⚛️ high-energy experiments

Optomechanical Accelerometer Search for Ultralight Dark Matter

This paper presents a resonant search for ultralight dark matter using a quantum-limited cavity optomechanical accelerometer that achieved a sensitivity of $\sim 10\;\text{n}g_0/\sqrt{\text{Hz}}$ but detected no signal, thereby validating the platform as a scalable approach for future competitive constraints on vector-mediated dark-matter interactions.

Original authors: M. Dey Chowdhury, J. P. Manley, C. A. Condos, A. R. Agrawal, D. J. Wilson

Published 2026-06-23

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: M. Dey Chowdhury, J. P. Manley, C. A. Condos, A. R. Agrawal, D. J. Wilson

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Big Picture: Hunting for Invisible Ghosts

Imagine the universe is filled with a mysterious, invisible substance called Dark Matter. We know it's there because it holds galaxies together with gravity, but we've never seen it, touched it, or figured out what it's made of. Scientists suspect some of this dark matter might be made of incredibly light, fast-moving particles that act more like a wave than a solid object.

The authors of this paper built a tiny, high-tech "ear" to listen for the faint hum of these dark matter waves. They didn't find any ghosts, but they proved their "ear" works perfectly and showed how to build better ones in the future.

The Tool: A Trampoline in a Freezer

To catch these invisible waves, the team built a cavity optomechanical accelerometer. That's a mouthful, so let's break it down:

The Trampoline: Inside a machine, they suspended a microscopic membrane (a thin sheet) made of silicon nitride. Think of it as a tiny, super-tight trampoline.
The Freezer: This trampoline is kept in a cryostat (a super-cold freezer) at 4 degrees above absolute zero. This is necessary to stop the trampoline from jiggling randomly due to heat, which would drown out any tiny signals.
The Laser: They shine a laser beam through the trampoline. Because the trampoline is part of a "cavity" (like a mirror box), the light bounces back and forth. If the trampoline moves even a tiny bit, the light changes. This allows them to measure the trampoline's movement with extreme precision.

The Theory: Why Would the Trampoline Move?

The team is looking for a specific type of dark matter that interacts with ordinary matter (like the atoms in the trampoline) in a very specific way.

The Analogy: Imagine you are standing on a boat (the trampoline) in a calm lake. Suddenly, a giant, invisible wind (the dark matter wave) blows.
The Twist: This wind doesn't push everything equally. It pushes the boat (made of silicon and copper) slightly differently than it pushes the water (or other materials).
The Result: Because the boat and the water react differently to the wind, the boat starts to rock back and forth.

In the experiment, the "boat" is the silicon chip, and the "trampoline" is the membrane attached to it. If the dark matter wave hits them, the difference in how the materials react causes the trampoline to vibrate at a very specific frequency.

The Challenge: The Noise Problem

The biggest problem is that the trampoline is so sensitive that it hears everything.

Thermal Noise: Even at near-freezing temperatures, atoms jiggle.
Vibrations: The building, the ground, and even the cooling machine itself create tiny shakes.

To fix this, the team built a Vibration Isolation System (VIS).

The Analogy: Imagine the trampoline is a delicate wine glass. To protect it from a shaking table, you hang the glass from a long, thin string (a pendulum) inside a soundproof box. The string is so long and loose that the shaking of the table doesn't reach the glass.
The Reality: They hung their detector on a 2-foot-long pendulum inside a vacuum chamber. This filters out almost all outside vibrations, leaving only the ultra-quiet environment needed to hear the dark matter.

The Experiment: Tuning the Radio

The team didn't know exactly what "note" (frequency) the dark matter would sing. So, they had to scan the radio dial.

Photothermal Tuning: They used a clever trick. By slightly heating the trampoline with a laser, they could change its natural vibration speed (frequency). It's like tightening or loosening a guitar string to change its pitch.
The Scan: They slowly "tuned" the trampoline across a range of frequencies (around 39,000 Hz) for about 800 seconds, listening for a signal that matched the expected shape of a dark matter wave.
The Feedback: They used a second laser to apply "radiation pressure" (a gentle push from light) to dampen the trampoline's natural jitters, making it even quieter and more sensitive.

The Results: Silence, but a Success

Did they find dark matter? No. They heard nothing but the expected background noise.

Is that a failure? No. In science, knowing what isn't there is just as important as knowing what is.

Setting the Limit: Because they heard nothing, they could draw a line in the sand. They proved that if dark matter exists with certain properties, it must be weaker than the level they could detect. This rules out a huge chunk of possibilities for what dark matter could be.
Proving the Concept: The most important result is that the machine worked exactly as designed. It operated at the "quantum limit" (the quietest possible state allowed by the laws of physics) and remained stable.

The Future: Building a Bigger Net

The paper concludes by saying this was just the "first generation" of the detector. It's like building a small, single-microphone radio.

To catch the dark matter signal, they propose:

Colder Temperatures: Going from 4 Kelvin to near 0 Kelvin to reduce noise.
Bigger Trampolines: Using larger, heavier membranes to catch more of the "wind."
Arrays: Instead of one detector, they suggest building a whole grid of them working together, like a choir of microphones, to amplify the signal.

In summary: The team built a super-sensitive, frozen trampoline to listen for the invisible wind of dark matter. They didn't hear the wind, but they proved their microphone is the best in the world and showed the blueprint for building a much better one to finally catch the ghost.

Technical Summary: Optomechanical Accelerometer Search for Ultralight Dark Matter

Problem Statement
Ultralight dark matter (UDM) with masses $m_{\text{DM}} \lesssim 10 \text{ eV}/c^2$ is hypothesized to form a coherent, oscillating field that exerts a differential acceleration on Standard Model (SM) objects. This acceleration depends on the objects' fractional charge difference (e.g., baryon minus lepton number, $B-L$ ) and the local dark matter density. Detecting this signal requires measuring minute differential accelerations between two free-falling bodies with different material compositions. While cavity optomechanical systems have been proposed as table-top detectors for such wave-like dark matter, realizing the necessary sensitivity at high frequencies ( $\sim$ kHz) remains a significant challenge due to thermal noise, ambient vibrations, and the difficulty of thermalizing micromechanical resonators to cryogenic temperatures while maintaining quantum-limited readout.

Methodology
The authors present a resonant search for UDM using a cryogenic cavity optomechanical accelerometer. The experimental setup consists of the following key components and protocols:

Device Architecture: The detector utilizes a pair of vertically integrated $\text{Si}_3\text{N}_4$ membranes (a "trampoline" and a square membrane) with different stiffnesses, mounted on a silicon chip which is fixed to a copper (Cu) plate. The Cu plate has a higher $B-L$ density than the $\text{Si}_3\text{N}_4$ and Si, creating a differential coupling to the dark matter field.
Cryogenic and Vibration Isolation: The device operates in a Bluefors LD-4K closed-cycle cryostat. To mitigate ambient vibrations, the detector is mounted on a custom 2-foot pendulum vibration isolation system (VIS) suspended from the mixing stage, providing an isolation factor of $\sim 10^{10}$ at the trampoline's resonance frequency ( $f_m \approx 39 \text{ kHz}$ ).
Readout and Control:
- Displacement Readout: A tunable diode laser (840–852 nm) probes the Fabry-Perot cavity formed by the membranes. The laser is tuned to the side of a fringe to maximize displacement sensitivity, achieving shot-noise-limited readout.
- Feedback Cooling: An auxiliary 658 nm laser provides radiation-pressure feedback to damp the mechanical resonance, reducing thermal noise and narrowing the effective bandwidth.
- Photothermal Tuning: The probe laser power is modulated to induce photothermal heating, allowing the mechanical resonance frequency ( $f_m$ ) to be scanned over a 25 Hz range. This enables a resonant search across multiple frequency bins without mechanical reconfiguration.
Data Analysis: The search employs matched-filter statistics. The acceleration power spectral density (PSD) is estimated by inverting the mechanical susceptibility of the closed-loop system. The data is analyzed against a model UDM signal lineshape ( $G_{\text{DM}}$ ) derived from the Maxwell-Boltzmann velocity distribution of dark matter in the Galactic halo. A composite estimator is constructed using weighted averaging over the frequency scan, and upper bounds are set using both frequentist and Bayesian frameworks, accounting for the "look-elsewhere effect" over the search span.

Key Contributions

Experimental Realization: The work demonstrates the first resonant search for UDM using a cavity optomechanical accelerometer operating at 4 K with a millimeter-scale $\text{Si}_3\text{N}_4$ trampoline ( $m \sim 10 \text{ ng}$ , $Q_m \sim 10^8$ ).
Performance Metrics: The system achieves an acceleration sensitivity of $10 \text{ ng}_0/\sqrt{\text{Hz}}$ over a 30 Hz bandwidth near resonance. The mechanical quality factor of the trampoline was measured to be $Q_m = 6 \times 10^7$ at 4 K.
Signal Processing: The paper details a robust data analysis pipeline involving photothermal frequency scanning and matched-filtering to extract upper bounds on the dark matter coupling strength, despite the absence of a detected signal.

Results

Search Outcome: No evidence of ultralight dark matter was observed. The distribution of the weighted power estimators was consistent with a near-Gaussian distribution expected from thermal and readout noise.
Constraints: The experiment sets an upper bound on the $B-L$ coupling constant of $g_{B-L} \approx 1 \times 10^{-17}$ at a frequency of $\sim 39 \text{ kHz}$ . This limit is consistent with the thermal noise floor of the experiment and sits above the constraints currently set by equivalence principle tests (e.g., Eöt-Wash) and LIGO at lower frequencies.
Noise Characterization: The acceleration noise was dominated by thermal noise ( $S_{\text{th}}^a$ ) and displacement imprecision ( $S_{\text{imp}}^x$ ). Photothermal heating was observed at a rate of $0.4 \text{ K}/\mu\text{W}$ , which was utilized for frequency tuning but resulted in a 3-fold increase in thermal noise at higher powers.

Significance and Claims
The authors frame this work as a "first generation" search that validates the working principle of a cavity optomechanical vector UDM detector. The paper claims modest success in demonstrating:

Stable, Quantum-Limited Operation: The ability to operate a high- $Q$ mechanical oscillator in a cryogenic environment with shot-noise-limited readout and radiation-pressure feedback cooling.
Scalability: The platform offers a path toward competitive constraints on vector-mediated dark matter interactions through future optimizations. Specifically, the authors note that reaching the theoretical sensitivity limits proposed in prior literature ( $g \sim 10^{-25}$ $g \sim 1 0^{- 25}$ ) would require:
- Lower operating temperatures (10 mK).
- Larger test masses (centimeter-scale membranes) with higher $m \times Q_m$ products.
- Higher cavity finesse ( $\mathcal{F} \sim 100$ ).
- Improved vibration isolation ( $10^{-12} \text{ g}_0/\sqrt{\text{Hz}}$ ).
- Longer integration times (up to a year).
Future Directions: The paper highlights that optimizing membrane geometry (e.g., 3D mass-loading) and utilizing multiplexed arrays probed by squeezed light could further enhance sensitivity, potentially making these detectors competitive with other leading dark matter search techniques.

The study concludes that while no signal was found, the experiment successfully demonstrates the necessary technologies and analysis routines for a scalable approach to resonant UDM detection.