How to Choose a Scintillator

The field now spans roughly twenty practical materials, from well-established inorganic crystals such as NaI(Tl) and BGO to advanced halides like LaBr3(Ce) and CLLBC. Each sits at a different point in a multidimensional tradeoff space. This note maps that space, explains the physics behind each dimension, and then walks through a structured decision process.

Energy resolution is the most visible figure of merit for gamma-ray spectroscopy. It is expressed as the full-width at half-maximum (FWHM) of a photopeak, divided by the peak energy, stated as a percentage. A lower number means the detector can distinguish gamma lines that are closer together in energy.

Resolution is governed by three factors working in combination: the total light yield of the crystal (more photons per keV means better counting statistics), the statistical spread in photon production (proportionality), and the noise of the photodetector. Materials with both high light yield and good proportionality achieve the best resolution. LaBr3(Ce) reaches 3.0% at 662 keV. CeBr3 delivers 4% at the same energy without the La-138 background that LaBr3 carries. SrI2(Eu), though slow, reaches about 3.5%. NaI(Tl) is at approximately 7% at 662 keV, which is adequate for a wide range of counting and identification tasks but insufficient for fine spectroscopy.

The figure below illustrates how FWHM resolution improves as photon yield rises, though the relationship is not perfectly proportional because proportionality effects and photodetector noise also contribute.

2. Density and Effective Atomic Number

Density and effective atomic number (Z) together determine stopping power: the ability of the crystal to absorb incident radiation within a given volume. This matters in two distinct ways. For gamma-ray applications, high density and high Z increase the photoelectric cross-section, making photopeak detection more efficient. For compact or portable instruments, high stopping power means a smaller crystal can achieve the same detection efficiency as a larger, cheaper alternative.

BGO (7.13 g/cm3), PbWO4 (8.28 g/cm3) and LYSO(Ce) (7.2 g/cm3), with CdWO4, are at the top of the density range. They are the materials of choice for PET scanners, high-energy physics calorimeters, and any application where detector volume is constrained. At the lower end, 6Li-glass (2.5 to 2.6 g/cm3) and CaF2(Eu) trade density for sensitivity to specific particle types, thermal neutrons and beta particles respectively.

3. Decay Time

Scintillation decay time sets the maximum rate at which a detector can process separate events. Fast materials allow high count rates without pulse pile-up. They also enable coincidence timing applications, where two detectors must record simultaneous events within a narrow time window.

BaF2 holds the record for conventional inorganic materials, with a fast component near 0.8 ns. YAP(Ce) and LYSO(Ce) both deliver decay times in the 20 to 50 ns range, fast enough for PET time-of-flight and physics coincidence work. CsI(undoped) is faster still. NaI(Tl) at 230 ns is adequate for most counting applications but limits throughput at high source activities. CsI(Tl) has a 600 ns principal component and a 3.4 us slow component, so its pulses need a long integration gate, and SrI2(Eu) at 1 to 5 us is the slowest of the common choices; neither suits count rates above a few tens of thousands per second.

4. Hygroscopicity

Hygroscopic crystals absorb moisture from ambient air, causing surface clouding and degraded optical coupling. This has two practical consequences. First, the crystal must be hermetically sealed, adding cost and mechanical complexity. Second, the seal must remain intact throughout the service life, which is a maintenance and reliability consideration for field-deployed instruments.

NaI(Tl), LaBr3(Ce), LBC, CeBr3, CLYC, CLLBC and SrI2(Eu) are hygroscopic and require sealed housings; CsI(Tl) is slightly hygroscopic and tolerates a lightly protected housing. BGO, GAGG(Ce), LYSO(Ce), PbWO4, BaF2, YAP(Ce), YAG(Ce) and 6Li-glass are not hygroscopic. For handheld instruments, outdoor monitoring stations, or any system that may experience condensation, non-hygroscopic materials simplify packaging significantly. When resolution demands the best hygroscopic materials, the cost of a well-designed hermetic housing is usually justified.

5. Cost and Availability

Crystal cost scales with raw material rarity, growth difficulty, and yield. NaI(Tl) and BGO are the most established and economical choices at scale. CsI(Tl) and CsI(Na) fall in a mid-cost band. LaBr3(Ce) and SrI2(Eu) carry a significant premium due to the difficulty of growing large, uniform crystals. CLYC and CLLBC are specialist materials with limited commercial production capacity; lead times and pricing reflect that. For high-volume applications such as medical imaging arrays or portal monitor panels, cost and availability often override resolution as the deciding factor.

These instruments require the best possible gamma resolution in a field-portable package. The detector must identify isotopes by their gamma-line signatures, which means resolving peaks separated by as little as 30 to 60 keV in the 200 to 1400 keV range. LaBr3(Ce) is the established premium choice. CeBr3 is preferred when the La-138 background at 1.46 MeV would interfere with environmental monitoring. CLLBC adds simultaneous neutron sensitivity. All three require hermetic sealed housings. If cost must be contained, NaI(Tl) remains adequate for coarse isotope discrimination.

Environmental and Low-Background Monitoring

Long-duration monitoring at low count rates places the emphasis on intrinsic background purity, resolution, and stability. CeBr3 leads here because it combines high light yield, good resolution, and the absence of the La-138 self-activity that burdens LaBr3 crystals. Large-volume NaI(Tl) detectors remain the workhorse for regulatory monitoring stations where the crystal geometry matters more than ultimate resolution.

Medical Imaging (PET and SPECT)

PET requires fast, dense crystals with good light yield. LYSO(Ce) dominates modern commercial PET because of its combination of high density (7.20 g/cm3), 50 ns decay, and non-hygroscopic nature. GAGG(Ce) is gaining adoption in SiPM-based small-animal and compact PET systems. BGO remains relevant for cost-sensitive large-panel designs. SPECT and CT applications often favor CsI(Tl) for its emission wavelength compatibility with silicon photodiodes and its structural robustness.

High-Energy Physics and Fast Timing

Calorimeter applications at particle accelerators prioritize stopping power above all else. PbWO4 is the established choice for electromagnetic calorimeters at the highest energies. Where some light yield can be traded for reduced hermeticity concerns, LYSO or GAGG are preferred. Timing applications, such as positron lifetime spectroscopy, demand BaF2 or YAP(Ce).

Geophysical and Harsh-Environment Logging

Downhole instruments operate at elevated temperatures and cannot be easily serviced. NaI(Tl) remains the most used logging crystal, with its hermetic housing engineered for the tool. BGO is used where density matters at moderate temperature, since its datasheet range ends at 100 C, and it is free of afterglow that would corrupt count-rate measurements in high-flux boreholes. CsI(Tl) (rated to 125 C) and GAGG(Ce) are strong alternatives; for geophysical applications requiring moderate resolution without hygroscopic complexity, GAGG is a current design choice.

General Laboratory Counting and Health Physics

When the application does not impose extreme constraints on resolution, density, or timing, NaI(Tl) is the clear starting point. It delivers the highest light yield of the established materials, an emission spectrum well matched to standard bialkali PMTs, and a decades-long supply chain. The hygroscopicity penalty is managed by standard aluminum-can housings, which are mature, inexpensive, and widely available.

PMT bialkali sensitivity peaks near 380 to 420 nm. Materials that emit in this window (NaI(Tl), LaBr3, LBC, CeBr3, CLYC, CLLBC, LYSO, 6Li-glass) couple efficiently to standard PMTs. BaF2 fast-component emission at 220 nm requires quartz-window PMTs with UV-extended photocathodes, which are more expensive and less common.

SiPM photon detection efficiency peaks near 420 to 450 nm and stays usable through the green, while silicon photodiodes peak further toward the red. CsI(Tl) at 550 nm and GAGG at 520 nm are well matched to silicon sensors, and LYSO and CLLBC at 420 nm to SiPMs. CeBr3 and LaBr3 at 370 nm work with SiPMs at reduced efficiency. YAP(Ce) at 350 nm is marginal for silicon; PMT readout is preferred.

Refractive index mismatch at the crystal-window interface causes internal reflection losses. BGO at 2.15 and PbWO4 at 2.16 have high refractive indices; coupling with optical grease helps, but not all index-matching fluids are compatible with all window materials.

For detailed guidance on PMT versus SiPM readout tradeoffs across the ScintIQ lineup, see the companion note SiPM vs. PMT Readout for Scintillation Detectors.

Contact the ScintIQ team directly:

• Email: info@berkeleynucleonics.com
• Phone: 800-234-7858