GEANT4 simulation for modelling the optics of LaBr scintillation imagers

GEANT4 SIMULATION FOR MODELLING THE OPTICS OF LABR3(CE) SCINTILLATION IMAGERS S. Lo Meo , G. Baldazzi, P. Bennati, D. Bollini, V. O. Cencelli, M. N. Cinti, N. Lanconelli, G. Moschini, F. L. Navarria, R. Pani, R. Pellegrini, A. Perrotta and F. Vittorini Abstract - The recent development of the LaBr3(Ce) crystals makes very attractive their application as a system of gamma imaging for SPET applications mainly due to their excellent scintillation properties. In this work, we use Monte Carlo simulations, in order to model the optical behavior of three continuous crystal configurations. Our goal is a better understanding of the intrinsic properties of a gamma camera based on LaBr3(Ce) crystals, coupled to a position sensitive photomultiplier tube. To this aim, the spatial and energy resolutions obtainable from optimum photodetection conditions are investigated. Index Terms— Monte Carlo, Geant4, LaBr3(Ce) crystal, Unified Optical Model, Penelope, Multi Anode PMT I. INTRODUCTION M ONTECARLO simulation techniques are becoming very common in the medical imaging community. Several topics are addressed by Monte Carlo simulations in the Nuclear Medicine field. Among these, it is worth remarking the optimization of imaging systems design (including detector, collimator, and shield design). GEANT4 [1,2] is an object oriented toolkit for simulation of current and next generation High-Energy Physics detectors. It is also a showcase example of technology transfer from particle physics to other fields such as medical science. GEANT4 permits an accurate modeling of radiation sources detectors and human body with easy configuration and friendly interface. At the same time GEANT4 makes it possible to follow with great precision the interactions within the different media. The recent introductions of the LaBr3(Ce) crystals makes very attractive their application as a system of gamma imaging for SPET applications. This is mainly due to their excellent scintillation properties, which offer the potential to replace the most widespread scintillation crystal: the NaI(Tl). In continuous shape, these crystals are able to provide spatial resolution values comparable to scintillation arrays or better. In this work, we use Monte Carlo simulations, in order to model the optical behavior of three continuous crystal configurations. Our goal is a better understanding of the intrinsic properties of a gamma camera based on LaBr3(Ce) crystals, coupled to a position sensitive photomultiplier tube. To this aim, the spatial and energy resolutions obtainable from optimum photodetection conditions are investigated. II. SIMULATION SETUP For the modelling of the electromagnetic interactions, the “Penelope” model [3] of GEANT4 is used. Such a model is best suited for energies ranging from a few hundred eV to about 1 GeV. Atomic relaxation following photoelectric effect, Compton scattering, ionization interactions, Rayleigh scattering, fluorescence photons or Auger electrons can be simulated; some of these processes are not included in the standard GEANT4 electromagnetic model. To improve the tracking of the electrons, a few parameters [4] have been tuned, relative to the numerical stability of the results, to the control of step and stopping power, and to the multiple scattering. Furthermore, GEANT4 allows also the simulation of the transport and boundary effects for the optical photons generated by the scintillating crystal. In this way, the whole process can be thoroughly simulated within that package. The simulation starts from the radioactive decay in the source of 99m Tc and halts when all optical photons reach the photomultiplier. The camera consists on a Multi Anodes-PMT (MA-PMT) Hamamatsu Flat Panel H8500 [5], coupled to a 50×50×4 mm3 continuous LaBr3(Ce) crystal [6][7] and a 3 mm glass window. The crystal (figure 1a) is surrounded by a thin layer of Aluminum. In the front crystal surface is covered by a very small layer of Teflon, acting as a Lambertian reflector. Figure 1b shows the charge distribution on 8 x 8 anode plane by a 140 energy keV photon absorbed into the LaBr3(Ce) crystal. Figure 1: (a) Geant4 simulation set up, (b) Photoelectrons distribution on 8 x 8 anode plane The boundary processes of all crystal surfaces follow the rules of the UNIFIED model, developed for the DETECT project [8]. The MA-PMT surface is modelled as a polished glass window, assuming an experimentally derived value for the photocathode quantum efficiency of the PMT itself. The Ground Air Gap Polished <Nphe> 1603 ± 1 1137 ± 1 1047 ± 1 ER % 5.9 ± 0.1 7.0 ± 0.1 7.2 ± 0.1 SR (mm) 0.78 ± 0.04 0.7 ± 0.1 0.77 ± 0.02 Table 1: Average Number of Photoelectrons, Spatial and Energy Resolution, obtained in function of the roughness of the crystal surfaces AIRGAP N phe 375 250 125 0 2 3 4 5 6 position (anodic unit) 7 2500 2000 1500 1000 500 0 250 300 350 400 Image Pixel Figure 3: Spatial Resolution comparison between Monte Carlo and experimental results. The experimental results are obtained with 2 mm step scanning 99mTc 0.4 mm collimated source. In conclusion the Monte Carlo carried out important informations about imaging potentials of LaBr3(Ce). Energy resolutions at 140 keV agree with the expected ones if we take into account the intrinsic energy resolution of 4% of LaBr3(Ce). Unfortunately from experimental data H8500 limits the energy resolution down to 9%. It could influence also the imaging spatial resolution considering that centroid method is utilized to determine position. Monte Carlo predictions allow to conclude that LaBr3(Ce) crystals with a ground treatment of surfaces and a PMT Photocathode with higher quantum efficiency could open the way to submillimeter spatial resolution imaging with high detection efficiency and energy resolution. [1] Polished 1 Exp. Results MC Ground MC Polished REFERENCES Ground 500 3000 Counts optical properties of the materials involved in the simulations (refraction index, absorption and scattering lengths) are gathered from literature. A scintillation light yield equal to 70000 photons/MeV is assumed for LaBr3(Ce). The scintillation photons are generated as a pure Poisson process, i.e., the intrinsic resolution of the crystal is not considered. Three types of optical configurations, as a function of the roughness of the crystal surfaces, are simulated: “Ground”, “Polished” and “Air Gap”. “Ground” and “Polished” refer to the status of the lateral surfaces of the crystals, while “Air gap” refers to an assembly consisting of a thin air interface between the crystal and the photodetector. The values of spatial, energy resolution and the average number of photoelectrons gained by anode are summarized in table 1. Figure 2 shows the charge projection along the X-direction as simulated for the three different crystal configurations and the same source irradiation. They are used to calculate position along Anger logic. The Monte Carlo confirm the increase in performance for the ground crystal configuration both energy and spatial resolution. By analysing the distribution of photoelectron we are able to appreciate the increase in the collection in the crystal’s centre but no so many differences in the boundary. 8 Figure 2: Average spread of Number of Photoelectrons in the anodic plane for different crystal configuration. The two dimensional distribution is projected along the X-direction (one anode is six millimeter wide). In figure 3 the spatial resolution analysis is shown. The simulation results are compared with experimental data. A compact gamma camera was scanned with 0.4 mm collimated 99m Tc spot source 2mm step. The simulation indicates the possibility of increasing in spatial resolution of about 30%. S. Agostinelli at al. “Geant4 a simulation toolkit” NIM section A, volume 506, Issue 3, 1/7/2003, Pages 250-303 [2] J. Allison at al. “Geant4 developments and applications”, Nuclear Science, IEEE Transactions on”, Volume 53, Issue 1, art 2, 1/2/2006 Pages 270 – 278 [3] F. Salvat, J. M. Fernández-Varea, E. Acosta, and J. Sempau, “PENELOPE: A code system for MonteCarlo simulation of electron and photon transport,” Workshop Proceedings, OECD Nuclear Energy Agency, Issy-les-Moulineaux, 2001. [4] S. Elles, V Ivanchenko, M. Marie, L. Urban, “Geant4 and Fano cavity test: where are we?” Monte Carlo Techniques in Radiotherapy delivery and verification: Third McGill International Workshop" Montreal, May 29-June1, 2007 (to be published in "Journal of Physics: Conference Series", IOP Publishing Limited). [5] http://www.hamamatsu.com [6] http://www.bicron.com/ [7] http://www.detectors.saint-gobain.com/ [8] A. Levin and C. Moisan, “A more physical approach to model the surface treatment of scintillation counters and its implementation into DETECT”, in Nuclear Science Symposium Conference Record, 1996 IEEE, vol. 2, 1996, pp. 702–706. [9] Pani, R.; Cinti, M.N.; Pellegrini, R.; et al., “LaBr3(Ce) scintillation camera” Nuclear Science Symposium Conference Record 2005 IEEE Volume 4, 2005 Page(s):2061 – 2065. [10] Pani, R.; Pellegrini, R.; Cinti, M.N.; et al., “Factors Affecting Hamamatsu H8500 Flat Panel PMT Calibration for Gamma Ray Imaging” Nuclear Science, IEEE Transactions on Volume 54, Issue 3, June 2007 Page(s): 438 – 443.