Fig. 3: Interferometric detection of squeezed thermomechanical noise in a nanomechanical oscillator.
From: Cavity electromechanics with parametric mechanical driving
a The thermal displacement fluctuations generate sidebands at ω = ωc + Ωm and ω = ωc − Ωm to a microwave tone sent to the cavity at ω = ωc as schematically shown in the inset. After down-conversion, we detect these sidebands and the corresponding power spectral density is shown for the parametric modulation switched off as blue line and with the parametric modulation switched on as red line. The black line is a Lorentzian fit to the data without parametric modulation. b shows the quadratures of the thermal displacement fluctuations vs. time in the top panels and as histograms (taken for 300 s of measurement time) in the bottom panels. Without parametric modulation, the thermal fluctuations are distributed equally in both quadratures (left side) and the quadrature histogram is a rotational symmetric Gaussian curve; with a parametric modulation applied, as shown on the right side, the fluctuations in one quadrature get amplified while the fluctuations in the second quadrature get de-amplified. The result is a squeezed thermal state. The colorscale represents histogram counts from low (dark) to high (orange) values. White pixels correspond to no recorded counts. The blue circles in the histogram plots are guides to the eye. In (c) we plot the distribution of X-quadrature values for the histograms shown in (b) as dots and Gaussian fits as lines. When the parametric modulation is switched on, the variance of the X-quadrature gets significantly decreased and the squeezing factor is approximately s = 0.49. The histograms are normalized to the total number of ~13,000 data points. Data were taken for V2Ω∕Vt = 0.67.
