Development of innovative detectors for Positron Emission Tomography

Positron emission tomography (PET) is a powerful nuclear imaging technique used widely nowadays in oncology, cardiology and neuropsychiatry.

The PET technology consists in injecting the patient with a radioactive tracer, of interest to probe a biochemical process. The decay of the tracer emits a positron which annihilate with an electron. As a result of the annihilation, two photons with energy 511 keV are emitted back-to-back and registered by the dedicated detectors. The line-of-response connects the two points where photons are detected and allow to reconstruct the tracer distribution when large statistics events have been accumulated.

The CaLIPSO group works on two projects :

  1. The first project, CaLIPSO, is a development of the high spatial resolution PET for the brain and pre-clinical studies.
  2. The PECHE project is a development of the fast light detection technique using Cherenkov effects to boost the time-of-flight (TOF) performance of PET scanners.

Motivations for CaLIPSO

The CaLIPSO project (French acronym for Calorimètre Liquide Ionisation Position Scintillation Organométallique) is based on the recent developments of the TriMethyl Bismuth (TMBi). This innovative liquid allows a very efficient and accurate detection of positron annihilation. Once fully developed and integrated with an acquisition system, it may become a key element of PET imagers. In particular, the detector performances would allow efficient 1 mm3 resolution PET imaging of the whole human brain, thus being an excellent complement of magnetic resonance imagers for diagnosis and research on neurodegenerative diseases. The development of such detector is a long term effort pursued at IRFU since 2009, partially funded by the Neuropôle de Recherche Francilien (NeRF), LabEx P2IO, and CEA interdisciplinary program TechnoSanté. This innovation is patent protected. The CaLIPSO detector uses the TriMethyl Bismuth to efficiently convert photons of energy less than 1 MeV, through the photoelectric effect. The photoelectric electron is relativistic in the liquid TMBi. Thus it produces a quasi-instantaneous flash of few tens of Cherenkov photons, as well as free charges in the liquid. Light produced is detected by an efficient photodetector and charges released drift along a strong electric field, pass through a Frisch grid, and are collected by a pixelated charge detector.


Scientific and technological challenges

The CaLIPSO project is very ambitious. At the project start, little was known about liquid TMBi for particle detection. Key science and technological issues were listed as:

  • TMBi physical properties: namely light production yield, light absorption length at optical and UV wavelength, charge mobility in TMBi. We measured these properties using tailored devices designed for handling our mildly-pyrophoric liquid.
  • The low light production efficiency does require careful optimization of the optical couplings. The first optical demonstrator, along with its detailed Monte Carlo simulation, allowed us to quantify the issue. The second optical prototype allowed us to demonstrate a very good efficiency of detection of 511 keV photons, especially, when they are converted to the electron via the photo-electric effect.
  • The electronic readout density: Integrating an electronic density approaching one preamplifier per mm2, has already been achieved in the laboratory (ASIC and 3D electronics), in the context of a spaceborne experiment. We have chosen to limit the resolution of our demonstrator to 2 x 2 mm2, in the charge collection plane, in order to avoid having to use 3D electronics and to focus in the short term our efforts on the detector development.
  • The TMBi must be ultra-purified to allow the charge drift and the measurement of gamma energies. We organized the detector cleaning and mounting work at the clean rooms. The materials, from which the detector is made, shall be carefully chosen to be compatible with ultra-high vacuum technologies.

Ultrapurification of TMBi

We need to extract electronegative molecules that trap free electrons in order to achieve free electron lifetime larger than 10 ?s. This will ensure effective charge drift and collection of the ionization signal. We use molecular sieves for purification because this technology has achieved an impurity level lower than 0.1 ppb O2 equivalent on TetraMethyl Silane in the past. We monitored TMBi purity level by measuring macroscopic ionization current yields, and measured significant increase of the ionization yields with purification. Big efforts are starting to improve the efficiency of this molecular sieve and to upgrade all the systems: vacuum system, distillation system, and purification system. And keep them clean.


Detection of the Cherenkov Light

To verify the efficiency of the Cherenkov light detection, we constructed and studied the optical prototype for the CaLIPSO project, which contains nine separate optical cells. According to Geant4 simulation, we detect in average 2.3 photo-electrons on the photocathode and register the converted gamma quanta with the 90% efficiency, if the 551 keV gamma are converted in TMBi via the photoelectric effect. The optical prototype has been tested using the radioactive positron source 22Na and a YAP:Ce detector, mounted in the back-to-back configuration on the test bench. The results of the test measurements show that the total efficiency of the prototype is about 35%, in a reasonable agreement with the simulation. The time resolution of the constructed prototype is limited by the time resolution of chosen PMT. We plan to improve it in the next prototype by using the micro-channel-plate PMT (MCP-PMT) and to achieve the time resolution of the order of 100 ps (FHWM).


Study the time-of-flight (TOF) technique limits with Cherenkov light.

In the PECHE project we are using the PbF2 crystals as a Cherenkov radiator attached to the MCP-PMT. This crystal produces no scintillation light, but only Cherenkov radiation. It is very dense (7.8 g/cm3 ) and has one of the highest photoelectric fraction, 46%. Due to these facts, it is possible to

create an efficient gamma detectors with a very small thickness of the order of 10 mm and hence minimize the dispersion of the photon trajectories. The ability to detect 511 keV photons in PbF2 crystal has been demonstrated in the literature, but detection efficiency is found to be is less than 10% , due to the limited number of photons produced by the Cherenkov effect. The low efficiency is a major limiting factor for making very fast TOF PET devices. To overtake this limitation we are improving the optical interface and using the customized MCP-PMT with sapphire window, (XP85012 PMT by Photonis) with a transit time spread of around 35 ps (RMS). Readout of MCP-PMT is realized, using ADC SAMPIC. This 16-channel module, developed jointly by IRFU-CEA and LAL-IN2P3, registers the digitized waveform shape with the sampling speed up to 10 GHz together with a precise time stamp of the signal arrival. It provides a time resolution better than 5 ps (RMS) and hence its contribution to the overall detector resolution is negligible. Summing up all contributions together, we expect to reach a time resolution of about 100 ps (FWHM) with an efficiency of the order of 25% for the conventional optical interface and 50% for the optimized one.

#3937 - Last update : 03/15 2017


Retour en haut