2
Gamma-Ray Measurements

Gamma rays emitted from the Hiroshima and Nagasaki explosions had two major sources. One was the nitrogen-capture gamma rays that arose from the capture of bomb neutrons when slowed in air to thermal energies to emit roughly 5 MeV gamma rays, which are very penetrating in air. These were proportional to the number of neutrons captured in nitrogen. The second source was the gamma rays emitted by the fission products in the fireball (which were proportional to the bomb yield). These gamma rays were emitted mostly between 1 and 10s after the explosion. A third, much smaller contribution was prompt gamma rays from the device itself.

THERMOLUMINESCENT MEASUREMENTS IN BRICK AND TILE

A most important feature of DS86 was that thermoluminescent (TL) measurements in brick and tile were used to verify the calculations of free-in-air kerma (FIA) from these gamma rays. The agreement in 1986 between the TL measurements and DS86 calculations was generally good and within the estimates of uncertainty for the DS86 calculations (Roesch 1987). Quartz, like any material used in thermoluminescent dosimetry, contains quantitative radiation-exposure history from the time of its initial annealing at high temperature. The method of retrieving the exposure history in quartz was based on a protocol first developed by Grogler and others (1960) for dating of ancient pottery fragments. The technique was improved by Ichikawa (1965) and Fleming and Thompson (1970), who used the TL method in archeological studies. Higashimura and others (1963) reported that the exposures and doses from gamma rays in Hiroshima and Nagasaki could be measured by this method. A substantial effort was then mounted to obtain suitable ma-



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 14
Status of the Dosimetry for the Radiation Effects Research Foundation (DS86) 2 Gamma-Ray Measurements Gamma rays emitted from the Hiroshima and Nagasaki explosions had two major sources. One was the nitrogen-capture gamma rays that arose from the capture of bomb neutrons when slowed in air to thermal energies to emit roughly 5 MeV gamma rays, which are very penetrating in air. These were proportional to the number of neutrons captured in nitrogen. The second source was the gamma rays emitted by the fission products in the fireball (which were proportional to the bomb yield). These gamma rays were emitted mostly between 1 and 10s after the explosion. A third, much smaller contribution was prompt gamma rays from the device itself. THERMOLUMINESCENT MEASUREMENTS IN BRICK AND TILE A most important feature of DS86 was that thermoluminescent (TL) measurements in brick and tile were used to verify the calculations of free-in-air kerma (FIA) from these gamma rays. The agreement in 1986 between the TL measurements and DS86 calculations was generally good and within the estimates of uncertainty for the DS86 calculations (Roesch 1987). Quartz, like any material used in thermoluminescent dosimetry, contains quantitative radiation-exposure history from the time of its initial annealing at high temperature. The method of retrieving the exposure history in quartz was based on a protocol first developed by Grogler and others (1960) for dating of ancient pottery fragments. The technique was improved by Ichikawa (1965) and Fleming and Thompson (1970), who used the TL method in archeological studies. Higashimura and others (1963) reported that the exposures and doses from gamma rays in Hiroshima and Nagasaki could be measured by this method. A substantial effort was then mounted to obtain suitable ma-

OCR for page 14
Status of the Dosimetry for the Radiation Effects Research Foundation (DS86) terial for reconstructing the dose as a function of distance from the bomb hypocenter. Samples of brick and roof tile were selected at homes, universities, temples, a hospital, and so on according to criteria for quantitative exposure measurements. The technique to obtain the TL data on brick and tile exposure was then developed and carried out with extraordinary care. Three methods of measurement, with variations, have been used. The first was the high-temperature technique, which is a straightforward measurement of the glow curve (luminescence with heating), from a sample of quartz. The sample is separated from the clay-tile matrix and, irradiated at a known exposure, and the glow curve is read. The main trap for electrons displaced by the gamma-ray exposure of interest is a relatively high energy trap of about 7 eV and requires near 500°C heating to erase the signal (by annealing) and up to about 300°C for the glow-curve (TL) measurement. The second, or additive, technique attempts to correct for an initial nonlinearity in the dose-response relationship of quartz by irradiating multiple samples of the extracted quartz at increasing doses, and correcting the measurements of the original sample in the nonlinear or supralinear region. The third, predose, technique and its variations use a lower-energy electron trap in quartz near 110°C, which fades rapidly and is not found in normal samples. The stored-dose information in the higher-energy trap can be explored with sufficient heating of the sample to permit transfer from the high-energy to the low-energy trap and then readout of the low-energy information. All three techniques have been used on specimens of brick and ceramic tiles in Hiroshima and Nagasaki and provide quantitative measurements of the cumulative gamma-ray dose histories of the quartz crystals in the samples. After conversion to estimates of free-in-air tissue kerma at the specimens’ locations at the time of the bombing, these measurements can be compared directly with estimates calculated from DS86. The details of the comparison procedure are shown in Figure 2–1. The gamma-ray kerma at both Hiroshima and Nagasaki was measured directly with the radiation signal stored in brick or tile within the range of the bomb radiation. The measurements show that, within the predictable error, the agreement with the DS86 calculation is satisfactory. However, there are at least three inherent sources of bias in making measurements with the TL method and comparing them with DS86 estimates. These are considered here to show the extent of agreement with the DS86 calculations: Fading. Fading of the radiation signal with time occurs readily in other TL materials, such as LiF, but quartz is considered to have a very stable high-energy trap. Fading of the signal occurs if heating, after exposure, occurs near the trap temperature region, 400–500°C. Backscatter medium. The response signal in a sample that is part of a large structure, such as a roof or wall with concomitant backscattered radiation from the surrounding material, is larger than the response to a calibration exposure carried out without backing material.

OCR for page 14
Status of the Dosimetry for the Radiation Effects Research Foundation (DS86) FIGURE 2–1 Measured and calculated quantities and related to gamma dose (free-in-air tissue kerma). Quartz energy response. At the same air kerma the energy signal in quartz at low photon energies per unit dose (or fluence) is greater than that at higher photon energy. Background issues. An additional consideration is the uncertainty in the measurement of the natural background radiation signal inherent in any brick or tile, which must be subtracted if bomb radiation is to be measured. These four factors and their effects on the quality of TL measurements are discussed in more detail in the following sections. Fading Figures 2–2 and 2–3, from The Effects of Nuclear Weapons (Glasstone and Dolan 1977), show the thermal effects of the Hiroshima bomb. Figure 2–2 shows the bubbling or blistering of common roof tile from the heat of the bomb. The ther-

OCR for page 14
Status of the Dosimetry for the Radiation Effects Research Foundation (DS86) FIGURE 2–2 Flash marks produced by thermal radiation on asphalt of bridge in Hiroshima. Where the railings served as protection from the radiation, there were no marks; the length and direction of “shadows” indicate the point of the bomb explosion (Glasstone and Dolan 1977). mal shadow of the railings on the asphalt bridge pavement was used to support the estimates of height of the detonation. The Effects of Nuclear Weapons provides evidence that blistering of tiles (Figure 2–3) was observed up to 980 m from the hypocenter. The radiant energy at this distance was estimated to be 45 cal/cm2. Experimental heating of similar tiles showed that an 1800°C impulse for 4 s produced the same blistering, although the heat penetration into the tile was thought to be greater than that in Hiroshima or Nagasaki. No further information on this point is given. Although temperature of that magnitude would certainly affect the TL signal after exposure, there are no data to confirm fading of a signal when the radiation exposure and the radiant energy occur at nearly the same moment. It is not known

OCR for page 14
Status of the Dosimetry for the Radiation Effects Research Foundation (DS86) FIGURE 2–3 Blistered surface of roof tile; left portion of the tile was shielded by an over-lapping one (0.37 mile from ground zero at Hiroshima) (Glasstone and Dolan 1977). what depth of penetration the radiant energy would have in a typical brick or tile. If fading due to annealing did occur in samples, the exposures from TL measurements would have been underestimated. The published studies contain statements suggesting that the tile samples collected had not been substantially heated by the explosion of the bomb or by accidents (Hoshi and others 1987). However, radiant energy as a function of distance from the hypocenter has not been addressed explicitly. Figure 2–4, also from The Effects of Nuclear Weapons, gives estimates of the radiant energy at ground distances from a hypothetical hypocenter for different weapon yields for a particular weather condition. Depending on the orientation of the tile with respect to the bomb, there was probably a potential for some annealing of the sample due to high temperatures at least out to 1000 m. Some of the TL measurements were performed by Edwin Haskell, who indicates that the predose technique used precludes any effect of fading (Haskell 2000). Moreover, if there were fading, the depth-dose curves in tile would not be consistent with the gamma-ray energy attenuation (Haskell 2000). Overall, the effect of signal fading in the samples measured was considered to be negligible, and there is no reason to disagree with what has been previously concluded. Backscatter Medium All calibration exposures were carried out with a 60Co calibration source (E=1.17, 1.33 MeV) except at the University of Utah, where a 137Cs source

OCR for page 14
Status of the Dosimetry for the Radiation Effects Research Foundation (DS86) FIGURE 2–4 Slant range for specified radiant exposures and energy yields (Glassstone and Dolan 1977).

OCR for page 14
Status of the Dosimetry for the Radiation Effects Research Foundation (DS86) (0.66 MeV, 30-keV x-rays) was used (Maruyama and Kuramoto 1987). Details of the exposures are provided by Ichikawa and others (1987). There was some ceramic material surrounding the sample to provide electron equilibrium and some backscatter medium. The in situ exposures from the burst undoubtedly included backscattered gamma rays different from those of the calibration exposures. Depending on the ambient scattering structures and gamma-ray energy, an increase in the gamma-ray signal due to scattered gamma rays might have been around 20%. The gamma-ray scatter to a single tile from adjacent tiles or structures from the broad-beam exposure from the burst is unknown. This was pointed out by Ichikawa and others (1987), who stated that the actual buildup of the gamma rays in the tiles could have been greater than that in the tiles exposed to 60Co gamma rays without the backscatter of concrete blocks. Two examples of backscatter are well known. In counting of gamma-emitting samples, a backscatter peak at 0.25 MeV is evident in the gamma-ray spectrum. That results from a 180° scatter of the photon originating from the source (sample) by the surrounding shielding. Thus, backscattered gamma rays contribute to the total signal measured. Similarly, in the calibrating of a pressurized ion chamber with a point source of photons, such as 60Co, scattered photons from the ambient structures or the ground surface contribute to the total signal. To avoid the effect, a shield, usually a lead block, is placed at the source to prevent line-of-sight radiation from the source (a shadow shield). It provides direct measurement of the scattered photon signal. The scattered signal is subtracted from the total to obtain the calibration signal. According to Kaul and Egbert (2000) the DS86 calculations included a component for backscatter and the backscatter signal enhancement is probably less than 5% at 1000 m ground distance. Quartz Energy Response All TL materials yield a different—that is increasing—light output with decreasing gamma ray energy. That is due to the higher energy absorption (a larger mass energy coefficient) in the medium with lower energy. Figure 2–5 shows the mass energy-transfer coefficient for SiO2 (taken for quartz) as a function of energy (Hubbell 1982). Below 0.1 MeV, there is an increase in the coefficient that would increase the energy signal deposited per unit kerma in free air. Maruyama (1983) stated that when the response from 60Co gamma rays was unity, that for 40 keV x-rays was about 3 at a depth of 0.5 cm in the brick. Figure 2–6 shows the cumulative distributions of fluence at Hiroshima, in the forward and downward directions, as functions of gamma-ray energy estimated in DS86 (Cullings 2000). About 30% of the gamma-ray fluence is below 0.1 MeV, so a signal enhancement of this order of magnitude is possible because of the physical absorption properties of quartz. Figure 2–6 also illustrates the angular dependence of the energy absorption of the brick or tile with respect to orientation to the burst, with about a 20% difference possible. The uncertainty due to orientation was also pointed out by Dean Kaul and colleagues (Roesch 1987).

OCR for page 14
Status of the Dosimetry for the Radiation Effects Research Foundation (DS86) FIGURE 2–5 Photon mass energy-transfer coefficient in SiO2 (Hubbell 1982). Figure 2–7 shows the cumulative gamma-ray fluence as a function of energy at Hiroshima with all angles combined for both prompt and delayed gamma rays. Delayed gamma rays contribute most of the exposure and dose from the burst. Figures 2–8 and 2–9 show the cumulative distributions of downward and forward fluence at several ground ranges. They illustrate the slight shift in the energy spectrum—with somewhat more abundant high energy—with increasing ground distance from the FIGURE 2–6 Hiroshima: cumulative gamma-ray fluence vs. energy at 1000 m downward and forward fluence (Cullings 2000).

OCR for page 14
Status of the Dosimetry for the Radiation Effects Research Foundation (DS86) FIGURE 2–7 Hiroshima: cumulative gamma-ray fluence with energy (prompt and delayed) at 1000 m; all angles combined (Cullings 2000). FIGURE 2–8 Hiroshima: cumulative gamma-ray vs. energy by distance (downward fluence) (Cullings 2000).

OCR for page 14
Status of the Dosimetry for the Radiation Effects Research Foundation (DS86) FIGURE 2–9 Hiroshma: cumulative gamma-ray fluence vs. energy by distance (forward fluence) (Cullings 2000).

OCR for page 14
Status of the Dosimetry for the Radiation Effects Research Foundation (DS86) burst. Again, in Figures 2–6 through 2–9, about 30% of the gamma-ray fluence is below 0.1 MeV, which would lead to a TL signal enhancement. Background Issues When a brick or tile is fabricated, the temperature is sufficient to erase or zero the TL radiation signal in the quartz. Later background radiation exposure is stored in the material, with bomb radiation adding to the total TL signal. The background radiation in the brick or tile is due to the uranium and thorium, their decay products, and potassium present in the material itself. External naturally occurring gamma rays and cosmic rays are also part of the background. The natural radioactivity inclusions in the material emit alpha, beta, and gamma rays. These were assumed to be constant over the lifetime of the sample being discussed here. That is roughly true, but in the initial firing of the material (1100–1200°C), a fraction of the 228Ra (a decay product in the thorium series) is volatilized, so the inherent background rate would increase over the approximately 40-year postfiring lifetime (Roesch 1987). The inherent background of the brick or tile was determined either by alpha counting or by measuring beta TL. Gamma-ray TL was used to determine the external gamma-ray and cosmic-ray background rates. By placing TL dosimeters on buildings where samples had been taken, the external gamma-ray and cosmic-ray background was measured. The total background exposure subtracted from the samples at Hiroshima and Nagasaki was 0.15–0.32 Gy. Thus, at 2000 m ground range at Hiroshima, for example, the background subtracted is several times that of the bomb signal, and the propagated error in the measurement is 100% (Roesch 1987). Thus background is extremely important especially at the larger distances. The most important of the other three factors that potentially affect the TL measurements is considered to be the increase in the magnitude (light output) of the TL in response to low gamma ray energy. Although the magnitude of the energy-response correction factor is not known precisely, a downward correction of 20% is plausible on the basis of the above considerations, and was used in the following analyses to derive the sensitivity comparisons. However, as will be shown later, the agreement is better without this correction. COMPARISON OF TL GAMMA-RAY MEASUREMENTS BETWEEN HIROSHIMA AND NAGASAKI Figures 2–10 and 2–11 summarize the TL measurements made to date at Hiroshima (Cullings 2000) as corrected to tissue dose (free-in-air [FIA] tissue kerma) for comparison with the DS86 calculations. The data are distinguished by type of reporting unit (roentgen, air dose, tissue dose, or quartz) but corrected in Figure 2–11 to FIA tissue dose so that any differences attributed to reporting units can be identified. None is evident. The DS86 FIA tissue-dose estimates are shown for comparison.

OCR for page 14
Status of the Dosimetry for the Radiation Effects Research Foundation (DS86) FIGURE 2–10 Hiroshima TL measurements. Original reported units—roentgens, tissue, air, quartz—converted to RA tissue kerma (Cullings 2000). FIGURE 2–11 Hiroshima TL measurements (Corrected). Original reported units— roentgen, tissue, air, quartz—converted to FIA tissue kerma (Cullings 2000).

OCR for page 14
Status of the Dosimetry for the Radiation Effects Research Foundation (DS86) FIGURE 2–12 Nagasaki TL measurements. Original measurements in reported units— roentgens, tissue, air, quartz—converted to FIA tissue kerma (Cullings 2000). In Hiroshima, there are replicate data at 2050 m. Their arithmetic mean and standard deviation are 0.058±0.089 Gy, as shown in Figure 2–10. The replicate data are important because they demonstrate the acknowledged precision of TL measurements, which, in general, is 20–30%. The mean of the measurements at 2050 m is nearly identical with the DS86 estimate of 0.06 Gy but with large uncertainties (Roesch 1987). Closer to the hypocenter, the agreement with DS86 appears good. Figure 2–11 shows the effect of a 20% correction (downward) to the reported measurements. The result does not appreciably improve the fit of the measurements to DS86. Figures 2–12 and 2–13 summarize the TL measurements made to date at Nagasaki (Cullings 2000). The data points are differentiated by marker style for the original reported units and corrected to FIA tissue kerma for comparison with DS86. At 2040 m there are replicate measurements, and the arithmetic mean and standard deviation are 0.18±0.04 Gy, compared with the DS86 estimate of 0.11 Gy (Roesch 1987). The same correction factor (0.8) would bring the 2040 m measurements essentially into agreement with DS86. Figure 2–13 shows the same data plotted with a correction factor of 0.8. The fit of the data to DS86 is not improved substantially, and there is a suggestion that a slightly shallower slope for DS86 would provide, in general, a somewhat better fit to the measurements. One test of the goodness of fit to the DS86 calculations is to examine residuals, that is differences between the data points and the DS86 calculated values. To reduce the scatter of the TL measurements, the measured TL data were averaged in intervals of 100 m. That compresses the number of data points and permits error terms to be calculated directly for the range interval. The interval average data for

OCR for page 14
Status of the Dosimetry for the Radiation Effects Research Foundation (DS86) FIGURE 2–13 Nagasaki TL measurements (corrected). Original reported units—roentgens, tissue, air, quartz—converted to FIA tissue kerma (Cullings 2000). Hiroshima, with error bars, are shown in Figure 2–14. The fractional residual is defined as residual=(mean of data interval at distance—DS86 value)/DS861 The fractional residual plots for Hiroshima are shown in Figure 2–15. The residual plots show both the original reported data and the data with a 20% (0.8) correction for the assumed energy-response correction. The residual plot in Figure 2–15 indicates that the original data as reported, are in somewhat better agreement with DS86 than the arbitrarily corrected value. The measured data for Nagasaki were averaged over 100-m distances. The average data, with error bars, are shown in Figure 2–16. The fractional residual plot for Nagasaki is shown in Figure 2–17 (similar to that for Hiroshima) for both the original reported data and the data with a 20% (0.8) correction for the assumed energy response. The residual plot in Figure 2–17 indicates that the original data as reported are in somewhat better agreement with DS86 than the arbitrarily corrected values. The residuals shown in Figures 2–15 and 2–17 represent the combined effects of errors, both systematic and random, in the DS86 estimates and in the measurements. Roughly stated, the relative magnitude of these combined errors is about 20% for comparatively high doses and about 100% for the natural background component, that is at 2000 m on the ground in Hiroshima. Nevertheless, the test of residuals indicates that because a reduction of 20% does not improve agreement between calculation and measurement, the agreement, fortuitously perhaps, might be to within about±10%. 1   The DS86 value is interpolated to the midpoint (midway between nearest and farthest measurements) of the distance.

OCR for page 14
Status of the Dosimetry for the Radiation Effects Research Foundation (DS86) FIGURE 2–14 Hiroshima TL measurements averaged over 100-m intervals. FIGURE 2–15 Hiroshima residuals (mean—DS86)/DS86; original measurement values and values corrected by 0.8.

OCR for page 14
Status of the Dosimetry for the Radiation Effects Research Foundation (DS86) FIGURE 2–16 Nagasaki TL measurements averaged over 100-m intervals. Figure 2–15 shows that the TLD measurements in Hiroshima appear to be about 20% lower than the DS86 calculation out to 1300 m, then 20% higher at ground distances greater than 1300 m. Nagatomo and others (1995) and Roesch (1987) pointed out this difference for distances greater than 1300 m. FIGURE 2–17 Nagasaki residuals (mean—DS86)/DS86; original measurement values and values corrected by 0.8.

OCR for page 14
Status of the Dosimetry for the Radiation Effects Research Foundation (DS86) Figure 2–17 shows that the TLD measurements in Nagasaki are uniformly about 20% lower than the DS86 calculation. The exception at both cities occurs in measurements at greater than 2000 m, where the error in the measurement technique is estimated to be of the order of 100%. Given the overall uncertainty of about 20% for the TLD technique (precision, sample orientation, etc.) there is good agreement of the gamma-ray measurements with DS86 calculations. Other sources of radiation might provide a small contribution to the TL signal. Neutron activation of short-lived nuclides, such as sodium and calcium, could have added to the TL signal. Fallout from the bomb might have contributed a small amount. There were two tornadoes at Hiroshima within a relatively short time after the burst, and fallout was probably washed away. If there are actually more fast neutrons at distances in Hiroshima than has been estimated in DS86 they should result in some increase also in the gamma-ray dose at distances because of neutron-capture reactions producing gamma rays. The four factors inherent in TL measurement uncertainty described cannot be evaluated quantitatively, but the agreement of the TL measurements with DS86 calculations is striking and provides support for the present estimates of gamma-ray dose. The committee suggests that the working group consider reevaluating the TL measurement data with particular emphasis on whether the energy response of the TL was properly accounted for and on background considerations.