Coherent Elastic Neutrino-Nucleus Scattering Explained

If you were designing a neutrino detector from scratch, the process you would most want to exploit is coherent elastic neutrino-nucleus scattering — CEvNS, pronounced “sevens”. It is the largest neutrino cross-section at the MeV energies relevant to reactor, solar, and supernova neutrinos, roughly two orders of magnitude above inverse beta decay. The catch, which kept the process unobserved for forty-three years after its prediction, is that the only detectable signal is a nuclear recoil of a few keV — too small for 1970s detector technology, and still a challenge today.

The prediction

In 1973, the Gargamelle bubble chamber at CERN detected weak neutral currents — neutrino-electron scattering mediated by the $Z^0$ boson without flavor change. The $Z$ couples to quarks as well as leptons, so neutrinos should also scatter neutrally off nucleons.

Daniel Freedman at MIT, writing in 1974, pointed out that at low momentum transfer the scattering would be coherent. When the momentum transfer $q$ is small compared to the inverse nuclear size $R^{-1}$, the neutrino’s de Broglie wavelength exceeds the nucleus and all the nucleons in the nucleus scatter in phase. The amplitudes add coherently rather than incoherently.

For weak neutral currents, the nuclear vector coupling is $Q_W=(1-4{\sin }^2{\theta }_W)Z-N$ The proton term $(1-4{\sin }^2{\theta }_W)\approx 0.075$ is small. The neutron term $N$ dominates. So $Q_W\approx -N$, and the cross-section is $\sigma \propto Q_W^2\approx N^2$ This quadratic enhancement in neutron number is the signature of CEvNS. For heavy nuclei with $N\sim 70$, it gives an enhancement of $\sim 5000$ over incoherent nucleon-level scattering — and the largest neutrino cross-section at energies below about 50 MeV.

Why it took 43 years to observe

The observable in CEvNS is a recoiling nucleus. Kinematics set the recoil energy at a few keV — two to three orders of magnitude below the threshold of the neutrino detectors of the 1970s, which were optimised for MeV-range electron or gamma signals. Detecting keV recoils demanded detector technologies that did not yet exist.

They emerged slowly, largely from the direct dark-matter detection community. Low-threshold germanium bolometers, silicon charge-coupled devices, and cryogenic scintillators matured through the 2000s. By 2015 the COHERENT collaboration, working at the Spallation Neutron Source at Oak Ridge, was ready.

The 2017 observation

The Spallation Neutron Source is a 1.4-megawatt pulsed proton accelerator built for materials-science neutron production. As a by-product, each 1-microsecond proton pulse on the mercury target produces stopped pions. Their decay chain, ${\pi }^{+}\to {\mu }^{+}+{\nu }_{\mu },{\mu }^{+}\to e^{+}+{\nu }_e+{\bar{\nu }}_{\mu }$ yields a well-characterised neutrino flux: prompt ${\nu }_{\mu }$ in the first microsecond, delayed ${\nu }_e$ and ${\bar{\nu }}_{\mu }$ following with the 2.2 μs muon lifetime. The pulsed timing and two-component decay profile give enormous background-rejection power.

COHERENT placed a 14.6-kilogram CsI[Na] scintillator crystal about 20 metres from the target in a shielded utility corridor called “Neutrino Alley”. The detector operated with a trigger threshold of a few photoelectrons, corresponding to nuclear recoils near 5 keV.

After 308 days of exposure, COHERENT observed 134 CEvNS candidate events above background at 6.7σ significance. The measured cross-section matched Standard Model expectations within uncertainties. Subsequent runs with liquid argon and germanium targets confirmed the predicted $N^2$ scaling across different nuclei.

What CEvNS enables

CEvNS is not just the completion of an old prediction. It opens an experimental sub-field.

Precision Standard Model tests. The cross-section depends on the weak mixing angle ${\sin }^2{\theta }_W$ at low momentum transfer — a regime not accessible to atomic parity violation or $Z$-pole measurements. COHERENT and the reactor-CEvNS program extract ${\sin }^2{\theta }_W$ at new kinematic points.

Non-standard interactions. Any beyond-Standard-Model modification to neutrino-quark couplings would distort the CEvNS rate and spectrum. Current measurements already constrain large regions of the non-standard-interaction parameter space at the tree level.

Nuclear structure. The coherent amplitude includes a form factor that falls off as $q$ approaches the inverse nuclear radius. Measuring the recoil spectrum across different targets maps the neutron distribution of nuclei — a largely independent probe of nuclear structure.

The neutrino floor. In direct dark-matter detection experiments, CEvNS from solar and atmospheric neutrinos sets a fundamental background: at sufficiently low WIMP cross-sections, neutrinos produce indistinguishable nuclear recoils. The “neutrino floor” defines the ultimate reach of current WIMP-search technology.

Applied research. The same physical process — a neutrino transferring keV-scale momentum to a target nucleus — is the foundation of applied energy-conversion research. The Master Equation framework expresses device output as an integral over an effective flux and an effective cross-section; for the neutrino contribution, that cross-section is fundamentally the CEvNS cross-section.

Reactor CEvNS

A parallel experimental program pursues CEvNS at nuclear reactors. Reactor antineutrinos have energies of 1–10 MeV — lower than the SNS beam and therefore producing even smaller recoils. The detection threshold demands are accordingly harder, but the cleaner flux composition (purely ${\bar{\nu }}_e$) and steady rate make reactor CEvNS particularly valuable for beyond-Standard-Model searches.

Experiments at reactor sites — CONUS, CONNIE, Dresden-II, RED-100, NUCLEUS — use germanium, silicon CCDs, and cryogenic calorimeters, aiming to push thresholds below 100 eV. A credible CEvNS observation at reactor energies would open the cleanest possible window onto non-standard neutrino physics.

Outlook

CEvNS is the last tree-level Standard Model neutrino process to be observed. Its characterisation is now an active frontier: multi-target measurements of the $N^2$ scaling, precision ${\sin }^2{\theta }_W$ extractions, NSI limits, and reactor-energy sensitivity are all ongoing. In a broader sense, the 43-year gap between Freedman’s prediction and COHERENT’s observation is a reminder that fundamental processes can be invisible for decades not because of any theoretical uncertainty but because the right detector does not yet exist. CEvNS took a generation of direct-dark-matter R&D to make possible. What other Standard Model processes are still waiting for the right instrument?