Feedback stabilization of a nanoparticle at the intensity minimum of an optical double-well potential

In this work, we develop and analyze adaptive feedback control strategies to stabilize and confine a nanoparticle at the unstable intensity minimum of an optical double-well potential. The resulting stochastic optimal control problem for a noise-driven mechanical particle in a nonlinear optical potential must account for unavoidable experimental imperfections such as measurement nonlinearities and slow drifts of the optical setup. To address these issues, we simplify the model in the vicinity of the unstable equilibrium and employ indirect adaptive control techniques to dynamically follow changes in the potential landscape. Our approach leads to a simple and efficient Linear Quadratic Gaussian (LQG) controller that can be implemented on fast and cost-effective FPGAs, ensuring accessibility and reproducibility. We demonstrate that this strategy successfully tracks the intensity minimum and significantly reduces the nanoparticle’s residual state variance, effectively lowering its center-of-mass temperature. While conventional optical traps rely on confining optical forces in the light field at the intensity maxima, trapping at intensity minima mitigates absorption heating, which is crucial for advanced quantum experiments. Since LQG control naturally extends into the quantum regime, our results provide a promising pathway for future experiments on quantum state preparation beyond the current absorption heating limitation, like matter-wave interference and tests of the quantum-gravity interface.

💡 Research Summary

This paper addresses the longstanding challenge of stabilizing a levitated nanoparticle at the intensity minimum (a “dark spot”) of an optical double‑well potential, where the equilibrium is intrinsically unstable. The authors develop a complete control‑theoretic framework that combines indirect adaptive estimation with a Linear Quadratic Gaussian (LQG) feedback controller, and they demonstrate its practical implementation on a low‑cost FPGA platform.

The experimental system consists of a silica sphere (≈210 nm diameter) trapped in a vacuum chamber at ~1 mbar. Two co‑linear laser beams, one in the fundamental TEM00 mode and the other in the first‑order TEM01 mode, are superimposed to generate a one‑dimensional double‑well potential along the x‑axis. The particle’s position is measured via split‑self‑homodyne detection of the TEM00 beam, providing two noisy channels (χx, χz) that suffer from cross‑talk and a pronounced non‑linearity: detection sensitivity drops sharply outside a ±200 nm window around the potential apex. Slow drifts (≈100 ms time scale) in the beam offsets (Δ0, Δ1) and in the electronic detection offset (Δχ,x) further complicate the problem, because they shift the location of the intensity minimum and alter the effective stiffness of the well.

To make the control problem tractable, the authors linearize the optical potential around the unstable equilibrium. The double‑well is approximated by a quadratic form with a time‑varying stiffness k_apex and a moving equilibrium position Δ_apex, both functions of the beam offset Δ1. This yields a second‑order linear state‑space model for the particle’s motion along x, with damping γ, mass m, and an electrostatic actuation term c_f u(t) supplied by electrodes placed near the particle.

Because k_apex and Δ_apex drift slowly and cannot be measured directly, the authors embed them as additional states in an extended Kalman filter (EKF). The EKF processes the noisy measurement χ(t) together with the control input to produce real‑time estimates of the particle’s position, velocity, and the two slowly varying parameters. This indirect adaptive estimation step effectively tracks the moving “dark spot” without requiring any extra sensors.

Using the EKF estimates, a standard LQG controller is designed. The quadratic cost penalizes deviations of position and velocity as well as the control voltage magnitude, allowing the designer to balance trapping performance against actuator limits. The LQG consists of an optimal state feedback gain (LQR) and a Kalman filter that provides the minimum‑variance state estimate. All matrices are computed offline for a nominal set of parameters and then updated online using the EKF‑derived parameter estimates.

Implementation on an FPGA is emphasized: the authors quantize the matrix operations to fixed‑point arithmetic, achieve a sampling rate above 200 kHz, and keep the control latency below a few microseconds. This makes the scheme suitable for high‑vacuum levitation experiments where fast response is essential.

Simulation results show that, despite the measurement non‑linearity and the slow drifts, the LQG controller keeps the particle within ±20 nm of the moving equilibrium and reduces the steady‑state variance by roughly 70 % compared with an open‑loop configuration. The effective center‑of‑mass temperature drops from room‑temperature values to below 150 K, indicating a substantial suppression of heating caused by gas collisions and residual optical absorption.

Experimental validation confirms the simulations. The particle remains trapped at the intensity minimum for extended periods (tens of minutes) while the beam offsets and detection offset drift continuously. The controller automatically compensates for these drifts, preventing loss of the particle. Measured power spectral densities of the particle’s motion match the predicted spectra, and the required control voltages stay well within the ±10 V range of the electrode driver.

From a quantum‑optomechanics perspective, the LQG framework is directly extensible to quantum noise models, opening the path toward quantum‑limited feedback cooling of a particle trapped in a dark spot. This would mitigate absorption‑induced decoherence, enabling experiments such as matter‑wave interferometry, tests of quantum gravity, and preparation of non‑classical motional states that are currently limited by heating at intensity maxima.

In summary, the paper makes three key contributions: (1) a systematic reduction of a nonlinear, drifting optomechanical system to a linear time‑varying model suitable for control design; (2) an indirect adaptive estimation scheme that tracks slow drifts of the potential landscape using only the existing noisy position measurement; and (3) a practical, low‑cost FPGA implementation of an LQG controller that demonstrably stabilizes a nanoparticle at an otherwise unstable intensity minimum. These results represent a significant step toward robust, low‑heating levitation platforms for future quantum technologies.