDC = 2\pi\times200 kHz at 100mW pump power, which is well within the initially required range and is in agreement with results from two other measurement methods. Optical resonators coupled via a beam-splitter interaction are studied theoretically and experimentally. In this work, the beam-splitter interaction of strength gBS was realised by a wave plate. A simplified experiment design enabled stabilisation of the coupled resonators. Our experimental observations accurately confirmed our theoretical predictions. The observed mode splitting yielded gBS = 2\pi\times235 kHz, within the updated requirements. Losses and a limited measurement strength will be the limiting factors for CQNC. The updated set of parameters, backed by the conducted experiments, paves the way towards a reduction of radiation-pressure noise of up to 4.8 dB.