Observation of Coherent Elastic Neutrino-Nucleus Scattering

Experimental setup

COHERENT exploits the intense pulsed neutrino flux produced as a by-product of the 1 MW Spallation Neutron Source at Oak Ridge National Laboratory. Each 600 ns proton pulse striking the mercury target produces stopped pions whose subsequent decay chain yields:

${\pi }^{+}\to {\mu }^{+}+{\nu }_{\mu }(\text{prompt, sub-μs})$ ${\mu }^{+}\to e^{+}+{\nu }_e+{\bar{\nu }}_{\mu }(\text{delayed, 2.2 μs exponential})$

The pulsed timing — 60 Hz at the SNS — combined with the two-component decay structure gives a powerful background rejection handle. “Neutrino Alley”, a utility corridor about 20 m from the target, provides shielding of ~12 m-water-equivalent and a quiet environment for detector deployment.

The 2017 measurement used a 14.6 kg CsI[Na] scintillator crystal. CsI has both Cs and I nuclei with similar $N$, allowing the aggregate coherent cross-section to be tested. The crystal was instrumented with a single photomultiplier, operating at a trigger threshold of ~5 photoelectrons corresponding to recoil energies of ~5 keV.

The result

After 308 live days of beam exposure, COHERENT observed 134 CEvNS candidate events over an expected Standard Model count of 173 ± 48 events — a 6.7σ observation. The significance accounts for timing, energy, and pulse-shape cuts that together reduce the beam-unrelated background to a manageable level.

The extracted cross-section is consistent with Standard Model expectations within experimental uncertainties. The fit also constrains non-standard neutrino-quark couplings, placing competitive limits in several orthogonal directions.

Technical challenges

CEvNS detection requires:

1. Very low recoil-energy thresholds (sub-keV to few-keV nuclear recoils)
2. Precise timing to separate signal from beam-unrelated backgrounds
3. Well-measured nuclear quenching factors — the fraction of recoil energy converted to scintillation
4. Low radioactive backgrounds in the relevant energy window

The CsI quenching factor was independently measured for this analysis, reducing a historically dominant source of systematic uncertainty.

Subsequent measurements

COHERENT has since extended the measurement to:

• Liquid argon (CENNS-10 detector, 2021) — confirming CEvNS on a light target with different $N$, testing the $N^2$ coherent scaling
• Germanium (2024) — lower threshold than CsI, accessing lower-energy recoils
• NaI[Tl] (planned multi-tonne scale) — higher statistics with both Na and I targets

Parallel programs at reactor sites — CONUS (Brokdorf/Leibstadt), CONNIE (Brazil), Dresden-II (USA), RED-100 (Russia) — pursue CEvNS with the cleaner reactor ${\bar{\nu }}_e$ flux, at recoil energies that demand even lower detector thresholds.

Significance

CEvNS is the largest neutrino cross-section at low energies and was the last unobserved tree-level Standard Model neutrino process. Its confirmation has implications across multiple programs:

• Precision SM tests of ${\sin }^2{\theta }_W$ at low momentum transfer
• Searches for beyond-SM physics — NSI, sterile neutrinos, neutrino magnetic moments
• Nuclear structure through form-factor dependence on neutron distribution
• Direct dark-matter detection, where CEvNS from astrophysical neutrinos sets the fundamental “neutrino floor”
• Applied research — the Master Equation framework for neutrinovoltaic conversion uses the CEvNS cross-section as the neutrino-channel σ_eff

The 43-year delay between Freedman’s 1974 prediction and COHERENT’s 2017 observation is one of the longer gaps between prediction and detection in particle physics, surpassed mainly by the gravitational-wave and Higgs-boson timelines. The delay reflects the engineering challenge of keV-scale recoil detection, not any uncertainty about the underlying physics.