Statistics of Daily Modulation in Dark Matter Direct Detection Experiments

This paper presents a statistical analysis demonstrating that daily modulation in dark matter direct detection experiments, particularly using anisotropic solid-state detectors like trans-stilbene, offers a discovery significance that scales with the square root of exposure time without saturating due to background systematics, and shows that optimizing detector orientations can reduce the required exposure by a factor of approximately five.

The Gist

The Big Picture: Hunting the Invisible Ghost

Imagine you are trying to find a specific type of ghost (Dark Matter) that is constantly flying through your house. You can't see it, but you have a special motion sensor (a detector) that clicks whenever a ghost bumps into it.

The problem? Your house is also full of other things that make noise: the fridge humming, the wind rattling the window, and the cat walking on the floor. These are your background noises. If you just count the total clicks, you can't tell if it's the ghost or just the cat.

For decades, scientists have looked for a "seasonal" pattern. Because Earth orbits the Sun, we move faster through the ghost wind in June and slower in December. This creates a yearly rhythm. But this paper is about a daily rhythm.

The Daily Rhythm: The Earth's Spin

As Earth spins on its axis once every 24 hours, our detector changes direction. Imagine holding a net up to catch rain. If you hold it flat, you catch a lot. If you tilt it, you catch less. As the Earth spins, our detector tilts relative to the "ghost wind."

If the detector is made of special materials (like trans-stilbene, a type of crystal), it becomes "directional." It catches ghosts better when facing one way and worse when facing another. This creates a daily modulation: the number of clicks goes up and down every 24 hours.

The Problem: The "Noisy" Background

The tricky part is that the background noise (the cat, the wind) might also have a daily rhythm. Maybe the temperature changes during the day, making the detector click more often at noon and less at midnight.

If the ghost signal and the background noise both wiggle up and down every 24 hours, they can hide inside each other. It's like trying to hear a specific violin note while a whole orchestra is playing the same melody.

The Solution: The "Three-Detector Orchestra"

The authors of this paper came up with a statistical recipe to solve this. Instead of using just one detector, they propose using multiple detectors (specifically, three) and arranging them in a clever way.

Here is the analogy:
Imagine you are trying to hear a specific song (the Dark Matter signal) in a room where a loud fan (the background noise) is also humming.

• Detector 1 is facing North.
• Detector 2 is facing East.
• Detector 3 is facing South.

Because the Earth is spinning, the "ghost wind" hits these three detectors at different times. The ghost signal will peak at different hours for each detector. However, the background noise (like the fan) is likely the same for all of them, or at least behaves similarly.

By optimizing the angles of these three detectors, the scientists can arrange them so that:

1. The Ghost Signal creates a complex, unique pattern across the three detectors that is hard to fake.
2. The Background Noise gets "cancelled out" or averaged out because the detectors are looking at different parts of the sky at different times.

The "Magic" Result

The paper does some heavy math (using something called "Fisher Information" and "Chi-square") to prove a very cool point:

If you arrange your detectors perfectly, you can find the ghost signal with 5 times less data than if you just used one detector or arranged them randomly.

Think of it like this:

• Bad Strategy: You try to listen to the ghost with one ear while the fan is humming. You need to listen for 5 years to be sure.
• Good Strategy: You use three ears (detectors) arranged in a circle. You can hear the ghost clearly in just 1 year.

Why This Matters

1. Saving Money and Time: Dark Matter detectors are incredibly expensive and heavy. If you can get the same result with 1/5th of the detector mass, you save millions of dollars and years of construction time.
2. Beating the Noise: Even if the background noise is huge and changes every day, this method can still separate the ghost signal from the noise.
3. The "Sidereal" Trick: The paper also mentions a clever time-trick. Because Earth orbits the Sun, the "ghost wind" shifts slightly every day. By stacking data over a year and adjusting for this shift (called "Sidereal Stacking"), the background noise cancels itself out almost perfectly, leaving only the ghost signal.

The Takeaway

This paper is a "user manual" for the next generation of Dark Matter hunters. It tells them: "Don't just build a bigger detector; build a smarter array of detectors."

By carefully rotating and angling multiple detectors, we can turn the Earth's daily spin into a super-powerful filter that separates the faint whisper of Dark Matter from the loud roar of everyday noise. It turns a difficult statistical puzzle into a solvable geometry problem.

Technical Summary

1. Problem Statement

Dark matter (DM) direct detection experiments rely on identifying a time-dependent modulation in the event rate to distinguish a DM signal from background noise. While annual modulation (due to Earth's orbit around the Sun) has been extensively studied, daily modulation (due to Earth's rotation) has recently gained attention, particularly for sub-GeV DM detection using anisotropic solid-state detectors (e.g., trans-stilbene crystals).

The core challenges addressed in this paper are:

• Unknown Backgrounds: Real-world experiments face unknown background rates that may be time-independent (flat) or time-dependent (modulating with a 24-hour period, potentially mimicking the DM signal).
• Statistical Optimization: How to maximize the statistical significance of a daily modulation signal in the presence of these unknown, potentially modulating backgrounds.
• Detector Configuration: How to optimally orient multiple anisotropic detectors to break degeneracies between the DM signal and background modulations.

2. Methodology

The authors develop a comprehensive statistical framework for analyzing daily modulation in counting experiments.

A. Statistical Formalism

• Likelihood Construction: They construct a binned Poisson likelihood function for a single detector and generalize it to an ensemble of $N_{det}$ detectors. The total rate is modeled as $r(t) = A s(t) + b(t)$, where $A$ is the signal strength, $s(t)$ is the time-dependent signal profile (dependent on detector orientation), and $b(t)$ is the background.
• Weak-Signal Limit: The analysis assumes the weak-signal limit ($As(t) \ll b(t)$), which is physically relevant for discovery scenarios and ensures the total rate remains positive even if the signal term is negative in a fit.
• Test Statistics: They utilize the profile likelihood ratio test statistic ($q$) and derive its behavior in the Gaussian limit (large counts) and via the Fisher Information Matrix (FIM).

B. Figure of Merit (FOM)

A primary theoretical result is the derivation of the scaling for discovery significance ($N_\sigma$). In the background-dominated regime with a flat background, the significance scales as:
$$N_\sigma \propto f_{RMS} \sqrt{T}$$
Where:

• $T$ is the total exposure time.
• $f_{RMS}$ is the root-mean-square modulation amplitude of the signal profile.
• Crucially, the significance does not saturate due to systematic uncertainties in the background rate; it continues to improve with exposure time.

C. Multi-Detector Optimization

The paper extends the formalism to multiple detectors with independent or correlated backgrounds.

• Fisher Information: They use the FIM to calculate the variance of the signal strength estimator ($\sigma_A^2$). Minimizing this variance maximizes sensitivity.
• Background Degeneracy: They demonstrate that if the background phase is unknown, a single detector cannot optimally separate the signal from a modulating background. However, by using multiple detectors with different orientations, the relative phases of the signal and background can be manipulated to break degeneracies.
• Optimization Strategy: The authors numerically optimize detector orientations ($\theta, \phi$) and relative phases ($\psi_t$) to minimize $\sigma_A$ (or the marginalized sensitivity $\bar{\sigma}_A$ if the background phase is unknown).

D. Sidereal Stacking

To handle the fact that the DM wind direction shifts slightly over the year (changing the phase of the daily modulation relative to the solar day), the authors propose a shifted sidereal stacking technique. This bins data based on the time since the DM wind was at the zenith ("DM noon") rather than the local solar time, allowing data from different days to be stacked without destructive interference.

3. Key Contributions

1. General Statistical Framework: Provided a prescriptive formalism for daily modulation analysis that handles arbitrary numbers of detectors, arbitrary binning (binned vs. unbinned), and both flat and time-varying backgrounds.
2. Scaling Law Validation: Confirmed that in the background-dominated regime, the discovery significance scales as $f_{RMS}\sqrt{T}$, proving that systematic uncertainties in the background rate do not cause the sensitivity to saturate.
3. Multi-Detector Optimization: Demonstrated that optimizing the relative orientations of multiple anisotropic detectors can significantly enhance sensitivity, even when the background modulation phase is unknown.
4. Monte Carlo Validation: Validated the asymptotic Fisher information results against Monte Carlo simulations, showing that the scaling laws hold even in the "narrow-bin" limit where Poisson fluctuations are large.
5. Sidereal Stacking Scheme: Introduced a method to isolate sidereal daily modulation from solar-day modulating backgrounds by shifting the time binning reference.

4. Results

The authors performed example analyses using trans-stilbene detectors for sub-GeV DM (specifically a 5 MeV DM candidate with a light mediator).

• Single Detector: For a flat background, optimizing the detector orientation can increase the modulation amplitude ($f_{RMS}$) by a factor of ~16 (heavy mediator) to ~7 (light mediator) compared to the worst orientation.
• Multi-Detector with Unknown Background Phase:
  • In scenarios with strongly modulating backgrounds (where the background amplitude is 90% of the flat rate), optimizing three detectors reduced the required exposure time by a factor of ~5 (or reduced the required detector mass by the same factor) to achieve a specific discovery significance compared to using a single optimized detector or randomly oriented detectors.
  • The optimal configuration for three detectors involves orienting them such that their signal phases are spread throughout the day (roughly $2\pi/3$ apart), effectively canceling out the background modulation while preserving the DM signal.
• Robustness: The optimal detector configuration for a strongly modulating background is very similar to that for a weakly modulating background, suggesting that a single experimental setup can be robust against varying background amplitudes.
• Sidereal Stacking: In a scenario with a 24-hour modulating background ($f_{RMS} \approx 64\%$) and a much weaker DM signal ($f_{RMS} \approx 6.7\%$), stacking data over a full year using shifted sidereal time allows the background to destructively interfere, leaving a clean DM signal.

5. Significance

This work provides a critical "recipe" for the next generation of directional dark matter experiments.

• Feasibility: It proves that daily modulation is a viable discovery channel even in the presence of large, time-varying backgrounds, provided the experiment utilizes anisotropic materials and multiple detectors.
• Design Guidance: It offers concrete guidance for experimentalists on how to arrange detector crystals to maximize sensitivity, potentially reducing the cost (mass) and time required for a discovery by a factor of 5.
• Background Rejection: It highlights the unique power of directional detectors to distinguish between DM signals and environmental backgrounds (which often modulate on a solar day) by exploiting the sidereal nature of the DM wind.

In conclusion, the paper establishes that daily modulation is not just a theoretical curiosity but a powerful, statistically robust tool for dark matter discovery, provided that the statistical analysis is tailored to the specific anisotropic properties of the detector materials and the nature of the background.