Water-Based Liquid Scintillator (WbLS)
- WbLS is a tunable hybrid medium that emulsifies a small organic scintillator fraction in ultra-pure water, preserving Cherenkov response alongside added scintillation light.
- Its composition can be finely adjusted through varying scintillator loading, fluor concentration, and surfactant chemistry to balance light yield, attenuation, and timing structure.
- WbLS supports diverse detector architectures—from ton-scale testbeds to kiloton observatories—enabling advanced neutrino detection, rare-event searches, and detailed calorimetry.
Searching arXiv for recent and foundational WbLS papers to ground the article with current literature. Water-based liquid scintillator (WbLS) is a surfactant-stabilized mixture in which a small fraction of organic liquid scintillator is emulsified in ultra-pure water, typically as micelles or nanoscopic droplets, so that the medium retains water-like transparency and Cherenkov response while adding scintillation light and low-threshold sensitivity (Alonso et al., 2014, Fischer, 2018). Across the literature, WbLS is treated as a tunable hybrid medium rather than a single fixed composition: the scintillator fraction, fluor content, surfactant chemistry, wavelength-shifting strategy, and possible isotope loading are all adjusted to trade light yield against attenuation, timing structure, and directional information. This tunability underlies its use in detector concepts ranging from ton-scale testbeds and segmented trackers to kiloton-scale observatories and multi-purpose concepts such as Theia and the ASDC (Bignell et al., 2015, Zhao et al., 2023).
1. Chemical formulations and compositional regimes
WbLS formulations span a broad range of organic loading and chemistries. Foundational concept papers describe “water-like” mixtures at about – scintillator in water, with higher-loading mixtures extending to scintillator in water, and emphasize that water-like behavior is retained at low loading whereas higher loading moves the medium toward an oil-like regime (Alonso et al., 2014, Seo, 2019). The solvent phase has included linear alkyl benzene (LAB), pseudocumene (PC), di-isopropylnaphthalene (DIN), phenylxylylethane, and phenylcyclohexane, typically with PPO as primary fluor and bis-MSB or MSB as secondary wavelength shifter when used (Alonso et al., 2014, Bignell et al., 2015, Collaboration et al., 2023).
Representative formulations illustrate the diversity of this design space.
| Study | Formulation | Purpose |
|---|---|---|
| (Bignell et al., 2015) | WbLS-1: PC by mass, PPO $0.4$ g/L, bis-MSB $3$ mg/L | Optical characterization and proton-beam modeling |
| (Bignell et al., 2015) | WbLS-2: PC by mass, PPO $1.36$ g/L, bis-MSB $7.48$ mg/L | Optical characterization and proton-beam modeling |
| (Callaghan et al., 2022) | LAB by volume with 0 g/L PPO, no additional secondary fluor | Proton light-yield measurement below 1 MeV |
| (Zhao et al., 2023) | 2 LAB by mass in water, with surfactant and anti-scattering additive, no separate wavelength shifter added | Ton-scale deployment and stability study |
| (Steiger et al., 2024) | 3 water, 4 Triton X-100, 5 LAB, PPO 6 g/L, vitamin C 7 mg/L | Novel formulation emphasizing stability and PSD |
| (Li et al., 15 Aug 2025) | 8 H9O and 0 LAB by mass, PPO 1 g/L, MSB 2 mg/L | 3D segmented tracker prototype |
Surfactant chemistry is central because no covalent reaction defines WbLS; the medium is produced by micellization driven by the hydrophobic effect (Steiger et al., 2024). Reported surfactants include linear alkyl sulfonate or similar anionic surfactants, modified polyethylene-glycol based surfactants, Triton X-100, IGEPAL CO-630, Pluronic F-127 or similar nonionic surfactants, and proprietary Brookhaven formulations (Fischer, 2018, Collaboration et al., 2023, Steiger et al., 2024, Onda et al., 25 Jul 2025, Andrade et al., 20 Mar 2026). Several studies explicitly connect surfactant choice to long-term emulsion stability, attenuation, and achievable organic loading (Zhao et al., 2023, Steiger et al., 2024, Onda et al., 25 Jul 2025).
A recurring advantage of WbLS is compatibility with dissolved or complexed additives. Concept and design papers cite soluble metal salts or complexes for neutron tagging and rare-event searches, including Gd, Li, Te, Xe, Pb, and Zr (Alonso et al., 2014, Fischer, 2018). Experimental programs have also developed Gd-compatible WbLS and measured light-yield response over 3–4 organic-plus-Gd loading, with no precipitation or phase separation observed up to 5 total organic+Gd loading on time scales of at least six months (Gwon et al., 17 Dec 2025).
2. Optical response, emission, attenuation, and light yield
The optical behavior of WbLS is formulation-dependent, but several regularities recur across measurements and simulations. Light yield generally increases with organic fraction, while attenuation length tends to decrease from water-like values toward those of pure scintillator (Alonso et al., 2014, Bignell et al., 2015, Kaptanoglu et al., 2021). Emission is typically governed by PPO and any added wavelength shifter, with PPO emission around 6–7 nm and shifted emission extending into the 8–9 nm region, which overlaps the quantum-efficiency or PDE maxima of bialkali PMTs, SiPMs, and Y11 wavelength-shifting fibers used in tracker concepts (Bignell et al., 2015, Li et al., 15 Aug 2025).
Attenuation is consistently described with Beer–Lambert behavior,
0
Measured or assumed attenuation properties vary widely with chemistry. For the 2015 PC-based WbLS-2, absorption above 1 nm fell to 2, implying attenuation lengths of tens of meters in the 3–4 nm region; a later 5 LAB-based WbLS formulation exhibited about two orders of magnitude longer attenuation length after material purification (Bignell et al., 2015). In the 2023 ton-scale detector, optical inputs used water attenuation 6 m in the 7–8 nm range and a fitted 9 WbLS light yield of $0.4$0 photons/MeV (Zhao et al., 2023). The Triton-X formulation reported $0.4$1 m (stat) $0.4$2 m (syst) and $0.4$3 m (stat) $0.4$4 m (syst), while dynamic light scattering gave a hydrodynamic micelle diameter of $0.4$5 nm and an estimated Rayleigh scattering length $0.4$6 m at $0.4$7 nm (Steiger et al., 2024).
Absolute light yield likewise covers a wide interval because the reported media are chemically distinct. The 2015 beam-and-simulation study found $0.4$8 (stat.) $0.4$9 (sys.) photons/MeV for WbLS-1 and $3$0 (stat.) $3$1 (sys.) photons/MeV for WbLS-2, compared with $3$2 (stat.) $3$3 (sys.) photons/MeV for pure liquid scintillator (Bignell et al., 2015). In the $3$4 ton-scale deployment, the weighted-average light yield was $3$5 photons/MeV, consistent with the earlier $3$6 photons/MeV small-cell result (Zhao et al., 2023). The BNL 1-ton DIN-based detector measured a non-Cherenkov yield of $3$7 photons/MeV at $3$8 WbLS (Xiang et al., 2024). The Triton-X formulation, measured by $3$9Cs Compton backscatter relative to EJ-309, yielded about 0 photons/MeV (Steiger et al., 2024).
Spectroscopic measurements show that WbLS does not necessarily inherit precursor scintillator properties in a trivial way. In X-ray studies of LAB/PPO-based WbLS prepared from a 1 g/L PPO precursor, the final 2, 3, and 4 WbLS samples all showed nearly identical spectra, and both their 5 nm intensity ratio and decay constants matched pure LAB with about 6 g/L PPO rather than the 7 g/L precursor. The paper states that this could indicate that the concentration of active PPO in the WbLS samples depends on their processing (Onken et al., 2020). This suggests that micellization and emulsification can alter the effective fluor concentration seen by the optical response, even when the precursor cocktail is nominally much more concentrated.
3. Time structure and Cherenkov–scintillation separation
The central operational idea of WbLS is that Cherenkov and scintillation light occupy different regions of time, angle, and wavelength space. Concept papers model the total time profile as a prompt Cherenkov term plus delayed scintillation exponentials, for example
8
or, in detector-level form,
9
with Cherenkov emission effectively prompt and scintillation delayed by rise and decay times of order nanoseconds to tens of nanoseconds (Fischer, 2018, Sawatzki et al., 2020). Because the scintillation component is isotropic while early Cherenkov photons retain ring-like directionality, timing cuts can be combined with topology and spectral information to isolate directional light.
Direct timing measurements quantify this separation. Using an LAPPD and a conventional PMT with effective resolution $1.36$0, three LAB+PPO-in-water mixtures at $1.36$1, $1.36$2, and $1.36$3 scintillator loading were fitted with a two-component exponential model plus nonzero rise time. The reported parameters were $1.36$4 ps, $1.36$5 ns, $1.36$6 ns for $1.36$7; $1.36$8 ps, $1.36$9 ns, $7.48$0 ns for $7.48$1; and $7.48$2 ps, $7.48$3 ns, $7.48$4 ns for $7.48$5 (Kaptanoglu et al., 2021). For those same mixtures, prompt timing cuts produced Cherenkov purity greater than $7.48$6 in all cases and greater than $7.48$7 for the $7.48$8 sample, with scintillation retention about $7.48$9 (Kaptanoglu et al., 2021).
Independent X-ray excitation measurements on LAB/PPO-based WbLS found similarly fast dominant scintillation. For 0, 1, and 2 LS fractions, the fastest component was about 3–4 ns with a weight fraction above 5, accompanied by a 6–7 ns component and a long 8 ns component (Onken et al., 2020). A distinct Triton-X formulation gave a two-component fluorescence fit with 9 ns, 00 ns carrying 01 of the light, and 02 ns carrying 03 (Steiger et al., 2024). These measurements reinforce the picture that WbLS timing is fast enough for sub-nanosecond photosensors to recover a useful prompt Cherenkov sample while preserving a delayed scintillation sample for calorimetry.
Detector studies translate this timing structure into reconstruction figures of merit. In MeV-scale simulations of 04, 05, and 06 WbLS in 07 kt and 08 kt detectors, the Cherenkov significance metric 09 peaked at later prompt windows for lower scintillator loading: in the 10 kt geometry at 11 MeV, the optimum was about 12 ns for 13 WbLS, 14 ns for 15, and 16 ns for 17 (Land et al., 2020). This documents the basic trade-off: increasing scintillator fraction improves calorimetric statistics but compresses the prompt window and degrades the purity of the earliest photons. The ASDC and Theia studies therefore emphasize fast timing photosensors, including LAPPDs with 18 ps or 19 ps, as enabling technology for event-by-event Cherenkov/scintillation separation in large detectors (Alonso et al., 2014, Fischer, 2018).
4. Ionization quenching, proton response, and pulse-shape discrimination
Although WbLS is often introduced through electron or minimum-ionizing-particle response, hadronic response is equally important for neutrino physics. A key milestone was the first measurement of the proton light yield of a 20 LAB-based WbLS below 21 MeV, performed with a double time-of-flight method at the LBNL 88-Inch Cyclotron (Callaghan et al., 2022). The WbLS formulation in that study contained 22 by volume LAB loaded with 23 g/L PPO and no additional secondary fluor, while the reference liquid scintillator was pure LAB + 24 g/L PPO + 25 mg/L bis-MSB, un-deoxygenated (Callaghan et al., 2022).
In that measurement, proton light yield 26 was defined relative to a 27 keV electron, with 28, and the target-PMT charge was calibrated in “MeV29” using the 30Cs Compton edge (Callaghan et al., 2022). Over 31–32 MeV proton kinetic energy, the WbLS proton light yield was about 33 lower than that of the LABPPO reference, uniformly across the measured range. At the top of the tabulated interval, for example, 34 in the 35–36 MeV bin was 37 for LABPPO and 38 for WbLS (Callaghan et al., 2022). The same work notes that WbLS has a somewhat larger contribution of prompt Cherenkov light because of the water matrix, affecting the electron calibration offset.
Quenching analysis in that study showed that Birks’ law alone was insufficient: 39 while the Chou extension,
40
described the data well when integrated numerically with SRIM stopping powers (Callaghan et al., 2022). For WbLS, the Birks-only fit gave 41 MeV42, 43 cm/GeV, and 44, whereas the Chou fit gave 45 MeV46, 47 cm/GeV, 48 cm49/GeV50, and 51 (Callaghan et al., 2022). The paper states that the best-fit 52 values are non-zero at more than 53, confirming the need for a second-order quenching term in both pure LABPPO and WbLS.
Higher-energy proton-beam studies also extracted Birks constants for low-loading PC-based WbLS. From 54 MeV data in Detector B, the reported values were 55 (stat.) 56 (sys.) mm/MeV for WbLS-1 and 57 (stat.) 58 (sys.) mm/MeV for WbLS-2, compared with 59 (stat.) 60 (sys.) mm/MeV for pure liquid scintillator (Bignell et al., 2015). This large formulation dependence indicates that “the quenching of WbLS” is not a universal material constant; it is specific to the solvent, fluor concentration, and emulsion chemistry used.
Pulse-shape discrimination adds another dimension to hadronic response. The Triton-X formulation was studied with neutron and gamma excitation from the 61 reaction, using a tail-to-total PSD variable
62
For this WbLS, the maximal separation was 63 at 64 ns, comparable to LAB+65 g/L PPO and below PC+66 g/L PPO, but sufficient for the paper to conclude that its PSD capability is comparable to fully-organic LAB-based scintillators (Steiger et al., 2024). This is directly relevant to neutron vetoes and to fast-neutron rejection in neutrino detectors.
5. Detector implementations from testbeds to segmented trackers
WbLS has been implemented in monolithic vessels, hybrid inserts, portable reactor concepts, and finely segmented trackers. The BNL 1-ton detector is an in-situ mixed UVT-acrylic cylinder holding about 67 m68 of liquid, instrumented with 69 PMTs: 70 71 PMTs beneath the tank and 72 73 PMTs on the side walls, digitized with CAEN V1730S boards at 74 MHz (Xiang et al., 2024). In that detector, pure water yielded 75 PE on bottom PMTs and 76 PE on side PMTs for crossing muons, while 77 WbLS yielded 78 PE and 79 PE respectively, enabling a fit-based extraction of scintillation and Cherenkov components (Xiang et al., 2024). The later Gd-compatible campaign on the same testbed measured intrinsic scintillation yield from 80 ph/MeV at 81 concentration to 82 ph/MeV at 83, with the data described by 84, 85, 86 (Gwon et al., 17 Dec 2025).
Scaling beyond bench and ton scale has also been demonstrated. A 87 L in-situ mixed detector reported a 88 WbLS light yield of 89 photons/MeV and stable operation over the L00–L02 run period, with no visible phase separation or cloudiness over three months of stable operation (Zhao et al., 2023). Brookhaven’s 30-ton prototype later extended this program to a 90L stainless-steel cylinder of radius 91 mm and half-height 92 mm, viewed by 93 94 Hamamatsu R16367 PMTs and equipped with recirculation systems for sequential exchange, nanofiltration, and future Gd-water band-pass filtering (Andrade et al., 20 Mar 2026). The 30-ton report states that baseline water runs over several months and post-injection WbLS runs over weeks exhibited stable light yield with less than 95 drift, and that month-averaged yield stability from a truncated-mean cosmic-muon analysis was 96 (Andrade et al., 20 Mar 2026).
A separate deployment strategy places WbLS inside an existing water Cherenkov detector. ANNIE installed a 97 L acrylic cylinder, the SANDI vessel, filled with 98 water and 99 by mass of a DIN+PPO organic scintillator stabilized by modified polyethylene-glycol based surfactants (Collaboration et al., 2023). In situ, the vessel showed a light increase factor of 00 for through-going muons and 01 for Michel electrons relative to pure water, while UV–vis transmission exhibited no statistically significant change before and after deployment (Collaboration et al., 2023). The paper reports this as a proof-of-concept demonstration of both Cherenkov light and scintillation from WbLS in a GeV neutrino beam environment.
Segmented WbLS trackers pursue a different optimization. One design encapsulates WbLS in 02 cm03 voxels read out by orthogonal Y11 wavelength-shifting fibers, achieving 04 water by mass in the active volume (Li et al., 15 Aug 2025). In its 05 cm06 matrix prototype, the WbLS version gave a most-probable light yield of 07 p.e./channel/MIP with mean 08 p.e., crosstalk 09, and timing resolution 10 ns per hit from the SiPM+FEB system (Li et al., 15 Aug 2025). A related WbLS tracking detector using reflective separators and three-directional fiber readout measured about 11 p.e./MeV per fiber in a 12 MeV positron beam, then improved the projected yield to about 13 p.e./MeV through a factor 14 increase from WbLS composition changes and a factor 15 increase from higher-reflectivity separators (Onda et al., 25 Jul 2025). In both cases, separator reflectivity, fiber trapping efficiency, and achievable scintillation yield are treated as the dominant engineering constraints.
6. Physics reach, detector trade-offs, and unresolved technical issues
The physics motivation for WbLS is the simultaneous pursuit of calorimetry, low threshold, neutron sensitivity, and directionality in one medium. Concept papers connect this to long-baseline oscillation measurements, solar and reactor neutrinos, diffuse supernova neutrinos, proton decay, neutrinoless double beta decay, and nonproliferation or reactor monitoring (Alonso et al., 2014, Fischer, 2018). In Theia, a 16-kiloton WbLS detector coupled to fast photosensors is proposed to address neutrino mass hierarchy, CP violation, solar and supernova neutrinos, 17, and proton decay, while the ASDC frames WbLS as the core of a 18–19 kiloton underground detector with broad isotope-loading capability (Fischer, 2018, Alonso et al., 2014).
Quantitative performance studies clarify where WbLS sits between pure water and pure scintillator. For 20 WbLS at 21 MeV with LAPPDs and 22 photocathode coverage, the MeV-scale reconstruction study reported 23 in a 24 kt detector and 25 in a 26 kt detector, corresponding to about 27 and 28, respectively (Land et al., 2020). The same study found that a high-coverage 29 kt detector would be capable of better than 30 precision on the CNO neutrino flux with a WbLS target in five years, while pure LS in the same framework reaches the 31 level (Land et al., 2020). For 32, that work quotes 33 yr at 34 CL for ten years of data taking with a Te-loaded target (Land et al., 2020).
For diffuse supernova neutrinos, a 35-loaded Theia configuration was modeled with roughly 36 scintillation photons/MeV, Cherenkov yield of order 37 photons/MeV, and attenuation length 38 m in the blue (Sawatzki et al., 2020). Using Cherenkov/scintillation separation, ring counting, and delayed-decay vetoes, that study finds signal efficiency of more than 39 and estimates that 40 kt41yr is sufficient for a 42 discovery of the DSNB under standard model assumptions (Sawatzki et al., 2020). Reactor-neutrino studies emphasize a different regime: a compact 43-ton WbLS detector at 44 m from the Akkuyu reactor, using 45, 46, or 47 LAB-based WbLS with Gd doping, was simulated to yield about 48 IBD events/day at 49 m for 50 GW51, with energy resolution around 52 at 53 MeV for the 54 WbLS + 55 Gd system (Bat et al., 2021).
Two persistent technical issues recur throughout the literature. The first is chemical and optical stability. Long-term detector operation depends on surfactant choice, oxygen control, purification compatibility, and impurity management. Stable behavior over months is reported in bench and prototype systems, but the 56 L study also documents a gradual 57–58/week decline after an accidental water top-up introduced impurities, confirmed by increased UV–vis absorbance (Zhao et al., 2023). The second is the balancing of light yield against directional information. Raising the scintillator fraction improves detection statistics and low-energy efficiency, but shortens the useful prompt window and can introduce non-local wavelength-shifted contributions that complicate calibration and reconstruction (Bignell et al., 2015, Land et al., 2020). This suggests that the optimal WbLS composition is experiment-specific: low-loading formulations favor Cherenkov purity and long attenuation; higher-loading formulations favor calorimetry and threshold; segmented detectors introduce an additional optimization over separator reflectivity and fiber geometry; and hadronic applications require careful quenching and PSD characterization rather than electron-based calibration alone (Callaghan et al., 2022, Onda et al., 25 Jul 2025, Li et al., 15 Aug 2025).
In that sense, WbLS is best understood not as a single detector medium but as a family of hybrid media whose water fraction, scintillator chemistry, and photodetector ecosystem are co-designed for the target physics program. The existing literature establishes that this family can be realized from liter to 59-ton scale, can support both monolithic and segmented readout schemes, and can preserve measurable Cherenkov information while adding scintillation light over a wide range of formulations (Zhao et al., 2023, Xiang et al., 2024, Andrade et al., 20 Mar 2026).