PERFORMANCE OF THE VARSKIN 5 (v5.3) ELECTRON DOSIMETRY MODEL

PERFORMANCE OF THE VARSKIN 5 (v5.3) ELECTRON DOSIMETRY MODEL Abstract A new electron skin dosimetry model was developed for the VARSKIN 5 tissue dosimetry code. This model employs energy deposition kernels that provides for improved accuracy of energy deposition at the end of electron tracks. The Monte Carlo code EGSnrc was utilized to develop these energy deposition kernels such that scaling of electron energy loss is dependent on effective atomic number and density of the source material, electron range and conservation of energy. This work contrasts VARSKIN’s electron dosimetry model to several existing deterministic and Monte Carlo dosimetry tools to determine the efficacy of these improvements. Comparison results are given for a wide range of scenarios that extend beyond the typical use of VARSKIN, including mono-energetic electrons and a homogenous water medium. For planar and point sources in contact with the skin, VARSKIN produces results equated to other dosimetry methods within 10%. However, it appears that VARSKIN is unable to account accurately for electron energy loss with the introduction of a cover material or an air gap. The comparisons herein confirm that VARSKIN provides accurate electron dose calculations for skin-contamination scenarios. INTRODUCTION VARSKIN 5 implements an enhanced electron dosimetry model that seeks to improve on several deficiencies of the previous VARSKIN software(1). This Monte Carlo-based dosimetry model better accounts for electron energy-loss through materials and offers a backscatter model estimated to provide a more accurate shallow dose simulation. Previous versions of VARSKIN incorporated Spencer’s moment-based energy dissipation model and a simplified handling of backscatter(2, 3). These models together did not provide a sufficient accounting of energy loss, especially when volumetric, high-Z sources were modeled(4). Previously, VARSKIN gave the user the ability to adjust source density to account for self-absorption of beta energy in volumetric sources(3). However, in extrapolation of that model, the density could be changed, but based on assumptions in developing the model, the effective Z was always that of water. Due to this limitation, the effects of self-absorption were not handled appropriately. VARSKIN 5 allows the user to select density and effective Z of the source material, thus better accounting for electron transport and energy loss through the source prior to skin entry. Another significant source of error in the previous VARSKIN was its handling of internal conversion electrons. Those nuclides not decaying by beta decay, but otherwise emitting electrons, were handled in a way that vastly under predicted electron dose, essentially limiting dose contributions to photon emissions. The new VARSKIN, via the SadCalc routine, generates an electron energy spectrum with a precise account of average electron energy and range. Herein, we provide dose estimates using the VARSKIN 5 electron dosimetry model, and compare those predictions with estimates in the literature of electron dose to skin. Some of the comparison manuscripts are historic, but have been included herein because of their significance and since many of their resulting factors are still in use today. We compare and contrast the shallow skin dose calculated by VARSKIN 5 to several different scenarios using both Monte Carlo simulation and deterministic methods. These comparisons suggest that the new VARSKIN model accounts for internal conversion electrons correctly and provides a very reliable approximation of shallow skin dose. A small number of manuscripts provide skin dosimetry for electrons, the majority of which come from the late 1980s and early 1990s, at a time when hot-particle dosimetry was a more prevalent topic. Most of these studies either mention the VARSKIN software or make comparison to the software version in existence at the time, VARSKIN(5), VARSKIN MOD2(2) or VARSKIN 3(3). The VARSKIN 4 software might be referenced for beta dosimetry, but its electron dose models are unchanged from VARSKIN 3(6). Historically, VARSKIN predictions of electron dose to skin were developed using a method of moment-based point-kernel integration with Berger’s scaled absorbed dose distributions (SADD), which in turn were based on energy dissipation distributions from Spencer(7). Berger’s distributions were density-scaled with homogeneous water and normalized to the radial distance at which 90% dissipation of electron energy occurs (the ×90 range). The electron dosimetry model of VARSKIN was redesigned in the early 2010s with the issuance of VARSKIN 5(1). In this current model, the energy deposition kernels are based on Monte Carlo simulations using EGSnrc(8) over a wide range of electron energy, source material density and atomic number. The major advantage of using a Monte Carlo-based method lies in the improved accuracy of energy deposition at the end of electron tracks(1). Scaling of energy loss in the new model is now dependent on material density and effective atomic number, electron range and conservation of energy. VARSKIN originally was designed as a dosimetry code for the prediction of radiation dose following skin contamination by beta emitters(5). Later, Durham developed the SadCalc routine to calculate the SAAD for beta emitters(2). For those nuclides not transitioning through beta decay, the code generated a ‘dummy’ electron energy distribution(3). The construction of these dummy distributions began with the tritium beta emission spectrum with an artificial yield of 0.1%, and then added conversion electron energies and yields to the spectrum. The SadCalc routine requires a beta distribution to calculate the SADD, but with such a low yield, the dummy tritium distribution is insignificant to the dose calculation(2, 3). From these spectra, an average electron energy and ×90 range was determined. With such low yields, the conversion electrons generally had very little influence on the beta dose estimation and hence the resulting SAAD(3). This, however, is not the case for nuclides emitting high-energy and high-yield conversion electrons. MATERIALS AND METHODS To demonstrate the accuracy of the new VARSKIN electron dosimetry model, we have compared version 5.3 shallow-dose estimates with similar predictions obtained by both deterministic and probabilistic methods. These methods of comparison, carried out for mono-energetic electrons and various electron-emitting nuclides, include the use of Berger’s scaled point kernels(9), Eltran3(10, 11), EGS4(12), MCNP5(4) and the Integrated TIGER Series(13, 14). Unless otherwise stated, results are obtained from calculations of skin dose per electron (for mono-energetic electrons) or per nuclear transition (nt, for electron-emitting nuclides) at the defined shallow depth of 7 mg/cm2. Shallow dose is averaged over an infinitely thin disk of 1 cm2 area at depth in tissue (or water). Source geometry is either a point, a 2D disk or an infinite plane. Models generally simulate the source resting on the skin surface with air or water backing the source and no gap between source and tissue. In one instance, a cotton cover with an air gap is included in the exposure scenario, with the cotton and the air gap being modeled between the source and the skin(12). We provide a list of comparison studies in Table 1. Table 1. Summary of those papers to which we compared. Reference Calculational Source Data Depths Tool Geometry Presentation1 (mg/cm2) Faw (1992) ITS2 Plane c, d 3–5, 5–10, 30–50 Rohloff/Heinzelmann (1986) MC8 Disk b 7 Kocher/Eckerman (1987) PK3 Plane c 4, 7, 8, 40 Chung et al. (1991) Eltran37 Point a 7 Sherbini et al. (2008) MCNP55 Various a 7–100 Covens et al. (2013) MCNPX6/ITS2 Disk b 7, 14, 22, 37 Taylor et al. (1997) EGS44 Point/cover c 7 Reference Calculational Source Data Depths Tool Geometry Presentation1 (mg/cm2) Faw (1992) ITS2 Plane c, d 3–5, 5–10, 30–50 Rohloff/Heinzelmann (1986) MC8 Disk b 7 Kocher/Eckerman (1987) PK3 Plane c 4, 7, 8, 40 Chung et al. (1991) Eltran37 Point a 7 Sherbini et al. (2008) MCNP55 Various a 7–100 Covens et al. (2013) MCNPX6/ITS2 Disk b 7, 14, 22, 37 Taylor et al. (1997) EGS44 Point/cover c 7 1: (a) By nuclide (beta only); (b) by nuclide (all electrons); (c) by mono-energetic calculation multiplied by nuclide electron emission distribution; and (d) by electron energy. 2: Integrated TIGER Series(15, 16). 3: Berger’s scaled point kernel(7). 4: EGS4(8). 5: MCNP5(17). 6: MCNPX(18). 7: Eltran3(10). 8: Monte Carlo program by Nuclear Research Centre Julich(19). Table 1. Summary of those papers to which we compared. Reference Calculational Source Data Depths Tool Geometry Presentation1 (mg/cm2) Faw (1992) ITS2 Plane c, d 3–5, 5–10, 30–50 Rohloff/Heinzelmann (1986) MC8 Disk b 7 Kocher/Eckerman (1987) PK3 Plane c 4, 7, 8, 40 Chung et al. (1991) Eltran37 Point a 7 Sherbini et al. (2008) MCNP55 Various a 7–100 Covens et al. (2013) MCNPX6/ITS2 Disk b 7, 14, 22, 37 Taylor et al. (1997) EGS44 Point/cover c 7 Reference Calculational Source Data Depths Tool Geometry Presentation1 (mg/cm2) Faw (1992) ITS2 Plane c, d 3–5, 5–10, 30–50 Rohloff/Heinzelmann (1986) MC8 Disk b 7 Kocher/Eckerman (1987) PK3 Plane c 4, 7, 8, 40 Chung et al. (1991) Eltran37 Point a 7 Sherbini et al. (2008) MCNP55 Various a 7–100 Covens et al. (2013) MCNPX6/ITS2 Disk b 7, 14, 22, 37 Taylor et al. (1997) EGS44 Point/cover c 7 1: (a) By nuclide (beta only); (b) by nuclide (all electrons); (c) by mono-energetic calculation multiplied by nuclide electron emission distribution; and (d) by electron energy. 2: Integrated TIGER Series(15, 16). 3: Berger’s scaled point kernel(7). 4: EGS4(8). 5: MCNP5(17). 6: MCNPX(18). 7: Eltran3(10). 8: Monte Carlo program by Nuclear Research Centre Julich(19). In addition to investigating skin dose from mono-energetic electrons between 50 keV and 5 MeV, we examined a comprehensive set of electron-emitting radionuclides (Table 2). These nuclides, with average energy between 16 and 935 keV, included primary electron emissions as beta decay, positron decay and conversion electrons. Table 2. List of electron-emitting nuclides considered herein. Nuclide V5.3 Average Primary electron Energy (MeV) Emission mechanism Tc-99m 0.0163 ce I-123 0.0282 ce In-111 0.0344 ce Ga-67 0.0356 ce Tl-201 0.0434 ce S-35 0.0489 β− C-14 0.0495 β− Pm-147 0.0620 β− Ca-45 0.0773 β− Co-60 0.0969 β− Cs-137 0.187 β− I-131 0.192 β− Sr-90 0.196 β− Tl-204 0.244 β− F-18 0.250 β+ Kr-85 0.251 β− Au-198 0.328 β− C-11 0.386 β+ I-132 0.496 β− Sr-89 0.584 β− K-40 0.585 β− P-32 0.695 β− Y-90 0.935 β− Nuclide V5.3 Average Primary electron Energy (MeV) Emission mechanism Tc-99m 0.0163 ce I-123 0.0282 ce In-111 0.0344 ce Ga-67 0.0356 ce Tl-201 0.0434 ce S-35 0.0489 β− C-14 0.0495 β− Pm-147 0.0620 β− Ca-45 0.0773 β− Co-60 0.0969 β− Cs-137 0.187 β− I-131 0.192 β− Sr-90 0.196 β− Tl-204 0.244 β− F-18 0.250 β+ Kr-85 0.251 β− Au-198 0.328 β− C-11 0.386 β+ I-132 0.496 β− Sr-89 0.584 β− K-40 0.585 β− P-32 0.695 β− Y-90 0.935 β− Table 2. List of electron-emitting nuclides considered herein. Nuclide V5.3 Average Primary electron Energy (MeV) Emission mechanism Tc-99m 0.0163 ce I-123 0.0282 ce In-111 0.0344 ce Ga-67 0.0356 ce Tl-201 0.0434 ce S-35 0.0489 β− C-14 0.0495 β− Pm-147 0.0620 β− Ca-45 0.0773 β− Co-60 0.0969 β− Cs-137 0.187 β− I-131 0.192 β− Sr-90 0.196 β− Tl-204 0.244 β− F-18 0.250 β+ Kr-85 0.251 β− Au-198 0.328 β− C-11 0.386 β+ I-132 0.496 β− Sr-89 0.584 β− K-40 0.585 β− P-32 0.695 β− Y-90 0.935 β− Nuclide V5.3 Average Primary electron Energy (MeV) Emission mechanism Tc-99m 0.0163 ce I-123 0.0282 ce In-111 0.0344 ce Ga-67 0.0356 ce Tl-201 0.0434 ce S-35 0.0489 β− C-14 0.0495 β− Pm-147 0.0620 β− Ca-45 0.0773 β− Co-60 0.0969 β− Cs-137 0.187 β− I-131 0.192 β− Sr-90 0.196 β− Tl-204 0.244 β− F-18 0.250 β+ Kr-85 0.251 β− Au-198 0.328 β− C-11 0.386 β+ I-132 0.496 β− Sr-89 0.584 β− K-40 0.585 β− P-32 0.695 β− Y-90 0.935 β− As seen in many of the figures throughout this article, comparison data are plotted on the ordinate with VARSKIN data plotted on the abscissa; therefore, a slope of one (represented by a dashed line) shows complete agreement between dose predictions. Data points appearing above the dashed line indicate that VARSKIN 5.3 is under predicting (or the comparison data are over predicting) radiation dose at shallow depths in tissue. Data points below a slope of one indicate the opposite. For the majority of comparisons, we have calculated dose per emitted electron (for mono-energetic sources) or dose per nuclear transition (for beta/electron emitters over a range of energies). The knowledgeable user can modify VARSKIN in a couple of ways for added utility: first, to model a mono-energetic electron; and second, to model a water/water (homogenous) interface at the skin surface. In the first case, the user should modify the ICRP38 data (the ICRP38.DAT file in the dat folder) for one of the nuclides. We have found that 7Be is one of the easiest to modify. Figure 1 shows the first few lines of ICRP38.DAT for the original data (left) and modified data (right). Figure 1. View largeDownload slide ICRP38.DAT original data (left) and modified data (right) for the consideration of mono-energetic electrons. Figure 1. View largeDownload slide ICRP38.DAT original data (left) and modified data (right) for the consideration of mono-energetic electrons. On examination of the original data (left), we see that 7Be has a half-life of 53.3 days and four major emissions per decay. The first in the list is a gamma ray (code 1) with a yield of 10.34% and an energy of 0.477605 MeV. The second and third lines show emissions of conversion electrons (code 6) with very low yields and roughly the same energies as the photon. The final emission is a low-energy X-ray (code 2) with a yield of 1.635%. We cannot simply remove three of the data lines and indicate that 7Be has only one emission; if we did, indices throughout the VARSKIN data files would be in disarray. Therefore, we will set yields and energies to zero (bolded data on the right) except for one entry. The modified entry shows the emission of an electron with 100% yield and an energy of 0.75 MeV. This will simulate the emission of a mono-energetic electron; no other alterations are necessary. The second modification allows the user to change the VARSKIN simulation (normally an air/water interface at the skin surface) to model a water/water interface. We accomplish this by modifying the.rad file (in the dat folder) for the nuclide of choice. The final 160 lines of data written to that file (by SadCalc) are backscatter factors for use in the half-space scenario (air/water interface). Manually setting each of those factors to ‘1’ removes consideration of half-space such that VARSKIN will model a homogeneous water medium (a water/water interface). RESULTS AND DISCUSSION We first examine VARSKIN 5.3 shallow dose predictions from exposure to mono-energetic electrons. Faw calculates absorbed dose to the skin at various depths using the CYLTRON and TIGER codes, Monte Carlo simulations with the Integrated Tiger series(13, 15, 16). He calculates dose to the skin for isotropic plane sources with mono-energetic electrons of 0.25 and 3.0 MeV. For 0.25 MeV electrons, Faw uses the TIGER code to simulate the source on the surface of skin, reportedly backed by air (i.e. air/water interface ‘with backscatter’) and then backed by a vacuum (i.e. vacuum/water interface ‘without backscatter’)(13). Skin dose estimates for the water/water and the air/water scenarios using VARSKIN 5.3 appear to contradict Faw’s explanation of ‘with and without backscatter’. A simple execution of MCNP, using a point source of 0.25 MeV electrons at a water interface backed in one case by air and in the other case by water, indicates that ‘without backscatter’ refers to the air/water scenario (Figure 2a) and ‘with backscatter’ refers to the water/water scenario (Figure 2b). Under these conditions, the VARSKIN data and Faw’s data are in extremely good agreement. Figure 2. View largeDownload slide VARSKIN 5.3 compared to Faw for 0.25 MeV mono-energetic electrons as a function of depth in tissue(13). (a) The electron source is assumed to be on the surface of skin with air above the skin (air/water interface). (b) The electron source is assumed to be on the surface of skin with water above the skin (water/water interface). Figure 2. View largeDownload slide VARSKIN 5.3 compared to Faw for 0.25 MeV mono-energetic electrons as a function of depth in tissue(13). (a) The electron source is assumed to be on the surface of skin with air above the skin (air/water interface). (b) The electron source is assumed to be on the surface of skin with water above the skin (water/water interface). For 3 MeV electrons (Figure 3), estimates from VARSKIN compare very well with CYLTRON(13), in an infinite water medium (water/water interface). VARSKIN dose estimates for the air/water interface are also in the figure, and indicate that the characteristics of the interface relative to electron backscatter are relatively unimportant for high-energy electrons beyond the shallow depths. Figure 3. View largeDownload slide VARSKIN 5.3 compared to Faw for 3 MeV mono-energetic electrons in an infinite water medium (w/w) and with an air/water (a/w) interface(13). Figure 3. View largeDownload slide VARSKIN 5.3 compared to Faw for 3 MeV mono-energetic electrons in an infinite water medium (w/w) and with an air/water (a/w) interface(13). Additionally, Faw calculates average electron dose as a function of energy at three volumetric depths (Figure 4), again for cases with reflection (water/water interface) and without reflection (air/water interface)(13). Skin dose estimates from VARSKIN 5.3 for the same scenario of an infinite plane source and an air/water interface, indicate that VARSKIN is a very good predictor of mono-energetic electron dose at various shallow depths over a wide range of electron energies. For electrons greater than ~2 MeV and depths between 5 and 50 mg/cm2, however, VARSKIN seems to over predict Faw by as much as 20%, possibly due to the method of handling backscatter at high energies. Figure 4. View largeDownload slide Dose per electron calculated by VARSKIN 5.3 compared with Faw at three averaging depths in tissue (3–5, 5–10 and 30–50 mg/cm2). The FAW data are assumed to consider dose for an air/water interface (see text)(13). Figure 4. View largeDownload slide Dose per electron calculated by VARSKIN 5.3 compared with Faw at three averaging depths in tissue (3–5, 5–10 and 30–50 mg/cm2). The FAW data are assumed to consider dose for an air/water interface (see text)(13). Rohloff and Heinzelmann, using their own Monte Carlo code, estimated electron dose to the skin for various beta-emitting nuclides over a wide range of average electron energy (35S, 14C, 147Pm, 45Ca, 60Co, 131I, 90Sr, 204Tl, 18F, 198Au, 89Sr, 32P and 90Y). They speak of calculating dose to skin, ‘both with and without tracking the beta particles backscattered from the surrounding air’(19, 20). Comparison with VARSKIN (Figure 5a and b) for an average energy between 49 keV (35S) and 0.935 MeV (90Y) and an air/water interface shows very good agreement at low energy, but indicates a consistent under prediction by as much as 10% from VARSKIN at the highest doses. Figure 5. View largeDownload slide (a) VARSKIN 5.3 compared to Rohloff (air/water interface) at 7 mg/cm2(19, 20). (b) The same comparison data plotted as a function of average beta energy. Figure 5. View largeDownload slide (a) VARSKIN 5.3 compared to Rohloff (air/water interface) at 7 mg/cm2(19, 20). (b) The same comparison data plotted as a function of average beta energy. Kocher and Eckerman estimated electron dose–rate factors for mono-energetic sources on the skin. They used the scaled point-kernel methods of Berger, inherently in a homogeneous water medium(7, 9). Comparison electron dose–rate factors for skin are duplicated in VARSKIN 5.3 (Figure 6). The Kocher/Eckerman dose factors were simulated in VARSKIN by assuming an infinitely large 2D disk source (15 cm2) on the skin surface with a uniformly distributed activity of 1 Bq/cm2 and all backscatter factors modified to unity (to mimic the water/water interface), as described above. Figure 6. View largeDownload slide Electron dose–rate factors from VARSKIN 5.3 (solid lines) compared to Kocher and Eckerman(9) (data points) with a water/water interface at three depths (4, 8 and 40 mg/cm2) for 13 radionuclides: 99mTc, 111In, 123I, 201Tl, 67Ga, 35S, 14C, 45Ca, 60Co, 131I, 90Sr, 18F and 32P. Figure 6. View largeDownload slide Electron dose–rate factors from VARSKIN 5.3 (solid lines) compared to Kocher and Eckerman(9) (data points) with a water/water interface at three depths (4, 8 and 40 mg/cm2) for 13 radionuclides: 99mTc, 111In, 123I, 201Tl, 67Ga, 35S, 14C, 45Ca, 60Co, 131I, 90Sr, 18F and 32P. We also compare the Kocher and Eckerman dose–rate factors to VARSKIN 5.3 (Figure 7) for specific nuclides at four depths beneath a water/water interface, including a composite of 13 nuclides from Table 1(9). The comparisons (Figures 6 and 7) show very good agreement between VARSKIN and the Kocher/Eckerman data for a homogeneous water medium. Figure 7. View largeDownload slide Skin beta dose–rate factors for VARSKIN 5.3 compared to Kocher and Eckerman(9) at four depths (4, 7, 8 and 40 mg/cm2) for 13 radionuclides: 99mTc, 111In, 123I, 201Tl, 67Ga, 35S, 14C, 45Ca, 60Co, 131I, 90Sr, 18F and 32P. Figure 7. View largeDownload slide Skin beta dose–rate factors for VARSKIN 5.3 compared to Kocher and Eckerman(9) at four depths (4, 7, 8 and 40 mg/cm2) for 13 radionuclides: 99mTc, 111In, 123I, 201Tl, 67Ga, 35S, 14C, 45Ca, 60Co, 131I, 90Sr, 18F and 32P. Chung et al. simulated point sources on the skin using a 2D Monte Carlo transport code, Eltran3, to estimate electron dose from six nuclides (and considered a seventh ‘nuclide’ as the sum of Sr-90 and Y-90, in equilibrium) to a depth of 7 mg/cm2(10, 11). VARSKIN 5.3 results are shown to be in good agreement with their data (Figure 8) for both air/water and water/water simulations. Figure 8. View largeDownload slide Point-source comparison between VARSKIN 5.3 and Eltran3(10) of beta/electron dose per transition (pGy/nt) at a depth of 7 mg/cm2 for six nuclides (and the sum of 90Sr/Y) with a boundary interface of both air/water and water/water. Figure 8. View largeDownload slide Point-source comparison between VARSKIN 5.3 and Eltran3(10) of beta/electron dose per transition (pGy/nt) at a depth of 7 mg/cm2 for six nuclides (and the sum of 90Sr/Y) with a boundary interface of both air/water and water/water. Covens et al.(14) have estimated skin dose at four depths for several nuclides using MCNPX. They assume an air/water interface with surface contamination areas normalized over 1 cm2. We compare results from similar calculations of seven nuclides using VARSKIN 5.3 (Figure 9). The Covens data and VARSKIN are in very good agreement. Figure 9. View largeDownload slide VARSKIN 5.3 compared to Covens et al.(14) using MCNPX at four different depths (7, 14, 22 and 37 mg/cm2) in tissue. Doses calculated with an air/water interface for 67Ga, 99mTc, 111In, 117mSn, 123I, 195mPt and 201Tl. Figure 9. View largeDownload slide VARSKIN 5.3 compared to Covens et al.(14) using MCNPX at four different depths (7, 14, 22 and 37 mg/cm2) in tissue. Doses calculated with an air/water interface for 67Ga, 99mTc, 111In, 117mSn, 123I, 195mPt and 201Tl. For an air/water interface, Taylor et al.(12) calculate dose for mono-energetic electrons and then multiply by electron emission spectra to estimate dose resulting from exposure to particular radionuclides. They used EGS4(8) to simulate beta skin dose for several different nuclides placed on a 26 mg/cm2 cotton cover (ρ = 0.7 g/cm3), with a 2 mm air gap over the skin(12). Average electron energies range from 0.097 (60Co) to 0.695 MeV (32P), with a comparison to VARSKIN (Figure 10) for 60Co, 137Cs, 131I, 90Sr, 85Kr, 11C, 132I, 89Sr and 32P, in order of increasing average electron energy (and generally in order of increasing dose per transition). Since the scenario is unchanged between nuclides, the only influence on shallow dose is the electron emission spectrum generally, where an increase in average energy equates to an increase in shallow skin dose. The results show fair agreement in the lower dose (lower energy) realm, but for a few nuclides of higher average energy (11C, 132I, 89Sr, 32P) VARSKIN under predicts dose by ~30–45%. Figure 10. View largeDownload slide Electron dose at 7 mg/cm2 in VARSKIN 5.3 compared to Taylor et al.(12). The scenario includes a cotton cover with an air gap. Figure 10. View largeDownload slide Electron dose at 7 mg/cm2 in VARSKIN 5.3 compared to Taylor et al.(12). The scenario includes a cotton cover with an air gap. When plotting results as a function of average beta energy (Figure 11), we can see that there is departure between the two estimates of shallow dose at ~250 keV. As surmised by Taylor et al.(12), and supported here, VARSKIN is apparently degrading electron energy more than is appropriate in its cover and air-gap model. Figure 11. View largeDownload slide VARSKIN 5.3 compared to Taylor et al.(12) for shallow dose at 7 mg/cm2 as a function of average beta energy. Figure 11. View largeDownload slide VARSKIN 5.3 compared to Taylor et al.(12) for shallow dose at 7 mg/cm2 as a function of average beta energy. In our last comparison figure, dose estimates from several of the scenarios discussed above are shown in one plot (Figure 12) to convey a general comparison of shallow dose across many different exposure scenarios. We developed the plot with data from seven different manuscripts (Table 1) and VARSKIN 5.3. The VARSKIN 5.3 dose estimates (solid blue line) considered a very simple 1 Bq point source for 23 electron-emitting radionuclides (Table 2) on the skin surface with an air/water interface. Dose was estimated over a 1 cm2 averaging disk at a depth of 7 mg/cm2 for an exposure time of 1 s (to result in units of dose per nuclear transition). We plot shallow dose by nuclide as a function of average beta/electron energy. The VARSKIN dose estimates appear to have discontinuities (at 40, 240 and 580 keV, for example) because of multiple dose values falling around the same average energy (for example, 89Sr and 40K at 0.584 and 0.585 MeV, respectively). Even though two nuclides emit the same average energy, their electron energy distributions are different, resulting in a different skin dose evaluation for those nuclides. Figure 12. View largeDownload slide Normalized shallow beta/electron dose from various scenarios. The data for Taylor w/ cover (dashes) has been modified to account for residual energy loss due to the cover/air layers, resulting in the ‘energy shifted’ data (solid squares). Figure 12. View largeDownload slide Normalized shallow beta/electron dose from various scenarios. The data for Taylor w/ cover (dashes) has been modified to account for residual energy loss due to the cover/air layers, resulting in the ‘energy shifted’ data (solid squares). For each comparison scenario (all data points), we normalized the ordinate around a dose of ~400 pGy/nt as calculated by VARSKIN such that all values fell on the plateau. This normalization was conducted by multiplying dose estimates for a given scenario by a simple scalar value. For example, the data of Rohloff and Heinzelmann(19, 20) were multiplied by a value of 100 for the normalization. An interesting feature of Figure 12 is that with this simple normalization, all scenarios tend to produce the same outcome, i.e. the same graphical shape. Electron shallow dose ranges about an order of magnitude over average electron energies from 16 keV up to ~1 MeV, with the dose plateau reached at ~200 keV. It is also interesting that the discontinuity in dose at ~30–50 keV is observable in all scenarios that track beta dose from low-energy emitters. The scenarios depicted by Sherbini et al. (pt)(4), Covens et al.(14) and Chung et al.(11) simulated point sources on the skin surface; whereas Faw(13), Sherbini et al. (pl)(4) and Kocher/Eckerman(9) simulated planar sources. The Taylor et al.(12) scenario considered a point source on a cotton cover with an air gap. Energy loss through the cotton and air was estimated such that the raw Taylor data (dashes) was shifted to the left (solid squares) based on a calculation of residual energy following electron passage through this material(21). The purpose of consolidating the data in Figure 12 is to demonstrate that VARSKIN’s approach to handling the physics of electron dosimetry is as effective as many of the Monte Carlo simulations to which we have compared. Backscatter, self-absorption, scaling of energy loss and energy deposition are all taken into account. CONCLUSION We have examined the efficacy of the VARSKIN electron dosimetry model for estimating shallow dose following the radiological contamination of skin. For point and planar sources on the skin surface, VARSKIN is shown to be within 10% of other deterministic and probabilistic methods. Discrepancies with other methods appear when VARSKIN models the presence of air-gaps or cover material (e.g. protective clothing). Compared to EGS4 results(12), VARSKIN underestimates by nearly a factor of two in one particular cover scenario with higher-energy electrons. This tends to suggest that energy loss in air and/or cover material is not handled properly in VARSKIN. The calculational structure of the VARSKIN model is shown to be consistent with other methods (Figure 12), even to the extent of predicting skin dose from low-energy electrons. The results presented herein provide additional data for building our confidence in the VARSKIN methods for beta/electron dosimetry. ACKNOWLEDGEMENTS The authors wish to thank Dr Colby Mangini for his insight on VARSKIN and the EGSnrc Monte Carlo code. FUNDING This work was supported by the Nuclear Regulatory Commission under contract NRC-HQ-60-15-R-0017. REFERENCES 1 Hamby , D. M. , Mangini , C. D. , Caffrey , J. A. and Tang , M. VARSKIN 5: A computer code for skin contamination dosimetry. NUREG/CR-6918, Rev. 2. U.S. Nuclear Regulatory Commission. Washington, DC (July 2014 ). 2 Durham , J. S. VARSKIN MOD2 and SADDE MOD2: computer codes for assessing skin dose from skin contamination. NUREG/CR-5873. U.S. Nuclear Regulatory Commission. Washington, DC (December 1992 ). 3 Durham , J. S. VARSKIN 3: a computer code for assessing skin dose from skin contamination. NUREG/CR-6918. U.S. Nuclear Regulatory Commission. Washington, DC (October 2006 ). 4 Sherbini , S. , DeCicco , J. , Gray , A. T. and Struckmeyer , R. Verification of the VARSKIN beta skin dose calculation computer code . Health Phys. 94 ( 6 ), 527 – 538 ( 2008 ). Google Scholar Crossref Search ADS PubMed 5 Traub , R. J. , Reece , W. D. , Scherpelz , R. I. and Sigalia , L. A. Dose calculation for contamination of the skin using the computer code VARSKIN. NUREG/CR/4418. US Nuclear Regulatory Commission. Washington, DC (August 1987 ). 6 Hamby , D. M. , Lodwick , C. J. , Palmer , T. S. , Reese , S. R. and Higley , K. A. VARSKIN 4: a computer code for skin contamination dosimetry. NUREG/CR-6918, Rev. 1. U.S. Nuclear Regulatory Commission. Washington, DC (June 2011 ). 7 Berger , M. J. Distributions of absorbed dose around point sources of electrons and beta particles in water and other media. (MIRD Pamphlet No. 7, Supplement No. 5) . J. Nucl. Med. 12 ( 5 ), 5 – 23 ( 1971 ). Google Scholar PubMed 8 Nelson , W. R. , Hirayama , H. and Rogers , D. W. The EGS4 code-system. Report no. SLAC-265. Stanford Linear Accelerator Center. Stanford University. Stanford, CA ( 1985 ). 9 Kocher , D. C. and Eckerman , K. F. Electron dose-rate conversion factors for external exposure of the skin from uniformly deposited activity on the body surface . Health Phys. 53 ( 2 ), 135 – 141 ( 1987 ). Google Scholar Crossref Search ADS PubMed 10 Chung , M. , Foderaro , A. H. , Jester , W. A. and Levine , S. H. Microcomputer Monte Carlo electron transport codes for beta skin dose calculations . IEEE Trans. Nucl. Sci. 18 ( 3 ), 936 – 941 ( 1991 ). Google Scholar Crossref Search ADS 11 Chung , M. , Levine , S. H. and Jester , W. A. Monte Carlo calculation and silicon detector measurement of the hot particle dose . Health Phys. 61 ( 6 ), 843 – 848 ( 1991 ). Google Scholar Crossref Search ADS PubMed 12 Taylor , D. C. , Hussein , E. M. A. and Yuen , P. S. Skin dose from radionuclide contamination on clothing . Health Phys. 72 ( 6 ), 835 – 841 ( 1997 ). Google Scholar Crossref Search ADS PubMed 13 Faw , R. E. Absorbed dose to skin from radionuclide sources on the body surface . Health Phys. 63 ( 4 ), 443 – 448 ( 1992 ). Google Scholar Crossref Search ADS PubMed 14 Covens , P. , Berus , D. , Caveliers , V. , Struelens , L. , Vanhavere , F. and Verellen , D. Skin dose rate conversion factors after contamination with radiopharmaceuticals: influence of contamination area, epidermal thickness and percutaneous absorption . J. Radiol. Prot. 33 , 381 – 393 ( 2013 ). Google Scholar Crossref Search ADS PubMed 15 Halbleib , J. A. and Melhorn , T. A. ITS: Integrated TIGER series of coupled electron/photon Monte Carlo transport codes. Report no. SAND 84–0573. Sandia National Laboratories. Albuquerque, NM ( 1984 ). 16 Halbleib , J. A. , Kensek , R. P. , Valdez , G. D. , Melhorn , T. A. , Seltzer , S. M. and Berger , M. J. ITS Version 3.0: the integrated TIGER series of coupled electron/photon Monte Carlo codes . IEEE Trans. Nucl. Sci. 39 , 1025 – 1030 ( 1992 ). Google Scholar Crossref Search ADS 17 X-5 Monte Carlo team . MCNP: a general Monte Carlo N-particle transport code. Version 5. Report No. LA-CP-03–0245, Los Alamos National Laboratory. Los Alamos, NM ( 2003 ). 18 Pelowitz , D. B. Ed. MCNPXTM User’s Manual. Version 2.6.0. Report No. LA-CP-07-1473 ( Los Alamos, NM : Los Alamos National Laboratory ) ( 2008 ). 19 Rohloff , F. A Monte Carlo program for the transport of beta particles through layers of different materials. In: Proceedings of the International Beta Dosimetry Symposium. February 15–18, 1983. Washington, DC. NUREG/CP-0050; p. 25 ( 1984 ). 20 Rohloff , F. and Heinzelmann , M. Calculation of dose rates for skin contamination by beta radiation . Radiat. Prot. Dosim. 14 ( 4 ), 279 – 287 ( 1986 ). Google Scholar Crossref Search ADS 21 Shultiz , J. K. and Faw , R. E. Radiation Shielding ( La Grange Park, IL : American Nuclear Society ) ( 2000 ) ISBN 0-89448-456-7. © The Author(s) 2018. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Radiation Protection Dosimetry Oxford University Press

PERFORMANCE OF THE VARSKIN 5 (v5.3) ELECTRON DOSIMETRY MODEL

Loading next page...
 
/lp/ou_press/performance-of-the-varskin-5-v5-3-electron-dosimetry-model-Mwn0eL3HXT
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com
ISSN
0144-8420
eISSN
1742-3406
D.O.I.
10.1093/rpd/ncx302
Publisher site
See Article on Publisher Site

Abstract

Abstract A new electron skin dosimetry model was developed for the VARSKIN 5 tissue dosimetry code. This model employs energy deposition kernels that provides for improved accuracy of energy deposition at the end of electron tracks. The Monte Carlo code EGSnrc was utilized to develop these energy deposition kernels such that scaling of electron energy loss is dependent on effective atomic number and density of the source material, electron range and conservation of energy. This work contrasts VARSKIN’s electron dosimetry model to several existing deterministic and Monte Carlo dosimetry tools to determine the efficacy of these improvements. Comparison results are given for a wide range of scenarios that extend beyond the typical use of VARSKIN, including mono-energetic electrons and a homogenous water medium. For planar and point sources in contact with the skin, VARSKIN produces results equated to other dosimetry methods within 10%. However, it appears that VARSKIN is unable to account accurately for electron energy loss with the introduction of a cover material or an air gap. The comparisons herein confirm that VARSKIN provides accurate electron dose calculations for skin-contamination scenarios. INTRODUCTION VARSKIN 5 implements an enhanced electron dosimetry model that seeks to improve on several deficiencies of the previous VARSKIN software(1). This Monte Carlo-based dosimetry model better accounts for electron energy-loss through materials and offers a backscatter model estimated to provide a more accurate shallow dose simulation. Previous versions of VARSKIN incorporated Spencer’s moment-based energy dissipation model and a simplified handling of backscatter(2, 3). These models together did not provide a sufficient accounting of energy loss, especially when volumetric, high-Z sources were modeled(4). Previously, VARSKIN gave the user the ability to adjust source density to account for self-absorption of beta energy in volumetric sources(3). However, in extrapolation of that model, the density could be changed, but based on assumptions in developing the model, the effective Z was always that of water. Due to this limitation, the effects of self-absorption were not handled appropriately. VARSKIN 5 allows the user to select density and effective Z of the source material, thus better accounting for electron transport and energy loss through the source prior to skin entry. Another significant source of error in the previous VARSKIN was its handling of internal conversion electrons. Those nuclides not decaying by beta decay, but otherwise emitting electrons, were handled in a way that vastly under predicted electron dose, essentially limiting dose contributions to photon emissions. The new VARSKIN, via the SadCalc routine, generates an electron energy spectrum with a precise account of average electron energy and range. Herein, we provide dose estimates using the VARSKIN 5 electron dosimetry model, and compare those predictions with estimates in the literature of electron dose to skin. Some of the comparison manuscripts are historic, but have been included herein because of their significance and since many of their resulting factors are still in use today. We compare and contrast the shallow skin dose calculated by VARSKIN 5 to several different scenarios using both Monte Carlo simulation and deterministic methods. These comparisons suggest that the new VARSKIN model accounts for internal conversion electrons correctly and provides a very reliable approximation of shallow skin dose. A small number of manuscripts provide skin dosimetry for electrons, the majority of which come from the late 1980s and early 1990s, at a time when hot-particle dosimetry was a more prevalent topic. Most of these studies either mention the VARSKIN software or make comparison to the software version in existence at the time, VARSKIN(5), VARSKIN MOD2(2) or VARSKIN 3(3). The VARSKIN 4 software might be referenced for beta dosimetry, but its electron dose models are unchanged from VARSKIN 3(6). Historically, VARSKIN predictions of electron dose to skin were developed using a method of moment-based point-kernel integration with Berger’s scaled absorbed dose distributions (SADD), which in turn were based on energy dissipation distributions from Spencer(7). Berger’s distributions were density-scaled with homogeneous water and normalized to the radial distance at which 90% dissipation of electron energy occurs (the ×90 range). The electron dosimetry model of VARSKIN was redesigned in the early 2010s with the issuance of VARSKIN 5(1). In this current model, the energy deposition kernels are based on Monte Carlo simulations using EGSnrc(8) over a wide range of electron energy, source material density and atomic number. The major advantage of using a Monte Carlo-based method lies in the improved accuracy of energy deposition at the end of electron tracks(1). Scaling of energy loss in the new model is now dependent on material density and effective atomic number, electron range and conservation of energy. VARSKIN originally was designed as a dosimetry code for the prediction of radiation dose following skin contamination by beta emitters(5). Later, Durham developed the SadCalc routine to calculate the SAAD for beta emitters(2). For those nuclides not transitioning through beta decay, the code generated a ‘dummy’ electron energy distribution(3). The construction of these dummy distributions began with the tritium beta emission spectrum with an artificial yield of 0.1%, and then added conversion electron energies and yields to the spectrum. The SadCalc routine requires a beta distribution to calculate the SADD, but with such a low yield, the dummy tritium distribution is insignificant to the dose calculation(2, 3). From these spectra, an average electron energy and ×90 range was determined. With such low yields, the conversion electrons generally had very little influence on the beta dose estimation and hence the resulting SAAD(3). This, however, is not the case for nuclides emitting high-energy and high-yield conversion electrons. MATERIALS AND METHODS To demonstrate the accuracy of the new VARSKIN electron dosimetry model, we have compared version 5.3 shallow-dose estimates with similar predictions obtained by both deterministic and probabilistic methods. These methods of comparison, carried out for mono-energetic electrons and various electron-emitting nuclides, include the use of Berger’s scaled point kernels(9), Eltran3(10, 11), EGS4(12), MCNP5(4) and the Integrated TIGER Series(13, 14). Unless otherwise stated, results are obtained from calculations of skin dose per electron (for mono-energetic electrons) or per nuclear transition (nt, for electron-emitting nuclides) at the defined shallow depth of 7 mg/cm2. Shallow dose is averaged over an infinitely thin disk of 1 cm2 area at depth in tissue (or water). Source geometry is either a point, a 2D disk or an infinite plane. Models generally simulate the source resting on the skin surface with air or water backing the source and no gap between source and tissue. In one instance, a cotton cover with an air gap is included in the exposure scenario, with the cotton and the air gap being modeled between the source and the skin(12). We provide a list of comparison studies in Table 1. Table 1. Summary of those papers to which we compared. Reference Calculational Source Data Depths Tool Geometry Presentation1 (mg/cm2) Faw (1992) ITS2 Plane c, d 3–5, 5–10, 30–50 Rohloff/Heinzelmann (1986) MC8 Disk b 7 Kocher/Eckerman (1987) PK3 Plane c 4, 7, 8, 40 Chung et al. (1991) Eltran37 Point a 7 Sherbini et al. (2008) MCNP55 Various a 7–100 Covens et al. (2013) MCNPX6/ITS2 Disk b 7, 14, 22, 37 Taylor et al. (1997) EGS44 Point/cover c 7 Reference Calculational Source Data Depths Tool Geometry Presentation1 (mg/cm2) Faw (1992) ITS2 Plane c, d 3–5, 5–10, 30–50 Rohloff/Heinzelmann (1986) MC8 Disk b 7 Kocher/Eckerman (1987) PK3 Plane c 4, 7, 8, 40 Chung et al. (1991) Eltran37 Point a 7 Sherbini et al. (2008) MCNP55 Various a 7–100 Covens et al. (2013) MCNPX6/ITS2 Disk b 7, 14, 22, 37 Taylor et al. (1997) EGS44 Point/cover c 7 1: (a) By nuclide (beta only); (b) by nuclide (all electrons); (c) by mono-energetic calculation multiplied by nuclide electron emission distribution; and (d) by electron energy. 2: Integrated TIGER Series(15, 16). 3: Berger’s scaled point kernel(7). 4: EGS4(8). 5: MCNP5(17). 6: MCNPX(18). 7: Eltran3(10). 8: Monte Carlo program by Nuclear Research Centre Julich(19). Table 1. Summary of those papers to which we compared. Reference Calculational Source Data Depths Tool Geometry Presentation1 (mg/cm2) Faw (1992) ITS2 Plane c, d 3–5, 5–10, 30–50 Rohloff/Heinzelmann (1986) MC8 Disk b 7 Kocher/Eckerman (1987) PK3 Plane c 4, 7, 8, 40 Chung et al. (1991) Eltran37 Point a 7 Sherbini et al. (2008) MCNP55 Various a 7–100 Covens et al. (2013) MCNPX6/ITS2 Disk b 7, 14, 22, 37 Taylor et al. (1997) EGS44 Point/cover c 7 Reference Calculational Source Data Depths Tool Geometry Presentation1 (mg/cm2) Faw (1992) ITS2 Plane c, d 3–5, 5–10, 30–50 Rohloff/Heinzelmann (1986) MC8 Disk b 7 Kocher/Eckerman (1987) PK3 Plane c 4, 7, 8, 40 Chung et al. (1991) Eltran37 Point a 7 Sherbini et al. (2008) MCNP55 Various a 7–100 Covens et al. (2013) MCNPX6/ITS2 Disk b 7, 14, 22, 37 Taylor et al. (1997) EGS44 Point/cover c 7 1: (a) By nuclide (beta only); (b) by nuclide (all electrons); (c) by mono-energetic calculation multiplied by nuclide electron emission distribution; and (d) by electron energy. 2: Integrated TIGER Series(15, 16). 3: Berger’s scaled point kernel(7). 4: EGS4(8). 5: MCNP5(17). 6: MCNPX(18). 7: Eltran3(10). 8: Monte Carlo program by Nuclear Research Centre Julich(19). In addition to investigating skin dose from mono-energetic electrons between 50 keV and 5 MeV, we examined a comprehensive set of electron-emitting radionuclides (Table 2). These nuclides, with average energy between 16 and 935 keV, included primary electron emissions as beta decay, positron decay and conversion electrons. Table 2. List of electron-emitting nuclides considered herein. Nuclide V5.3 Average Primary electron Energy (MeV) Emission mechanism Tc-99m 0.0163 ce I-123 0.0282 ce In-111 0.0344 ce Ga-67 0.0356 ce Tl-201 0.0434 ce S-35 0.0489 β− C-14 0.0495 β− Pm-147 0.0620 β− Ca-45 0.0773 β− Co-60 0.0969 β− Cs-137 0.187 β− I-131 0.192 β− Sr-90 0.196 β− Tl-204 0.244 β− F-18 0.250 β+ Kr-85 0.251 β− Au-198 0.328 β− C-11 0.386 β+ I-132 0.496 β− Sr-89 0.584 β− K-40 0.585 β− P-32 0.695 β− Y-90 0.935 β− Nuclide V5.3 Average Primary electron Energy (MeV) Emission mechanism Tc-99m 0.0163 ce I-123 0.0282 ce In-111 0.0344 ce Ga-67 0.0356 ce Tl-201 0.0434 ce S-35 0.0489 β− C-14 0.0495 β− Pm-147 0.0620 β− Ca-45 0.0773 β− Co-60 0.0969 β− Cs-137 0.187 β− I-131 0.192 β− Sr-90 0.196 β− Tl-204 0.244 β− F-18 0.250 β+ Kr-85 0.251 β− Au-198 0.328 β− C-11 0.386 β+ I-132 0.496 β− Sr-89 0.584 β− K-40 0.585 β− P-32 0.695 β− Y-90 0.935 β− Table 2. List of electron-emitting nuclides considered herein. Nuclide V5.3 Average Primary electron Energy (MeV) Emission mechanism Tc-99m 0.0163 ce I-123 0.0282 ce In-111 0.0344 ce Ga-67 0.0356 ce Tl-201 0.0434 ce S-35 0.0489 β− C-14 0.0495 β− Pm-147 0.0620 β− Ca-45 0.0773 β− Co-60 0.0969 β− Cs-137 0.187 β− I-131 0.192 β− Sr-90 0.196 β− Tl-204 0.244 β− F-18 0.250 β+ Kr-85 0.251 β− Au-198 0.328 β− C-11 0.386 β+ I-132 0.496 β− Sr-89 0.584 β− K-40 0.585 β− P-32 0.695 β− Y-90 0.935 β− Nuclide V5.3 Average Primary electron Energy (MeV) Emission mechanism Tc-99m 0.0163 ce I-123 0.0282 ce In-111 0.0344 ce Ga-67 0.0356 ce Tl-201 0.0434 ce S-35 0.0489 β− C-14 0.0495 β− Pm-147 0.0620 β− Ca-45 0.0773 β− Co-60 0.0969 β− Cs-137 0.187 β− I-131 0.192 β− Sr-90 0.196 β− Tl-204 0.244 β− F-18 0.250 β+ Kr-85 0.251 β− Au-198 0.328 β− C-11 0.386 β+ I-132 0.496 β− Sr-89 0.584 β− K-40 0.585 β− P-32 0.695 β− Y-90 0.935 β− As seen in many of the figures throughout this article, comparison data are plotted on the ordinate with VARSKIN data plotted on the abscissa; therefore, a slope of one (represented by a dashed line) shows complete agreement between dose predictions. Data points appearing above the dashed line indicate that VARSKIN 5.3 is under predicting (or the comparison data are over predicting) radiation dose at shallow depths in tissue. Data points below a slope of one indicate the opposite. For the majority of comparisons, we have calculated dose per emitted electron (for mono-energetic sources) or dose per nuclear transition (for beta/electron emitters over a range of energies). The knowledgeable user can modify VARSKIN in a couple of ways for added utility: first, to model a mono-energetic electron; and second, to model a water/water (homogenous) interface at the skin surface. In the first case, the user should modify the ICRP38 data (the ICRP38.DAT file in the dat folder) for one of the nuclides. We have found that 7Be is one of the easiest to modify. Figure 1 shows the first few lines of ICRP38.DAT for the original data (left) and modified data (right). Figure 1. View largeDownload slide ICRP38.DAT original data (left) and modified data (right) for the consideration of mono-energetic electrons. Figure 1. View largeDownload slide ICRP38.DAT original data (left) and modified data (right) for the consideration of mono-energetic electrons. On examination of the original data (left), we see that 7Be has a half-life of 53.3 days and four major emissions per decay. The first in the list is a gamma ray (code 1) with a yield of 10.34% and an energy of 0.477605 MeV. The second and third lines show emissions of conversion electrons (code 6) with very low yields and roughly the same energies as the photon. The final emission is a low-energy X-ray (code 2) with a yield of 1.635%. We cannot simply remove three of the data lines and indicate that 7Be has only one emission; if we did, indices throughout the VARSKIN data files would be in disarray. Therefore, we will set yields and energies to zero (bolded data on the right) except for one entry. The modified entry shows the emission of an electron with 100% yield and an energy of 0.75 MeV. This will simulate the emission of a mono-energetic electron; no other alterations are necessary. The second modification allows the user to change the VARSKIN simulation (normally an air/water interface at the skin surface) to model a water/water interface. We accomplish this by modifying the.rad file (in the dat folder) for the nuclide of choice. The final 160 lines of data written to that file (by SadCalc) are backscatter factors for use in the half-space scenario (air/water interface). Manually setting each of those factors to ‘1’ removes consideration of half-space such that VARSKIN will model a homogeneous water medium (a water/water interface). RESULTS AND DISCUSSION We first examine VARSKIN 5.3 shallow dose predictions from exposure to mono-energetic electrons. Faw calculates absorbed dose to the skin at various depths using the CYLTRON and TIGER codes, Monte Carlo simulations with the Integrated Tiger series(13, 15, 16). He calculates dose to the skin for isotropic plane sources with mono-energetic electrons of 0.25 and 3.0 MeV. For 0.25 MeV electrons, Faw uses the TIGER code to simulate the source on the surface of skin, reportedly backed by air (i.e. air/water interface ‘with backscatter’) and then backed by a vacuum (i.e. vacuum/water interface ‘without backscatter’)(13). Skin dose estimates for the water/water and the air/water scenarios using VARSKIN 5.3 appear to contradict Faw’s explanation of ‘with and without backscatter’. A simple execution of MCNP, using a point source of 0.25 MeV electrons at a water interface backed in one case by air and in the other case by water, indicates that ‘without backscatter’ refers to the air/water scenario (Figure 2a) and ‘with backscatter’ refers to the water/water scenario (Figure 2b). Under these conditions, the VARSKIN data and Faw’s data are in extremely good agreement. Figure 2. View largeDownload slide VARSKIN 5.3 compared to Faw for 0.25 MeV mono-energetic electrons as a function of depth in tissue(13). (a) The electron source is assumed to be on the surface of skin with air above the skin (air/water interface). (b) The electron source is assumed to be on the surface of skin with water above the skin (water/water interface). Figure 2. View largeDownload slide VARSKIN 5.3 compared to Faw for 0.25 MeV mono-energetic electrons as a function of depth in tissue(13). (a) The electron source is assumed to be on the surface of skin with air above the skin (air/water interface). (b) The electron source is assumed to be on the surface of skin with water above the skin (water/water interface). For 3 MeV electrons (Figure 3), estimates from VARSKIN compare very well with CYLTRON(13), in an infinite water medium (water/water interface). VARSKIN dose estimates for the air/water interface are also in the figure, and indicate that the characteristics of the interface relative to electron backscatter are relatively unimportant for high-energy electrons beyond the shallow depths. Figure 3. View largeDownload slide VARSKIN 5.3 compared to Faw for 3 MeV mono-energetic electrons in an infinite water medium (w/w) and with an air/water (a/w) interface(13). Figure 3. View largeDownload slide VARSKIN 5.3 compared to Faw for 3 MeV mono-energetic electrons in an infinite water medium (w/w) and with an air/water (a/w) interface(13). Additionally, Faw calculates average electron dose as a function of energy at three volumetric depths (Figure 4), again for cases with reflection (water/water interface) and without reflection (air/water interface)(13). Skin dose estimates from VARSKIN 5.3 for the same scenario of an infinite plane source and an air/water interface, indicate that VARSKIN is a very good predictor of mono-energetic electron dose at various shallow depths over a wide range of electron energies. For electrons greater than ~2 MeV and depths between 5 and 50 mg/cm2, however, VARSKIN seems to over predict Faw by as much as 20%, possibly due to the method of handling backscatter at high energies. Figure 4. View largeDownload slide Dose per electron calculated by VARSKIN 5.3 compared with Faw at three averaging depths in tissue (3–5, 5–10 and 30–50 mg/cm2). The FAW data are assumed to consider dose for an air/water interface (see text)(13). Figure 4. View largeDownload slide Dose per electron calculated by VARSKIN 5.3 compared with Faw at three averaging depths in tissue (3–5, 5–10 and 30–50 mg/cm2). The FAW data are assumed to consider dose for an air/water interface (see text)(13). Rohloff and Heinzelmann, using their own Monte Carlo code, estimated electron dose to the skin for various beta-emitting nuclides over a wide range of average electron energy (35S, 14C, 147Pm, 45Ca, 60Co, 131I, 90Sr, 204Tl, 18F, 198Au, 89Sr, 32P and 90Y). They speak of calculating dose to skin, ‘both with and without tracking the beta particles backscattered from the surrounding air’(19, 20). Comparison with VARSKIN (Figure 5a and b) for an average energy between 49 keV (35S) and 0.935 MeV (90Y) and an air/water interface shows very good agreement at low energy, but indicates a consistent under prediction by as much as 10% from VARSKIN at the highest doses. Figure 5. View largeDownload slide (a) VARSKIN 5.3 compared to Rohloff (air/water interface) at 7 mg/cm2(19, 20). (b) The same comparison data plotted as a function of average beta energy. Figure 5. View largeDownload slide (a) VARSKIN 5.3 compared to Rohloff (air/water interface) at 7 mg/cm2(19, 20). (b) The same comparison data plotted as a function of average beta energy. Kocher and Eckerman estimated electron dose–rate factors for mono-energetic sources on the skin. They used the scaled point-kernel methods of Berger, inherently in a homogeneous water medium(7, 9). Comparison electron dose–rate factors for skin are duplicated in VARSKIN 5.3 (Figure 6). The Kocher/Eckerman dose factors were simulated in VARSKIN by assuming an infinitely large 2D disk source (15 cm2) on the skin surface with a uniformly distributed activity of 1 Bq/cm2 and all backscatter factors modified to unity (to mimic the water/water interface), as described above. Figure 6. View largeDownload slide Electron dose–rate factors from VARSKIN 5.3 (solid lines) compared to Kocher and Eckerman(9) (data points) with a water/water interface at three depths (4, 8 and 40 mg/cm2) for 13 radionuclides: 99mTc, 111In, 123I, 201Tl, 67Ga, 35S, 14C, 45Ca, 60Co, 131I, 90Sr, 18F and 32P. Figure 6. View largeDownload slide Electron dose–rate factors from VARSKIN 5.3 (solid lines) compared to Kocher and Eckerman(9) (data points) with a water/water interface at three depths (4, 8 and 40 mg/cm2) for 13 radionuclides: 99mTc, 111In, 123I, 201Tl, 67Ga, 35S, 14C, 45Ca, 60Co, 131I, 90Sr, 18F and 32P. We also compare the Kocher and Eckerman dose–rate factors to VARSKIN 5.3 (Figure 7) for specific nuclides at four depths beneath a water/water interface, including a composite of 13 nuclides from Table 1(9). The comparisons (Figures 6 and 7) show very good agreement between VARSKIN and the Kocher/Eckerman data for a homogeneous water medium. Figure 7. View largeDownload slide Skin beta dose–rate factors for VARSKIN 5.3 compared to Kocher and Eckerman(9) at four depths (4, 7, 8 and 40 mg/cm2) for 13 radionuclides: 99mTc, 111In, 123I, 201Tl, 67Ga, 35S, 14C, 45Ca, 60Co, 131I, 90Sr, 18F and 32P. Figure 7. View largeDownload slide Skin beta dose–rate factors for VARSKIN 5.3 compared to Kocher and Eckerman(9) at four depths (4, 7, 8 and 40 mg/cm2) for 13 radionuclides: 99mTc, 111In, 123I, 201Tl, 67Ga, 35S, 14C, 45Ca, 60Co, 131I, 90Sr, 18F and 32P. Chung et al. simulated point sources on the skin using a 2D Monte Carlo transport code, Eltran3, to estimate electron dose from six nuclides (and considered a seventh ‘nuclide’ as the sum of Sr-90 and Y-90, in equilibrium) to a depth of 7 mg/cm2(10, 11). VARSKIN 5.3 results are shown to be in good agreement with their data (Figure 8) for both air/water and water/water simulations. Figure 8. View largeDownload slide Point-source comparison between VARSKIN 5.3 and Eltran3(10) of beta/electron dose per transition (pGy/nt) at a depth of 7 mg/cm2 for six nuclides (and the sum of 90Sr/Y) with a boundary interface of both air/water and water/water. Figure 8. View largeDownload slide Point-source comparison between VARSKIN 5.3 and Eltran3(10) of beta/electron dose per transition (pGy/nt) at a depth of 7 mg/cm2 for six nuclides (and the sum of 90Sr/Y) with a boundary interface of both air/water and water/water. Covens et al.(14) have estimated skin dose at four depths for several nuclides using MCNPX. They assume an air/water interface with surface contamination areas normalized over 1 cm2. We compare results from similar calculations of seven nuclides using VARSKIN 5.3 (Figure 9). The Covens data and VARSKIN are in very good agreement. Figure 9. View largeDownload slide VARSKIN 5.3 compared to Covens et al.(14) using MCNPX at four different depths (7, 14, 22 and 37 mg/cm2) in tissue. Doses calculated with an air/water interface for 67Ga, 99mTc, 111In, 117mSn, 123I, 195mPt and 201Tl. Figure 9. View largeDownload slide VARSKIN 5.3 compared to Covens et al.(14) using MCNPX at four different depths (7, 14, 22 and 37 mg/cm2) in tissue. Doses calculated with an air/water interface for 67Ga, 99mTc, 111In, 117mSn, 123I, 195mPt and 201Tl. For an air/water interface, Taylor et al.(12) calculate dose for mono-energetic electrons and then multiply by electron emission spectra to estimate dose resulting from exposure to particular radionuclides. They used EGS4(8) to simulate beta skin dose for several different nuclides placed on a 26 mg/cm2 cotton cover (ρ = 0.7 g/cm3), with a 2 mm air gap over the skin(12). Average electron energies range from 0.097 (60Co) to 0.695 MeV (32P), with a comparison to VARSKIN (Figure 10) for 60Co, 137Cs, 131I, 90Sr, 85Kr, 11C, 132I, 89Sr and 32P, in order of increasing average electron energy (and generally in order of increasing dose per transition). Since the scenario is unchanged between nuclides, the only influence on shallow dose is the electron emission spectrum generally, where an increase in average energy equates to an increase in shallow skin dose. The results show fair agreement in the lower dose (lower energy) realm, but for a few nuclides of higher average energy (11C, 132I, 89Sr, 32P) VARSKIN under predicts dose by ~30–45%. Figure 10. View largeDownload slide Electron dose at 7 mg/cm2 in VARSKIN 5.3 compared to Taylor et al.(12). The scenario includes a cotton cover with an air gap. Figure 10. View largeDownload slide Electron dose at 7 mg/cm2 in VARSKIN 5.3 compared to Taylor et al.(12). The scenario includes a cotton cover with an air gap. When plotting results as a function of average beta energy (Figure 11), we can see that there is departure between the two estimates of shallow dose at ~250 keV. As surmised by Taylor et al.(12), and supported here, VARSKIN is apparently degrading electron energy more than is appropriate in its cover and air-gap model. Figure 11. View largeDownload slide VARSKIN 5.3 compared to Taylor et al.(12) for shallow dose at 7 mg/cm2 as a function of average beta energy. Figure 11. View largeDownload slide VARSKIN 5.3 compared to Taylor et al.(12) for shallow dose at 7 mg/cm2 as a function of average beta energy. In our last comparison figure, dose estimates from several of the scenarios discussed above are shown in one plot (Figure 12) to convey a general comparison of shallow dose across many different exposure scenarios. We developed the plot with data from seven different manuscripts (Table 1) and VARSKIN 5.3. The VARSKIN 5.3 dose estimates (solid blue line) considered a very simple 1 Bq point source for 23 electron-emitting radionuclides (Table 2) on the skin surface with an air/water interface. Dose was estimated over a 1 cm2 averaging disk at a depth of 7 mg/cm2 for an exposure time of 1 s (to result in units of dose per nuclear transition). We plot shallow dose by nuclide as a function of average beta/electron energy. The VARSKIN dose estimates appear to have discontinuities (at 40, 240 and 580 keV, for example) because of multiple dose values falling around the same average energy (for example, 89Sr and 40K at 0.584 and 0.585 MeV, respectively). Even though two nuclides emit the same average energy, their electron energy distributions are different, resulting in a different skin dose evaluation for those nuclides. Figure 12. View largeDownload slide Normalized shallow beta/electron dose from various scenarios. The data for Taylor w/ cover (dashes) has been modified to account for residual energy loss due to the cover/air layers, resulting in the ‘energy shifted’ data (solid squares). Figure 12. View largeDownload slide Normalized shallow beta/electron dose from various scenarios. The data for Taylor w/ cover (dashes) has been modified to account for residual energy loss due to the cover/air layers, resulting in the ‘energy shifted’ data (solid squares). For each comparison scenario (all data points), we normalized the ordinate around a dose of ~400 pGy/nt as calculated by VARSKIN such that all values fell on the plateau. This normalization was conducted by multiplying dose estimates for a given scenario by a simple scalar value. For example, the data of Rohloff and Heinzelmann(19, 20) were multiplied by a value of 100 for the normalization. An interesting feature of Figure 12 is that with this simple normalization, all scenarios tend to produce the same outcome, i.e. the same graphical shape. Electron shallow dose ranges about an order of magnitude over average electron energies from 16 keV up to ~1 MeV, with the dose plateau reached at ~200 keV. It is also interesting that the discontinuity in dose at ~30–50 keV is observable in all scenarios that track beta dose from low-energy emitters. The scenarios depicted by Sherbini et al. (pt)(4), Covens et al.(14) and Chung et al.(11) simulated point sources on the skin surface; whereas Faw(13), Sherbini et al. (pl)(4) and Kocher/Eckerman(9) simulated planar sources. The Taylor et al.(12) scenario considered a point source on a cotton cover with an air gap. Energy loss through the cotton and air was estimated such that the raw Taylor data (dashes) was shifted to the left (solid squares) based on a calculation of residual energy following electron passage through this material(21). The purpose of consolidating the data in Figure 12 is to demonstrate that VARSKIN’s approach to handling the physics of electron dosimetry is as effective as many of the Monte Carlo simulations to which we have compared. Backscatter, self-absorption, scaling of energy loss and energy deposition are all taken into account. CONCLUSION We have examined the efficacy of the VARSKIN electron dosimetry model for estimating shallow dose following the radiological contamination of skin. For point and planar sources on the skin surface, VARSKIN is shown to be within 10% of other deterministic and probabilistic methods. Discrepancies with other methods appear when VARSKIN models the presence of air-gaps or cover material (e.g. protective clothing). Compared to EGS4 results(12), VARSKIN underestimates by nearly a factor of two in one particular cover scenario with higher-energy electrons. This tends to suggest that energy loss in air and/or cover material is not handled properly in VARSKIN. The calculational structure of the VARSKIN model is shown to be consistent with other methods (Figure 12), even to the extent of predicting skin dose from low-energy electrons. The results presented herein provide additional data for building our confidence in the VARSKIN methods for beta/electron dosimetry. ACKNOWLEDGEMENTS The authors wish to thank Dr Colby Mangini for his insight on VARSKIN and the EGSnrc Monte Carlo code. FUNDING This work was supported by the Nuclear Regulatory Commission under contract NRC-HQ-60-15-R-0017. REFERENCES 1 Hamby , D. M. , Mangini , C. D. , Caffrey , J. A. and Tang , M. VARSKIN 5: A computer code for skin contamination dosimetry. NUREG/CR-6918, Rev. 2. U.S. Nuclear Regulatory Commission. Washington, DC (July 2014 ). 2 Durham , J. S. VARSKIN MOD2 and SADDE MOD2: computer codes for assessing skin dose from skin contamination. NUREG/CR-5873. U.S. Nuclear Regulatory Commission. Washington, DC (December 1992 ). 3 Durham , J. S. VARSKIN 3: a computer code for assessing skin dose from skin contamination. NUREG/CR-6918. U.S. Nuclear Regulatory Commission. Washington, DC (October 2006 ). 4 Sherbini , S. , DeCicco , J. , Gray , A. T. and Struckmeyer , R. Verification of the VARSKIN beta skin dose calculation computer code . Health Phys. 94 ( 6 ), 527 – 538 ( 2008 ). Google Scholar Crossref Search ADS PubMed 5 Traub , R. J. , Reece , W. D. , Scherpelz , R. I. and Sigalia , L. A. Dose calculation for contamination of the skin using the computer code VARSKIN. NUREG/CR/4418. US Nuclear Regulatory Commission. Washington, DC (August 1987 ). 6 Hamby , D. M. , Lodwick , C. J. , Palmer , T. S. , Reese , S. R. and Higley , K. A. VARSKIN 4: a computer code for skin contamination dosimetry. NUREG/CR-6918, Rev. 1. U.S. Nuclear Regulatory Commission. Washington, DC (June 2011 ). 7 Berger , M. J. Distributions of absorbed dose around point sources of electrons and beta particles in water and other media. (MIRD Pamphlet No. 7, Supplement No. 5) . J. Nucl. Med. 12 ( 5 ), 5 – 23 ( 1971 ). Google Scholar PubMed 8 Nelson , W. R. , Hirayama , H. and Rogers , D. W. The EGS4 code-system. Report no. SLAC-265. Stanford Linear Accelerator Center. Stanford University. Stanford, CA ( 1985 ). 9 Kocher , D. C. and Eckerman , K. F. Electron dose-rate conversion factors for external exposure of the skin from uniformly deposited activity on the body surface . Health Phys. 53 ( 2 ), 135 – 141 ( 1987 ). Google Scholar Crossref Search ADS PubMed 10 Chung , M. , Foderaro , A. H. , Jester , W. A. and Levine , S. H. Microcomputer Monte Carlo electron transport codes for beta skin dose calculations . IEEE Trans. Nucl. Sci. 18 ( 3 ), 936 – 941 ( 1991 ). Google Scholar Crossref Search ADS 11 Chung , M. , Levine , S. H. and Jester , W. A. Monte Carlo calculation and silicon detector measurement of the hot particle dose . Health Phys. 61 ( 6 ), 843 – 848 ( 1991 ). Google Scholar Crossref Search ADS PubMed 12 Taylor , D. C. , Hussein , E. M. A. and Yuen , P. S. Skin dose from radionuclide contamination on clothing . Health Phys. 72 ( 6 ), 835 – 841 ( 1997 ). Google Scholar Crossref Search ADS PubMed 13 Faw , R. E. Absorbed dose to skin from radionuclide sources on the body surface . Health Phys. 63 ( 4 ), 443 – 448 ( 1992 ). Google Scholar Crossref Search ADS PubMed 14 Covens , P. , Berus , D. , Caveliers , V. , Struelens , L. , Vanhavere , F. and Verellen , D. Skin dose rate conversion factors after contamination with radiopharmaceuticals: influence of contamination area, epidermal thickness and percutaneous absorption . J. Radiol. Prot. 33 , 381 – 393 ( 2013 ). Google Scholar Crossref Search ADS PubMed 15 Halbleib , J. A. and Melhorn , T. A. ITS: Integrated TIGER series of coupled electron/photon Monte Carlo transport codes. Report no. SAND 84–0573. Sandia National Laboratories. Albuquerque, NM ( 1984 ). 16 Halbleib , J. A. , Kensek , R. P. , Valdez , G. D. , Melhorn , T. A. , Seltzer , S. M. and Berger , M. J. ITS Version 3.0: the integrated TIGER series of coupled electron/photon Monte Carlo codes . IEEE Trans. Nucl. Sci. 39 , 1025 – 1030 ( 1992 ). Google Scholar Crossref Search ADS 17 X-5 Monte Carlo team . MCNP: a general Monte Carlo N-particle transport code. Version 5. Report No. LA-CP-03–0245, Los Alamos National Laboratory. Los Alamos, NM ( 2003 ). 18 Pelowitz , D. B. Ed. MCNPXTM User’s Manual. Version 2.6.0. Report No. LA-CP-07-1473 ( Los Alamos, NM : Los Alamos National Laboratory ) ( 2008 ). 19 Rohloff , F. A Monte Carlo program for the transport of beta particles through layers of different materials. In: Proceedings of the International Beta Dosimetry Symposium. February 15–18, 1983. Washington, DC. NUREG/CP-0050; p. 25 ( 1984 ). 20 Rohloff , F. and Heinzelmann , M. Calculation of dose rates for skin contamination by beta radiation . Radiat. Prot. Dosim. 14 ( 4 ), 279 – 287 ( 1986 ). Google Scholar Crossref Search ADS 21 Shultiz , J. K. and Faw , R. E. Radiation Shielding ( La Grange Park, IL : American Nuclear Society ) ( 2000 ) ISBN 0-89448-456-7. © The Author(s) 2018. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Journal

Radiation Protection DosimetryOxford University Press

Published: Oct 1, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off