Andreas Thon

The digital Silicon Photomultiplier — A novel sensor for the detection of scintillation light

We developed a fully digital Silicon Photomultiplier (dSiPM) of 3.8 mm × 3.3 mm in size containing 8188 individual Geiger-mode cells. Each detected photon is directly converted into a digital signal in each of the Geiger-mode cells of the sensor. In addition, the complete trigger logic and a time-to-digital converter are integrated into the sensor. To show the performance of the sensor, LYSO crystals of different sizes were coupled to the sensor. The coincidence timing resolution for 3 mm × 3 mm × 5 mm LYSO crystals using a 22Na source amounts to 153 ps FWHM. The energy resolution at 511 keV was determined to be 10.7 % for 4 mm × 4 mm × 22 mm crystals. It is shown that saturation correction can be done without prior need for sensor calibration. The temperature dependence of the photon detection efficiency was found to be -0.6 %/°C including the temperature variation of the light output of LYSO.

Performance evaluation of a prototype Positron Emission Tomography scanner using Digital Photon Counters (DPC)

We show performance results of a prototype Positron Emission Tomography scanner based on digital SiPMs, or Digital Photon Counters (DPC), developed by the Philips Digital Photon Counting unit. The scalability of the DPC technology is demonstrated by an excellent system coincidence timing resolution of 266 ps FWHM and an energy resolution of 10.7 % FWHM. Even while using 4 mm × 4 mm × 22 mm LYSO crystals, the spatial resolution is close to 2.4 mm. Although not optimized yet, the image homogeneity is 5.8 %. We show that the system performance is maintained even at highest count rates encountered in PET scans.

Solid-state detector stack for ToF-PET/MR

The integration of PET and MR imaging requires a novel type of highly integrated PET detector. To cope with geometric constraints and MR compliance a very compact detector stack was built within the HYPERImage consortium. This allows a four side buttable detector module design with a low dead space in between. The scintillation light coming from a LYSO array is converted in a SiPM sensor tile with a high packing fraction and a high photo detection efficiency to provide sub-ns time-of-flight timing resolution. The analog signals coming from the SiPM elements are digitized close to the sensor to minimize potential crosstalk. A custom mixed-signal ASIC was integrated on a 64 channel sensor stack which is powered and controlled by an FPGA interface board. The complete sensor stack is assembled and characterized to extract the PET relevant parameters, in particular energy, timing and spatial resolution for clinical and pre-clinical PET/MR applicationsB.

Compact SiPM based detector module for time-of-flight PET/MR

We present a compact detector module for γ detection in the PET part of a simultaneous ToF-PET/MR system. The module covers an area of 3:3 cm × 3:3 cm with 64 SiPM based readout channels. It is composed of a stack of three PCBs of identical size: The SiPMs on the topmost PCB are read out by two full-custom ASICs located on a second PCB located underneath. A third PCB at the bottom of the stack contains a local voltage regulator, an FPGA for ASIC control and data processing, and DACs to generate bias voltages for the readout ASICs and the SiPM devices. An LYSO scintillator block is optically coupled to the SiPMs for gamma to light conversion.

Assessment of photodiodes as a light detector for PET scanners

Current PET systems based on pixelated scintillator arrays coupled to photomultiplier tubes suffer from pile-up and electronics dead time at high count rates. With a pixelated readout, i.e. one-to-one coupling of a scintillator crystal to a photo detector, these effects can be strongly reduced. Recent developments of high light output scintillators like LYSO and LaBr3 in combination with very low noise amplifiers based on modern CMOS processes make it possible to use high quantum efficiency blue-sensitive PIN photodiodes as a light detector. To explore the potential of this approach, a model of the signal detection chain was implemented. It comprises the scintillation light pulse and its quantum noise, optical coupling, charge conversion in the diode, noise sources of the integrating amplifier, shaper circuits for the energy and timing channel, and the discriminator for the timing channel. The model is verified using off-the-shelf PIN photodiodes and a dedicated CMOS preamplifier excited by a picosecond laser as well as scintillator pulses. The model predicts that with high light output scintillators, high quantum efficiency photodiodes and optimized preamplifiers, a pixelated PET readout with very good energy resolution and sufficient timing resolution can be realized. To complement the study, an APD-based readout is also considered and the related signal to noise issues are discussed

Assessment of the spatial resolution of PET scanners using a Geant4-based Monte Carlo tool

In this paper, the spatial resolution of PET scanners is assessed by means of a Geant4-based Monte Carlo tool. To obtain a high level of accuracy, all relevant contributions like /spl beta//sup +/ range and gamma noncollinearity, crystal size and material, intercrystal scatter, and scanner geometry are implemented in detail. For a system based on Anger logic, the scintillator light yield, the photon statistics and the PMT characteristics are also taken into account. In the simulation, Monte Carlo and analytical algorithms share a common setup, which allows for the necessary computation speed to generate clinical count statistics without compromising accuracy. In the first part, the major contributions to the spatial resolution are analyzed separately. The impact on the total system resolution is then illustrated by investigating three scanner designs with different gantry sizes, crystal sizes, crystal materials and /spl beta//sup +/ tracers.

Exact modeling of analog pulses for PET detector modules

The quality of PET images depends on the position, energy and time resolution of the gamma detector, which usually consists of scintillation crystals coupled to an array of photomultiplier tubes (PMTs). We have developed a simulation tool which models the conversion of scintillation photons into PMT waveforms, thereby bridging the gap between Monte Carlo simulations of the gamma interaction and pure electronics simulations of the data acquisition system. In our model, we track the scintillation photons individually from the crystal through the light guide to the photocathode of the PMT. The PMT characteristics are treated on a single photoelectron basis, incorporating pulse height distribution and single-photon response as well as spatially dependent quantum efficiency, transit time and transit time spread. From the PMT waveforms, we derive the position, energy and time resolution. The time resolution depends on the scintillator material, the PMT properties and the time-stamping method. By separating the different contributions, the limiting factors in the performance of a detector can be identified, facilitating its optimization. A comparison with measured data is included.

Rate-dependence of the key performance parameters in an Anger logic based PET detector

The image quality of a PET scanner depends strongly on the spatial, energy, and time resolution of the detector. These parameters are usually specified for the case of low count rates. At higher rates, signals from subsequent events can overlap, and this pile-up deteriorates the resolution. In an Anger based detector, the high encoding ratio of scintillator pixels to photo multiplier tubes (PMTs) facilitates pile-up. The probability of pile-up depends on the detector geometry, the scintillator material, and the trigger scheme, and it increases with the count rate. In this paper, the key performance parameters of a whole-body PET detector are investigated as a function of the rate. Experimentally, we use scintillator arrays which can be coupled to PMT arrays with PMTs of different size, but otherwise similar characteristics. First results are presented which show-that signal pile-up leads to only small performance deterioration in typical /sup 18/F oncology studies, but has to be considered when using short lived tracers, e.g. based on /sup 11/C or /sup 15/O.

Impact of the light detection chain on the NEC in a full-body PET scanner

In an ideal PET scanner, the count rate behavior depends on the true sensitivity, the scatter fraction, the singles rate and the coincidence timing window. However, a detailed system analysis shows that due to signal pile-up, the design of the light detection chain (light output, light spread, light detection) also has a strong impact on the resolution and performance of the system. In this paper, the dependence between the light detector size and the count rate behavior is analyzed, and the results are illustrated in terms of the noise equivalent count rate (NEC). Using a suite of system simulation tools, the performance of a PET scanner is modeled down to the level of single optical photons and photo electrons, allowing the count rate behavior to be derived from first principles. Hereby, scintillator materials with different decay times (LYSO, LaBr3) are considered. The results show that for activity levels encountered with short half-life tracers like 11C and 15O, the system NEC can differ by a factor of more than two for systems with identical scintillator geometries, but different light detector sizes