The Neutrino Floor: Where WIMP Searches Hit Their Bedrock

For four decades, direct dark-matter detection has been guided by a single principle: build the most sensitive, most background-free detector possible, look for nuclear recoils from WIMP interactions, and push the WIMP cross-section limit downward with every new generation. The progression has been steady: from the early NaI and Ge detectors of the 1990s with sensitivity at $10^{-40}$ cm² to today’s ton-scale liquid-xenon TPCs reaching $10^{-48}$ cm². The ten-orders-of-magnitude improvement in sensitivity is one of the most impressive sustained experimental advances in fundamental physics.

But this progression cannot continue indefinitely. As WIMP cross-sections approach $10^{-49}$ cm² for masses around 10 GeV, the detectors begin to see a population of nuclear recoils that are not from dark matter. They are from neutrinos — solar, atmospheric, and supernova — coherently scattering off detector nuclei via the same CEvNS process that COHERENT discovered in 2017. These neutrino-induced recoils are kinematically indistinguishable from WIMP recoils. They cannot be subtracted by better shielding, lower radioactivity, or any other detector improvement. They are the bedrock.

This neutrino floor is the most consequential intersection between neutrino physics and dark-matter physics. After 30 years of viewing the two fields as separate, the experimental programmes are now converging — and direct dark-matter detection is becoming, simultaneously, low-energy neutrino physics.

How the floor arises

The same coherent elastic neutrino-nucleus scattering that COHERENT discovered at the SNS produces nuclear recoils at any low-threshold detector exposed to a neutrino source. The Earth has many such sources:

• Solar neutrinos: ${}^8$B and ${}^7$Be neutrinos producing recoils up to ~5 keV at xenon
• Atmospheric neutrinos: GeV-energy neutrinos producing recoils up to ~100 keV
• Diffuse supernova neutrino background: tens-of-MeV neutrinos producing recoils up to ~10 keV
• Geological antineutrinos: low-energy contribution at threshold-eligible energies

For a WIMP detector with several keV recoil-energy threshold, the dominant contributor is solar ${}^8$B neutrinos, which produce the keV-scale recoils that match the WIMP signature for a ~5 GeV mass.

The CEvNS rate per detector mass is calculable from the well-measured neutrino fluxes and cross-sections: $\frac{dN}{d{E}_{R}dtdM}=\int d{E}_{\nu }\Phi (E_{\nu })\frac{d{\sigma }_{CEvNS}}{d{E}_R}$ For a xenon-based experiment, the integrated solar-neutrino-induced rate above 1 keV is approximately 1 event per ton-year per cm²-equivalent dark-matter cross-section of $4\times 10^{-49}$ cm². This is the floor for liquid-xenon WIMP detection in the few-GeV mass range.

Mass-dependent floor structure

The neutrino floor is not a single number; it depends on WIMP mass.

Low WIMP mass ($\lesssim 5$ GeV): ${}^8$B solar neutrinos dominate. The floor is at $\sim 4\times 10^{-49}$ cm² for a xenon detector.

Mid WIMP mass ($5-30$ GeV): Lower-energy solar neutrinos contribute, but the floor is somewhat lower because the WIMP recoil spectrum extends higher than the solar-neutrino-induced one. The floor is around $10^{-49}$ cm².

High WIMP mass ($\gtrsim 100$ GeV): Atmospheric neutrinos take over as the dominant background. The floor is approximately $10^{-50}$ cm² for these masses.

These numbers are for spin-independent WIMP-nucleon cross-section in standard-halo-model assumptions. For spin-dependent WIMP couplings, the floor structure differs because nuclear matrix elements vary with target.

The “fog” rather than “floor”

The classic description as a “floor” implied a hard lower limit below which no WIMP signal could be distinguished. In practice, the situation is more nuanced.

The neutrino fog, as the modern terminology describes it, is a region of cross-section parameter space where the WIMP signal becomes increasingly contaminated by neutrino background. The contamination depends on:

1. Detector exposure (longer exposure = more neutrino events = more confusion)
2. Energy threshold (lower threshold = more solar-neutrino events)
3. Statistical analysis sophistication (likelihood ratio tests with full energy/angular spectra can extract a WIMP signal even slightly below the naive floor)

A 2018 reanalysis by Billard, Strigari, and Figueroa-Feliciano clarified that the floor is closer to a “saturating sensitivity” than a hard cutoff. With sufficient exposure, an experiment can in principle detect WIMPs below the naive floor — but the discovery significance grows extremely slowly with exposure (square-root scaling becomes much weaker). For practical purposes, the cross-section reach saturates within a factor of 3-5 of the floor.

Current experimental status

Several experiments are operating at or near the floor:

LZ (LUX-ZEPLIN) at the Sanford Underground Research Facility, South Dakota — 7 tonnes of xenon (5.5-tonne fiducial). First WIMP results in 2022 set limits at $\sim 10^{-47}$ cm² for 30-GeV WIMPs. Continued running through 2025-2026 should reach within a factor of 5 of the floor.

XENONnT at the Gran Sasso National Laboratory, Italy — 8.6 tonnes of xenon (5.9-tonne fiducial). Similar sensitivity to LZ, operating in parallel for cross-checks.

PandaX-4T at the China Jinping Underground Laboratory — 4 tonnes of xenon. Slightly less sensitive than LZ/XENONnT but provides an independent measurement.

SuperCDMS (Sudbury, Canada) — Cryogenic Ge bolometers, sensitivity peaks at 1-10 GeV WIMPs. Already within range of solar-neutrino backgrounds for low-mass searches.

By the late 2020s, all current experiments will have either detected the solar-neutrino CEvNS background as a “first detection of the neutrino floor” or set limits at it. Either outcome is scientifically valuable: the first directly demonstrates the floor’s existence; the second establishes the experiments’ calibration of the floor.

Strategies to push past the floor

Three approaches address the neutrino-fog limit:

Directional detection. WIMPs in the galactic halo move at $\sim 220$ km/s, much slower than the Earth’s rotation around the Sun. Their incoming direction in Earth-frame coordinates correlates strongly with the constellation Cygnus (the direction of the solar system’s motion through the halo). Solar neutrinos, by contrast, come from the Sun. Atmospheric neutrinos arrive isotropically from above. A directional detector — typically a low-pressure gas time-projection chamber that can resolve recoil tracks — could distinguish WIMP-from-Cygnus recoils from neutrino-induced ones. The DRIFT, NEWAGE, and DMTPC programmes are pursuing this approach. Sensitivity is currently far below conventional detectors but can in principle exceed them at the floor.

Spectral discrimination. WIMPs have a known recoil spectrum (Maxwellian galactic halo, exponentially falling). Solar neutrinos produce a different spectrum (peaked at ${}^8$B’s endpoint kinematics). With enough events and high resolution, the two can be statistically separated. This works best when the WIMP cross-section is well above the floor; near the floor, statistical separation becomes prohibitively expensive in exposure.

Alternative dark-matter candidates. If WIMPs aren’t dark matter, the neutrino floor matters less. Sub-GeV dark matter (where the WIMP recoil energies are below the solar-neutrino-recoil range), axions and axion-like particles (which couple electromagnetically rather than via WIMP scattering), and primordial black holes (different signatures entirely) are all alternatives whose backgrounds are distinct from the WIMP neutrino floor.

CEvNS as both background and signal

The same physics that produces the neutrino floor also enables CEvNS measurement at COHERENT, CONUS, RED-100, and others. Direct dark-matter experiments at the neutrino floor are simultaneously neutrino experiments, observing solar neutrinos through their CEvNS recoils with high statistics.

This dual role has been recognised by the community. LZ’s run plan explicitly includes a “neutrino physics” analysis: the rate of ${}^8$B-induced CEvNS events. With sufficient exposure (~10 tonne-years), LZ should detect ${}^8$B CEvNS at greater than 5σ significance independently of any dark-matter consideration. This would be a precision measurement of the solar ${}^8$B neutrino flux through a completely different channel than Super-Kamiokande, providing an independent cross-check of the Standard Solar Model.

The same applies to atmospheric neutrinos. At very high WIMP masses, the floor is dominated by atmospheric neutrinos. Detecting these via CEvNS at multi-tonne dark-matter detectors would give a direct measurement of the atmospheric neutrino flux at energies below the muon-tracking threshold of conventional atmospheric-neutrino detectors.

Implications

The neutrino floor reshapes the dark-matter detection landscape in several ways.

For experiment design: New detectors must integrate a “neutrino physics” mode. The floor is no longer a problem to overcome but a calibration target and physics opportunity.

For physics interpretation: A WIMP signal near the floor would require careful statistical separation from the irreducible neutrino background. Discovery claims will need spectral information, not just rate excesses.

For community structure: The previously separate dark-matter and neutrino-physics communities are now overlapping at the experimental level. Conferences increasingly feature joint sessions; experimental papers increasingly address both topics.

For theoretical priorities: The clean WIMP search regime is ending. Either dark matter is found at masses where the floor is approachable (in which case CEvNS-like events confirm it), or it lies in regions where the floor is irrelevant (sub-GeV, axion, etc.). Either outcome reframes the search programmes.

The neutrino floor is, in physics terms, a measurement that has been waiting to happen for decades. Now that the floor is being approached, the detection of solar-neutrino CEvNS events at LZ or XENONnT will be one of the most quietly important measurements of the late 2020s — establishing both the floor’s existence and the bridge between two fields that have long converged on the same instrument.