⚛️ phenomenology

Scalable Dark Matter Searches Using Integrated Photonics

This paper proposes a scalable, integrated photonics-based experimental approach using refractive index-modulated resonators coupled to single-photon detectors to search for sub-eV dark matter candidates, such as axion-like particles and dark photons, by leveraging nanophotonic confinement to explore previously uncharted parameter space.

Original authors: Nikita Blinov, Christina Gao, Roni Harnik, Ryan Janish, Neil Sinclair

Published 2026-02-23

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Nikita Blinov, Christina Gao, Roni Harnik, Ryan Janish, Neil Sinclair

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

Imagine the universe is filled with a ghostly, invisible substance called Dark Matter. We know it's there because it holds galaxies together with its gravity, but we have no idea what it actually is. It's like trying to find a specific type of invisible fish in the ocean without knowing if it's big, small, fast, or slow.

For decades, scientists have been fishing for these "ghost particles" using massive, expensive, and very slow equipment. They usually build huge, hollow metal boxes (like giant tuning forks) and slowly tune them up and down, one frequency at a time, hoping to catch a signal. It's like trying to find a specific radio station by turning the dial one tiny notch at a time, waiting years to scan the whole dial.

This paper proposes a revolutionary new way to fish: instead of one giant, slow tuning fork, let's build a massive, high-tech fishing net made of millions of tiny, microscopic sensors.

Here is the breakdown of their idea, using simple analogies:

1. The Problem: The "One-Station-at-a-Time" Bottleneck

Current experiments are like having a single, very sensitive microphone in a huge stadium. If a singer (Dark Matter) is singing a specific note, the microphone might hear it. But if the singer changes the note, you have to physically move the microphone or retune it. Since Dark Matter could be singing any note in a huge range (specifically the "eV" range, which corresponds to light frequencies), scanning one by one would take forever.

2. The Solution: The "Smartphone Chip" Approach

The authors suggest using Integrated Photonics. Think of this as the technology inside your smartphone or the internet's fiber-optic cables, but shrunk down to the size of a grain of sand.

Instead of building one giant box, they propose printing hundreds of thousands of tiny resonators (like microscopic drums or rings) onto a single silicon chip, similar to how computer chips are made.

The Analogy: Imagine a piano. Current experiments are like having one piano and trying to play every note one by one. This new idea is like building a piano with 100,000 keys, where every key is a tiny, perfect sensor, all connected to a single recording device.

3. The Magic Trick: "Frequency Multiplexing"

Here is the clever part. If you put 100,000 identical drums next to each other, they might interfere with each other and cancel out the signal (like noise-canceling headphones).

To fix this, the team designs the drums so that every single one is tuned to a slightly different note.

The Analogy: Imagine a choir where every singer is singing a different note at the same time. Because they are all different, they don't clash. The "Dark Matter" (the ghostly wind) blows through the room. If the wind happens to match the pitch of one specific singer, that singer starts to vibrate loudly.
Because the singers are all different, the system can listen to hundreds of different "notes" (Dark Matter masses) simultaneously. It's not scanning; it's a broadband spectrum analyzer that watches the whole radio dial at once.

4. Catching the Ghost: The "Micro-Detector"

When a Dark Matter particle hits one of these tiny resonators, it doesn't create a loud bang. It creates a single, tiny flash of light (a photon).

The chip is connected to ultra-sensitive detectors (like a super-powered camera pixel or a superconducting wire) that can see a single photon.
If the "ghost" is there, one of the millions of tiny drums will ring, sending a flash of light to the detector.

5. Why This Changes Everything

Scalability: Because these chips are made using standard factory techniques (like making computer processors), we can mass-produce them. We can put millions of these sensors on a single wafer.
Speed: Instead of taking 10 years to scan a range of masses, this system could scan the entire "eV" range in a few years or even months.
Cost: It uses the mature, cheap technology of the semiconductor industry rather than building unique, custom-made physics labs for every experiment.

The Bottom Line

The authors are saying: "Stop trying to catch the invisible fish with one giant, slow net. Let's build a massive, automated factory of tiny, high-speed nets that can catch the fish no matter what size it is, all at the same time."

If successful, this technology could finally reveal the nature of Dark Matter, potentially solving one of the biggest mysteries in physics, all by using the same manufacturing tricks that put the internet in our pockets.

1. The Problem

The microscopic nature of Dark Matter (DM) remains unknown. While ultralight DM candidates (masses $\ll$ electron mass) are theoretically well-motivated (e.g., axions, axion-like particles [ALPs], and dark photons [DPs]), current experimental searches face a critical gap in the electronvolt (eV) mass range (corresponding to optical frequencies, 0.1–3 eV).

Limitations of Current Methods: Traditional "haloscope" experiments (like ADMX) use high-quality microwave cavities to resonantly convert DM into photons. These are highly effective in the $\mu$ eV range (GHz frequencies) but cannot scale to the eV range due to the physical size constraints of optical cavities and the inability to scan frequencies rapidly.
The Coherence Challenge: Wave-like DM has a finite spatial coherence length ( $\lambda_{dB} \approx 0.4$ mm for eV-scale DM). If a detector array exceeds this size, signals from different parts of the detector interfere destructively, preventing the simple summation of signals from many resonators to boost sensitivity.
The Gap: The eV mass range is largely unexplored, representing a significant discovery opportunity that complements searches for solar axions and stellar evolution constraints.

2. Methodology

The authors propose a frequency-multiplexed integrated photonic approach to search for DM in the 0.1–3 eV range. The core concept involves fabricating arrays of thousands of microscopic optical resonators on semiconductor wafers.

A. Physical Mechanism

DM as a Source: In models where DM couples to electromagnetism (axions via $g_{a\gamma}$ or dark photons via kinetic mixing $\chi$ ), the ambient DM field acts as a classical source current ( $J_D$ ) in Maxwell's equations.
Resonant Excitation: When the DM mass ( $m_D$ ) matches the resonant frequency of an optical mode, it drives that mode, generating signal photons.
Phase Matching: A major hurdle is the momentum mismatch between non-relativistic DM ( $v \sim 10^{-3}c$ ) and relativistic optical photons. The authors propose using periodically modulated refractive index structures (e.g., etched microrings or photonic crystal slabs). These structures provide the necessary momentum kick to satisfy conservation laws, enabling efficient coupling.
Mode Selection: The design utilizes Bound States in the Continuum (BIC) or similar modes that offer high quality factors ( $Q$ ) and strong overlap with the DM field.

B. The Scaling Solution: Frequency Multiplexing

To overcome the coherence length limitation, the authors reject the strategy of coupling many identical resonators to a single readout (which leads to destructive interference and signal saturation). Instead, they propose:

Frequency Multiplexing: Fabricating an array of resonators, each with a distinct resonant frequency ( $\omega_R, \omega_R(1+\delta), \dots$ ).
Independent Operation: Because each resonator targets a different DM mass, they operate independently. There is no interference between them, even if they are coupled to a single waveguide bus.
Readout: Signal photons from the resonators are guided via on-chip waveguides to low-noise single-photon detectors (Superconducting Nanowire Single-Photon Detectors [SNSPDs] for $<1.12$ eV and Skipper-CCDs for $>1.12$ eV).
Scalability: This approach leverages semiconductor manufacturing to create hundreds of thousands of resonators on a single wafer, effectively creating a broadband "spectrum analyzer" that monitors hundreds of DM masses simultaneously.

3. Key Contributions

Novel Experimental Paradigm: Shifts DM detection from sequential scanning of single bulky cavities to parallel monitoring of mass-produced photonic chips.
Theoretical Framework for Coupled Arrays: Developed a coupled-mode theory for arrays of resonators driven by incoherent DM sources. They demonstrated that while same-frequency arrays suffer from coherence limits, frequency-diverse arrays allow for linear scaling of the monitored bandwidth without signal suppression.
Phase Matching via Nanophotonics: Demonstrated that periodic photonic structures (like grooved microrings or patterned slabs) can resolve the momentum mismatch problem, achieving overlap factors ( $\eta$ ) of 0.01–0.1.
Sensitivity Projections: Provided concrete sensitivity estimates for both Axion/ALP and Dark Photon models across the 0.1–3 eV range.

4. Results and Sensitivity Projections

The paper projects that this approach can probe previously unexplored parameter space:

Axion/ALP Sensitivity:
- Assuming a 9.4 T MRI magnet bore filled with photonic arrays, the experiment can reach axion-photon couplings of $g_{a\gamma} \sim 10^{-10} \text{ GeV}^{-1}$ .
- This sensitivity reaches into the theoretically motivated QCD axion band.
- The search strategy involves two phases: Phase 1 (0.1–1.12 eV) using SNSPDs and Phase 2 (1.12–3 eV) using CCDs, requiring approximately 2 years for full coverage.
Dark Photon Sensitivity:
- Dark photon searches do not require external magnetic fields, making them a more accessible "proof-of-principle" entry point.
- Even with modest active volumes (0.1 cm³), the setup can achieve kinetic mixing sensitivities of $\chi \sim 10^{-13}$ , surpassing current limits from experiments like LAMPOST and XENON1T.
Scalability: A single 15 cm (6-inch) wafer can accommodate $\sim 10^6$ resonators. By combining multiple wafers, the active volume can be scaled up to $\sim 1000 \text{ cm}^3$ , limited only by magnet bore size (for axions) or lab space (for dark photons).

5. Significance

Bridging the Gap: This work addresses the most significant blind spot in current DM searches: the eV mass range.
Technological Leverage: It utilizes the maturity, precision, and scalability of the integrated photonics industry (currently driving AI and data center growth) for fundamental physics. This allows for the mass production of detectors that would be impossible to assemble individually.
Feasibility: The required quality factors ( $Q \sim 500-5000$ ) and fabrication tolerances are well within current state-of-the-art capabilities (silicon, silicon nitride, lithium niobate platforms).
Roadmap to Discovery: The paper outlines a clear path: start with dark photon searches (no magnets needed) to validate the technology, then scale up to axion searches using large-bore magnets. This approach could potentially open a new window on dark matter discovery within the next decade.

In summary, the paper proposes a transformative, scalable architecture for dark matter detection that overcomes the fundamental coherence limits of traditional methods by utilizing frequency-multiplexed integrated photonics, offering a realistic path to exploring the elusive eV-scale dark matter regime.