Surface-Scintillator Detectors (SSD)

• Surface-Scintillator Detectors (SSDs) are scintillator-based active surfaces that capture energy deposited by charged particles and distinguish electromagnetic, muonic, and neutron components.
• They convert energy deposits into optical signals via wavelength-shifting fibers or direct coupling, playing a key role in ultra-high-energy cosmic-ray and neutrino detection systems.
• Innovative designs including precise calibration, 3D segmentation, and additive manufacturing enhance SSD performance for improved particle composition analyses.

Searching arXiv for recent SSD-related papers to ground the article. [Tool use unavailable in this interface; proceeding with the provided arXiv records and citing them directly.] Surface-Scintillator Detectors (SSDs) are scintillator-based detectors deployed at a physical surface or boundary, where charged particles deposit energy in a scintillating medium and the resulting optical signal is transported to a photodetector by wavelength-shifting fibers or direct optical coupling. In ultra-high-energy cosmic-ray instrumentation, the archetypal SSD is the thin plastic scintillator module mounted on top of a water-Cherenkov detector, introduced to exploit the different detector responses to the electromagnetic and muonic components of extensive air showers; in low-background noble-liquid instrumentation, the same term has been used for a thin slow scintillating layer at the liquid-argon vessel boundary that tags surface events through pulse-shape discrimination (Collaboration, 10 Jul 2025, Conte, 12 Jul 2025, Boulay et al., 2019). Across these implementations, the defining feature is not a single geometry but an active scintillating surface whose optical, temporal, and calibration properties are engineered to isolate surface-adjacent particle populations.

1. Functional definition and detector physics

In air-shower arrays, SSDs are thin plastic scintillators whose response is essentially proportional to the charged-particle energy deposition per unit path length, $dE/dx$, and therefore more uniform for electrons, positrons, and muons traversing the scintillator than the response of a water-Cherenkov detector. This contrast underlies the layered-detector strategy of AugerPrime, where the SSD response in MIP units is combined with the WCD response in VEM units to disentangle electromagnetic and muonic shower components station by station (Conte, 12 Jul 2025, Collaboration, 10 Jul 2025).

The same surface-scintillator logic appears in the planned IceCube surface enhancement, but with a different systems objective. There, plastic scintillator panels installed above the snow are intended to calibrate the effect of snow accumulation on IceTop tanks, improve cosmic-ray measurements, and enhance atmospheric background rejection for high-energy astrophysical neutrino detection. The proposed surface enhancement comprises 256 plastic scintillation detectors arranged as 32 stations with 8 scintillator panels of $1.5\,\text{m}^2$ each (Leszczyńska et al., 2019).

In liquid argon detectors, the term denotes a different geometry but a closely related function. A thin scintillating layer inserted between the acrylic vessel wall and a TPB wavelength shifter converts the detector boundary itself into an SSD: surface $\alpha$ decays deposit energy in a scintillator with a much slower time profile than liquid argon, so their prompt fraction is driven away from the nuclear-recoil region of interest (Boulay et al., 2019, Gallacher et al., 2019). This suggests that “surface-scintillator detector” is best understood as a functional category rather than a unique hardware format.

2. Canonical architectures in large air-shower arrays

The mature observatory-scale implementation is the AugerPrime Scintillator Surface Detector. Each module is a rectangular box of external dimensions $3800\,\text{mm} \times 1280\,\text{mm}$ with total active scintillator area $3.84\,\text{m}^2$. The detector is composed of two scintillator panels; each panel contains 24 extruded bars of size $$1.6\,\text{m} \times 5\,\text{cm} \times 1\,\text{cm}$$, giving 48 bars per module. The bars are based on FNAL-NICADD STYRON 663-W polystyrene, doped with $1\%$ PPO and $0.03\%$ POPOP, co-extruded with a $0.25\,\text{mm}$ TiO$_2$ reflective coating, and instrumented with Kuraray Y11(300)M S-type wavelength-shifting fibers. All 96 fiber ends are bundled into a PMMA “cookie” and read out by a single Hamamatsu R9420 PMT. The PMT signal is digitized by a 12-bit FADC at 120 MHz, with SSD high- and low-gain channels in a gain ratio of 128, and the channel is designed to sustain signals above $1.5\,\text{m}^2$0 MIP before saturation (Collaboration, 10 Jul 2025, Conte, 12 Jul 2025).

The AugerPrime mechanical design is explicitly field-hardened. The housing is an aluminium box with a sandwich bottom panel of aluminium and extruded polystyrene foam, internal EPS filling to reduce free air volume, pressure equalization through a porous but light-tight sintered metal plate, and a corrugated aluminium sunroof mounted about $1.5\,\text{m}^2$1 above the cover. Production QA used the ratio $1.5\,\text{m}^2$2 as a light-yield proxy; the design requirement was $1.5\,\text{m}^2$3 SPE/MIP, while final SSDs typically achieved $1.5\,\text{m}^2$4 (Collaboration, 10 Jul 2025).

The IceCube surface scintillator design retains the thin-bar plastic-scintillator paradigm but uses SiPM readout. Each panel consists of 16 Fermilab plastic scintillator bars of dimensions $1.5\,\text{m}^2$5, coated with a reflective TiO$1.5\,\text{m}^2$6 dye and arranged to form an active area of $1.5\,\text{m}^2$7. A $1.5\,\text{m}^2$8 diameter Kuraray Y11(300) multi-cladding fiber is embedded in the bars and coupled by optical gel to a Hamamatsu S13360-6025PE SiPM. The assembly is supported by styrofoam and plywood and enclosed in an aluminum casing; at the trigger level, the simulation imposes $1.5\,\text{m}^2$9 photoelectrons within 200 ns and a signal cut of $\alpha$0 VEM (Leszczyńska et al., 2019).

Long-bar scintillator R&D provides the underlying optical benchmark for many SSD bar geometries. A 16 m Kuraray Y11(200) S-type WLS fiber read out at both ends by Geiger-mode multi-pixel photodiodes yielded about 15 p.e./MIP for a 1 cm wide bar and about 10 p.e./MIP for a 4 cm wide bar at the worst point, corresponding to a minimum light yield of 7.2 p.e./MeV for the 4 cm bar (Mineev et al., 2011). For surface arrays, such measurements define the feasible region for channel reduction without losing MIP efficiency.

3. Calibration, signal units, and response models

Operational SSD calibration in AugerPrime is based on atmospheric muons. Calibration histograms of peak and charge are accumulated every 61 seconds, and SSD muon purity is improved by requiring a muon-like signal in the associated WCD. The core correction is the omnidirectional-to-vertical conversion

$\alpha$1

so the vertical-equivalent calibration unit is obtained from $\alpha$2. AugerPrime also developed an independent online calibration based on a single-bin threshold trigger tuned to 70 Hz; the empirical relation

$\alpha$3

provides a real-time estimate of the MIP peak from the stabilized threshold $\alpha$4. Shower signals are then expressed in MIP units and fitted with an SSD lateral distribution function using $\alpha$5, $\alpha$6, and $\alpha$7 as the shower-size parameter (Conte, 12 Jul 2025).

At the model level, SSD response has been embedded in a universality-based parameterization that treats the ground signal as a sum of four components: $\alpha$8, $\alpha$9, $3800\,\text{mm} \times 1280\,\text{mm}$0, and $3800\,\text{mm} \times 1280\,\text{mm}$1. For each component,

$3800\,\text{mm} \times 1280\,\text{mm}$2

In this parameterization, the EM component has $3800\,\text{mm} \times 1280\,\text{mm}$3, while the muonic and muon-related components scale approximately linearly with $3800\,\text{mm} \times 1280\,\text{mm}$4. The reference SSD signal is parameterized directly in MIP units, with the detector response derived from full detector simulation in Auger Offline (Stadelmaier et al., 2024).

IceCube uses a complementary but more detector-explicit simulation strategy. A Geant4.10 model includes scintillator bars, TiO$3800\,\text{mm} \times 1280\,\text{mm}$5 reflective coating, WLS fibers, and realistic SiPM geometry and optical coupling. Position-dependent photon yield and the time of the first detected photon are parameterized from detailed optical simulations and then folded into large-scale CORSIKA shower simulations, permitting array-level reconstruction without full optical tracking of every photon (Leszczyńska et al., 2019).

4. Observatory-scale implementations and measured performance

The AugerPrime SSD system has moved from design to stable array operation. More than 1400 SSD units were deployed, and the full array was completed in 2023. Phase-II monitoring shows a Gaussian MIP charge distribution over the array, with only $3800\,\text{mm} \times 1280\,\text{mm}$6 of calibrated signals outside $3800\,\text{mm} \times 1280\,\text{mm}$7; the outliers are associated with a small set of detectors operating away from the nominal high-voltage range of $3800\,\text{mm} \times 1280\,\text{mm}$8. The array-averaged MIP charge exhibits an annual modulation with dominant period of approximately 362 days and amplitude of about $3800\,\text{mm} \times 1280\,\text{mm}$9, while hourly studies show MIP variation up to $3.84\,\text{m}^2$0 throughout a day and peak-to-peak variation of about $3.84\,\text{m}^2$1, more pronounced in summer. The upgraded WCD+SPMT+SSD system extends the usable dynamic range to several tens of thousands of VEM/MIP without significant saturation (Conte, 12 Jul 2025).

AugerPrime SSDs have also been used to isolate the neutron component of extensive air showers. Using the WCD-defined start time $3.84\,\text{m}^2$2, the SSD trace is analyzed in relative time $3.84\,\text{m}^2$3. The majority of prompt shower pulses occur within approximately $3.84\,\text{m}^2$4, whereas neutron candidates are selected with $3.84\,\text{m}^2$5. Pulse finding requires a start above $3.84\,\text{m}^2$6, an end defined by two consecutive sub-threshold bins, and a minimum duration of $3.84\,\text{m}^2$7. In the $3.84\,\text{m}^2$8 bin, late pulses appear in $3.84\,\text{m}^2$9 of traces at $$1.6\,\text{m} \times 5\,\text{cm} \times 1\,\text{cm}$$0 and $$1.6\,\text{m} \times 5\,\text{cm} \times 1\,\text{cm}$$1 of traces at $$1.6\,\text{m} \times 5\,\text{cm} \times 1\,\text{cm}$$2; in the $$1.6\,\text{m} \times 5\,\text{cm} \times 1\,\text{cm}$$3 bin, the corresponding fractions are $$1.6\,\text{m} \times 5\,\text{cm} \times 1\,\text{cm}$$4 and $$1.6\,\text{m} \times 5\,\text{cm} \times 1\,\text{cm}$$5. This extends SSD utility from prompt EM–muon separation to delayed hadronic diagnostics (Schulz, 23 Jul 2025).

IceCube’s proposed surface array shows the same instrument class operating in a different energy and environmental regime. The array trigger is defined by $$1.6\,\text{m} \times 5\,\text{cm} \times 1\,\text{cm}$$6 scintillator detectors with signal $$1.6\,\text{m} \times 5\,\text{cm} \times 1\,\text{cm}$$7 VEM within a 1.5 $$1.6\,\text{m} \times 5\,\text{cm} \times 1\,\text{cm}$$8s window. For protons up to $$1.6\,\text{m} \times 5\,\text{cm} \times 1\,\text{cm}$$9 zenith, trigger efficiency reaches about $1\%$0 at $1\%$1; reconstruction efficiency reaches about $1\%$2 for protons and about $1\%$3 for iron. Direction resolution is a few degrees below $1\%$4 and improves to better than $1\%$5 above $1\%$6. A simple two-parameter composition analysis using the scintillator LDF slope $1\%$7 and the tank-to-scintillator signal ratio at 200 m yields a proton–iron figure of merit greater than 1.3 in both zenith bins considered (Leszczyńska et al., 2019).

These results contradict the notion that SSDs are merely auxiliary particle counters. In both AugerPrime and IceCube, the SSD is a calibrated composition-sensitive observable, and in AugerPrime it has already become a probe of delayed neutron activity as well.

5. Segmentation, additive manufacturing, and non-canonical SSDs

A major recent extension of the SSD concept is the move from thin, quasi-two-dimensional modules to fully three-dimensional segmented scintillators. The 3DET collaboration’s additive-manufactured “SuperCube” is a $1\%$8 matrix of optically isolated scintillating voxels, for a total of 125 voxels, fabricated by Fused Injection Molding (FIM). Horizontal inter-voxel walls are $1\%$9 thick, vertical walls are $0.03\%$0, and metal rods of diameter $0.03\%$1 are inserted during fabrication to define continuous fiber channels. The reflective frame is printed in a white polycarbonate filament mixed with PTFE; the cavities are then filled with molten polystyrene scintillator doped with pTP and POPOP, using CFD-optimized parameters $0.03\%$2 and $0.03\%$3. Readout uses Kuraray Y11 double-cladding WLS fibers and Hamamatsu MPPC S13360-1325CS SiPMs with photon detection efficiency $0.03\%$4 (Kose, 2024).

Performance of the FIM SuperCube is already within the regime relevant to surface scintillator instrumentation. The measured light yield is about 28 photoelectrons per channel for a minimum-ionizing particle crossing a voxel; within one cube the non-uniformity is about $0.03\%$5, while across the five central cubes the variation is less than $0.03\%$6. Cube-to-cube light leakage is $0.03\%$7, compared with $0.03\%$8 in earlier custom-reflector prototypes. For surface-scintillator contexts, the paper explicitly notes that 3D segmentation could be translated into stacking multiple such cubes into a large plane with fine transverse and longitudinal segmentation, improving shower imaging at the surface (Kose, 2024). This suggests that future SSDs need not remain limited to bar-and-panel geometries.

Large-area optical R&D has pursued a different path: clear fibres mounted perpendicular on the surface of the scintillator as a matrix providing uniform surface reduction. The design target is time resolution of the order 100–200 ps while avoiding a number of channels that grows proportional to the surface. For a representative case with scintillator thickness $0.03\%$9, fiber length $0.25\,\text{mm}$0, and fiber density $0.25\,\text{mm}$1, direct-photon time spread is about 112 ps, while diffuse reflection can increase the light collection efficiency by about a factor $0.25\,\text{mm}$2 relative to direct photons alone (Grabski, 2019). The optical principle is directly relevant to SSDs whenever timing, uniformity, and channel economy are co-equal constraints.

In low-background argon detectors, layered wavelength-shifting and scintillating films realize an SSD at the detector boundary. A 50 micron scintillator layer with decay time approximately 300 ns or greater, placed behind a 3 μm TPB film, can provide discrimination power greater than $0.25\,\text{mm}$3 against surface $0.25\,\text{mm}$4 events while preserving about $0.25\,\text{mm}$5 nuclear-recoil acceptance in the defined region of interest (Boulay et al., 2019). Argon-1 implements this concept in a 30-kg single-phase LAr detector using two Hamamatsu S14161-3050HS-08 SiPM arrays with 128 total channels; surface events are separated with the usual $0.25\,\text{mm}$6 variable and a topological statistic

$0.25\,\text{mm}$7

which is larger for localized surface scintillation than for bulk events (Gallacher et al., 2019).

6. Systematic limitations and future directions

The dominant SSD systematics in deployed surface arrays are now well characterized. In AugerPrime, the omnidirectional-to-vertical muon correction contributes a $0.25\,\text{mm}$8 uncertainty through the measured $0.25\,\text{mm}$9 ratio; temperature drives seasonal MIP variations of about $_2$0 and daily variations up to $_2$1, with peak-to-peak changes of about $_2$2. In the late-pulse neutron analysis, non-hadronic contamination in the $_2$3 window ranges from $_2$4 to $_2$5 depending on energy, zenith angle, distance, and pulse charge, while electronics undershoot and baseline saturation below about 250 m from the shower core suppress measurable late-pulse rates (Conte, 12 Jul 2025, Schulz, 23 Jul 2025).

Model-space limitations are equally specific. The universality-based SSD response model is tuned and validated for zenith angles up to $_2$6 and energies between roughly $_2$7 and $_2$8; at larger angles, the factorized parameterization breaks down. The IceCube enhancement remains conditioned by snow-related systematics in the tanks and by the need to validate SiPM response and optical parameterizations under South Pole conditions. In additive-manufactured 3D scintillators, current crosstalk of $_2$9, intra-voxel non-uniformity at the $1.5\,\text{m}^2$00 level, technical attenuation length of about 19 cm, and the absence of detailed long-term aging, radiation-hardness, and environmental-stability studies define the present R&D boundary (Stadelmaier et al., 2024, Leszczyńska et al., 2019, Kose, 2024).

The near-term trajectory is therefore less about proving the detector class than about tightening its control parameters. AugerPrime is already exploiting redundant histogram-based and rate-based calibrations, and future analyses will use combined SSD+WCD+FD observables for composition and hadronic-interaction studies. IceCube’s program points toward a hybrid surface detector combining scintillator, tank, and radio response. The 3DET collaboration is pursuing fully automated 3D printing, improved white reflector materials, and even additive-manufactured sampling calorimeters with metal filaments (Conte, 12 Jul 2025, Leszczyńska et al., 2019, Kose, 2024). A plausible implication is that SSDs are evolving from thin charged-particle counters into a broader family of active surfaces: composition-sensitive planes in air-shower arrays, neutron-sensitive delayed-signal detectors, and optically instrumented boundary layers in low-background cryogenic detectors.