Coherent Quantum Noise Cancellation (CQNC)

In this area of research we are investigating the coherent cancellation of backaction noise in optomechanical force measurements, in particular for interferometric gravitational wave detection.

Optomechanical detectors have reached the standard quantum limit in position and force sensing where measurement backaction noise starts to be the limiting factor for the sensitivity. A strategy to circumvent measurement backaction and surpass the standard quantum limit has been suggested by M. Tsang and C. M. Caves [Phys. Rev. Lett. 105, 123601 (2010)].

Theoretical schematic for CQNC (from M. Tsang and C. M. Caves, Phys. Rev. Lett. 105, 123601 (2010))
Potential experimental schematic for CQNC, consisting of a polarisation non-degenerate sub-threshold optical parametric oscillator and beamsplitter.



We conducted a detailed analysis of this method and assessed its benefits, requirements, and limitations [M. H. Wimmer et al, Phys. Rev A 89, 053836 (2014)
]. Based on this analysis we concluded that a proof-of-principle demonstration (realized as a micro-optomechanical system) is demanding, but possible. However, for parameters relevant to interferometric gravitational wave detection, the requirements for backaction evasion appear to be prohibitive.

We intend to reduce the deleterious effect of radiation pressure noise (RPN) by means of destructive interference with an anti-noise profess: By coupling a second (“ancilla”) cavity to the main (“meter”) cavity via a variable beamsplitter interaction and parametric down conversion we can realise a “negative mass oscillator”.

To achieve perfect noise cancellation, exact matching of the respective coupling strengths is required. Additionally, the linewidths of ancilla cavity and mechanical oscillator need to be matched, and the ancilla cavity needs to be sideband-resolved. Under these conditions, an improvement in sensitivity of one over twice the mechanical quality factor is possible. However, under experimentally more realistic constraints (such as imperfect matching) we still expect a significant off-resonant improvement in sensitivity.