Case Study 2: CE$\nu$NS — When Neutrinos Bounce Off Nuclei

A 43-Year Prediction

In 1973, Daniel Freedman at Los Alamos published a two-page paper in Physical Review D with a deceptively simple title: "Coherent effects of a weak neutral current." The paper predicted that neutrinos should scatter coherently off entire nuclei via the neutral current ($Z^0$ exchange), producing a tiny nuclear recoil. The cross section, Freedman noted, would be proportional to $N^2$ — the square of the neutron number — making it enormous by neutrino standards.

The prediction was immediately recognized as important, but it came with a catch that Freedman himself acknowledged: the only observable signal is a nuclear recoil of a few to a few tens of keV. In 1973, no detector technology existed that could see such small recoils above backgrounds. The Standard Model's prediction for coherent elastic neutrino-nucleus scattering (CE$\nu$NS, pronounced "sevens") would remain unconfirmed for 43 years.

The Physics of Coherence

Why $N^2$?

The coherent enhancement in CE$\nu$NS has the same origin as the coherent enhancement in X-ray diffraction from crystals or Rayleigh scattering of light from small particles. When the wavelength of the probe is larger than the target, the scattering amplitudes from individual scatterers add coherently.

For CE$\nu$NS, the relevant condition is:

$$\lambda_{\text{probe}} = \frac{\hbar}{q} \gg R_{\text{nucleus}}$$

where $q$ is the momentum transfer and $R$ is the nuclear radius. When this condition holds, the $Z^0$ boson "sees" the entire nucleus as a single scatterer. The amplitude is the sum of weak charges:

$$\mathcal{A} \propto \sum_{i=1}^{A} Q_{W,i} = Z Q_W^p + N Q_W^n$$

Since $Q_W^p = 1 - 4\sin^2\theta_W \approx 0.07$ and $Q_W^n = -1$:

$$\mathcal{A} \propto -N + 0.07Z \approx -N$$

The cross section, proportional to $|\mathcal{A}|^2$, scales as $N^2$. For cesium ($N = 78$), this means the CE$\nu$NS cross section is roughly $78^2 \approx 6,000$ times larger than the corresponding single-neutron cross section.

The Cross Section

The differential cross section for CE$\nu$NS is:

$$\frac{d\sigma}{dT_R} = \frac{G_F^2 M}{4\pi} Q_W^2 \left(1 - \frac{MT_R}{2E_\nu^2}\right) F^2(q^2)$$

where $T_R$ is the nuclear recoil kinetic energy, $M$ is the nuclear mass, $Q_W \approx N$ is the weak nuclear charge, and $F(q^2)$ is the nuclear form factor.

The maximum recoil energy is:

$$T_R^{\max} = \frac{2E_\nu^2}{M + 2E_\nu} \approx \frac{2E_\nu^2}{M}$$

For ${}^{133}\text{Cs}$ and $E_\nu = 30\,\text{MeV}$ (typical of stopped-pion neutrinos):

$$T_R^{\max} \approx \frac{2 \times (30)^2}{133 \times 931} \approx 14.5\,\text{keV}$$

The total cross section at $E_\nu = 30\,\text{MeV}$ is approximately $3 \times 10^{-39}\,\text{cm}^2$ — tiny by nuclear physics standards, but enormous by neutrino physics standards. The inverse beta decay cross section at the same energy is $\sim 10^{-41}\,\text{cm}^2$, roughly 300 times smaller.

The COHERENT Experiment

The Setting: Oak Ridge

The COHERENT (Coherent Elastic Neutrino Nucleus Interaction Experiment) collaboration chose the Spallation Neutron Source (SNS) at Oak Ridge National Laboratory in Tennessee. The SNS is primarily a materials science facility: it produces the most intense pulsed neutron beam in the world by slamming 1-GeV protons into a liquid mercury target. But the proton-mercury collisions also produce copious pions, which stop in the mercury target and decay:

$$\pi^+ \to \mu^+ + \nu_\mu \quad (\tau_\pi = 26\,\text{ns})$$ $$\mu^+ \to e^+ + \nu_e + \bar{\nu}_\mu \quad (\tau_\mu = 2.2\,\mu\text{s})$$

The result is an intense flux of neutrinos in three flavors: a prompt monoenergetic $\nu_\mu$ (29.8 MeV) from pion decay, and delayed $\nu_e$ and $\bar{\nu}_\mu$ with continuous spectra up to 52.8 MeV from muon decay.

The COHERENT detector was positioned 19.3 meters from the SNS target in a basement corridor called "Neutrino Alley," shielded from the spallation neutrons by several meters of steel, concrete, and gravel.

The Detector: 14.6 kg of CsI[Na]

The first COHERENT detector was a 14.6-kg sodium-doped cesium iodide (CsI[Na]) scintillation crystal — a technology borrowed from medical imaging and particle physics calorimetry. When a neutrino scatters coherently off a cesium or iodine nucleus, the recoiling nucleus deposits its kinetic energy in the crystal, producing a tiny flash of scintillation light.

The key challenge was the detection threshold. The nuclear recoils of interest produce only $\sim 5$–$30$ photoelectrons in the CsI[Na] crystal — right at the edge of detectability. The scintillation efficiency for nuclear recoils (the "quenching factor") is much lower than for electron recoils: only about 8.7% of the energy deposited by a nuclear recoil produces scintillation light, compared to $\sim 100\%$ for an electron of the same energy.

The Timing Trick

The feature that made the measurement possible was the pulsed time structure of the SNS. The proton beam arrives in $\sim 700\,\text{ns}$ bursts at 60 Hz, giving a duty factor of only $\sim 4 \times 10^{-5}$. This means that in a random trigger, the probability of coinciding with a beam pulse is only $10^{-4}$ — a powerful handle for rejecting steady-state backgrounds.

The analysis exploited two distinct timing windows:

  1. Prompt window ($0$–$1\,\mu\text{s}$ after beam pulse): dominated by the monoenergetic $\nu_\mu$ from pion decay.
  2. Delayed window ($1$–$6\,\mu\text{s}$ after beam pulse): dominated by $\nu_e$ and $\bar{\nu}_\mu$ from muon decay.

By comparing the event rate in these windows to the rate outside any beam window (the "anti-coincidence" sample), the CE$\nu$NS signal could be extracted with high confidence.

The Result

In August 2017, the COHERENT collaboration published their result in Science:

$$\text{Observed: } 134 \pm 22\,\text{events}$$ $$\text{Standard Model prediction: } 173 \pm 48\,\text{events}$$

The measurement rejected the background-only hypothesis at $6.7\sigma$, confirming the existence of CE$\nu$NS. The result was consistent with the Standard Model prediction within uncertainties (the ratio was $0.77 \pm 0.16$), though the central value was slightly below the prediction.

The paper received enormous attention: it was one of the top-cited physics papers of 2017 and opened an entirely new experimental channel for neutrino physics.

Subsequent Measurements and Improvements

Argon Measurement (CENNS-10)

In 2020, COHERENT published the first CE$\nu$NS measurement using a 24-kg liquid argon (LAr) detector, observing $159 \pm 43$ events (SM prediction: $\sim 190$). The argon measurement was important because it confirmed the $N^2$ scaling: argon ($N = 22$) has a much smaller weak charge than cesium ($N = 78$), so the cross section per atom is $\sim (22/78)^2 \approx 0.08$ times smaller. The observed ratio between CsI and Ar event rates was consistent with this prediction.

Germanium

COHERENT has also deployed a germanium detector array at the SNS. Germanium offers superb energy resolution and a well-characterized response function. Combined with the CsI and Ar measurements, the three targets provide a powerful test of the $N^2$ scaling across the nuclear chart.

Reactor Measurements

Several groups are pursuing CE$\nu$NS measurements using reactor antineutrinos ($E_\nu \sim 1$–$8\,\text{MeV}$). At these lower energies, the recoil energies are sub-keV, requiring detectors with extremely low thresholds (germanium bolometers, CCD sensors). Reactor measurements would test CE$\nu$NS in a different energy regime and provide complementary constraints on NSI.

Implications and Applications

Testing the Standard Model

The CE$\nu$NS cross section is a clean prediction of the Standard Model, depending only on $G_F$, $\sin^2\theta_W$, and the nuclear form factor. Deviations from the SM prediction could indicate:

  • Nonstandard neutrino interactions (NSI): New mediators (light $Z'$ bosons, scalar particles) that modify the effective neutrino-quark couplings.
  • Neutrino magnetic moment: An anomalously large magnetic moment would produce an additional electromagnetic scattering component that grows at low recoil energies.
  • Sterile neutrinos: If neutrinos oscillate into sterile states on the $\sim 20$-meter baseline from source to detector, the CE$\nu$NS rate would be reduced.

Current COHERENT data constrain NSI couplings at the $\sim 10\%$ level, competitive with constraints from global oscillation data for certain coupling combinations.

Nuclear Structure

Because $Q_W \approx -N$, the CE$\nu$NS cross section is sensitive to the neutron density distribution through the nuclear form factor $F(q^2)$. Precise CE$\nu$NS measurements on heavy nuclei could provide a neutrino-based determination of the neutron distribution — completely independent of the electron scattering (PREX/CREX) measurements discussed in Section 32.2.2.

This connection between CE$\nu$NS and nuclear structure is a vivid illustration of how frontier neutrino physics depends on basic nuclear science.

Dark Matter Detection

Perhaps the most consequential practical impact of CE$\nu$NS is on dark matter searches. The process produces nuclear recoils that are experimentally identical to WIMP-nucleus scattering. As dark matter detectors improve in sensitivity, CE$\nu$NS from solar and atmospheric neutrinos will become an irreducible background — the "neutrino floor."

For xenon-based detectors, the neutrino floor from ${}^8\text{B}$ solar neutrinos lies at WIMP-nucleon cross sections of $\sim 10^{-48}\,\text{cm}^2$ for WIMP masses around 6 GeV. The next generation of experiments (XLZD) will approach this floor within the next decade.

The COHERENT measurement was the first direct demonstration that the neutrino floor is real — that neutrinos actually do produce coherent nuclear recoils at the predicted rate.

Supernova Neutrino Detection

CE$\nu$NS provides a flavor-blind channel for detecting supernova neutrinos — all six neutrino species ($\nu_e$, $\bar{\nu}_e$, $\nu_\mu$, $\bar{\nu}_\mu$, $\nu_\tau$, $\bar{\nu}_\tau$) scatter coherently off nuclei. The burst of $\sim 10^{58}$ neutrinos from a Galactic supernova at 10 kpc would produce $\sim 1000$ CE$\nu$NS events per tonne of xenon. Large dark matter detectors like LZ and XENONnT are already configured to trigger on supernova bursts, providing a new supernova observatory that complements existing neutrino detectors (Super-Kamiokande, IceCube).

Reflections: From Prediction to Discovery

The CE$\nu$NS story encapsulates several important themes in the relationship between nuclear physics and particle physics:

  1. Theory can outrun experiment by decades. Freedman's 1973 prediction was made one year after the discovery of neutral currents at CERN — it was a straightforward consequence of the Standard Model. But detection required detector technology that did not exist for 43 years.

  2. Opportunistic use of existing infrastructure. The SNS was built for materials science, not neutrino physics. The COHERENT collaboration recognized that it also provided an ideal neutrino source and a location (Neutrino Alley) with manageable backgrounds. The total cost of the CsI detector was a tiny fraction of the SNS's construction cost.

  3. Nuclear physics is essential for interpretation. The CE$\nu$NS cross section, the form factor, the quenching factor — all depend on nuclear physics. Without accurate nuclear structure input, the measurement cannot be interpreted in terms of fundamental physics.

  4. One measurement opens many doors. From a single cross section measurement, CE$\nu$NS connects to neutrino properties, nuclear structure, dark matter backgrounds, supernova physics, and reactor monitoring. The physics return per dollar invested is extraordinary.

Discussion Questions:

  1. The COHERENT CsI result was $77\% \pm 16\%$ of the Standard Model prediction. If this deficit is real, what are the possible explanations? How would you design a follow-up experiment to distinguish between them?

  2. Compare the advantages and disadvantages of CsI, liquid argon, and germanium as CE$\nu$NS detectors. Consider: coherent enhancement ($N^2$), recoil energy, energy threshold, energy resolution, scalability, and background rejection.

  3. Reactor antineutrinos have much lower energies ($\sim 1$–$8\,\text{MeV}$) than SNS neutrinos ($\sim 15$–$53\,\text{MeV}$). Calculate the maximum nuclear recoil energy for a 5-MeV reactor antineutrino on germanium ($A = 73$). Why does this make reactor CE$\nu$NS experiments much more challenging?