SHADES Hybrid Neutron Detector

• SHADES hybrid neutron detector is a purpose-built underground system that combines EJ-309 liquid scintillators and 3He proportional counters to directly measure the 22Ne(α,n)25Mg cross section in the stellar Gamow window.
• The detector employs a modular design integrating active moderation, pulse-shape discrimination (enhanced by machine learning), and delayed neutron capture timing to achieve sensitivity improvements by at least two orders of magnitude.
• Prototype tests validated the design by demonstrating effective energy calibration, neutron–gamma discrimination, and a clear 3–7 μs coincidence signature, establishing its potential for precise astrophysical measurements.

Searching arXiv for the SHADES detector papers and closely related neutron-detector references. The SHADES hybrid neutron detector is a purpose-built, deep-underground neutron-detection system developed to enable direct measurements of the $^{22}\text{Ne}(\alpha,n)^{25}\text{Mg}$ cross section in the stellar Gamow window $450\,\text{keV}<E_\text{cm}<750\,\text{keV}$, where previous direct experiments provided only upper limits and the reaction-rate uncertainty reaches about $500\%$ at $T\approx 0.2$–$0.3$ GK. Conceived within the ERC project SHADES (“Scintillator–$^{3}$He Array for Deep-underground Experiments on the S-process”), the detector combines EJ-309 liquid scintillators, $^{3}$He proportional counters, borated polyethylene moderation and shielding, and operation in the ultra-low-background environment of LNGS in order to improve sensitivity by at least two orders of magnitude over the state of the art (Rapagnani et al., 6 Oct 2025). A prototype module has been characterized with low-energy neutrons and constitutes the validated building block of the larger array (Chillery et al., 6 Nov 2025).

1. Astrophysical role and measurement objective

SHADES was designed for a specific nuclear-astrophysics problem: the direct determination of the $^{22}\text{Ne}(\alpha,n)^{25}\text{Mg}$ reaction rate in the energy region relevant to the s-process. In stars, the slow neutron-capture process is driven predominantly by two $\alpha$-induced neutron-source reactions, $^{13}\text{C}(\alpha,n)^{16}\text{O}$ and $450\,\text{keV}<E_\text{cm}<750\,\text{keV}$0. In massive stars with $450\,\text{keV}<E_\text{cm}<750\,\text{keV}$1, $450\,\text{keV}<E_\text{cm}<750\,\text{keV}$2 is the main neutron source in the mass region $450\,\text{keV}<E_\text{cm}<750\,\text{keV}$3–90, while in AGB stars it contributes, together with $450\,\text{keV}<E_\text{cm}<750\,\text{keV}$4, to the synthesis of s-process nuclei up to $450\,\text{keV}<E_\text{cm}<750\,\text{keV}$5–209 (Rapagnani et al., 6 Oct 2025).

The experimental target is defined by helium-burning temperatures $450\,\text{keV}<E_\text{cm}<750\,\text{keV}$6–$450\,\text{keV}<E_\text{cm}<750\,\text{keV}$7 GK, which map onto a center-of-mass Gamow window of $450\,\text{keV}<E_\text{cm}<750\,\text{keV}$8. In exactly this interval, only upper limits from direct measurements were available, indirect constraints were strongly model dependent, and the relevant cross section was described as extremely small. The consequence is a regime in which reaction yield is tiny while environmental, cosmic-ray-induced, and beam-induced backgrounds dominate. Existing detectors either lacked sufficient efficiency or could not discriminate neutrons from background with the required fidelity in this region (Rapagnani et al., 6 Oct 2025).

Within that context, SHADES is not a general-purpose neutron detector. It is a detector optimized for a narrow but consequential measurement objective: direct low-energy $450\,\text{keV}<E_\text{cm}<750\,\text{keV}$9 data in the stellar window, plus reduction of uncertainties in higher-energy resonance parameters. This specialized purpose explains the emphasis on deep-underground operation, hybrid moderation-and-capture logic, and differential sensitivity to neutron energies relevant to the reaction of interest.

2. Array architecture and deployment context

The full SHADES setup centers on a high-purity extended $500\%$0 gas target surrounded by a nearly $500\%$1 hybrid detector array and embedded in borated polyethylene that acts both as mechanical support and neutron shield. The gas target is described as approximately 30 keV thick in beam-energy loss, operated at about 5 mbar, and housed in a compact differentially pumped chamber. Around it, the detector arrangement is designed to maximize solid-angle coverage, provide sufficient moderator thickness for fast neutrons emitted over a broad angular distribution, and suppress environmental neutron backgrounds (Rapagnani et al., 6 Oct 2025).

Configuration	Core components	Function
Full SHADES array	12 EJ-309 liquid scintillators, 18 $500\%$2He counters, high-purity $500\%$3Ne gas target, borated polyethylene	Direct $500\%$4 measurements at LNGS
Prototype module	One 12.7 cm $500\%$5 12.7 cm EJ-309 scintillator and six $500\%$6He counters	Characterization of quenching, PSD, and neutron-capture coincidences

The project literature places SHADES within the deep-underground accelerator infrastructure of LNGS. One overview associates the measurement campaign with the new LUNA-MV accelerator, a 3.5 MV Singletron with an ECR source delivering $500\%$7 beams up to 3.5 MeV and currents up to about 500 $500\%$8A, with high long-term energy and current stability. The prototype report identifies the characterized module as the basis of the array being prepared for an upcoming underground study at the Bellotti Ion Beam facility (Rapagnani et al., 6 Oct 2025, Chillery et al., 6 Nov 2025).

The prototype geometry clarifies the modular logic of the larger detector. It consists of one cylindrical EJ-309 liquid scintillator, 12.7 cm in diameter and 12.7 cm in length, read by a 10-stage ETL9390 PMT at $500\%$9 V, and surrounded by six GE Reuter Stokes RS-P4-0810-250 $T\approx 0.2$0He proportional counters filled to 10 bar and biased at $T\approx 0.2$1 V. The six counters are arranged in a hexagonal pattern around the scintillator with a center-to-center distance of 9.76 cm. In the FRANZ characterization, the scintillator front face was 51.9 cm from the $T\approx 0.2$2Li target and three lead bricks provided a total of 7.5 cm shielding between target and scintillator (Chillery et al., 6 Nov 2025).

This modularity is central to the detector concept. A single EJ-309-plus-$T\approx 0.2$3He cell can be characterized in detail, while the full SHADES array scales the same moderated-capture logic to a twelve-scintillator, eighteen-counter system around the astrophysical target.

3. Hybrid detection principle

SHADES is “hybrid” in the strict instrumental sense that it merges two complementary neutron-detection media into a single integrated system. The EJ-309 liquid scintillators act as active moderators and as fast detectors with neutron–gamma discrimination through pulse-shape discrimination. The $T\approx 0.2$4He proportional counters provide high-probability thermal-neutron capture and a distinct ionization signal through the reaction

$T\approx 0.2$5

The supporting borated polyethylene adds a second capture channel for external thermalized neutrons through

$T\approx 0.2$6

thereby suppressing environmental backgrounds before they reach the $T\approx 0.2$7He tubes (Rapagnani et al., 6 Oct 2025).

The detection chain is sequential. A neutron produced in the $T\approx 0.2$8Ne gas target leaves the interaction region as a fast neutron, enters the EJ-309 and surrounding moderator, and undergoes multiple elastic scattering interactions that generate recoil protons visible in the scintillator. After losing energy and thermalizing, it diffuses until capture in a $T\approx 0.2$9He tube. A capture signal, optionally in time correlation with preceding scintillation pulses, identifies a neutron that originated in the moderated region surrounding the target (Rapagnani et al., 6 Oct 2025).

This architecture gives SHADES several concurrent functionalities. It provides high intrinsic neutron efficiency by combining large solid-angle coverage with multiple $0.3$0He counters. It provides neutron–gamma discrimination through EJ-309 PSD. It adds a topological and temporal signature—moderation in scintillator followed by delayed capture in $0.3$1He—that is unavailable to a pure scintillator spectrometer or a pure thermal-counter array. The project overview explicitly states that “energy sensitivity” in SHADES is not neutron spectroscopy in the narrow sense, but differential sensitivity to neutrons produced by $0.3$2 in the Gamow window versus neutrons from other sources. A plausible implication is that the detector is optimized to separate reaction-like and background-like neutron populations rather than to reconstruct arbitrary neutron spectra with high resolution (Rapagnani et al., 6 Oct 2025).

A common misconception is therefore that SHADES is simply a $0.3$3He long counter surrounded by moderator. It is not. Its defining feature is the combined use of active EJ-309 moderation, PSD, segmented geometry, and delayed thermal-neutron capture. Conversely, it is not merely a scintillator neutron spectrometer, because the event definition relies critically on moderation-and-capture logic in $0.3$4He.

4. Prototype characterization and measured performance

The prototype module was tested at FRANZ with the $0.3$5 reaction, producing neutrons across 50–720 keV, a range chosen to overlap the neutron energies relevant to future s-process measurements. For proton energies from 1900 to 2450 keV, the mean neutron energy at the EJ-309 position and $0.3$6 was reported as 51, 134, 202, 262, 321, 380, 440, 497, 552, 605, 659, and 713 keV; the 7.5 cm lead shielding slightly reduced these means (Chillery et al., 6 Nov 2025).

The EJ-309 energy scale was calibrated in electron-equivalent energy using Compton edges from $0.3$7Cs and $0.3$8Co. Its energy resolution was parameterized as

$0.3$9

with $^{3}$0, $^{3}$1, and $^{3}$2. The neutron quenching relation $^{3}$3 was then obtained by correlating the neutron locus in EJ-309 with the mean neutron energies from the $^{3}$4Li(p,n) kinematics. Rational, quadratic, and exponential parameterizations were fitted; the quadratic form was adopted for subsequent neutron-energy calibration, while the quadratic and exponential descriptions were identified as the best over the low-energy range (Chillery et al., 6 Nov 2025).

The pulse-shape discrimination analysis used the standard charge-integration definition

$^{3}$5

with an 80 ns short gate and a 480 ns long gate. For the highest-energy neutron run, the EJ-309 neutron locus appeared at PSD $^{3}$6, whereas the broader $^{3}$7-ray population occupied lower PSD. At lower neutron energies, the neutron and $^{3}$8 loci began to overlap below about 60 keVee, and traditional PSD became insufficient for clean separation. To extend discrimination, the collaboration implemented a Gaussian-Mixture Variational Autoencoder with an auxiliary classifier and loss

$^{3}$9

using $^{3}$0, $^{3}$1, and $^{3}$2. This ML-based approach pushed effective neutron/$^{3}$3 discrimination down to a lowest detectable neutron energy of 163 keV (Chillery et al., 6 Nov 2025).

The prototype also demonstrated the moderated-capture timing signature that underlies the full SHADES concept. The coincidence variable was defined as

$^{3}$4

and, for the 713 keV neutron run, a clear peak appeared between 3 and 7 $^{3}$5s in EJ-309 neutron-like events followed by $^{3}$6He captures. Selecting that 3–7 $^{3}$7s window removed 98.9% of all EJ-309 $^{3}$8-like events and 98.4% of random EJ-309 neutron-like events. On the counter side, gating on EJ-309 coincidences reduced the low-energy $^{3}$9 region of the summed $^{22}\text{Ne}(\alpha,n)^{25}\text{Mg}$0He spectrum by a factor 4.5 and reduced the total counter rate by 48%, while leaving the 764 keV neutron-capture peak clearly visible. A GEANT4 v11.3.0 simulation, using FTFP_BERT_HP, G4ThermalNeutrons, and G4StoppingPhysics, reproduced the measured 3–7 $^{22}\text{Ne}(\alpha,n)^{25}\text{Mg}$1s coincidence structure after inclusion of charge-collection and timing smearing terms (Chillery et al., 6 Nov 2025).

These results establish the prototype not only as a proof of hardware functionality but as a quantified demonstration of the detector logic: quenching-based energy sensitivity in EJ-309, neutron–gamma separation, and delayed capture confirmation in $^{22}\text{Ne}(\alpha,n)^{25}\text{Mg}$2He.

5. Background suppression and low-background design

Background reduction is fundamental to SHADES because the intended $^{22}\text{Ne}(\alpha,n)^{25}\text{Mg}$3 signal lies in a regime of extremely low reaction yield. The project overview identifies five principal background classes: cosmic-ray-induced neutrons, environmental neutrons, environmental $^{22}\text{Ne}(\alpha,n)^{25}\text{Mg}$4 rays, beam-induced neutrons from impurities—especially carbon and boron—and activation products (Rapagnani et al., 6 Oct 2025).

The first layer of mitigation is location. LNGS drastically suppresses the cosmic-ray-induced neutron flux, and comparison spectra from $^{22}\text{Ne}(\alpha,n)^{25}\text{Mg}$5He proportional counters show that underground the environmental neutron component is reduced so strongly that the intrinsic radioactivity of aluminum detector materials becomes the dominant background. That observation is significant because it shifts the limiting background from external neutron fields to detector-material selection and cleanliness (Rapagnani et al., 6 Oct 2025).

The second layer is structural shielding. The SHADES array is held by borated polyethylene, which is simultaneously a moderator, a support structure, and an environmental-neutron shield. Fast external neutrons are slowed in hydrogen-rich material and then preferentially captured by boron before they can reach the $^{22}\text{Ne}(\alpha,n)^{25}\text{Mg}$6He counters. Additional polyethylene shielding surrounds the main array, and the future underground configuration described in the prototype work includes 2-inch-thick borated polyethylene with 5% boron around the neutron detectors (Rapagnani et al., 6 Oct 2025, Chillery et al., 6 Nov 2025).

The third layer is event topology. Many environmental $^{22}\text{Ne}(\alpha,n)^{25}\text{Mg}$7 rays interact in EJ-309, but PSD rejects them, and the definition of a reaction neutron as a moderation signal in the scintillator followed by a delayed $^{22}\text{Ne}(\alpha,n)^{25}\text{Mg}$8He capture suppresses accidental and non-neutron backgrounds. The prototype data show that this logic is already effective before underground deployment, which suggests still stronger rejection under LNGS conditions where the long-time neutron background is expected to fall substantially (Chillery et al., 6 Nov 2025).

The fourth layer is beam-line cleanliness. Because carbon and boron have comparatively large $^{22}\text{Ne}(\alpha,n)^{25}\text{Mg}$9 cross sections, SHADES places special emphasis on avoiding those contaminants in the gas-target system. The measures explicitly listed include carbon-free pumps, gaskets, and sealings, a gas purifier, and high-flow differential pumping that maintains ultra-high vacuum $\alpha$0 mbar in the beam line a few centimeters from the $\alpha$1 mbar target region. The target region is kept compact to minimize the amount of material traversed by the beam at elevated pressure (Rapagnani et al., 6 Oct 2025).

Taken together, these design choices define SHADES as a low-background system in the strong experimental sense: its sensitivity gain is expected to come from both higher true-neutron efficiency and a large reduction in false-neutron yield. The project overview states that the envisioned SHADES background rate is about two orders of magnitude lower, in reaction-yield units, than earlier surface or less-optimized underground measurements (Rapagnani et al., 6 Oct 2025).

6. Scientific impact, constraints, and development trajectory

The immediate scientific purpose of SHADES is to convert the unexplored part of the $\alpha$2 Gamow window from a domain of upper limits into one of direct cross-section data. The project overview states that with the planned sensitivity improvement, actual cross section values—not just upper limits—can be measured in $\alpha$3. Those data are then intended for R-matrix analysis and for astrophysical evaluation with NuGrid stellar models across a range of masses and metallicities (Rapagnani et al., 6 Oct 2025).

The detector’s significance therefore extends beyond hardware performance. Direct low-energy cross sections constrain resonance strengths and energies of the low-energy states that dominate the stellar rate, reduce the current $\alpha$4 uncertainty, and propagate into more precise Maxwellian-averaged reaction rates for weak- and main-s-process nucleosynthesis. This suggests that SHADES is best understood as an enabling instrument for a specific astrophysical inference chain: low-background neutron detection, direct cross-section extraction, R-matrix analysis, and stellar-model recalculation.

The same sources also make clear that the detector is not free of unresolved constraints. Even underground, intrinsic radioactivity of detector materials, especially aluminum in $\alpha$5He counters, can be limiting. Residual beam-induced background from contaminants may persist despite carbon- and boron-free design choices. The overall neutron-detection efficiency $\alpha$6 must be controlled at the few-percent level to avoid dominating the cross-section error budget. Long runs, potentially lasting months, require high beam stability and stable target conditions. The approximately 30 keV target thickness introduces an effective energy spread that must be folded into the analysis, and pressure and temperature stability must be monitored continuously (Rapagnani et al., 6 Oct 2025).

The development trajectory presented in the SHADES literature proceeds from completed detector and gas-target design, through characterization on the surface at the University of Naples “Federico II” and underground at LNGS, to installation and commissioning of the gas target, then full underground measurements, followed by R-matrix and astrophysical analysis. The prototype paper adds the experimentally grounded result that the basic cell already provides the combination of EJ-309 quenching calibration, neutron–gamma discrimination, and $\alpha$7He coincidence timing required for the full array (Chillery et al., 6 Nov 2025).

Within the broader landscape of hybrid neutron detectors, SHADES belongs to the class of moderated-capture systems that combine a fast-neutron-sensitive scintillator with a thermal-neutron capture channel, but its optimization target is unusually specific: deep-underground measurements of very small $\alpha$8 cross sections in stellar-energy windows. That specialized optimization—rather than hybridization alone—is what distinguishes the SHADES hybrid neutron detector as an instrument.