Society for Radiological Protection
Journal of Radiological Protection
J. Radiol. Prot. 34 (2014) 347–361
doi:10.1088/0952-4746/34/2/347
Evidence for age-related performance degradation of 241Am foil sources commonly used in UK schools R Whitcher1 , R D Page2 and P R Cole3 1 CLEAPSS, The Gardiner Building, Brunel Science Park, Uxbridge UB8 3PQ,
UK 2 Nuclear Physics Group, Department of Physics, University of Liverpool, L69 3BX, UK 3 Radiation Protection Office, University of Liverpool, L69 3BX, UK E-mail: [email protected] Received 6 October 2013, revised 27 January 2014 Accepted for publication 30 January 2014 Published 4 April 2014 Abstract
The characteristics of alpha radiation have for decades been demonstrated in UK schools using small sealed 241 Am sources. There is a small but steady number of schools who report a considerable reduction in the alpha count rate detected by an end-window GM detector compared with when the source was new. This cannot be explained by incorrect apparatus or set-up, foil surface contamination, or degradation of the GM detector. The University of Liverpool and CLEAPSS collaborated to research the cause of this performance degradation. The aim was to find what was causing the performance degradation and the ramifications for both the useful and safe service life of the sources. The research shows that these foil sources have greater energy straggling with a corresponding reduction in spectral peak energy. A likely cause for this increase in straggling is a significant diffusion of the metals over time. There was no evidence to suggest the foils have become unsafe, but precautionary checks should be made on old sources. Keywords: radioactivity, foil sources, americium-241, americium
1. Introduction
Small sealed metal foil sources have been used for decades in UK schools to demonstrate the characteristics of alpha, beta and gamma radiation, including the demonstration that alpha radiation is blocked by a sheet of paper. In the past few years CLEAPSS4 has received a 4 CLEAPSS (the name originating from Consortium of Local Educational Authorities for the Provision of Science
Services) is an advisory service providing support in science and technology for a consortium of local authorities and their schools in England, Wales and Northern Ireland. 0952-4746/14/020347+15$33.00
c 2014 IOP Publishing Ltd
Printed in the UK
347
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
Figure 1. Diagram of source holder and foil.
small but steady number of reports from schools of older 241 Am foil sources where the alpha count rate detected by an end-window GM detector, typically the ZP1481 GM detector manufactured by Centronic, has dropped considerably compared with when the source was new. This manifests when paper is placed between the source and GM detector window to demonstrate the blocking of alpha radiation. The paper has little effect as if there were no significant alpha emission, only gamma emission, from the source. This could not be explained by incorrect apparatus or set-up, foil surface contamination, or degradation of the GM detector. Each of these possible causes was eliminated by careful examination of the source surface and set-up, and substitution of the GM detector with a known working detector. In contrast, the sources still performed as expected when used with spark detectors and open ionisation chambers and showed the blocking of alpha radiation by paper with no problem. The University of Liverpool and CLEAPSS collaborated to research the cause of this performance degradation. The object of the study was to determine the cause and assess the ramifications for both the useful and safe service life of the sources. 2. Source design
The most common source design is the ‘cup-style’ which has changed little since its introduction in the early 1960s. The radioactive foil, typically 185 kBq, is mounted in a small plated brass cup with a spigot at the back of the cup to enable the source to be handled with a tool. The foil is held in place by a circlip and in some also with epoxy adhesive. The foil is protected by a wire mesh at the front held by a metal cap that is pressed on. The cap press fit is tight and not designed to be removed. See figure 1. The recess from the front of the cup to the face of the foil is about 4 mm. The radionuclide in the alpha foil is usually 241 Am (although 239 Pu and 226 Ra have also been used). 2.1. Construction of the foils
The original manufacturer of the active foil was Atomic Energy Authority Radiochemical Centre. This has changed ownership many times, becoming Amersham, then NycomedAmersham. Currently the sealed-source manufacturing is owned by Eckert and Ziegler, GmbH. The foil production method was described by the National Radiological Protection Board (NRPB, now part of Public Health England) in their report NRPB-R28 (Williams 1974) and again in an Amersham instruction document (Amersham 1991). According to correspondence with the UK distributor, High Technology Sources Ltd (HTSL) this is still the current method used for the production of 241 Am foils. The NRPB report explains that the radioactive compound (AmO2 in this case) is mixed with gold and sintered at high temperature into a briquette. The sintered block is encased in silver with one side open. A layer of gold or gold–palladium alloy is forged onto the open side and the active material is now fully 348
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
encapsulated; this layer becomes the source face. There are several formulations used for the source face, the common face materials are gold–palladium alloys that comprise about 90% palladium. The briquette is cold-rolled to produce a sheet foil. The foil thickness is 0.15–0.2 mm thick, comprising a surface layer, beneath which is the active matrix layer, and then a thicker silver substrate. The NRPB gives the thickness of the active matrix as roughly between 1 and 2 µm thick and the source face up to 3 µm thick. The later Amersham instruction document states it differently: the active matrix is given as ∼1 µm and the face thickness as ∼2 µm. The thin face layer allows the penetration by alpha radiation while retaining the active material bound in the gold active matrix. Discs of diameter about 10 mm are stamped from the foil—the stamping effectively cold-welds the edge of the disc—and the foil disc is mounted into the plated brass holder. The manufacturer’s data state the activity of the disc is 185 kBq ± 20%. The specific activity of 241 Am oxide (AmO2 ) is 111 GBq g−1 , so it comprises about 0.1% of the active matrix assuming the active matrix is 10 mm diameter and 1 µm thick. The alpha radiation will lose energy in penetrating the foil face layer and there will be a spread in the energy of the alpha particles as they emerge from the face. This is termed energy straggling, and a measure of straggle is the full width half maximum (FWHM)—the energy change between half the spectral peak value either side of the peak. As shown by Comfort et al (1966) and others, the spectral peak alpha energy reduces and the energy straggling increases as the thickness of the material through which the alpha radiation travels increases. From information supplied by HTSL, the foils designed for the cup sources are designed to give a spectral peak alpha emission of 4.5 MeV to a tolerance of 10%, and with energy straggling FWHM not exceeding 0.5 MeV. HTSL also believed this was the specification of 241 Am foils made in previous decades but there are no records available to verify this. 2.2. Americium-241 as an alpha source for teaching
Americium-241 has a useful long half-life (432.6 years) and predominant alpha spectral peak emissions with high enough energy—5.486 MeV 84.8%, 5.443 MeV 13.1% and 5.388 1.66% (National Nuclear Data Center)—to be detected by a thin window GM detector. The decay product, 237 Np, has a half-life of 2.144 × 106 years so the decay chain products in the source during its service life are negligible. However, 241 Am is not an ideal radionuclide to demonstrate the characteristics of alpha radiation because there are significant gamma and x-ray photons in the radioactive decay (x-rays arising from internal conversion) which will be detected by a GM detector such as the ZP1481. This does not affect the use of 241 Am with school spark counters and open ionisation chambers which are relatively insensitive to gamma and x-ray photons, but it does affect the demonstration of blocking alpha emissions with paper in front of a GM detector window. 239 Pu would be better than 241 Am in this respect, but the plutonium foil cup source has not been available since about 1980. There are conversion electrons and Auger electrons emitted in the 241 Am decay, but mainly below 55 keV and few are likely to be detected by the ZP1481; the combined areal density of the foil surface layer, the air gap and the GM detector window will amount to at least 5 mg cm−2 (refer to electron ranges in Radiological Health Handbook 1970). 3. Methods and results 3.1. Selection and inspection of sources
Schools were contacted to obtain sources that had been reported as showing the reduced detection rate where the school was confident that the source had performed satisfactorily in 349
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
Table 1. Americium-241 sources used in this study.
Source identifier
Date purchased
Original supplier
Nominal activity (kBq)
Front face (from surface colour)
Reported showing reduced count rate
S1
1990
185
Gold
Yes
S2
∼1980
185
Gold/palladium
Yes
S3
1987
185
Gold/palladium
No
S4
∼1980
185
Gold/palladium
No
S5
2006
185
Gold/palladium
No
S6
2006
185
Gold/palladium
No
S7
2013 (new)
Philip Harris Griffin and George Griffin and George Phillip Harris Philip Harris Philip Harris HTSL
185
Gold/palladium
No
the past. The aim was to obtain sources to investigate this phenomenon. A school was also contacted that had relatively new 241 Am cup sources. CLEAPSS holds cup sources for training purposes and one 241 Am source was found to show a similar low alpha count rate using a GM detector. From these, six sources were selected (see table 1); S1 and S2 were reported showing the reduced alpha detection rate, S3, S4, S5 and S6 were not reported as showing it. S5 and S6 were newer, purchased in 2006. Additionally, source S7 was a new source loaned by HTSL. This had the same foil as supplied in school cup sources, but it had a different mount. 3.2. The condition of the sources
The sources were checked for integrity by a dry-wipe leak test. The accessible surface of each source was wiped with a small piece of dry cellulose filter paper. The paper was then tested for contamination over a 100 s counting period with a contamination monitor. No contamination was found. The faces of the foils were carefully examined using a digital camera with a macro lens, 12 megapixel resolution. The faces were highly reflective with no visual evidence of surface degradation or any coating. See figure 2 for an example. The foil face on each test source appeared in good condition with no observable surface defects. In source S1, the foil was not quite concentric with the brass cup and a foil edge could just be observed on one side. 3.3. Verification of the reduced detected alpha count rate
The aim was to reproduce the results from the sources reported as showing a reduced alpha count rate, including the low reduction in count rate using a paper block. For sources S1–S6, a test was carried out to measure the variation in detection rate with distance of the foil from the GM detector window, with and without a paper block. This was to look at the way in which the foil alpha-particle count rate falls with distance from the GM detector, and how the count rate reduces with the paper block. S7 was excluded from this test because of an incompatible mount. The sources S1–S6 were each set coaxially to a ZP1481 end-window GM detector (connected to a Philip Harris Digicounter) using a radioactivity bench, and the count recorded over a 100 s interval at various distances from the detector window to the foil front face, first with paper between the source and detector, and then without. The GM detector was operated at 450 V; its protective plastic grill cap was removed for maximum efficiency. (Note that in 350
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
Figure 2. Photograph of the front of source S2 (the source shown is held by forceps). The foil is held in place by a circlip. On some sources there are flat marks on the cup edge where the cap was pressed on in manufacture. Table 2. Photon count rates at 25 mm separation between the ZP1481 GM detector
window and the foil surface, and a comparison with S6 giving a rough indication of comparative source activity. Source identifier
Photon count rate at 25 mm separation (s−1 )
Count rate ratio (compared with S6)
S1 S2 S3 S4 S5 S6
11.47 12.18 8.31 17.49 16.41 16.38
0.70 0.74 0.51 1.06 1.00 1.00
schools it is common to leave the cap on to protect the GM detector.) The cup to GM detector distance was measured with a vernier caliper to an accuracy of 0.1 mm. (The recess of the cup and the recess of the GM window were measured separately with a plastic depth-gauge vernier caliper to avoid damage. These were added to the cup–detector separation to obtain the foil to window separation.) The count rate was corrected for background radiation, but not corrected for detector dead time (about 120 µs). Figure 3 shows the variation in detection rate for photon emissions (gamma radiation and x-rays) with distance between the detector window and foil face. The alpha emissions were blocked using a sheet of paper, 0.1 mm thick. The graphs point to S1, S2 and S3 having a noticeably lower activity than S4, S5 and S6. See table 2. The current manufacturer’s source activity tolerance is specified as 20%, but there is supporting evidence that the tolerance was greater in earlier decades; investigations by Lucas (1966) with 226 Ra foils used in school sources showed a 50% variation from the nominal activity of 185 kBq. The alpha count rate by the GM detector was deduced from the difference between count rates without and with the paper block in front of the detector window (see figure 4). The graph for S5 is a broken line only to distinguish it from the other graphs. 351
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
Graph of photon count rate variation (using paper to block the alpha radiation—paper thickness 0.1 mm) with distance of foil to detector window. The trend lines are drawn to guide the eye; they are not lines of best fit. Figure 3.
For all the sources, the total count rate with the GM detector window close to the source, in the region of a few mm from the front gauze, is dominated by the alpha radiation. The GM detector counting efficiency for alpha particles that penetrate the GM detector window into the active volume is essentially 100%. In contrast, the counting efficiency for photons is low—rarely above a few per cent (Knoll 2010). However, alpha particles have a definite range and beyond this the detected count rate is just from the photon emissions. A large drop in count rate when paper is placed between the source and detector is needed for an effective demonstration of alpha-particle characteristics—the actual alpha count rate is less important. In this respect, S1 and S2 were poor compared with the newer sources at the same source–detector separation; S3 was acceptable. See figure 5 which shows the variation in count rate ratio (without and with paper) with distance of foil to detector window of sources S1 and S2 compared with the newer source S6. (Error bars have been included to indicate the uncertainty arising from the counting period and randomness in the radiation emissions.) Table 3 gives the count rate ratio without and with paper at a GM detector window to foil surface separation of 10 mm (i.e. about 5 mm between the front of the GM detector and the source gauze). 3.4. Gamma spectrometry
A gamma spectral analysis was taken of source S1 to look for any anomalies in radionuclide composition. The energy spectrum was measured with a calibrated high purity germanium 352
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
Figure 4. Graph of alpha-particle count rate with distance of foil to detector window. Table 3. Ratio of count rate without and with the paper block at a GM window to foil
separation of 10 mm. Source identifier
Ratio of count rate without and with paper (at 10 mm)
S1 S2 S3 S4 S5 S6
3.8 11.8 23.1 34.0 34.8 32.0
(HPGe) liquid nitrogen cooled detector connected to an Ortec ASPec-927 16k multichannel buffer, and calibrated at two photon energies with a 74 kBq 152 Eu calibration source, using the 121.7817 and 1408.006 keV gamma emissions. The 241 Am spectrum from the school source was compared with a 185 kBq 241 Am reference source, see figure 6. The comparison is shown over the energy range 8–70 keV. For the sampling period chosen there were no significant characteristic spectral peaks above this energy from the sources. The difference in spectral peak structures in the 20–30 keV part of the spectrum is accounted for by the x-ray fluorescence from source S1. The 241 Am emissions cause x-ray fluorescence from the silver in the substrate, the average energy for Kα transitions for silver being 22.1 keV and for Kβ transitions, 25.0 keV (Deslattes et al 2005). 353
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
Figure 5. Count rate ratio (without and with paper) with distance of foil to detector
window for sources S1 and S2 compared with the newer source S6. The paper thickness was 0.1 mm. 3.5. Alpha spectrometry
A possible cause of the sources S1 and S2 producing the lower alpha count rates is energy attenuation of the alpha emissions. To investigate this, the alpha energy spectrum of the sources was analysed using a Canberra 7401 system with a new 450 mm2 Canberra A450-18AM Alpha Analyst PIPS detector connected to an Ortec ASPec-927 16k multichannel buffer. This was calibrated with an open mixed alpha reference disc source (233 U, 238 Pu, 239 Pu and 244 Cm, activity about 100 Bq. The alpha energies used for calibration were the 244 Cm 5804.77 keV spectral peak and the 233 U 4824.2 keV spectral peak). The energy spectra from the cup sources are shown in figure 7. A distribution curve was drawn through data points using a moving average (over a 10 keV interval) and Bezier curve to obtain a good fit. The graphs for S4 and S5 are broken lines only to distinguish them from the other graphs. The larger energy straggling and lower spectral peak energy can clearly be seen for sources S1 and S2. See table 4. Just discernible was a small peak at the higher end of S4 graph, which indicated trace levels of 241 Am at or very near the foil surface. 354
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
Figure 6. Comparison of gamma spectra from an 241 Am reference source and source
S1 (count normalised to the same total).
Figure 7. Alpha-particle energy spectra of foil sources (count normalised to same total). Table 4. Spectral peak alpha energy of the sources in this study.
Source identifier
Peak energy (MeV)
FWHM (MeV)
S1 S2 S3 S4 S5 S6 S7
3.77 4.00 4.15 4.54 4.70 4.68 4.39
0.60 0.58 0.41 0.69 0.24 0.25 0.39
The spectral peak energies of S1 and S2 are outside of the 10% tolerance of 4.5 MeV spectral peak alpha energy. S1, S2 and S4 have a FWHM greater than 0.5 MeV. 355
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
Figure 8. SRIM simulation with uniform distribution and normal distribution of 241 Am, compared with measured straggle, for source S7. Solid line—measured
straggle. Dashed line—calculated straggle based on normal distribution. Dotted line—calculated straggle based on uniform distribution. 3.6. Monte Carlo modelling of the energy straggling
To gauge the thickness of the layers in the sources, S1 and S2 in particular, the Monte Carlo ion transport program SRIM (Ziegler et al 2008) was used to calculate the energy straggle of various configurations of foil layers. Many simulations were carried out with different thickness values for the active matrix layer and face layer, starting with the foil manufacturer’s indicative data. The results were compared to the measured data to find good fits and the best fit layer thickness values were recorded in table 6. Each simulation used 104 ‘emissions’; the energies of the emissions were set in the same values and proportions as the three principal alpha energies of 241 Am. The initial emission point was randomly selected within the active matrix and the initial direction it travelled was randomly selected from a uniform distribution of directions in a cone of apex angle 30◦ towards the detector, simulating the geometry of the detected alpha paths through the foil. When the initial position of the alpha emission in the active matrix was randomly selected to model a uniform distribution of 241 Am, the comparison between the calculated and measured energy straggles was poor. However, when the initial position was randomly selected to model a normal distribution of 241 Am within the active matrix, in a direction normal to the face of the foil with the maximum concentration at the centre of the layer, the comparison was much better. Figure 8 shows as an example the new source, S7. The input data used for the SRIM simulations are given in table 5. One of the outcomes of the simulations points to the concentration of 241 Am through the active matrix layer becoming more concentrated towards the layer centre during the sintering and rolling process. Figure 9 shows the calculated energy straggle by SRIM compared with the measured straggle based on a normal random distribution of 241 Am in the active matrix. The input data used for the SRIM simulations are given in table 6. S1 and S2 are the sources reported as showing the reduced detection count rate, S6 and S7 are newer sources not exhibiting this effect. The input data for layer thicknesses of S6 and S7 broadly agree with the manufacturer’s indicative data. However, for S1 and S2, while the face layer thicknesses are in agreement with 356
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
Table 5. SRIM input for calculated straggle in figure 8.
Graph
Active matrix thickness
Distribution of 241 Am in the active matrix
Dotted
2.0 µm Au
Uniform distribution
Dashed
2.0 µm Au
Normal distribution, 99.9% within matrix thickness
Face layer 1.5 µm 90% Pd, 10% Au 1.5 µm 90% Pd, 10% Au
Figure 9. SRIM calculation with normal random distribution of 241 Am, compared with
measured straggle. The SRIM calculated straggle is shown in dotted lines. Table 6. SRIM input for calculated energy straggle shown in figure 9.
Source identifier
Active matrix thickness
Distribution of 241 Am in the active matrix
S1
3.1 µm Au
S2
2.5 µm Au
S3
2.0 µm Au
S4
4.0 µm Au
S5 (for clarity, not shown because of similarity to S6) S6
1.2 µm Au
Normal, 99.9% layer thickness Normal, 99.9% layer thickness Normal, 99.9% layer thickness Normal, 99.9% layer thickness Normal, 99.9% layer thickness
1.25 µm Au
S7
2.0 µm Au
Face layer thickness
within
2.0 µm 100% Au
within
2.0 µm 90% Pd, 10% Au
within
2.0 µm 90% Pd, 10% Au
within
0.2 µm 90% Pd, 10% Au
within
1.25 µm 90% Pd, 10% Au
Normal, 99.9% within layer thickness Normal, 99.9% within layer thickness
1.25 µm 90% Pd, 10% Au 1.5 µm 90% Pd, 10% Au
the manufacturer’s indicative data, the active matrix layers are quite thicker, 3.1 µm and 2.5 µm respectively. Caution is needed in the use of these SRIM results because of the assumptions made in the input data and simulation: the layer thicknesses were assumed uniform, the density of the sintered gold was assumed to be that of ordinary solid gold, the gold–palladium alloy was assumed to be 90% Pd, 10% gold, and the alloy was modelled as two separate layers. 357
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
Table 7. Data for calculating depth of S1 face layer using equation (1).
Counts from Lβ1-6 (MCA channels 102–126). Mean energy 17.4364 keV Counts from 59.5409 keV gamma peak (MCA channels 403–425) Mass attenuation coefficients (cm2 g−1 ) Density (g cm−3 ) of face layer Thickness of face layer (cm)
Calibration source (source A)
Source S1 (source B)
7810
5982
94 296
153 642
µ1 = 0.193 089 µ2 = 0.842 940 1.18
µ3 = 4.653 49 µ4 = 122.345 19.5
0.1
The validity of the SRIM modelling was checked with the evidence from the S1 source photon energy spectrum. Refer to the earlier diagram figure 6. The Np L x-ray spectrum photons from the 241 Am decay (which follow from the emission of conversion electrons) will be attenuated by the foil face gold layer to a greater degree than the 59.5409 keV gamma spectral peak energy photons For two sources A and B, let Ra and Rb be the corresponding ratios of the E 1 energy photon count rate to the E 2 energy photon count rate by a detector. Let µ1 and µ2 be the mass attenuation coefficients at E 1 and E 2 for source A face layer material of density ρ1 and thickness d1 , and similarly let µ3 and µ4 be the corresponding mass attenuation coefficients for source B face layer material of density ρ2 and thickness d2 . Assuming narrow beam geometry, then: Ra e(µ1 −µ2 )ρ1 d1 = Rb e(µ3 −µ4 )ρ2 d2 .
(1)
Using the data in table 7, this gave a mean depth of 241 Am beneath the foil surface of 3.66×10−4 cm, i.e. 3.66 µm, which compares to the estimated midpoint by SRIM of 3.55 µm for the active matrix layer. The calibration source is a sealed source encased within a polymethyl methacrylate protective layer. The mass attenuation coefficient data was interpolated from data retrieved from the NIST database (Hubbell and Seltzer 2004). The x-ray energies for the x-ray photons were retrieved from Ekstr¨om and Firestone (2004) online database. 4. Discussion and conclusion
The reduction in spectral peak energy for sources S1 and S2 explains why schools are observing sources where the alpha count rate detected by their end-window GM detector has dropped considerably compared with when the source was new; much of the emitted alpha radiation will not penetrate the GM detector window. Many teachers’ guide books that describe the demonstration of blocking alpha radiation by paper are not specific about the distance between the detector and the front of the 241 Am source, saying place the source close to the detector but not specifying a numerical value. The teachers’ guide books that do give practical details often recommend a distance of around 10 mm between the detector and the front of the cup source, which is unsatisfactory for sources such as S1 and S2. To make matters worse, in a typical demonstration set-up in schools, the protective GM detector cap is normally kept in place, reducing the alpha detection efficiency and preventing the cup getting very close to the GM detector window. 358
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
The ZP1481 has a mica window with an areal density between 2.5 and 3.0 mg cm−2 (Centronic 1997), the thickness being a compromise between detector efficiency and robustness. The alpha energy required to penetrate the window differs slightly from one GM detector to another owing to the variation in the mica areal density. An estimate can be made by modelling the alpha penetration of mica using the SRIM; for mica in the range 2.5–3.0 mg cm−2 the alpha energy needs to be at least 3.0 MeV for the 2.5 mg cm−2 and 3.4 MeV to penetrate the 3.0 mg cm−2 window. In practice there will be a small air gap between the source and window, so the actual minimum energy will be a little higher. The recess design of the cup sources means there will be at least a 4 mm air gap between the foil and window, so for alpha radiation travelling the minimum distance from the source to the detector window, the practical minimum alpha energy for detection is going to be in the range 3.6–3.9 MeV depending on window thickness. There have been enough independent reports from schools of 241 Am foil sources where the alpha count rate detected by a GM tube has dropped considerably compared to when the source was new to be confident that this is not caused by a batch of foils with a reduced energy spectrum at the time of manufacture. The condition of the sample sources and leak tests are evidence the source foils are in good condition, so surface degradation is most unlikely to be a factor. Additionally, a low atomic mass material on the foil surface would not produce the same reduced ratio of the count rate from the Np Lβ x-rays to the count rate from the 59.54 keV gamma emission that was observed. The reduced alpha count rate has been observed both in gold and gold–palladium faces, so the source face composition does not seem to be a determining factor. The energy straggling and reduced spectral peak energy of S1 and S2 is consistent with a net migration of the 241 Am away from the foil face (although it may not necessarily be the americium which has migrated) and an effective increase of the active matrix layer thickness; this is supported by the SRIM calculations. A likely hypothesis for this reduction in spectral peak energy and increase in straggling in some 241 Am foils is a significant diffusion of the metals over time. For example, silver can diffuse through gold films at room temperatures (Liu et al 1985) and diffusion occurs more rapidly when a metal has been cold worked (Dini 1993), which would be from the cold rolling in this case. The research by Liu et al indicates significant diffusion over months. However, the diffusion in the 241 Am foils appears to be a slower rate, over years. The calculations from SRIM predict the migration of material in the active matrix need only be of the order of 1 µm to produce the significant energy degradation that is being seen here. The calculations also point to the active layer in sources S1 and S2 effectively increasing in the direction of the substrate. If so, this does not raise a concern for the safe condition of the sources because the 241 Am will remain within the foil. S4 shows greater energy straggling than the newer sources, but unlike S1 and S2, the spectral peak energy is higher at 4.7 MeV so the increased energy straggle will not greatly affect the blocking of alpha radiation by paper demonstration. S4 was not reported as showing a reduced count, so there is uncertainty whether the energy straggle has actually increased with age. If originally the foil layer dimensions met the manufacturer’s indicative data, it would suggest a net migration of 241 Am away from the original active matrix outward towards the face and towards the substrate of the source. If this suggestion is correct, this would have a bearing on the safe condition of the foil if eventually the americium migrated to the source surface. The small alpha energy peak of 5.48 MeV is from 241 Am near or at the surface, plausibly 241 Am at the site of a microscopic pit in the gold/palladium layer. If there was diffusion of 241 Am to the top surface, the alpha energy distribution curve would be unlikely to drop to zero before 5.48 MeV. Testing the diffusion hypothesis needs further investigation by surface analysis, e.g. by sputtering. However, it would be difficult to remove the foils of the sample sources for surface 359
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
analysis because the foil would likely become damaged by the removal. Should sources be found where the foil can be removed from its holder without damage, this will permit further investigation. The question remains why this increased energy straggling over time appears in some 241 Am foil sources and less so in others. For example, in contrast to sources S1 and S2, the source S3 has a spectral peak energy and FWHM still within the manufacturer’s indicative data. There is a difference in the evidence on the construction of the foils: the Amersham packing note (1991) shows a diagram where there is a gold backing layer between the silver substrate and the active gold matrix, but the NRPB report (Williams 1974) does not mention this. HTSL explained that the foil manufacturer at that time—Amersham—produced a variety of alpha foils that had different ratio of materials in the surface layers, consequently there is today some uncertainty of the composition of the foils used in school sources from that period. It is therefore a reasonable hypothesis that there are variations in foil design of which some are more predisposed to this change in energy straggling over time. If the diffusion hypothesis is correct, Dini (1993) states that while diffusion can improve the adherence between the layers of metal, in contrast, where inter-diffusion between two metals is uneven, it is possible that a weakness can be created that eventually could cause the delamination of the gold or gold–palladium face layer. There is no evidence from the 241 Am foils in this study that there is any delamination of the face. Dini’s statement was in the context of electrodeposition whereas the foil face is hot-forged, but the principle remains.
5. Advice for educational establishments
For schools and other educational establishments that have 241 Am foil sources exhibiting the greater energy straggling and reduced spectral peak energy, there is no evidence currently to suggest the foils have become unsafe, but they may be of limited use as alpha emitters when detected by an end-window GM detector such as the commonly used ZP1481. The sources can still be used with spark counters and ionisation chambers including the demonstration of blocking alpha radiation by paper, because these detectors have no solid window. If the protective cap of the ZP1481 detector is removed and the gap between the source cup and GM detector window is kept as small as practicable, 2 mm or less, then the blocking of alpha radiation by paper demonstration using the GM detector may still be effective if the reduction in spectral peak energy has not become too great. But if it has, the source will need replacing for this application. Care needs to be taken with the source so close to the unprotected GM detector window; the window would be broken easily if the source were to be knocked into it, damaging the detector permanently. Although this study looked at just a small sample of sources, a sensible precautionary measure for this type of foil source would be to carry out leak tests at least yearly with remote inspection of the face to check the surface remains in good condition, and with no discolouration which could be from gold or 241 Am that has migrated through the face layer.
Acknowledgments
The authors would like to thank Oriel High School, Crawley, and St Benedict’s Roman Catholic School, Bury St Edmunds, who loaned their americium sources for this study, and High Technology Sources Ltd (HTSL) for the loan of a new source and the information supplied. 360
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
References Amersham 1991 Safety instructions for unpacking and use of alpha foil sources. HI 041 Issue 1 Centronic 1997 Geiger Muller Tubes Databook (New Addington, UK: Centronic) Comfort J R, Decker J F, Lynk E T, Scully M O and Quinton A R 1966 Energy loss and straggling of alpha particles in metal foil Phys. Rev. 150 249–56 Deslattes R D et al 2005 X-Ray Transition Energies (version 1.2) (Gaithersburg, MD: National Institute of Standards and Technology) [Online] Available at http://physics.nist.gov/XrayTrans (Accessed 19 July 2013) Dini J W 1993 Electrodeposition: The Materials Science of Coatings and Substrates (New Jersey: Noyes Publications) Ekstr¨om L P and Firestone R B 2004 Table of radioactive isotopes, Database Version 2.1 LBNL Isotopes Project—LUNDS Universitet [Online] Available at http://ie.lbl.gov/toi/index.htm (Accessed 10 January 2014) Hubbell J H and Seltzer S M 2004 Tables of X-Ray Mass Attenuation Coefficients and Mass Energy-Absorption Coefficients (version 1.4) (Gaithersburg, MD: National Institute of Standards and Technology) [Online] Available at http://physics.nist.gov/xaamdi (Accessed 21 January 2014) Knoll G F 2010 Radiation Detection and Measurement 4th edn (New York: Wiley) pp 216–8 Liu Y, Wang L, Zhan J and Chui L 1985 An investigation of silver diffusion through gold films by Auger electron spectroscopy Vacuum 35 537–8 Lucas J W 1966 Precautions in the use of radioactive sealed sources in schools School Sci. Rev. ASE 47 19–27 National Nuclear Data Center, Information accessed from the NuDat 2 database, Brookhaven National Laboratory, New York [Online] Available at www.nndc.bnl.gov/nudat2/ (Accessed 23 September www.nndc.bnl.gov/nudat2/ (Accessed 23 September 2013) Radiological Health Handbook 1970 Publication No 2016 (Rockville, MD: Bureau of Radiological Health) p 123 Williams T G 1974 The Use of Plutonium-239 Sources in Schools and Other Educational Establishments NRPB-R28 (Harwell: HMSO) Ziegler J F, Ziegler M D and Biersack J P 2010 SRIM—the stopping and range of ions in matter Nucl. Instrum. Methods Phys. Res. B 268 1818–23
361
Copyright of Journal of Radiological Protection is the property of IOP Publishing and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
Journal of Radiological Protection
J. Radiol. Prot. 34 (2014) 347–361
doi:10.1088/0952-4746/34/2/347
Evidence for age-related performance degradation of 241Am foil sources commonly used in UK schools R Whitcher1 , R D Page2 and P R Cole3 1 CLEAPSS, The Gardiner Building, Brunel Science Park, Uxbridge UB8 3PQ,
UK 2 Nuclear Physics Group, Department of Physics, University of Liverpool, L69 3BX, UK 3 Radiation Protection Office, University of Liverpool, L69 3BX, UK E-mail: [email protected] Received 6 October 2013, revised 27 January 2014 Accepted for publication 30 January 2014 Published 4 April 2014 Abstract
The characteristics of alpha radiation have for decades been demonstrated in UK schools using small sealed 241 Am sources. There is a small but steady number of schools who report a considerable reduction in the alpha count rate detected by an end-window GM detector compared with when the source was new. This cannot be explained by incorrect apparatus or set-up, foil surface contamination, or degradation of the GM detector. The University of Liverpool and CLEAPSS collaborated to research the cause of this performance degradation. The aim was to find what was causing the performance degradation and the ramifications for both the useful and safe service life of the sources. The research shows that these foil sources have greater energy straggling with a corresponding reduction in spectral peak energy. A likely cause for this increase in straggling is a significant diffusion of the metals over time. There was no evidence to suggest the foils have become unsafe, but precautionary checks should be made on old sources. Keywords: radioactivity, foil sources, americium-241, americium
1. Introduction
Small sealed metal foil sources have been used for decades in UK schools to demonstrate the characteristics of alpha, beta and gamma radiation, including the demonstration that alpha radiation is blocked by a sheet of paper. In the past few years CLEAPSS4 has received a 4 CLEAPSS (the name originating from Consortium of Local Educational Authorities for the Provision of Science
Services) is an advisory service providing support in science and technology for a consortium of local authorities and their schools in England, Wales and Northern Ireland. 0952-4746/14/020347+15$33.00
c 2014 IOP Publishing Ltd
Printed in the UK
347
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
Figure 1. Diagram of source holder and foil.
small but steady number of reports from schools of older 241 Am foil sources where the alpha count rate detected by an end-window GM detector, typically the ZP1481 GM detector manufactured by Centronic, has dropped considerably compared with when the source was new. This manifests when paper is placed between the source and GM detector window to demonstrate the blocking of alpha radiation. The paper has little effect as if there were no significant alpha emission, only gamma emission, from the source. This could not be explained by incorrect apparatus or set-up, foil surface contamination, or degradation of the GM detector. Each of these possible causes was eliminated by careful examination of the source surface and set-up, and substitution of the GM detector with a known working detector. In contrast, the sources still performed as expected when used with spark detectors and open ionisation chambers and showed the blocking of alpha radiation by paper with no problem. The University of Liverpool and CLEAPSS collaborated to research the cause of this performance degradation. The object of the study was to determine the cause and assess the ramifications for both the useful and safe service life of the sources. 2. Source design
The most common source design is the ‘cup-style’ which has changed little since its introduction in the early 1960s. The radioactive foil, typically 185 kBq, is mounted in a small plated brass cup with a spigot at the back of the cup to enable the source to be handled with a tool. The foil is held in place by a circlip and in some also with epoxy adhesive. The foil is protected by a wire mesh at the front held by a metal cap that is pressed on. The cap press fit is tight and not designed to be removed. See figure 1. The recess from the front of the cup to the face of the foil is about 4 mm. The radionuclide in the alpha foil is usually 241 Am (although 239 Pu and 226 Ra have also been used). 2.1. Construction of the foils
The original manufacturer of the active foil was Atomic Energy Authority Radiochemical Centre. This has changed ownership many times, becoming Amersham, then NycomedAmersham. Currently the sealed-source manufacturing is owned by Eckert and Ziegler, GmbH. The foil production method was described by the National Radiological Protection Board (NRPB, now part of Public Health England) in their report NRPB-R28 (Williams 1974) and again in an Amersham instruction document (Amersham 1991). According to correspondence with the UK distributor, High Technology Sources Ltd (HTSL) this is still the current method used for the production of 241 Am foils. The NRPB report explains that the radioactive compound (AmO2 in this case) is mixed with gold and sintered at high temperature into a briquette. The sintered block is encased in silver with one side open. A layer of gold or gold–palladium alloy is forged onto the open side and the active material is now fully 348
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
encapsulated; this layer becomes the source face. There are several formulations used for the source face, the common face materials are gold–palladium alloys that comprise about 90% palladium. The briquette is cold-rolled to produce a sheet foil. The foil thickness is 0.15–0.2 mm thick, comprising a surface layer, beneath which is the active matrix layer, and then a thicker silver substrate. The NRPB gives the thickness of the active matrix as roughly between 1 and 2 µm thick and the source face up to 3 µm thick. The later Amersham instruction document states it differently: the active matrix is given as ∼1 µm and the face thickness as ∼2 µm. The thin face layer allows the penetration by alpha radiation while retaining the active material bound in the gold active matrix. Discs of diameter about 10 mm are stamped from the foil—the stamping effectively cold-welds the edge of the disc—and the foil disc is mounted into the plated brass holder. The manufacturer’s data state the activity of the disc is 185 kBq ± 20%. The specific activity of 241 Am oxide (AmO2 ) is 111 GBq g−1 , so it comprises about 0.1% of the active matrix assuming the active matrix is 10 mm diameter and 1 µm thick. The alpha radiation will lose energy in penetrating the foil face layer and there will be a spread in the energy of the alpha particles as they emerge from the face. This is termed energy straggling, and a measure of straggle is the full width half maximum (FWHM)—the energy change between half the spectral peak value either side of the peak. As shown by Comfort et al (1966) and others, the spectral peak alpha energy reduces and the energy straggling increases as the thickness of the material through which the alpha radiation travels increases. From information supplied by HTSL, the foils designed for the cup sources are designed to give a spectral peak alpha emission of 4.5 MeV to a tolerance of 10%, and with energy straggling FWHM not exceeding 0.5 MeV. HTSL also believed this was the specification of 241 Am foils made in previous decades but there are no records available to verify this. 2.2. Americium-241 as an alpha source for teaching
Americium-241 has a useful long half-life (432.6 years) and predominant alpha spectral peak emissions with high enough energy—5.486 MeV 84.8%, 5.443 MeV 13.1% and 5.388 1.66% (National Nuclear Data Center)—to be detected by a thin window GM detector. The decay product, 237 Np, has a half-life of 2.144 × 106 years so the decay chain products in the source during its service life are negligible. However, 241 Am is not an ideal radionuclide to demonstrate the characteristics of alpha radiation because there are significant gamma and x-ray photons in the radioactive decay (x-rays arising from internal conversion) which will be detected by a GM detector such as the ZP1481. This does not affect the use of 241 Am with school spark counters and open ionisation chambers which are relatively insensitive to gamma and x-ray photons, but it does affect the demonstration of blocking alpha emissions with paper in front of a GM detector window. 239 Pu would be better than 241 Am in this respect, but the plutonium foil cup source has not been available since about 1980. There are conversion electrons and Auger electrons emitted in the 241 Am decay, but mainly below 55 keV and few are likely to be detected by the ZP1481; the combined areal density of the foil surface layer, the air gap and the GM detector window will amount to at least 5 mg cm−2 (refer to electron ranges in Radiological Health Handbook 1970). 3. Methods and results 3.1. Selection and inspection of sources
Schools were contacted to obtain sources that had been reported as showing the reduced detection rate where the school was confident that the source had performed satisfactorily in 349
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
Table 1. Americium-241 sources used in this study.
Source identifier
Date purchased
Original supplier
Nominal activity (kBq)
Front face (from surface colour)
Reported showing reduced count rate
S1
1990
185
Gold
Yes
S2
∼1980
185
Gold/palladium
Yes
S3
1987
185
Gold/palladium
No
S4
∼1980
185
Gold/palladium
No
S5
2006
185
Gold/palladium
No
S6
2006
185
Gold/palladium
No
S7
2013 (new)
Philip Harris Griffin and George Griffin and George Phillip Harris Philip Harris Philip Harris HTSL
185
Gold/palladium
No
the past. The aim was to obtain sources to investigate this phenomenon. A school was also contacted that had relatively new 241 Am cup sources. CLEAPSS holds cup sources for training purposes and one 241 Am source was found to show a similar low alpha count rate using a GM detector. From these, six sources were selected (see table 1); S1 and S2 were reported showing the reduced alpha detection rate, S3, S4, S5 and S6 were not reported as showing it. S5 and S6 were newer, purchased in 2006. Additionally, source S7 was a new source loaned by HTSL. This had the same foil as supplied in school cup sources, but it had a different mount. 3.2. The condition of the sources
The sources were checked for integrity by a dry-wipe leak test. The accessible surface of each source was wiped with a small piece of dry cellulose filter paper. The paper was then tested for contamination over a 100 s counting period with a contamination monitor. No contamination was found. The faces of the foils were carefully examined using a digital camera with a macro lens, 12 megapixel resolution. The faces were highly reflective with no visual evidence of surface degradation or any coating. See figure 2 for an example. The foil face on each test source appeared in good condition with no observable surface defects. In source S1, the foil was not quite concentric with the brass cup and a foil edge could just be observed on one side. 3.3. Verification of the reduced detected alpha count rate
The aim was to reproduce the results from the sources reported as showing a reduced alpha count rate, including the low reduction in count rate using a paper block. For sources S1–S6, a test was carried out to measure the variation in detection rate with distance of the foil from the GM detector window, with and without a paper block. This was to look at the way in which the foil alpha-particle count rate falls with distance from the GM detector, and how the count rate reduces with the paper block. S7 was excluded from this test because of an incompatible mount. The sources S1–S6 were each set coaxially to a ZP1481 end-window GM detector (connected to a Philip Harris Digicounter) using a radioactivity bench, and the count recorded over a 100 s interval at various distances from the detector window to the foil front face, first with paper between the source and detector, and then without. The GM detector was operated at 450 V; its protective plastic grill cap was removed for maximum efficiency. (Note that in 350
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
Figure 2. Photograph of the front of source S2 (the source shown is held by forceps). The foil is held in place by a circlip. On some sources there are flat marks on the cup edge where the cap was pressed on in manufacture. Table 2. Photon count rates at 25 mm separation between the ZP1481 GM detector
window and the foil surface, and a comparison with S6 giving a rough indication of comparative source activity. Source identifier
Photon count rate at 25 mm separation (s−1 )
Count rate ratio (compared with S6)
S1 S2 S3 S4 S5 S6
11.47 12.18 8.31 17.49 16.41 16.38
0.70 0.74 0.51 1.06 1.00 1.00
schools it is common to leave the cap on to protect the GM detector.) The cup to GM detector distance was measured with a vernier caliper to an accuracy of 0.1 mm. (The recess of the cup and the recess of the GM window were measured separately with a plastic depth-gauge vernier caliper to avoid damage. These were added to the cup–detector separation to obtain the foil to window separation.) The count rate was corrected for background radiation, but not corrected for detector dead time (about 120 µs). Figure 3 shows the variation in detection rate for photon emissions (gamma radiation and x-rays) with distance between the detector window and foil face. The alpha emissions were blocked using a sheet of paper, 0.1 mm thick. The graphs point to S1, S2 and S3 having a noticeably lower activity than S4, S5 and S6. See table 2. The current manufacturer’s source activity tolerance is specified as 20%, but there is supporting evidence that the tolerance was greater in earlier decades; investigations by Lucas (1966) with 226 Ra foils used in school sources showed a 50% variation from the nominal activity of 185 kBq. The alpha count rate by the GM detector was deduced from the difference between count rates without and with the paper block in front of the detector window (see figure 4). The graph for S5 is a broken line only to distinguish it from the other graphs. 351
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
Graph of photon count rate variation (using paper to block the alpha radiation—paper thickness 0.1 mm) with distance of foil to detector window. The trend lines are drawn to guide the eye; they are not lines of best fit. Figure 3.
For all the sources, the total count rate with the GM detector window close to the source, in the region of a few mm from the front gauze, is dominated by the alpha radiation. The GM detector counting efficiency for alpha particles that penetrate the GM detector window into the active volume is essentially 100%. In contrast, the counting efficiency for photons is low—rarely above a few per cent (Knoll 2010). However, alpha particles have a definite range and beyond this the detected count rate is just from the photon emissions. A large drop in count rate when paper is placed between the source and detector is needed for an effective demonstration of alpha-particle characteristics—the actual alpha count rate is less important. In this respect, S1 and S2 were poor compared with the newer sources at the same source–detector separation; S3 was acceptable. See figure 5 which shows the variation in count rate ratio (without and with paper) with distance of foil to detector window of sources S1 and S2 compared with the newer source S6. (Error bars have been included to indicate the uncertainty arising from the counting period and randomness in the radiation emissions.) Table 3 gives the count rate ratio without and with paper at a GM detector window to foil surface separation of 10 mm (i.e. about 5 mm between the front of the GM detector and the source gauze). 3.4. Gamma spectrometry
A gamma spectral analysis was taken of source S1 to look for any anomalies in radionuclide composition. The energy spectrum was measured with a calibrated high purity germanium 352
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
Figure 4. Graph of alpha-particle count rate with distance of foil to detector window. Table 3. Ratio of count rate without and with the paper block at a GM window to foil
separation of 10 mm. Source identifier
Ratio of count rate without and with paper (at 10 mm)
S1 S2 S3 S4 S5 S6
3.8 11.8 23.1 34.0 34.8 32.0
(HPGe) liquid nitrogen cooled detector connected to an Ortec ASPec-927 16k multichannel buffer, and calibrated at two photon energies with a 74 kBq 152 Eu calibration source, using the 121.7817 and 1408.006 keV gamma emissions. The 241 Am spectrum from the school source was compared with a 185 kBq 241 Am reference source, see figure 6. The comparison is shown over the energy range 8–70 keV. For the sampling period chosen there were no significant characteristic spectral peaks above this energy from the sources. The difference in spectral peak structures in the 20–30 keV part of the spectrum is accounted for by the x-ray fluorescence from source S1. The 241 Am emissions cause x-ray fluorescence from the silver in the substrate, the average energy for Kα transitions for silver being 22.1 keV and for Kβ transitions, 25.0 keV (Deslattes et al 2005). 353
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
Figure 5. Count rate ratio (without and with paper) with distance of foil to detector
window for sources S1 and S2 compared with the newer source S6. The paper thickness was 0.1 mm. 3.5. Alpha spectrometry
A possible cause of the sources S1 and S2 producing the lower alpha count rates is energy attenuation of the alpha emissions. To investigate this, the alpha energy spectrum of the sources was analysed using a Canberra 7401 system with a new 450 mm2 Canberra A450-18AM Alpha Analyst PIPS detector connected to an Ortec ASPec-927 16k multichannel buffer. This was calibrated with an open mixed alpha reference disc source (233 U, 238 Pu, 239 Pu and 244 Cm, activity about 100 Bq. The alpha energies used for calibration were the 244 Cm 5804.77 keV spectral peak and the 233 U 4824.2 keV spectral peak). The energy spectra from the cup sources are shown in figure 7. A distribution curve was drawn through data points using a moving average (over a 10 keV interval) and Bezier curve to obtain a good fit. The graphs for S4 and S5 are broken lines only to distinguish them from the other graphs. The larger energy straggling and lower spectral peak energy can clearly be seen for sources S1 and S2. See table 4. Just discernible was a small peak at the higher end of S4 graph, which indicated trace levels of 241 Am at or very near the foil surface. 354
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
Figure 6. Comparison of gamma spectra from an 241 Am reference source and source
S1 (count normalised to the same total).
Figure 7. Alpha-particle energy spectra of foil sources (count normalised to same total). Table 4. Spectral peak alpha energy of the sources in this study.
Source identifier
Peak energy (MeV)
FWHM (MeV)
S1 S2 S3 S4 S5 S6 S7
3.77 4.00 4.15 4.54 4.70 4.68 4.39
0.60 0.58 0.41 0.69 0.24 0.25 0.39
The spectral peak energies of S1 and S2 are outside of the 10% tolerance of 4.5 MeV spectral peak alpha energy. S1, S2 and S4 have a FWHM greater than 0.5 MeV. 355
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
Figure 8. SRIM simulation with uniform distribution and normal distribution of 241 Am, compared with measured straggle, for source S7. Solid line—measured
straggle. Dashed line—calculated straggle based on normal distribution. Dotted line—calculated straggle based on uniform distribution. 3.6. Monte Carlo modelling of the energy straggling
To gauge the thickness of the layers in the sources, S1 and S2 in particular, the Monte Carlo ion transport program SRIM (Ziegler et al 2008) was used to calculate the energy straggle of various configurations of foil layers. Many simulations were carried out with different thickness values for the active matrix layer and face layer, starting with the foil manufacturer’s indicative data. The results were compared to the measured data to find good fits and the best fit layer thickness values were recorded in table 6. Each simulation used 104 ‘emissions’; the energies of the emissions were set in the same values and proportions as the three principal alpha energies of 241 Am. The initial emission point was randomly selected within the active matrix and the initial direction it travelled was randomly selected from a uniform distribution of directions in a cone of apex angle 30◦ towards the detector, simulating the geometry of the detected alpha paths through the foil. When the initial position of the alpha emission in the active matrix was randomly selected to model a uniform distribution of 241 Am, the comparison between the calculated and measured energy straggles was poor. However, when the initial position was randomly selected to model a normal distribution of 241 Am within the active matrix, in a direction normal to the face of the foil with the maximum concentration at the centre of the layer, the comparison was much better. Figure 8 shows as an example the new source, S7. The input data used for the SRIM simulations are given in table 5. One of the outcomes of the simulations points to the concentration of 241 Am through the active matrix layer becoming more concentrated towards the layer centre during the sintering and rolling process. Figure 9 shows the calculated energy straggle by SRIM compared with the measured straggle based on a normal random distribution of 241 Am in the active matrix. The input data used for the SRIM simulations are given in table 6. S1 and S2 are the sources reported as showing the reduced detection count rate, S6 and S7 are newer sources not exhibiting this effect. The input data for layer thicknesses of S6 and S7 broadly agree with the manufacturer’s indicative data. However, for S1 and S2, while the face layer thicknesses are in agreement with 356
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
Table 5. SRIM input for calculated straggle in figure 8.
Graph
Active matrix thickness
Distribution of 241 Am in the active matrix
Dotted
2.0 µm Au
Uniform distribution
Dashed
2.0 µm Au
Normal distribution, 99.9% within matrix thickness
Face layer 1.5 µm 90% Pd, 10% Au 1.5 µm 90% Pd, 10% Au
Figure 9. SRIM calculation with normal random distribution of 241 Am, compared with
measured straggle. The SRIM calculated straggle is shown in dotted lines. Table 6. SRIM input for calculated energy straggle shown in figure 9.
Source identifier
Active matrix thickness
Distribution of 241 Am in the active matrix
S1
3.1 µm Au
S2
2.5 µm Au
S3
2.0 µm Au
S4
4.0 µm Au
S5 (for clarity, not shown because of similarity to S6) S6
1.2 µm Au
Normal, 99.9% layer thickness Normal, 99.9% layer thickness Normal, 99.9% layer thickness Normal, 99.9% layer thickness Normal, 99.9% layer thickness
1.25 µm Au
S7
2.0 µm Au
Face layer thickness
within
2.0 µm 100% Au
within
2.0 µm 90% Pd, 10% Au
within
2.0 µm 90% Pd, 10% Au
within
0.2 µm 90% Pd, 10% Au
within
1.25 µm 90% Pd, 10% Au
Normal, 99.9% within layer thickness Normal, 99.9% within layer thickness
1.25 µm 90% Pd, 10% Au 1.5 µm 90% Pd, 10% Au
the manufacturer’s indicative data, the active matrix layers are quite thicker, 3.1 µm and 2.5 µm respectively. Caution is needed in the use of these SRIM results because of the assumptions made in the input data and simulation: the layer thicknesses were assumed uniform, the density of the sintered gold was assumed to be that of ordinary solid gold, the gold–palladium alloy was assumed to be 90% Pd, 10% gold, and the alloy was modelled as two separate layers. 357
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
Table 7. Data for calculating depth of S1 face layer using equation (1).
Counts from Lβ1-6 (MCA channels 102–126). Mean energy 17.4364 keV Counts from 59.5409 keV gamma peak (MCA channels 403–425) Mass attenuation coefficients (cm2 g−1 ) Density (g cm−3 ) of face layer Thickness of face layer (cm)
Calibration source (source A)
Source S1 (source B)
7810
5982
94 296
153 642
µ1 = 0.193 089 µ2 = 0.842 940 1.18
µ3 = 4.653 49 µ4 = 122.345 19.5
0.1
The validity of the SRIM modelling was checked with the evidence from the S1 source photon energy spectrum. Refer to the earlier diagram figure 6. The Np L x-ray spectrum photons from the 241 Am decay (which follow from the emission of conversion electrons) will be attenuated by the foil face gold layer to a greater degree than the 59.5409 keV gamma spectral peak energy photons For two sources A and B, let Ra and Rb be the corresponding ratios of the E 1 energy photon count rate to the E 2 energy photon count rate by a detector. Let µ1 and µ2 be the mass attenuation coefficients at E 1 and E 2 for source A face layer material of density ρ1 and thickness d1 , and similarly let µ3 and µ4 be the corresponding mass attenuation coefficients for source B face layer material of density ρ2 and thickness d2 . Assuming narrow beam geometry, then: Ra e(µ1 −µ2 )ρ1 d1 = Rb e(µ3 −µ4 )ρ2 d2 .
(1)
Using the data in table 7, this gave a mean depth of 241 Am beneath the foil surface of 3.66×10−4 cm, i.e. 3.66 µm, which compares to the estimated midpoint by SRIM of 3.55 µm for the active matrix layer. The calibration source is a sealed source encased within a polymethyl methacrylate protective layer. The mass attenuation coefficient data was interpolated from data retrieved from the NIST database (Hubbell and Seltzer 2004). The x-ray energies for the x-ray photons were retrieved from Ekstr¨om and Firestone (2004) online database. 4. Discussion and conclusion
The reduction in spectral peak energy for sources S1 and S2 explains why schools are observing sources where the alpha count rate detected by their end-window GM detector has dropped considerably compared with when the source was new; much of the emitted alpha radiation will not penetrate the GM detector window. Many teachers’ guide books that describe the demonstration of blocking alpha radiation by paper are not specific about the distance between the detector and the front of the 241 Am source, saying place the source close to the detector but not specifying a numerical value. The teachers’ guide books that do give practical details often recommend a distance of around 10 mm between the detector and the front of the cup source, which is unsatisfactory for sources such as S1 and S2. To make matters worse, in a typical demonstration set-up in schools, the protective GM detector cap is normally kept in place, reducing the alpha detection efficiency and preventing the cup getting very close to the GM detector window. 358
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
The ZP1481 has a mica window with an areal density between 2.5 and 3.0 mg cm−2 (Centronic 1997), the thickness being a compromise between detector efficiency and robustness. The alpha energy required to penetrate the window differs slightly from one GM detector to another owing to the variation in the mica areal density. An estimate can be made by modelling the alpha penetration of mica using the SRIM; for mica in the range 2.5–3.0 mg cm−2 the alpha energy needs to be at least 3.0 MeV for the 2.5 mg cm−2 and 3.4 MeV to penetrate the 3.0 mg cm−2 window. In practice there will be a small air gap between the source and window, so the actual minimum energy will be a little higher. The recess design of the cup sources means there will be at least a 4 mm air gap between the foil and window, so for alpha radiation travelling the minimum distance from the source to the detector window, the practical minimum alpha energy for detection is going to be in the range 3.6–3.9 MeV depending on window thickness. There have been enough independent reports from schools of 241 Am foil sources where the alpha count rate detected by a GM tube has dropped considerably compared to when the source was new to be confident that this is not caused by a batch of foils with a reduced energy spectrum at the time of manufacture. The condition of the sample sources and leak tests are evidence the source foils are in good condition, so surface degradation is most unlikely to be a factor. Additionally, a low atomic mass material on the foil surface would not produce the same reduced ratio of the count rate from the Np Lβ x-rays to the count rate from the 59.54 keV gamma emission that was observed. The reduced alpha count rate has been observed both in gold and gold–palladium faces, so the source face composition does not seem to be a determining factor. The energy straggling and reduced spectral peak energy of S1 and S2 is consistent with a net migration of the 241 Am away from the foil face (although it may not necessarily be the americium which has migrated) and an effective increase of the active matrix layer thickness; this is supported by the SRIM calculations. A likely hypothesis for this reduction in spectral peak energy and increase in straggling in some 241 Am foils is a significant diffusion of the metals over time. For example, silver can diffuse through gold films at room temperatures (Liu et al 1985) and diffusion occurs more rapidly when a metal has been cold worked (Dini 1993), which would be from the cold rolling in this case. The research by Liu et al indicates significant diffusion over months. However, the diffusion in the 241 Am foils appears to be a slower rate, over years. The calculations from SRIM predict the migration of material in the active matrix need only be of the order of 1 µm to produce the significant energy degradation that is being seen here. The calculations also point to the active layer in sources S1 and S2 effectively increasing in the direction of the substrate. If so, this does not raise a concern for the safe condition of the sources because the 241 Am will remain within the foil. S4 shows greater energy straggling than the newer sources, but unlike S1 and S2, the spectral peak energy is higher at 4.7 MeV so the increased energy straggle will not greatly affect the blocking of alpha radiation by paper demonstration. S4 was not reported as showing a reduced count, so there is uncertainty whether the energy straggle has actually increased with age. If originally the foil layer dimensions met the manufacturer’s indicative data, it would suggest a net migration of 241 Am away from the original active matrix outward towards the face and towards the substrate of the source. If this suggestion is correct, this would have a bearing on the safe condition of the foil if eventually the americium migrated to the source surface. The small alpha energy peak of 5.48 MeV is from 241 Am near or at the surface, plausibly 241 Am at the site of a microscopic pit in the gold/palladium layer. If there was diffusion of 241 Am to the top surface, the alpha energy distribution curve would be unlikely to drop to zero before 5.48 MeV. Testing the diffusion hypothesis needs further investigation by surface analysis, e.g. by sputtering. However, it would be difficult to remove the foils of the sample sources for surface 359
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
analysis because the foil would likely become damaged by the removal. Should sources be found where the foil can be removed from its holder without damage, this will permit further investigation. The question remains why this increased energy straggling over time appears in some 241 Am foil sources and less so in others. For example, in contrast to sources S1 and S2, the source S3 has a spectral peak energy and FWHM still within the manufacturer’s indicative data. There is a difference in the evidence on the construction of the foils: the Amersham packing note (1991) shows a diagram where there is a gold backing layer between the silver substrate and the active gold matrix, but the NRPB report (Williams 1974) does not mention this. HTSL explained that the foil manufacturer at that time—Amersham—produced a variety of alpha foils that had different ratio of materials in the surface layers, consequently there is today some uncertainty of the composition of the foils used in school sources from that period. It is therefore a reasonable hypothesis that there are variations in foil design of which some are more predisposed to this change in energy straggling over time. If the diffusion hypothesis is correct, Dini (1993) states that while diffusion can improve the adherence between the layers of metal, in contrast, where inter-diffusion between two metals is uneven, it is possible that a weakness can be created that eventually could cause the delamination of the gold or gold–palladium face layer. There is no evidence from the 241 Am foils in this study that there is any delamination of the face. Dini’s statement was in the context of electrodeposition whereas the foil face is hot-forged, but the principle remains.
5. Advice for educational establishments
For schools and other educational establishments that have 241 Am foil sources exhibiting the greater energy straggling and reduced spectral peak energy, there is no evidence currently to suggest the foils have become unsafe, but they may be of limited use as alpha emitters when detected by an end-window GM detector such as the commonly used ZP1481. The sources can still be used with spark counters and ionisation chambers including the demonstration of blocking alpha radiation by paper, because these detectors have no solid window. If the protective cap of the ZP1481 detector is removed and the gap between the source cup and GM detector window is kept as small as practicable, 2 mm or less, then the blocking of alpha radiation by paper demonstration using the GM detector may still be effective if the reduction in spectral peak energy has not become too great. But if it has, the source will need replacing for this application. Care needs to be taken with the source so close to the unprotected GM detector window; the window would be broken easily if the source were to be knocked into it, damaging the detector permanently. Although this study looked at just a small sample of sources, a sensible precautionary measure for this type of foil source would be to carry out leak tests at least yearly with remote inspection of the face to check the surface remains in good condition, and with no discolouration which could be from gold or 241 Am that has migrated through the face layer.
Acknowledgments
The authors would like to thank Oriel High School, Crawley, and St Benedict’s Roman Catholic School, Bury St Edmunds, who loaned their americium sources for this study, and High Technology Sources Ltd (HTSL) for the loan of a new source and the information supplied. 360
J. Radiol. Prot. 34 (2014) 347
R Whitcher et al
References Amersham 1991 Safety instructions for unpacking and use of alpha foil sources. HI 041 Issue 1 Centronic 1997 Geiger Muller Tubes Databook (New Addington, UK: Centronic) Comfort J R, Decker J F, Lynk E T, Scully M O and Quinton A R 1966 Energy loss and straggling of alpha particles in metal foil Phys. Rev. 150 249–56 Deslattes R D et al 2005 X-Ray Transition Energies (version 1.2) (Gaithersburg, MD: National Institute of Standards and Technology) [Online] Available at http://physics.nist.gov/XrayTrans (Accessed 19 July 2013) Dini J W 1993 Electrodeposition: The Materials Science of Coatings and Substrates (New Jersey: Noyes Publications) Ekstr¨om L P and Firestone R B 2004 Table of radioactive isotopes, Database Version 2.1 LBNL Isotopes Project—LUNDS Universitet [Online] Available at http://ie.lbl.gov/toi/index.htm (Accessed 10 January 2014) Hubbell J H and Seltzer S M 2004 Tables of X-Ray Mass Attenuation Coefficients and Mass Energy-Absorption Coefficients (version 1.4) (Gaithersburg, MD: National Institute of Standards and Technology) [Online] Available at http://physics.nist.gov/xaamdi (Accessed 21 January 2014) Knoll G F 2010 Radiation Detection and Measurement 4th edn (New York: Wiley) pp 216–8 Liu Y, Wang L, Zhan J and Chui L 1985 An investigation of silver diffusion through gold films by Auger electron spectroscopy Vacuum 35 537–8 Lucas J W 1966 Precautions in the use of radioactive sealed sources in schools School Sci. Rev. ASE 47 19–27 National Nuclear Data Center, Information accessed from the NuDat 2 database, Brookhaven National Laboratory, New York [Online] Available at www.nndc.bnl.gov/nudat2/ (Accessed 23 September www.nndc.bnl.gov/nudat2/ (Accessed 23 September 2013) Radiological Health Handbook 1970 Publication No 2016 (Rockville, MD: Bureau of Radiological Health) p 123 Williams T G 1974 The Use of Plutonium-239 Sources in Schools and Other Educational Establishments NRPB-R28 (Harwell: HMSO) Ziegler J F, Ziegler M D and Biersack J P 2010 SRIM—the stopping and range of ions in matter Nucl. Instrum. Methods Phys. Res. B 268 1818–23
361
Copyright of Journal of Radiological Protection is the property of IOP Publishing and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
