Interference-Based 3D Optical Cold Damping of a Levitated Nanoparticle

Abstract

Achieving efficient three-dimensional feedback cooling of levitated nanoparticles is a key requirement for precision sensing and quantum control in levitated optomechanics. Here we demonstrate three-dimensional optical feedback cooling of a levitated nanoparticle using an interference-enhanced optical force generated within a single beam path. In this scheme, a weak auxiliary field co-propagates with the trapping tweezer and interferes with it to produce a tunable optical force that enables cold damping along all three center-of-mass motional axes without additional beam paths or trap reconfiguration. Using this approach, we cool a 142-nm-diameter silica nanoparticle in high vacuum to effective temperatures of 625.8, 711.6, and 19.9 mK along the $x$, $y$, and $z$ directions, respectively, at a pressure of $8.5\times10^{-6}$ mbar. The cooling dynamics and their dependence on feedback gain and pressure are well described by a cold-damping model. Because the feedback force is generated optically, the scheme does not rely on electrical actuation and is directly compatible with neutral particles. These results establish interference-based optical forces as a simple and broadly applicable mechanism for three-dimensional feedback control in levitated optomechanics, with a clear pathway toward the quantum regime under improved vacuum and detection conditions.

Figures (4)

• Figure 1: Experimental setup. A silica nanoparticle is trapped in high vacuum by a 1064 nm optical tweezer formed by NA$=0.8$ objective. Backscattered light is collected by the same objective and routed via a Faraday rotator (FR) and a polarizing beam splitter (PBS) to the detection path, where balanced homodyne detection measures the axial ($z$) motion and split photodetectors measure the transverse ($x,y$) motions. An auxiliary beam derived at BS1 passes through a fiber-coupled amplitude modulator (AM) and a fiber stretcher (FS), and is then recombined with the trapping beam at BS2 to co-propagate to the trap. A slight displacement between the two beams gives rise to an interference-induced gradient force with nonzero components along all three spatial directions (inset). Feedback is implemented by electronically processing the measured motional signals with appropriate gains and delays and applying the resulting drive to the AM.
• Figure 2: (a) Power spectral densities (PSDs) of the particle motion along the $z$ axis, measured at a pressure of $2.3\times10^{-5}$ mbar for different feedback gains, corresponding to effective damping rates $\Gamma_{\mathrm{eff}} \approx 0.16$ kHz (orange), $0.52$ kHz (green), and $2.66$ kHz (blue). Solid lines are fits to Eq. \ref{['eq:motion_fft_IL_PSD']}. (b) Extracted effective temperature $T_{\mathrm{eff}}$ of the $z$-axis motion as a function of $\Gamma_{\mathrm{eff}}$. Each data point represents the mean of ten independent measurements (1 s acquisition time each), with error bars indicating the standard deviation. The solid line shows the theoretical prediction based on Eq. (\ref{['eq:Teff']}), while the shaded region indicates the uncertainty due to the imprecision of the pressure gauge ($\approx 30\%$).
• Figure 3: Extracted effective temperatures $T_{\mathrm{eff}}$ of the particle motion along the $x$ (red), $y$ (green), and $z$ (blue) axes as a function of background pressure. Each data point represents the mean temperature extracted from 12 consecutive PSDs, each calculated from a 50 ms time trace, with error bars indicating the standard deviation. As the pressure is reduced, $T_{\mathrm{eff}}$ decreases linearly (dashed lines), consistent with the expected pressure scaling of the gas damping rate. At lower pressures, deviations from linearity appear and $T_{\mathrm{eff}}$ saturates as the contribution of measurement noise (second term in Eq. (\ref{['eq:Teff']})) becomes significant. Solid lines show the theoretical prediction based on Eq. (\ref{['eq:Teff']}), and the shaded regions indicate the uncertainty due to the pressure gauge imprecision.
• Figure 4: PSDs of the particle motion along the $x$, $y$, and $z$ axes, measured using split detections for the transverse directions ($x$, $y$) and homodyne detection for the axial direction ($z$). Each panel compares the PSDs measured without feedback cooling at a pressure of 2 mbar (red) and with simultaneous cooling of all three motions activated at a pressure of $8.6\times10^{-6}$ mbar (blue). Feedback cooling strongly suppresses the mechanical resonances, leading to effective temperatures of $(625.8,\, 711.6,\, 19.9)$ mK for the $x$, $y$, and $z$ motions, respectively. For the transverse directions, the cooled resonances approach the detection noise floor. The resonance frequencies are indicated by arrows.