Tl-Halide Scintillators for TOF-PET - Project Summary
The current detector technology in PET requires scintillation that has fast response, excellent timing
resolution, high detection sensitivity, good energy resolution, and last but not least acceptable cost. At present,
most PET systems use crystals of LSO (Lu2SiO5:Ce) or its analog LYSO, which satisfy many of the listed
requirements. But LSO, after years of development, has reached its performance limit, especially for the
scanners with a long axial field of view (AFOV) that are currently being developed. The goal of these scanners
is to increase the geometrical coverage and significantly increase detection sensitivity (by a factor of 30-40),
thereby reducing the scanning times (30-40 times faster) or the patient’s radiation exposure.
However, long AFOV scanners face two main challenges: greater depth-of-interaction (DOI) effects, which
increase blurring and noise; and an increase in the volume required for the constituent crystals, which make
up some 50% of the cost of the entire scanner. The use of shorter crystals can counteract both the DOI effects
and the increased crystal volume (hence cost), but with a major loss of detection efficiency, defeating the
original purpose. Another approach for reducing DOI effects is a double-ended read-out but this increases
both cost and system complexity. Therefore, to achieve viable and affordable long AFOV scanners, a new
scintillation material is required that would provide higher stopping power than LSO, with similar or better timing
properties, and at a lower cost.
These requirements can be met by a scintillator based on TlCl. This host has a comparable density (7.0
g/cm3 vs. 7.4 g/cm3 for LSO) and a much higher effective Z of 77 (vs. 65 to LSO.) When double doped with Be
and I, it exhibits a fast donor-acceptor type luminescence with decay time <10 ns along with ultra-fast Cherenkov
component (vs. 40 ns of LSO.) While the material’s light yield is low, its timing properties are excellent with better
than 300 ps resolution FWHM when paired with LYSO, without any significant optimization. The only property
where the material is deficient is its energy resolution (>15% at 511 keV, due to its low light yield). The key metric
for TlCl is also its very low cost, in terms of raw materials as well as production – estimated 50% lower than LSO.
In this project, we plan to optimize the doping content of TlCl to maximize its scintillation properties and
achieve energy resolution of about 10% at 511 keV and timing resolution of 200 ps. In Phase II we will increase
the volumes of produced material, develop cost reduction schemes, and produce and evaluate PET detection
modules with the same performance goals.