The FastScan whole body counter: efficiency as a function of BOMAB phantom size and energy modelled by MCNP

Transcription

1 The FastScan whole body counter: efficiency as a function of BOMAB phantom size and energy modelled by MCNP Gary H. Kramer and Jeannie Fung Human Monitoring Laboratory, Radiation Protection Bureau, 775 Brookfield Road, Ottawa, Ontario K1A 1C1, Canada (gary_h_kramer@hc-sc.gc.ca, Abstract. The Canberra FastScan whole body counter has been investigated using Monte Carlo simulations. In this manner, the counter has been virtually calibrated using detectors and phantoms of different sizes. The results have been compared with the five FastScan counters currently deployed in different locations throughout Canada. The results show that when a person is measured in a FastScan, and their body size is different from that of the Reference Man calibration phantom, the error introduced into the activity estimate is small with a maximum value of about 10%, either as an over- or under-estimate. If the two smaller phantoms are ignored (the four and ten year old) then the size dependency remains at about 10%, but only as an overestimate. The results also shows that the current commercial version of the FastScan counter has the best counting efficiency of the three system simulated up to 1173 kev. 1.0 Introduction There are five FastScan whole body counters in Canada. Their performance characteristics were reported elsewhere [1] and it was apparent that these instruments seemed to suffer less size dependency than other types of whole body counters (i.e., chair, bed, scanning bed). To further investigate this phenomenon, the FastScan whole body counter was modelled and its counting efficiency simulated for a variety of BOMAB phantom sizes using MCNP4A. The use of Monte Carlo simulations to establish calibration data has several advantages as the purchase of a BOMAB phantom family can be both expensive and its use time consuming. The establishment of all the calibration curves for a FastScan whole body counter would require the HML to: purchase radioactive standards that cover the required energy range (100-2,000 kev), fill each phantom with a single radionuclide, visit one of the FastScan sites for a long enough period to perform all the calibrations (this would be problematic as they are in routine and continual use at the Canadian Power Generating Stations), develop the calibration sets. The work would take several weeks to cover all the phantom sizes and photon energies and would create many litres of radioactive waste. The alternative, Monte Carlo simulations, suffers from none of the above disadvantages. 2.0 Methods and Materials 2.1 Modelling of the BOMAB phantoms: The HML uses BOMAB phantoms that simulate: A Reference four year old child (P4), a Reference ten year old child (P10), a five-percentile male (PM5), a Reference Female (PF), Reference Male (PM), and a 95-percentile male (PM95). The Reference phantoms (P4, P10, PF, and PM) were originally developed from data contained in Reference Man [2]. The other phantoms (PM5 and PM95) were designed using selected anthropometry data [3] (Nutrition Canada 1980). The phantoms were measured in the HML to determine the height, semi-major, and semi-minor axis of each phantom. The thickness of the high density polyethylene wall was 0.25 cm except at the filling cap end where it was 1.5 cm. This data was used to construct a set of virtual BOMAB phantoms. 2.2 The FastScan Counter: The FastScan uses two NaI (Tl) detectors, configured in a linear array on a common vertical axis. Originally the detectors were each 10.2 cm x 10.2 cm x 40.6 cm with a single photo-multiplier tube at the end; a later design changed the detector to 7.6 cm x 12.7 cm x 40.6 cm. 1

2 The subject stands inside the shield facing the detectors. The detectors are shielded by 10 cm of low background steel and the whole unit weighs 4,500 kg. The counter occupies floor space of only 1.3 m x 1 m, and is less than 2.1 m in height. The vertical detector placement and interior shield dimensions were chosen based upon anthropometric data of a working population [4] of both adult males and females. The locations of the internal organs were obtained from the Synder-MIRD Mathematical Phantom [5]. Canberra supplied some more detailed diagrams of detector locations, and shield composition so that the FastScan could be modelled. This information was used to model the detectors, local shielding and support structure. Three sizes of detectors were modelled: 10.2 cm x 10.2 cm x 40.6 cm, 7.6 cm x 12.7 cm x 40.6 cm., and 10.2 cm x 12.7 cm x 40.6 cm (width, depth and height respectively). The photomultiplier tubes were not modelled. For ease of identification in the discussion below, the three counting systems will be labelled as System 1, System 2, and System 3, respectively. 2.3 The Simulations: The following energies were simulated: 126, 280, 364, 468, 662, 834, 1173, 1332, 1460, 1836 and 2754 kev. Many of these photon energies were chosen to represent radionuclides frequently used in calibrations: 57 Co, 131 I, 137 Cs, 60 Co, 54 Mn, 40 K, 88 Y, 24 Na. The remaining energies are interpolations to fill in areas so that efficiency versus energy curves can be constructed. Each photon energy was run independently and were mono-energetic for each phantom size. The photons that interacted with the upper and lower detectors and deposited their full energy were tallied so that a detector efficiency was obtained for the upper detector and lower detector. The array (upper and lower combined) efficiency was obtained simply by summing the two detectors. The Monte Carlo code used for the simulations has been described in detail elsewhere [6]. The authors of MCNP consider that a relative error value of suggests that the tally result is questionable [7]. Tally results for which the relative error is above 0.2 are not likely to be meaningful, but are generally reliable for a relative error less than 0.1. The number of photons used in the simulations was 10 7 so that the relative error varied from to for the lower detector and to for the upper detector. 2.4 Benchmarking: The five Canadian facilities that own and operate a FastScan whole body counter supplied their efficiency data to the HML in the form of efficiency equations. The form of the efficiency equations in use are: Pt Lepreau: 2 3 ( a2+ a3. x+ a4. x + a5. x ) Eff= e and x = Ln( a1 / energy) Other stations: ( b1+ b2. y+ b3. y + b4 y + b5 y + b6 y ) Eff= e and y = Ln( energy) where a1' to a5' and b1' to b6' are regression coefficients determined from the Canberra software. 3.0 Results and Discussion 3.1 Counting Efficiency: The counting efficiencies of the virtual BOMAB phantoms as a function of photon energy are shown in Tables 1-9. The tables give efficiencies for the upper detector, lower detector, and detector array for each of the simulated detector sizes (Systems 1-3). 2

3 Table 1: Counting efficiency (count/photon) of the upper detector (System 3) of a FastScan WBC using different sized virtual BOMAB phantoms as a function of photon energy. P4 = four year old phantom, P10 = 10 year old phantom, PM5 = five percentile phantom, PF = female phantom, PM = male phantom, PM95 = ninety fifth percentile phantom. P x x x x x x 10-4 P x x x x x x 10-3 PM x x x x x x 10-3 PF 2.21 x x x x x x 10-3 PM 2.44 x x x x x x 10-3 PM x x x x x x 10-3 P x x x x x 10-4 P x x x x x 10-3 PM x x x x x 10-3 PF 2.09 x x x x x 10-3 PM 2.36 x x x x x 10-3 PM x x x x x 10-3 Table 2: Counting efficiency (count/photon) of the lower detector (System 3) of a FastScan WBC using different sized virtual BOMAB phantoms as a function of photon energy. P x x x x x x 10-3 P x x x x x x 10-3 PM x x x x x x 10-3 PF 4.04 x x x x x x 10-3 PM 3.49 x x x x x x 10-3 PM x x x x x x 10-3 P x x x x x 10-3 P x x x x x 10-3 PM x x x x x 10-3 PF 3.81 x x x x x 10-3 PM 3.44 x x x x x 10-3 PM x x x x x 10-3 Table 3: Counting efficiency (count/photon) of the detector array (System 3) of a FastScan WBC using different sized virtual BOMAB phantoms as a function of photon energy. 126 kev 280 kev 364 kev 468 kev 662 kev 834 kev P x x x x x x 10-3 P x x x x x x 10-3 PM x x x x x x 10-3 PF 6.25 x x x x x x 10-3 PM 5.94 x x x x x x 10-3 PM x x x x x x 10-3 P x x x x x 10-3 P x x x x x 10-3 PM x x x x x 10-3 PF 5.90 x x x x x 10-3 PM 5.80 x x x x x 10-3 PM x x x x x

4 Table 4: Counting efficiency (count/photon) of the upper detector (System 2) of a FastScan WBC using different sized virtual BOMAB phantoms as a function of photon energy. P x x x x x x 10-4 P x x x x x x 10-3 PM x x x x x x 10-3 PF 2.51 x x x x x x 10-3 PM 2.84 x x x x x x 10-3 PM x x x x x x 10-3 P x x x x x 10-4 P x x x x x 10-4 PM x x x x x 10-3 PF 1.94 x x x x x 10-3 PM 2.24 x x x x x 10-3 PM x x x x x 10-3 Table 5: Counting efficiency (count/photon) of the lower detector (System 2) of a FastScan WBC using different sized virtual BOMAB phantoms as a function of photon energy. P x x x x x x 10-3 P x x x x x x 10-3 PM x x x x x x 10-3 PF 4.93 x x x x x x 10-3 PM 4.25 x x x x x x 10-3 PM x x x x x x 10-3 P x x x x x 10-3 P x x x x x 10-3 PM x x x x x 10-3 PF 3.92 x x x x x 10-3 PM 3.54 x x x x x 10-3 PM x x x x x 10-3 Table 6: Counting efficiency (count/photon) of the detector array (System 2) of a FastScan WBC using different sized virtual BOMAB phantoms as a function of photon energy. 126 kev 280 kev 364 kev 468 kev 662 kev 834 kev P x x x x x x 10-3 P x x x x x x 10-3 PM x x x x x x 10-3 PF 7.44 x x x x x x 10-3 PM 7.09 x x x x x x 10-3 PM x x x x x x 10-3 P x x x x x 10-3 P x x x x x 10-3 PM x x x x x 10-3 PF 5.87 x x x x x 10-3 PM 5.78 x x x x x 10-3 PM x x x x x

5 Table 7: Counting efficiency (count/photon) of the upper detector (System 1) of a FastScan WBC using different sized virtual BOMAB phantoms as a function of photon energy. P x x x x x x 10-4 P x x x x x x 10-3 PM x x x x x x 10-3 PF 2.13 x x x x x x 10-3 PM 2.38 x x x x x x 10-3 PM x x x x x x 10-3 P x x x x x 10-4 P x x x x x 10-4 PM x x x x x 10-3 PF 1.84 x x x x x 10-3 PM 2.11 x x x x x 10-3 PM x x x x x 10-3 Table 8: Counting efficiency (count/photon) of the lower detector (System 1) of a FastScan WBC using different sized virtual BOMAB phantoms as a function of photon energy. P x x x x x x 10-3 P x x x x x x 10-3 PM x x x x x x 10-3 PF 4.13 x x x x x x 10-3 PM 3.57 x x x x x x 10-3 PM x x x x x x 10-3 P x x x x x 10-3 P x x x x x 10-3 PM x x x x x 10-3 PF 3.60 x x x x x 10-3 PM 3.26 x x x x x 10-3 PM x x x x x 10-3 Table 9: Counting efficiency (count/photon) of the detector array (System 1) of a FastScan WBC using different sized virtual BOMAB phantoms as a function of photon energy. 126 kev 280 kev 364 kev 468 kev 662 kev 834 kev P x x x x x x 10-3 P x x x x x x 10-3 PM x x x x x x 10-3 PF 6.25 x x x x x x 10-3 PM 5.95 x x x x x x 10-3 PM x x x x x x 10-3 P x x x x x 10-3 P x x x x x 10-3 PM x x x x x 10-3 PF 5.44 x x x x x 10-3 PM 5.36 x x x x x 10-3 PM x x x x x

6 Eff (cnt/photon) The differences between the three crystal sizes are shown in Fig. 1 using the PM phantom. Other phantoms gave similar results to Fig. 1. At low energies Systems 1 and 3 perform identically. As the energy rises the efficiency of System 3 rises above that of System 1 due to the greater thickness of the detector crystal. System 2 has the highest efficiency at low energies due to the increased width of the crystal compared with the other systems. Its counting efficiency falls below that of System 3 at about 1200 kev but remains above System 1 over the range simulated (126 kev kev). This indicates 0.01 System 1 System 2 System Fig 1: Efficiency of the PM BOMAB phantom as a function of photon energy for three detector crystal sizes. that of the three systems System 2 is the best choice in terms of counting efficiency for expected nuclides. The upper and lower detector are both geometry dependent as can be seen from Tables 1-9. The smallest phantom, P4, has the highest counting efficiency when measured by the lower detector whereas the largest phantom, PM95, has similar efficiencies in both upper and lower detectors. Fortunately, when the two detectors are summed to give a two-detector array the counting efficiencies are almost independent of phantom size. The biggest difference would be between the P4 and PM95 phantoms and for the lower, upper and array these differences at 126 kev are a factor of 0.54, 3.4, and 0.90, respectively. The size dependency of System 2 is illustrated in Fig. 2 where it can be seen that the counting efficiency of the smaller phantoms drops below that of the bigger phantoms as the energy (kev)

7 Eff (cnt/photon) 0.01 P4 P10 PM PF PM PM (kev) Fig. 2: Counting Efficiency of all the phantoms in System 2. rises. For example, the P4 phantom becomes less efficient than the PM95 phantom at energies > 468 kev. This is likely due to the lessening importance of self-attenuation of the emitted photons by the stable components of the phantoms. 3.2 Minimum Detectable Activity: Background is proportional to detector volume [8]. The larger the volume the larger the background. The volumes of Systems 1-3 are: 4182 cm 3, 3880 cm 3, 5207 cm 3, respectively. System 2, therefore, is the best choice as the background would be expected to be the lowest and the efficiency is the highest. 3.3 Benchmarking: The regression parameters that describe the FastScan efficiency equations are given in Table 10. The efficiency equations are also shown in Fig. 3 where it can be seen that four of the five agree quite well while the fifth exhibits lower efficiencies. The latter whole body counter is of the first generation and has smaller detectors than the other four. The disagreement of the counting efficiencies of the remaining four counting systems in the lower energy range is due to the different facilities choice of fitting algorithm - see above and Table 10.The efficiency equations were developed using the Canberra Transfer Phantom (CTF) and standard sources. The CTF can simulate a lung, GI, whole body, or thyroid geometry and is easy to use because it requires no assembly and uses a single source in a 20 ml liquid scintillation vial for all calibration geometries. The manufacturer states that this phantom is not appropriate for low energy measurements, or geometries where the detector is behind the subject, or for counters other than Canberra linear geometry counters. 7

8 Eff (cnt/photon) Table 10: Regression parameters for the FastScan efficiency equations. Facility Regression parameter a1 a2 a3 a4 a5 Pt Lepreau b1 b2 b3 b4 b5 b6 Pickering Darlington Bruce A Bruce B Pickering (new detectors) Pt Lepreau Pickering Darlington Bruce A Bruce B Pick. New Det (kev) Fig. 3: Plot of the five Canadian FastScan whole body counters. The efficiency response with this phantom has been designed to replicate the ANSI N13.30 phantoms (Livermore Lung, BOMAB total body and ANSI-N-44.3 Thyroid) for the Canberra counters; however, the simulations have been performed with BOMAB phantoms. Fig. 4 shows the Reference Man BOMAB counting efficiencies compared to the Hydro One FastScan counters. The agreement is excellent considering the scatter in the observed data. The Pt Lepreau data has been omitted. 8

9 CE(cnt/photon) 3.4 Size Dependency: Over the energies and phantom sizes simulated the size dependency of System 2 is found to be between 0.92 and 1.10 of the counting efficiency of the PM phantom (RefMan). Systems 1 and 3 perform similarly. This means that when a person is measured in a FastScan, and their body size is different from that of the Reference Man calibration phantom, the error introduced into the activity estimate is small with a maximum value of about 10%, either as an over- or underestimate. If the two smaller phantoms are ignored (the four and ten year old) then the size dependency remains at about 10%, but only as an overestimate. Looking more closely at the data in Table 6 one sees that the counting efficiency of the PM phantom is 0.01 System 2 Pickering Darlington Bruce B Bruce A (kev) Fig. 4: Comparison of measured and predicted efficiency curves for the FastScan Whole Body Counter using the PM phantom (RefMan). the lowest of the series PM5, PF, PM, and PM95 (except at 126 kev) indicating that any deviation from the Reference Man size will lead to an over-estimate of the body burden and, therefore, an overestimate of the dose estimate. This finding suggests that the FastScan should be calibrated using a phantom that is representative of the median (height, weight) of the population being measured to avoid biasing the whole body counting results. 9

10 4.0 Conclusions The response of the FastScan has been modelled with three detector sizes, including the one currently sold by Canberra. The results have also been compared to the efficiency data of the five FastScan counters currently in use across Canada to validate the results of the modelling. The results show that Monte Carlo can be used to perform a primary calibration that would otherwise be both expensive and difficult. 5.0 Acknowledgements The Human Monitoring Laboratory wishes to thank the FastScan operators at Pt. Lepreau NGS, Darlington NGS, Pickering NGS and the Bruce NGS for supplying the efficiency data used in this report. References 1. Kramer GH. Performance Testing of the Canberra FastScan Whole Body Counters in Canada. Rad. Prot. Manag. 17(1): 31-38, (2000). 2. International Commission on Radiological Protection. Report of the Task Group on Reference Man. Oxford: Pergamon Press; ICRP Publication 23, (1975). 3. Nutrition Canada. Anthropometry Report: Height, Weight and Body Dimensions. Ottawa: Health and Welfare Canada, (1980). 4. Panero J, Zelnick M. Human Dimension and Interior Space. Whitney Library of Design, New York, (1979). 5. Snyder WS, Ford MR, Warner GG, Fisher HL. Estimates of absorbed fractions for monoenergetic photon sources uniformly distributed in various organs of a heterogeneous phantom. J Nucl Med Sup No. 3: 7-52, (1969). 6. Kramer GH, Burns LC, Guerriere S. Monte Carlo simulation of a scanning detector whole body counter and the effect of BOMAB phantom size on the calibration. Health Phys 83(4): , (2002). 7. Briesmeister JF. MCNP - A general Monte Carlo code for neutron and photon transport. Los Alamos, NM: Los Alamos National Laboratory; LA-7396-M, Rev. 2, (1986). 8. Keyser RM, Wagner S. Using the IEC standard to describe low-background detectors - what may one expect. Proceedings of the winter meeting of the American Nuclear Society, Washington DC, (1998). 10