MINER Experiment: Neutrino Precision Studies

• The MINER Experiment comprises two distinct neutrino programs: MINERνA for few-GeV neutrino–nucleus scattering and reactor MINER for low-energy CEνNS and ALP searches.
• MINERνA employs a fine-grained scintillator detector with multiple nuclear targets to provide precise cross-section measurements essential for reducing uncertainties in oscillation experiments.
• Reactor MINER utilizes ultra-low-threshold cryogenic detectors and aggressive shielding near research reactors to explore coherent elastic neutrino–nucleus scattering and potential dark-sector signatures.

MINER is a name used in the arXiv literature for two distinct neutrino experimental programs. MINER$\nu$A (Main INjEctoR $\nu$-A) is an on-axis, few-GeV neutrino–nucleus scattering experiment in the Fermilab NuMI beamline, built around a fine-grained scintillator detector and multiple nuclear targets for precision cross-section measurements and nuclear-effect studies. The Mitchell Institute Neutrino Experiment at Reactor (MINER) is a short-baseline reactor experiment at Texas A&M University, later extended in projection to HFIR, using cryogenic detectors to pursue coherent elastic neutrino–nucleus scattering (CE$\nu$NS) and related rare-event searches such as axion-like particles (ALPs). Although the two programs operate in very different energy regimes and with different detector technologies, both are motivated by precision weak-interaction measurements and by the control of systematics relevant to neutrino physics more broadly (Perdue, 2011, Collaboration et al., 2016, Mondal et al., 14 Oct 2025).

1. Nomenclature and experimental scope

In usage within neutrino physics, “MINER” denotes two separate experimental efforts rather than a single apparatus. MINER$\nu$A is a few-GeV neutrino cross-section experiment in the NuMI beam at Fermilab, whereas the reactor MINER program is a low-threshold cryogenic experiment near research reactors, oriented toward CE$\nu$NS and beyond-the-Standard-Model searches (Perdue, 2011, Collaboration et al., 2016, Mirzakhani et al., 29 Apr 2025).

Program	Expansion	Primary regime
MINER$\nu$A	Main INjEctoR $\nu$-A	Few-GeV neutrino–nucleus scattering
MINER	Mitchell Institute Neutrino Experiment at Reactor	MeV reactor antineutrinos and sub-keV to keV recoils

The distinction is substantive. MINER$\nu$A was optimized for exclusive and inclusive $\nu$ and $\bar\nu$ interactions on CH, C, Fe, and Pb, with emphasis on charged-current quasielastic (CCQE), resonance, coherent, and deep inelastic channels, together with nuclear-target comparisons. Reactor MINER was designed around ultra-low-threshold cryogenic detectors close to a 1 MW TRIGA reactor core, initially with germanium and silicon concepts and later with sapphire calorimeters, to access CE$\nu$0NS and low-mass ALP signatures (Perdue, 2011, Collaboration et al., 2016, Mirzakhani et al., 29 Apr 2025).

2. MINER$\nu$1A in the NuMI neutrino program

MINER$\nu$2A is a dedicated neutrino–nucleus scattering experiment operating on-axis in the NuMI beamline at Fermilab. It is optimized for the few-GeV range, where quasielastic scattering, resonance production, and shallow/deep inelastic scattering all contribute significantly to the total cross section. Its central motivation is to measure neutrino cross sections and final-state kinematics on several nuclei with sufficient precision to reduce interaction-model and nuclear-physics uncertainties in oscillation experiments, while also providing a “pure weak probe” of nuclear structure complementary to charged-lepton scattering (Perdue, 2011).

The connection to oscillation physics is direct. For a typical two-flavor $\nu$3 disappearance analysis, the survival probability is

$\nu$4

Experiments do not observe $\nu$5 directly; they observe visible energy from final-state particles. That visible energy depends on the incident neutrino flux, the interaction cross section, detector response, and intranuclear final-state interactions such as re-scattering, absorption, charge exchange, and pion production or absorption. MINER$\nu$6A was therefore designed to provide precise differential cross-section data and nuclear-target comparisons for experiments such as MINOS, NOvA, T2K, and, by extension, DUNE (Perdue, 2011).

The beamline uses 120 GeV protons from the Main Injector striking a graphite target, with horn focusing configured in Low Energy mode. Two principal beam modes are used: Forward Horn Current, which focuses $\nu$7 and yields predominantly $\nu$8, and Reverse Horn Current, which focuses $\nu$9 and yields predominantly $\nu$0. A notable NuMI capability is variable focusing through target-position changes and horn-current variation; MINER$\nu$1A exploited these configurations as a data-driven reweighting strategy to reduce flux systematics by constraining hadron production in $\nu$2 space (Perdue, 2011).

Using GENIE 2.6.2, the collaboration quoted the following charged-current inclusive event yields for $\nu$3 POT in Low Energy mode; these are raw generator predictions and not acceptance corrected (Perdue, 2011).

Material	$\nu$4 POT (LE $\nu$5)	$\nu$6 POT (LE $\nu$7)
Carbon target	10,800	3,400
Iron target	64,500	19,200
Lead target	68,400	10,800
Scintillator (CH) tracker	409,000	134,000

These projected samples already indicate the role of MINER$\nu$8A as a high-statistics benchmark for generator tuning and for the study of nuclear dependence in weak interactions.

3. Detector architecture, calibration, and data acquisition in MINER$\nu$9A

The MINER$\nu$0A detector is a horizontal stack of modular planes, each module weighing about 2 tons. Each module contains an Inner Detector built from triangular extruded plastic scintillator strips and an Outer Detector consisting of a steel frame instrumented with scintillator bars. Upstream lie dedicated nuclear targets of carbon, iron, and lead interleaved with scintillator planes; the central tracker is composed almost entirely of scintillator planes; downstream are electromagnetic and hadronic calorimeter sections; and the MINOS Near Detector sits downstream to analyze exiting muons (Perdue, 2011).

The active tracker defines the experiment’s fine-grained core. For a 90 cm radius cut, the fiducial CH mass is 6.43 tons, although stricter fiducial cuts make the effective mass closer to 5 tons in many analyses. Each plane contains 127 triangular scintillator strips. Charge sharing between adjacent strips provides sub-strip spatial resolution, and through-going muons exhibit tracking residuals just over 3 mm. The planes are arranged in repeating stereo orientations: X planes have vertical strips, while U and V planes are rotated by $\nu$1, enabling full 3D track and vertex reconstruction (Perdue, 2011).

The dedicated target masses for a 90 cm radius cut are 0.17 tons of carbon, 0.97 tons of iron, and 0.98 tons of lead. Their placement in a common detector and common beam permits relative A-dependent measurements with reduced systematics. Muons exiting MINER$\nu$2A are measured in the magnetized MINOS Near Detector, which provides sign selection and momentum reconstruction by curvature, so the MINER$\nu$3A+MINOS system combines fine-grained vertexing and calorimetry with downstream muon spectrometry (Perdue, 2011).

The hadronic calibration strategy included a dedicated Test Beam Experiment using quarter-sized MINER$\nu$4A planes with the same scintillator geometry, PMTs, electronics, and DAQ. Two configurations were used, “20 Tracker – 20 ECAL” and “20 ECAL – 20 HCAL,” and particle identification in the test beam showed good separation of pions, protons, and muons. This underpinned the modeling of hadronic response in GEANT4 and the conversion of visible energy into reconstructed hadronic energy (Perdue, 2011).

The data acquisition system described for MINER$\nu$5A reflects the same emphasis on full event characterization. The detector sits about 105 m underground in the MINOS Near Detector Hall. Readout is based on Hamamatsu R7600 64-channel multi-anode PMTs and Front-End Boards hosting TriP-t ASICs. A 53.1 MHz master clock is multiplied by two on the FEBs, giving a basic tick of about 9.4 ns; discriminator fine timing reaches about 2.4 ns through quarter-tick subdivision. By the summer of 2010 and later, the system comprised 15 CROCs, 509 FEBs, and 32,576 channels. In standard NuMI running, the DAQ handled about 1 MB per spill, with total live time of 96.4% and live time exceeding 99% during normal operations (Perdue et al., 2012).

4. Event reconstruction, CCQE analysis, and nuclear-model issues in MINER$\nu$6A

MINER$\nu$7A event reconstruction proceeds through hit formation and clustering, 3D track finding in the X/U/V views, vertex reconstruction, muon identification through matching into MINOS, and calorimetric reconstruction of hadronic energy. For CCQE kinematics, the experiment uses the standard bound-nucleon-at-rest approximation with a binding-energy correction. For neutrino QE, the reconstructed neutrino energy is written as (Perdue, 2011)

$\nu$8

with $\nu$9 interchanged for antineutrino QE, and

$\nu$0

A preliminary antineutrino CCQE analysis used $\nu$1 POT in Reverse Horn Current Low Energy mode with 2.86 tons of fiducial CH in the tracker during construction. The selection required a well-reconstructed $\nu$2 track matched to MINOS with identified positive charge and imposed a low recoil-energy requirement. The recoil energy was defined as all energy outside a 5 cm radial cylinder around the muon track within a 100 ns time window around the interaction; a 2D cut in recoil energy versus $\nu$3 was then used to purify a QE-like sample (Perdue, 2011).

Early data–simulation comparisons already exposed the importance of flux and nuclear modeling. For QE-selected candidates, the paper noted that “the event deficit is flat in $\nu$4, but not neutrino energy,” indicating that the $\nu$5 shape was similar to the model while the energy dependence was not. A plausible implication is that flux modeling, rather than only form-factor modeling, required further tuning, which is consistent with the horn-current and target-position flux program (Perdue, 2011).

The later GiBUU analysis of MINER$\nu$6A sharpened these issues substantially. For the MINER$\nu$7A CH target and flux in the range $\nu$8 GeV, true CCQE contributes only about one third of the total charged-current cross section, while deep inelastic scattering becomes dominant for $\nu$9. In the 0-pion sample that MINER$\nu$0A uses to define “QE-like,” the total is still about 1.5 times the true QE contribution because $\nu$1 excitation with pionless decay, higher resonances, 2p–2h processes, and DIS followed by final-state interactions all feed the same topology (Mosel et al., 2014).

The reconstruction consequences are nontrivial. In the GiBUU study, the reconstructed event distribution for 0-pion samples acquired substantial strength below 1.5 GeV even though the true incoming flux was zero there, and the reconstructed energy peak was shifted downward by about 400 MeV relative to the true energy. For the flux-averaged $\nu$2 distribution, the reconstructed spectrum exceeded the true one up to $\nu$3; near the peak at $\nu$4, it was about 25% larger. The reconstructed slope between $\nu$5 and $\nu$6 was steeper than the true slope, corresponding to an apparent smaller axial mass $\nu$7 if interpreted in a pure QE framework (Mosel et al., 2014).

This analysis also argued that a more robust QE selection would be a muon plus exactly one proton, any number of neutrons, and no other hadrons; in GiBUU this “1p only + 0$\nu$8” channel is 90% dominated by true QE, albeit with a loss of about one third of the QE cross section. More generally, the MINER$\nu$9A case established that channel definitions by final-state topology must be interpreted through realistic transport calculations including Fermi motion, binding, Pauli blocking, multinucleon effects, and full hadronic final-state interactions (Mosel et al., 2014).

5. Reactor MINER: site, shielding concept, and background program

The Mitchell Institute Neutrino Experiment at Reactor was conceived as a CE$\nu$0NS experiment near the Nuclear Science Center at Texas A&M University, using the 1 MW TRIGA reactor as an intense, very short-baseline $\nu$1 source. The proposed deployment placed detectors in the Thermal Column cavity about 2–2.3 m from the core face. In the Standard Model, the anticipated CE$\nu$2NS signal in the 10–1000 eV$\nu$3 window was estimated at $\nu$4–20 events/kg/day, which immediately set the design background target (Collaboration et al., 2016).

The initial detector concept used low-threshold cryogenic germanium and silicon detectors exploiting Neganov–Luke phonon amplification, with a target recoil threshold near 10 eV$\nu$5. The simulated payload comprised four Ge and four Si detectors, each modeled as a cylinder of 100 mm diameter and 33 mm thickness. The central experimental problem was not signal generation but background reduction, because acceptable backgrounds in the CE$\nu$6NS window were of order $\nu$7 events/kg/day, only moderately above the predicted signal (Collaboration et al., 2016).

The background studies combined in situ measurements and transport simulations based on MCNP and GEANT4. At the core, the gamma flux was estimated as $\nu$8; fast neutrons above 100 keV were estimated at $\nu$9; thermal neutrons below 0.625 eV were estimated at $\nu$0. Gamma measurements in the cavity with an HPGe detector showed that the event rate decreased by a factor of about 3.5 for each 0.5 m increase in distance between core and detector in that configuration. Copper-foil activation measured a thermal neutron flux of $\nu$1 at the foil position inside the cavity, while the muon-rate measurements indicated roughly a factor-of-two reduction inside the cavity relative to open sky (Collaboration et al., 2016).

The preliminary shielding design placed 1.38 m of 5% borated high-density polyethylene and 30.5 cm of lead between the reactor-side graphite block and the detector payload, with additional lead and polyethylene layers, copper around the detectors, and plastic scintillator for cosmic vetoing. In full GEANT4 simulations of this geometry, the estimated background rate in the 10–1000 eV$\nu$2 window was compatible with the target of $\nu$3 events/kg/day, while the total rate outside that window was about 30 Hz summed over all detectors. This established experimental feasibility for a first CE$\nu$4NS detection at the NSC site, albeit with a challenging signal-to-background ratio (Collaboration et al., 2016).

6. Sapphire phase, CE$\nu$5NS search, ALP limits, and the HFIR program

The later reactor MINER program used cryogenic sapphire $\nu$6 calorimeters with Transition Edge Sensors to pursue both CE$\nu$7NS and ALP searches. In the ALP study, three cylindrical sapphire detectors were arranged in a vertical tower inside a copper housing: two outer detectors of 181 g each and one middle 4 mm detector of 72 g, for a total target mass of 435 g. The detector tower was cooled in a BlueFors dilution refrigerator to $\nu$8 mK, and the 4 mm detector achieved a baseline energy resolution of $\nu$9 eV, corresponding to an effective threshold of about 105 eV when combining channels (Mirzakhani et al., 29 Apr 2025).

The ALP analysis was carried out with 59.5 hours of reactor-on data and 163.8 hours of reactor-off data, using energy deposits above 3 keV because deposits below 3 keV remained blinded for the CE$\nu$0NS program. The theoretical framework used the interaction terms

$\nu$1

with production by Primakoff and Compton-like processes in the reactor and detection through scattering or decay in the detector. In this first result, MINER excluded ALPs with couplings as small as $\nu$2 and $\nu$3, demonstrating that the same low-threshold infrastructure developed for CE$\nu$4NS could probe low-mass dark-sector signatures (Mirzakhani et al., 29 Apr 2025).

The dedicated CE$\nu$5NS search at TRIGA used a primary sapphire detector mass of 72 g with baseline energy resolution of about 40 eV. Using exposures of 158 g-days reactor-on and 381 g-days reactor-off, the analysis performed statistical background subtraction in the 0.25–3 keV region of interest. The Standard Model CE$\nu$6NS differential cross section was written as

$\nu$7

and the expected signal rate in the TRIGA configuration was only $\nu$8, far below the measured low-energy background. After subtracting reactor-induced background with the aid of GEANT4, the best-fit ratio of observed to Standard Model CE$\nu$9NS rate was

$\bar\nu$0

with a significance of

$\bar\nu$1

These values indicate that the TRIGA CE$\bar\nu$2NS data were dominated by low-energy background rather than by a statistically resolvable CE$\bar\nu$3NS excess (Mondal et al., 14 Oct 2025).

The planned response is relocation to the 85 MW$\bar\nu$4 High Flux Isotope Reactor at Oak Ridge National Laboratory, at a baseline of about 5 m. The HFIR projections retain the MINER strategy of low-threshold cryogenic detection but pair it with higher antineutrino flux, improved shielding, and increased detector mass. Under the assumptions used in the projection, the upgraded setup is expected to achieve a 3$\bar\nu$5 CE$\bar\nu$6NS detection within about 30 kg·days of exposure; the same study notes that a 3$\bar\nu$7 result with 14 kg·days would require backgrounds below about $\bar\nu$8 (Mondal et al., 14 Oct 2025).

Across both MINER$\bar\nu$9A and reactor MINER, the recurring experimental theme is the same: precision weak-interaction measurements are limited less by counting statistics than by the control of flux, detector response, and nuclear or material-dependent final-state effects. MINER$\nu$00A addressed those issues in the few-GeV regime through fine-grained tracking, multiple nuclear targets, and external muon spectrometry; reactor MINER addresses them in the MeV regime through cryogenic calorimetry, aggressive shielding, and reactor-on/reactor-off subtraction. The shared acronym thus connects two experimentally distinct, but methodologically related, neutrino programs (Perdue, 2011, Mondal et al., 14 Oct 2025).