Opaque Scintillation Detectors

• Opaque scintillation detectors are radiation sensors that harness deliberately engineered opacity to confine scintillation light locally, preserving the topology of energy deposition.
• They integrate embedded wavelength-shifting fibers with compact photosensors to efficiently extract and transport photons over short distances for sub-millimeter spatial resolution.
• Performance evaluations report up to 90% light confinement within 5 cm, demonstrating scalable designs for applications in neutron detection, neutrino physics, and medical imaging.

Searching arXiv for recent and foundational papers on opaque scintillation detectors, LiquidO, and fiber-readout architectures. Opaque scintillation detectors are radiation detectors in which the scintillating medium is intentionally used in a strongly scattering or otherwise nontransparent optical regime, so that scintillation light is not transported ballistically over macroscopic distances to peripheral photosensors but is instead collected locally by embedded or closely coupled optical structures such as wavelength-shifting fibers. In this class of detector, opacity is not merely tolerated; it is used to preserve the spatial topology of energy deposition, to enable self-segmentation without physical cell walls, or to make otherwise difficult scintillators practically readable. The concept spans several distinct architectures, including ZnS(Ag)/$^{6}$LiF neutron detectors with embedded wavelength-shifting fibers and silicon photomultipliers for local light extraction (Mosset et al., 2013), LiquidO-style highly scattering liquid detectors in which stochastic confinement creates localized “light balls” around interaction sites (Cabrera et al., 2019), and opaque water-based liquid scintillator systems that demonstrate both spectroscopy and topological reconstruction through fiber lattices (Collaboration, 2024).

1. Definition and operating principle

An opaque scintillation detector differs from a conventional transparent scintillation detector in its optical transport regime. In a conventional detector, scintillation photons are expected to propagate through the bulk medium to sensors located mainly at the detector boundary. In an opaque detector, the medium is used so that photons do not retain long-range line-of-sight transport. Instead, the light is confined locally by strong scattering, harvested near the point of creation by embedded wavelength-shifting fibers, or extracted through distributed local collectors when the scintillator itself is poorly suited to bulk optical transport (Cabrera et al., 2019).

Two physically distinct forms of “opacity” appear in the literature. One is opacity by scattering, in which the medium remains weakly absorbing enough that photons survive but undergo a random walk near their creation point. LiquidO exemplifies this regime: the medium is engineered to have a short scattering length and a long absorption length, producing a localized optical response that earlier work described as a “light-ball” (Cabrera et al., 2019). The other is effective opacity by highly scattering detector structure, as in ZnS(Ag)/$^{6}$LiF neutron screens, which are not optically transparent detector blocks and therefore do not support efficient bulk light collection to a remote sensor (Mosset et al., 2013).

The readout strategy follows from that optical regime. Rather than treating the scintillator volume as a transparent light guide, opaque detectors move the optical transport function into local structures. The most common implementation is a lattice or embedded network of wavelength-shifting fibers, which absorb nearby scintillation photons, re-emit at longer wavelength, and transport the shifted light to SiPMs or PMTs (Mosset et al., 2013, Collaboration, 2024). This suggests a general principle: when transparency is unavailable or deliberately suppressed, the detector must be read out through local capture rather than global collection.

The concept is broader than a single detector family. A neutron-sensitive ZnS(Ag)/$^{6}$LiF sandwich with embedded fibers and SiPM-oriented photon counting is a focused example of an opaque scintillation detector optimized for local extraction from a highly scattering scintillator (Mosset et al., 2013). LiquidO and related opaque liquid scintillator systems instead make the entire active volume scattering-dominated so that event topology is retained in the fiber hit pattern (Collaboration et al., 4 Mar 2025, Collaboration et al., 18 Jul 2025).

2. Optical transport regimes and light confinement

The central optical distinction is between scattering-dominated confinement and absorption-dominated loss. LiquidO requires short scattering length and low enough absorption; if absorption dominated, photons would simply die before reaching sensors (Collaboration et al., 18 Jul 2025). This distinction is stated repeatedly in the opaque scintillator literature because detector performance depends on local redistribution of light rather than mere suppression of total light.

A representative LiquidO formulation demonstrates this explicitly. In the 10 litre prototype, NoWaSH-Opaque at $5^\circ$C is described with a scattering length $\sim 0.001$ m = 1 mm, while absorption lengths remain $>0.5$ m / $>10$ m at 370 / 430 nm (Collaboration et al., 4 Mar 2025). In that system, the measured confinement profile shows 50% within 2 cm, 80% within 4 cm, and 90% within 5 cm of the point-like energy deposition, with negligible detection beyond about 10 cm (Collaboration et al., 4 Mar 2025). The closely related 10 litre LiquidO report states that 90% (80%) of the light [is] confined within a radius of 5 cm (4 cm) when the scattering length is on the order of a few millimetres (Navas-Nicolás et al., 14 Mar 2025). These observations are presented as direct evidence that opacity can preserve local topology rather than erase it.

In smaller prototypes, the same mechanism is observed on shorter scales. A 30 mm cubic LiquidO detector filled with a wax-based opaque scintillator having a scattering length of approximately 0.5 mm and instrumented with 64 wavelength-shifting fibres arranged in an $8\times 8$ grid at 3.2 mm pitch shows that a through-going muon produces a much sharper track-like signal pattern in the opaque fill than in a transparent fill (Collaboration et al., 18 Jul 2025). The paper interprets this as “stochastic confinement” on top of “direct confinement” from the dense fibre lattice. A plausible implication is that the useful optical localization scale is set jointly by scattering length and fiber pitch rather than by either parameter alone.

Opaque water-based liquid scintillator systems show a similar transport regime. In a 1 liter cube filled with oWbLS2, the reduced scattering length was optimized to 5.7 mm and the effective absorption length to 0.169 m in a Geant4-based optical fit (Collaboration, 2024). In a later 3D-projection detector, measured transverse charge confinement was tighter than a Geant4 simulation with a 2 cm scattering length, placing the effective scattering length well below 2 cm (Che et al., 11 May 2026). Although these values are not directly interchangeable across media, they collectively support the same detector logic: short scattering length produces localized light collection, provided absorption is not dominant.

The opaque medium itself may be produced in different ways. NoWaSH uses paraffin wax in LAB so that the transparency and viscosity of the material can be tuned by temperature adjustment; it is colorless and transparent around 40C and has a milky wax structure below 20C (Buck et al., 2019). Opaque WbLS uses water-induced micellar structure deliberately as scattering centers (Collaboration, 2024). Microcrystal scintillators obtain opacity from multiple scattering off microscopic grains of inorganic crystals suspended in an organic scintillating carrier (Wagner et al., 2018). These different materials strategies share the same detector objective: localization of scintillation light on a scale shorter than the detector size and commensurate with the readout pitch.

3. Readout architectures and local light extraction

The dominant readout architecture for opaque scintillation detectors is distributed wavelength-shifting fiber readout. In a highly scattering medium, emitted photons quickly lose directional information and are better collected by nearby fibers than by remote boundary sensors (Mosset et al., 2013). The fibers then transport the light to compact photosensors, most often SiPMs.

This architecture is explicit in neutron-sensitive opaque scintillators. The POLDI upgrade detector uses ZnS(Ag)/$^{6}$LiF scintillator screens read out by embedded Kuraray Y11(200)MC fibers of $250~\mu\mathrm{m}$ diameter at 0.6 mm pitch, coupled ultimately to SiPMs (Mosset et al., 2013). The single-channel prototype contains one scintillator/fiber sandwich with two scintillating layers and four WLS fibers, the lower layer being $^{6}$0 thick with four machined grooves and the upper layer $^{6}$1 thick (Mosset et al., 2013). This geometry places the fibers $^{6}$2 from the top surface and $^{6}$3 from the bottom surface, which directly affects collected light yield (Mosset et al., 2013).

In LiquidO-style detectors, the same principle is extended to 2D or 3D fiber lattices. The 30 mm LiquidO “Cube” uses 64 Saint-Gobain BCF-91A double-clad wavelength-shifting fibres, each of 1 mm diameter, at a pitch of 3.2 mm and read out at both ends, giving 128 optical channels (Collaboration et al., 18 Jul 2025). The 10 litre Mini-LiquidO prototype is traversed by 208 Kuraray B-3 WLS fibres, of which 56 fibres are coupled to Hamamatsu S13360--1350PE SiPMs on one end (Navas-Nicolás et al., 14 Mar 2025). The 1 liter opaque WbLS detector uses 32 WLS fibers total, arranged as 16 fibers in each of two vertical planes, read out by a Hamamatsu H12700A 64-channel MAPMT (Collaboration, 2024). The later pilot 3D-projection oWbLS detector uses 320 fibers in three orthogonal planes at 1 cm pitch with Hamamatsu S13360-1325CS MPPCs (Che et al., 11 May 2026).

Photosensor choice is similarly adapted to the architecture. In the POLDI neutron detector, SiPMs are selected instead of PMTs because of their high packing fraction and insensitivity to magnetic fields (Mosset et al., 2013). In LiquidO prototypes, SiPM arrays are used at both fiber ends or on distributed boards to preserve compactness and exploit single-photon sensitivity (Collaboration et al., 18 Jul 2025, Che et al., 11 May 2026). A transferable design lesson appears in cryogenic fiber-relay systems as well: once light is captured locally by wavelength-shifting fibers, long-range transport can be decoupled from the scintillator bulk and bridged mechanically by free-space optics or other interfaces if needed (Stoykov et al., 2011).

Some architectures are related rather than strictly identical. ArCLight is not an opaque scintillator medium, but it is a thin, dielectric, reflective, wavelength-shifting light-trap panel with edge-mounted SiPMs that collects scintillation light non-imagingly over broad surfaces (Auger et al., 2017). Its relevance is architectural: it demonstrates that distributed collection and internal trapping can substitute for conventional large-area PMT coverage in environments where direct line-of-sight readout is undesirable.

4. Materials systems and representative detector implementations

Opaque scintillation detectors are defined as much by material choice as by readout geometry. Several material families appear in the literature.

ZnS(Ag)/$^{6}$4LiF neutron scintillators

ZnS(Ag)/$^{6}$5LiF is a neutron-sensitive, highly scattering scintillator used in the POLDI diffractometer upgrade (Mosset et al., 2013). The neutron conversion proceeds through

$^{6}$6

with the resulting triton carrying 2.73 MeV and the alpha 2.05 MeV (Mosset et al., 2013). The scintillator is relatively slow: decay measurements fitted with

$^{6}$7

give $^{6}$8, $^{6}$9, $^{6}$0, and $^{6}$1, with photon fractions 2%, 5%, 41%, and 52% respectively (Mosset et al., 2013). Thus 93% of the detected photons belong to components with decay times of 1.224 and 15.2 $^{6}$2 (Mosset et al., 2013). This slow temporal structure is a major reason photon counting is favored over analog integration.

NoWaSH and opaque organic liquid scintillators

NoWaSH is a LAB-based wax scintillator developed specifically for opaque detector concepts (Buck et al., 2019). Tested formulations include NoWaSH-10, NoWaSH-15, and NoWaSH-20, with 10, 15, and 20 wt.% wax respectively (Buck et al., 2019). At 30°C after controlled cooling, NoWaSH-20 is already opaque for distances greater than 1 cm, while at 25°C all three appear opaque (Buck et al., 2019). For wax concentrations above 10 wt.%, the scattering length observed is less than 2 mm at 400 nm (Buck et al., 2019). The material can be filled warm and operated cool, and its low-temperature opacity is tied to wax crystallization rather than strong absorption.

The 10 litre LiquidO prototype employs NoWaSH-20, described as LAB + paraffin wax + PPO at 3 g/l, with 20% wax loading by weight (Navas-Nicolás et al., 14 Mar 2025). At $^{6}$3, it is opaque; at $^{6}$4, it is transparent (Navas-Nicolás et al., 14 Mar 2025). The prototype uses this temperature dependence to demonstrate progressive confinement as the scattering length decreases with decreasing temperature (Navas-Nicolás et al., 14 Mar 2025).

Opaque water-based liquid scintillators

Opaque water-based liquid scintillator extends the WbLS concept into a strongly scattering regime. In the 1 liter detector, oWbLS1, oWbLS2, and oWbLS3 were produced with 1.5%, 2.0%, and 2.5% water, respectively (Collaboration, 2024). The organic phase is a mix of modified polyethylene glycol-based surfactants in $^{6}$5 DIN liquid scintillator with PPO at 3 g/L and bis-MSB at 15 mg/L (Collaboration, 2024). Measured light yields are $^{6}$6, $^{6}$7, and $^{6}$8 photons/MeV for oWbLS1, oWbLS2, and oWbLS3, compared with $^{6}$9 photons/MeV for LAB+PPO (Collaboration, 2024). This shows that strong scattering need not imply low intrinsic scintillation yield.

A later pilot 3D-projection oWbLS detector uses an $5^\circ$0 acrylic vessel instrumented with three orthogonal planes of Kuraray Y11 multi-clad wavelength-shifting fibers and CITIROC-based front-end boards (Che et al., 11 May 2026). The paper reports three-dimensional event displays of cosmic muons and proton beam candidates, as well as tight transverse confinement and single-channel timing resolution $5^\circ$1–$5^\circ$2 ns with good photostatistics (Che et al., 11 May 2026).

Microcrystal scintillators

A more general material concept is the hybrid organic/inorganic microcrystal scintillator, in which microscopic grains of inorganic scintillating crystals are suspended in an organic scintillating carrier (Wagner et al., 2018). The intended grain size range is approximately $5^\circ$3 to $5^\circ$4 (Wagner et al., 2018). For a representative simulation with 8 $5^\circ$5m grains, 50/50 volume fraction, $5^\circ$6, $5^\circ$7, and wavelength 400 nm, the mean free path between two scattering processes is $5^\circ$8, and photons remain largely confined within roughly 5 cm of the origin (Wagner et al., 2018). This suggests that opaque scintillation need not be limited to liquids or wax-based systems; it can be engineered through particulate media as well.

5. Signal formation, reconstruction, and timing

The opaque detector regime changes not only optics but also signal processing. Because light collection is localized and often sparse, reconstruction proceeds from the spatial and temporal structure of the local fiber hit pattern rather than from a global flash.

In LiquidO-style detectors, the most direct observable is the pattern of light among nearby fibers. In the 30 mm Cube, each row of 8 fibres gives a 1D transverse position estimate via a weighted centroid,

$5^\circ$9

where $\sim 0.001$0 is the signal weight, taken either as raw ToT or as estimated p.e. summed from both fibre ends (Collaboration et al., 18 Jul 2025). The row resolution is then estimated from the geometric mean of residual widths,

$\sim 0.001$1

Using this method, the opaque detector achieves $\sim 0.001$2 per row in central rows, compared with $\sim 0.001$3 for the transparent fill (Collaboration et al., 18 Jul 2025). This is sub-pitch localization at 3.2 mm pitch.

The 1 liter opaque WbLS detector uses a simpler center-of-mass-like reconstruction. For example,

$\sim 0.001$4

with analogous expressions for $\sim 0.001$5 and $\sim 0.001$6, followed by empirical polynomial remapping from reconstructed CoM space to true position (Collaboration, 2024). Within the central fiducial region, the mean reconstruction error is 4.4 mm for 1.6 MeV-equivalent events and 7.4 mm for 0.8 MeV-equivalent events (Collaboration, 2024). This suggests that even a simple charge-weighted reconstruction can recover mm-scale position information from a sparse opaque-fiber detector.

The 3D-projection oWbLS detector extends this to voxelized reconstruction from orthogonal fiber planes. Candidate voxels are formed by matching hits from XZ and YZ views at common $\sim 0.001$7, with voxel charge assigned as

$\sim 0.001$8

Track extraction then uses a PE-weighted covariance matrix,

$\sim 0.001$9

followed by principal component analysis (Che et al., 11 May 2026). This is a distinctly imaging-oriented use of scintillation readout: the medium is continuous, yet the detector behaves as if segmented by optical confinement.

Timing can be equally important. In POLDI, the slow scintillator motivates an innovative photon-counting signal processing system based on repeated 200 ns counting windows at 5 MHz, with a rolling

$>0.5$0

and FPGA logic for neutron/noise and neutron/gamma discrimination (Mosset et al., 2013). For trigger condition $>0.5$1, the simulated trigger efficiency on neutrons is 90% and the noise-trigger rate is 25 Hz; applying a second threshold on $>0.5$2 of 10 yields 98% background rejection and 90% neutron detection efficiency for triggered events, giving $>0.5$3 and residual background rate 0.5 Hz (Mosset et al., 2013). The important point is architectural: in opaque scintillator readout, front-end logic is tightly coupled to the scintillator temporal structure.

LiquidO timing work addresses a different question: whether nanosecond-scale timing survives strong scattering well enough to separate prompt Cherenkov and delayed scintillation light. In the 10 litre prototype, custom front-end and digitisation hardware with time resolution below 100 ps demonstrates clear separation between Cherenkov and scintillation pulses in suitable media (Navas-Nicolás et al., 14 Mar 2025). The related confinement paper reports a water Cherenkov pulse FWHM of about 8 ns and shows that opacity in NoWaSH induces at most an average event time shift of $>0.5$4 ns relative to transparent LAB+PPO-like operation (Collaboration et al., 4 Mar 2025). This suggests that stochastic confinement does not necessarily destroy timing-based particle-identification handles.

6. Performance, applications, and trade-offs

Opaque scintillation detectors have been proposed or demonstrated for neutron diffraction, neutrino physics, antineutrino detection, rare-event searches, calorimetry, medical imaging, and compact tracking devices (Mosset et al., 2013, Cabrera et al., 2019, Collaboration, 2024, Navas-Nicolás et al., 14 Mar 2025). Their shared attraction is the combination of scintillation light yield with localized topology information.

The performance envelope varies strongly by implementation. The POLDI detector specification targets 400 angular channels, 2.5 mm spatial resolution, 65% detection efficiency at 1.2 $>0.5$5, time resolution $>0.5$6, sustainable count rate 4 kHz/channel, gamma sensitivity $>0.5$7, and intrinsic noise $>0.5$8 (Mosset et al., 2013). The 1 liter opaque WbLS detector demonstrates both spectroscopy and topological reconstruction of point-like events, with energy resolution $>0.5$9 at 1.6 MeV-equivalent and $>10$0 at 0.8 MeV-equivalent (Collaboration, 2024). The LiquidO Cube demonstrates sub-millimetre tracking with 450 $>10$1m per-row position resolution and about 170 p.e./MeV detected light for the opaque fill, compared with 140 p.e./MeV in the transparent fill, albeit with large absolute uncertainty (Collaboration et al., 18 Jul 2025). The 10 litre Mini-LiquidO prototype measures nominal response 8.4 PE/MeV in NoWaSH-Opaque, but data-driven scaling suggests optimized small detectors could reach $>10$2–500 PE/MeV (Collaboration et al., 4 Mar 2025).

These results point to a recurring trade-off structure. Stronger scattering sharpens localization but can complicate absolute energy response and calibration. Denser fiber lattices improve collection and interpolation but increase inactive material, complexity, and cost (Collaboration et al., 18 Jul 2025). In neutron-sensitive ZnS(Ag)-based systems, slow decay and afterglow require careful retrigger suppression and digital logic (Mosset et al., 2013). In opaque WbLS systems, measured energy resolution remains modest in early prototypes because total response includes collection fluctuations, detector nonuniformity, and prototype inefficiencies (Collaboration, 2024). In hybrid opaque concepts for double beta plus decay, improved topology and blob-dependent Cherenkov/scintillation information are accompanied by large embedded fiber inventories and demanding radiopurity requirements (Collaboration et al., 2024).

Applications often depend on those trade-offs. For neutrino detectors, opacity is attractive because event topology is preserved and heavy loading is easier than in transparent scintillators (Cabrera et al., 2019, Collaboration et al., 2024). For low-energy neutron detectors, opacity-tolerant local fiber capture provides a route away from $>10$3He while retaining thin, segmented channels (Mosset et al., 2013). For 3D tracking calorimeters, oWbLS with orthogonal fiber planes offers a fully active alternative to mechanically segmented scintillator cubes (Che et al., 11 May 2026). A plausible implication is that opaque scintillation is especially useful where topology matters as much as calorimetry.

7. Historical development, controversies, and future directions

The opaque scintillation detector concept emerged as a deliberate break with the long-standing paradigm of transparency in scintillator detectors. “Neutrino Physics with an Opaque Detector” framed this explicitly as a concept that “breaks with the conventional paradigm of transparency” by confining and collecting light near its creation point with an opaque scintillator and a dense array of optical fibres (Cabrera et al., 2019). Early material work on NoWaSH then established that a paraffin-wax/LAB system could be switched between transparent and opaque states and could produce local confinement in a small LiquidO test cell (Buck et al., 2019). Subsequent work progressively moved from proof of principle to detector characterization, quantitative confinement studies, topological reconstruction, and particle tracking (Collaboration, 2024, Collaboration et al., 4 Mar 2025, Collaboration et al., 18 Jul 2025, Che et al., 11 May 2026).

One recurrent misconception is that an opaque scintillator is simply a poor scintillator with too much absorption. The literature consistently rejects this interpretation. LiquidO requires short scattering length and long enough absorption length; the observed increase in nearby-fibre yield at low temperature in NoWaSH rules out an absorption-only explanation (Navas-Nicolás et al., 14 Mar 2025). Likewise, microcrystal scintillator work emphasizes that the medium is turbid or diffusive because of repeated Mie scattering, not because scintillation photons are immediately lost (Wagner et al., 2018).

Another important distinction is between opaque scintillation media and merely non-imaging light collectors. ArCLight, for example, is highly relevant as a non-imaging, internally trapping scintillation-light collector, but it is not itself an opaque scintillator medium (Auger et al., 2017). Conversely, transparent monolithic scintillators with engineered optical boundaries can exploit restricted transport and sparse readout for localization, yet they are not opaque detectors in the LiquidO sense (Simhony et al., 2024). The category therefore includes both strict opaque-media systems and closely related architectures that solve similar optical-transport problems by local extraction.

Current research directions are defined by scaling and optimization. The LiquidO program is moving from point-like demonstrations toward larger prototypes and detailed particle-identification studies (Navas-Nicolás et al., 14 Mar 2025). Opaque WbLS detectors are progressing from 1 liter two-plane systems to three-view 3D-projection detectors with beam-test validation (Collaboration, 2024, Che et al., 11 May 2026). Fiber-collection modeling has advanced from qualitative arguments to experimental validation and diffusion-based reduced models (Wilhelm et al., 2023). Hybrid concepts now combine opacity with Cherenkov/scintillation discrimination and blob-based reconstruction for positron-rich signatures such as double beta plus decay (Collaboration et al., 2024). This suggests that opaque scintillation detectors are developing not as a single device type but as a detector paradigm centered on local optical collection in media where long-range transparency is unavailable, undesirable, or deliberately suppressed.