Vincent Dumont, Hoi-Kwan Lau, Aashish A. Clerk, Jack C. Sankey
9 min
Quadratic dispersive coupling (QDC) is a powerful tool in optomechanics, enabling quantum nondemolition (QND) measurements of phonon numbers and the generation of non-Gaussian quantum states. However, practical QDC setups—typically based on a membrane-in-the-middle (MIM) geometry—are inherently plagued by linear dissipative coupling. This coupling introduces quantum radiation force noise (QRFN), which severely limits the device's performance in the quantum regime.
This paper introduces a strategy to suppress this parasitic noise. By shifting the membrane from the central (symmetric) position characteristic of standard MIM systems toward one of the cavity mirrors, the authors create an asymmetric cavity configuration. The researchers show that at specific membrane positions, the linear dissipative coupling—and the associated QRFN—is dramatically reduced through destructive interference. Crucially, this geometric change does not degrade the strength of the quadratic dispersive coupling, meaning the desired quantum functionality is preserved while the noise floor is lowered.
The authors provide a complete quantum analysis for both ideal single-port cavities and realistic systems with finite losses. In an ideal setup, this asymmetry can reduce QRFN by several orders of magnitude. In lossy systems, they identify an optimal membrane location that balances the subcavity dissipation rates to minimize noise. They specifically demonstrate how this approach improves the performance of cavity-assisted optical levitation and QND phonon number measurements, showing that the standard performance limits for these technologies are not fundamental but rather specific to the conventional symmetric MIM design.
This work provides a simple, robust design principle that is immediately applicable to existing experimental platforms. By reducing the noise that hinders QDC-based quantum protocols, this approach broadens the viability of next-generation optomechanical sensors and quantum information processors. It fundamentally redefines the noise constraints previously assumed for these systems.
As the field of optomechanics advances, quadratic dispersive coupling (QDC) represents an increasingly feasible path toward qualitatively new functionality. However, the leading QDC geometries generate linear dissipative coupling and an associated quantum radiation force noise that is detrimental to QDC applications. Here, we propose a simple geometry that dramatically reduces this noise without altering the QDC strength. We identify optimal regimes of operation, and discuss advantages within the examples of optical levitation and nondestructive phonon measurement.
Alex: That escaping information comes with quantum radiation force noise, meaning quantum fluctuations in the force light exerts on the membrane. Those fluctuations can kick the mechanical oscillator between states. A nominally quadratic measurement can therefore disturb the quantity it aims to read.
Sam: What was the previous design, and why did it leave this problem unresolved?
Alex: The standard design is called membrane-in-the-middle. The membrane divides the cavity into equal compartments, whose linear frequency responses cancel. Earlier quantum analyses found a strict backaction limit for phonon measurements in that arrangement, even with otherwise ideal optics.
Sam: Was moving the membrane already known to help, or is that the new suggestion?
Alex: Earlier classical calculations suggested that moving it toward the back mirror could reduce displacement-dependent light leakage. But that did not prove suppression of quantum force noise. This paper supplies the full quantum analysis and extends it to cavities with internal losses.
Sam: Walk us through the design change. Moving away from the middle sounds like it would destroy the cancellation.
Alex: Simply moving it anywhere would not work. The authors make the compartments unequal, then tune the membrane to a quadratic point. That is an operating position where one optical mode still has no leading linear frequency response.
Sam: So unequal compartment lengths do not mean abandoning the useful symmetry in the force balance.
Alex: At that operating point, the circulating optical powers on either side balance, so there is no net force. Their model shows that the quadratic coupling depends on total cavity length, not on how that length is divided. The unwanted noise, however, does depend on the division.
Sam: Give me the simplest case where that distinction produces a quieter system.
Alex: Consider an ideal single-port cavity, where all light exchange occurs through the input mirror. The back mirror lets nothing escape. Moving the membrane close to that back mirror greatly reduces how strongly its motion changes the cavity’s total light leakage.
Sam: The back end is already closed, so the membrane has less leverage over how light gets out.
Alex: In that ideal case, the calculated force noise falls with the square of the remaining back-compartment length, relative to total length. Meanwhile, the quadratic coupling stays unchanged. The analysis also includes destructive interference between noise pathways, which suppresses the low-frequency force noise.
Sam: An ideal sealed back mirror is useful for understanding the mechanism. What happens when photons can be lost elsewhere?
Alex: Then shrinking the back compartment indefinitely stops helping. Loss channels introduce force noise that lacks the same low-frequency cancellation. Motion also changes the relative power escaping from the compartments, carrying position information even if total leakage barely changes.
Sam: That suggests measuring only the change in total cavity loss could miss the disturbance.
Alex: The authors explicitly make that point. A zero gradient in total loss does not guarantee zero quantum force noise. Their optimal geometry balances the compartments’ photon decay rates, but residual noise remains because information can escape through separate channels.
Sam: What is the size of the improvement once those losses are included?
Alex: In one illustrative calculation, the cavity is ten centimeters long. The assumed round-trip loss in each compartment is one part per million. With their chosen input transmission and mechanical frequency, the predicted force noise is reduced by a factor of 250 compared with the centered membrane.
Sam: That is a calculated comparison using feasible parameters, not a measured reduction across fabricated devices?
Alex: Correct. The evidence is analytical quantum modeling and calculated parameter comparisons, not a new experiment. I’d regard it as a strong theoretical case within its assumptions. It does not establish that an operating apparatus will achieve the predicted noise floor.
Sam: Which assumptions should a device researcher check before applying that improvement?
Alex: The simple expressions require the optical modes to be sufficiently separated compared with their decay rates and the relevant mechanical frequency. The largest benefits also favor the resolved-sideband regime, where mechanical oscillation is fast compared with the cavity’s optical decay. Deep in the opposite regime, the improvement diminishes.
Sam: Bring that back to the applications. Does less force noise actually make individual vibration states easier to distinguish?
Alex: The paper compares the measurement rate with the rate of noise-induced jumps between phonon states. Reading the state requires measuring faster than backaction changes it. In one near-single-port example, the minimum zero-point motion variance needed for ground-state resolution improves by a factor of 25 over the centered design.
Sam: And that is a different performance measure from the force-noise reduction, not another estimate of the same number.
Alex: It is a measurement threshold, incorporating both readout and disturbance. For optical trapping, the authors make another comparison: their optimized cavity can produce less force noise than a free-space trap with the same optical spring strength. That benefit requires sufficiently small losses from the back compartment.
Sam: Who should read the full paper, and where should they begin?
Alex: Anyone designing quadratic optomechanical readout or cavity-assisted optical traps. Start with “Dissipative force noise,” especially the lossy-cavity analysis and its comparison with the centered geometry. Then read the application section relevant to your device, checking the optical-mode separation and sideband assumptions.
Sam: And for researchers outside that community, what is the line to remember?
Alex: A quieter quantum measurement need not mean a weaker interaction. Here, changing the geometry reduces the unwanted disturbance while preserving the useful coupling—within the model’s operating limits.
Sam: Thanks for listening.