Journal of Environmental Radioactivity 96 (2007) 103e109 www.elsevier.com/locate/jenvrad

Measurements of short-lived cosmic-ray-produced radionuclides in rainwater K. Komura*, Y. Kuwahara, T. Abe, K. Tanaka, Y. Murata, M. Inoue Low Level Radioactivity Laboratory, Institute of Nature and Environmental Technology (K-INET), Kanazawa University, Wake, Tatsunokuchi, Ishikawa 923-1224, Japan Accepted 15 January 2007 Available online 10 April 2007

Abstract Cosmic-ray-produced (CP) nuclides with half-lives shorter than 21 h were measured in rainwater by ultra-low-background gamma spectrometry at Ogoya Underground Laboratory. As levels of CP nuclides are extremely low and their half-lives are very short, quick sampling methods for a large volume of rainwater and rapid chemical separations by ion exchange method were developed. The nuclides measured were shortlived 24Na, 28Mg, 38S, 38Cl, 39Cl, as well as nuclides with longer half-lives 7Be and 22Na. The number of atoms of CP nuclides in rainwater were evaluated to range from 30 to 1500 L1 for 24Na (n ¼ 16, mean; 520 [6.7 mBq L1]), 80 to 600 L1 for 28Mg (n ¼ 13, mean; 260 [2.4 mBq L1]), 400 to 1900 L1 for 39Cl (n ¼ 6, mean; 1200 [250 mBq L1]), 1  106 to 4  107 L1 for 7Be (n ¼ 16, mean; 7  106 [1.05 Bq L1]) and 2  103 to 1  105 L1 for 22Na (n ¼ 9, mean; 2  104 [0.2 mBq L1]). Ó 2007 Elsevier Ltd. All rights reserved. Keywords: Cosmic ray produced nuclide; Rain water; Berrylium-7; Sodoim-22; Sodium-24; Magnesium-28; Chlorine38; Chlorine-39; Sulfur-38; Gamma ray spectrometry; Underground laboratory; Ogoya underground laboratory

1. Introduction During the 1950s and 1960s, analysis of freshly precipitated rainwater resulted in the discovery of more than 20 radionuclides (Arnold and Ali Al-Salih, 1955; Marquez and Costa, 1955; Davis and Schaeffer, 1955; Arnold, 1956; Goel, 1956; Winsberg, 1956; Lal et al., 1957, 1960; * Corresponding author. Tel.: þ81 761 51 4440; fax: þ81 761 51 5528. E-mail address: [email protected] (K. Komura). 0265-931X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.jenvrad.2007.01.022

104

K. Komura et al. / J. Environ. Radioactivity 96 (2007) 103e109

Marquez et al., 1957; Ro¨del, 1963; Perkins et al., 1965). These nuclides are called cosmicray-produced (CP) nuclides (or cosmic-ray-induced radionuclides) because they are produced by nuclear reactions between high-energy cosmic rays and nitrogen, oxygen and argon atoms in the air. Some of the CP nuclides such as 7Be, 3H and 14C with longer half-lives and relatively high production rates have been used widely as useful tracers to investigate geochemical processes occurring in atmosphere and for radioactive dating. On the other hand, most CP nuclides derived from 40Ar have very short half-lives and their production rates are extremely low. Therefore these nuclides have not been used yet as geochemical or hydrological tracers due to difficulties with their measurements. Ultra-low-background Ge detectors placed in the Ogoya Underground Laboratory (OUL) (Komura, 1997; Hamajima and Komura, 2004; Komura and Hamajima, 2004) have made it possible to measure extremely low levels of radionuclides, which cannot be detected by ordinary low-background detectors in a surface laboratory. One of the applications in low-level counting carried out at OUL has been to detect CP nuclides in rainwater with half-lives shorter than 1 d, i.e. 38S (T1/2: 2.83 h), 38Cl (T1/2: 37.2 m), 39Cl (T1/2: 55.6 m), 18F (T1/2: 109.7 m), 24Na (T1/2: 14.96 h) and 28Mg (T1/2: 20.9 h). In this paper, we describe detection methods for the analysis of CP nuclides in rainwater by rapid sampling of a large volume of rainwater and rapid separation using ion exchange resin. Results obtained from the rain samples collected at Tatsunokuchi, Ishikawa Prefecture, Japan, are presented and discussed. 2. Experimental In order to measure extremely low levels of CP nuclides with short half-lives, the following three barriers must be overcome: (1) a rapid sampling of large volume of rainwater, (2) a rapid chemical separation of CP nuclides, and (3) a rapid measurement using an extremely low-background detector. Such analyses have been successfully conducted as described below. 2.1. Rapid sampling of large volume of rainwater As activity levels of short-lived CP nuclides in rainwater are extremely low, a large volume (w50 L) of rainwater must be collected in a short time. At first, we tried to collect rainwater using 4e6 large buckets set at the roof of Low Level Radioactivity Laboratory (LLRL), Kanazawa University, in Tatsunokuchi, located north of central Japan. Using this method, however, only 10 L of rainwater could be collected during 30 min of exposure even in a heavy rain. After various trials, a part of the roof of LLRL (29 m2) was used for rainwater collection. Using this method, we succeeded in collecting 50 L of rainwater within 5 min in heavy rain, and in w30 min during ordinary rain. Since half-lives of CP nuclides are short, special care to avoid the influence of past rains in collected water was not necessary. 2.2. Rapid separation of CP nuclides In order to separate short-lived CP nuclides from rainwater, ion exchange techniques are considered to be the best choice because these methods are very simple and rapid enough to quantitatively adsorb trace amounts of radionuclides in large volumes of water samples. POWDEXÒ-PCH and POWDEXÒ-PAO, which are widely used for chemical processing of contaminated water at nuclear power stations, have been used in this study. It has been confirmed in preliminary experiments that Naþ, Mg2þ and Cl ions in large water samples can be quantitatively adsorbed within 10 min if POWDEXÒ-PCH and POWDEXÒ-PAO are added at a concentration of 2 g per 10 L of rainwater.

K. Komura et al. / J. Environ. Radioactivity 96 (2007) 103e109

105

A 10-g aliquot both of POWDEXÒ-PCH and POWDEXÒ-PAO were added simultaneously to a 50-L rainwater sample collected in a 70-L bucket. CP nuclides were quantitatively separated onto cation and anion ion exchange resin mixture by 10e15 min of stirring using a propeller-type stirrer. Ion exchange resins were settled at the bottom of bucket and then collected through a filter paper. The resin was packed in a 5 cm  5 cm polyethylene bag to make gamma-ray counting source. Total time required for chemical separation and source preparation was w30 min. 2.3. Gamma-ray measurements at the Ogoya Underground Laboratory (OUL) 2.3.1. Ogoya Underground Laboratory In order to measure extremely low levels of radionuclides, OUL was constructed in 1995 in a tunnel of the former Ogoya copper mine located 21 km from LLRL (Komura, 1997). Thickness of the rock (tuff) overburden is 135 m, which corresponds to 270 m water equivalent and intensities of cosmic-ray muons and neutrons are w1/200 and w1/300 of those at ground surface, respectively. Sixteen ultra-low-background Ge detectors are equipped in OUL as of October 2006. Background levels of these Ge detectors are w2 orders of magnitude lower than corresponding size of Ge detectors set in an overground counting room (Hamajima and Komura, 2004). 2.3.2. Measurements of CP nuclides Gamma-ray measurements could start 1.5e2 h after the sampling. The counting source was set directly on the end-cap of large volume coaxial type Ge detectors with 93.5% and 91% of detection efficiencies relative to a 7.6 cm diameter  7.6 cm high NaI(Tl) detector. Gamma-ray spectra were recorded on an 8K multi-channel analyzer at 0.4 keV/channel, so that the 2754 keV gamma rays from 24Na could be registered. Table 1 summarizes half-lives, energies and abundance of major gamma rays from CP nuclides measured in this study. Gamma rays emitted from short-lived CP nuclides are higher energies, therefore large volume coaxial HPGe detectors have been used. In an ordinary experiment, gamma-ray spectra were recorded at 20e 30 min intervals during the first 2e4 h with the aim of measuring very short-lived CP nuclides, i.e. 38Cl (38S) and 39Cl. Then, measurement intervals of several hours continued over 1e2 d to analyze 24Na and 28Mg.

3. Results and discussion 3.1. Gamma-ray spectra of rainwater sample The first experiment was made in the middle of June in 2004. Since then, more than 50 measurements were carried out at LLRL. Fig. 1 shows an example of gamma-ray spectra taken for Table 1 Half-lives, energies and abundances of major gamma rays emitted from CP nuclides studied in this work Nuclide

Half-life

Measured gamma ray (keV)

Abundance (%)

38

S 38 Cl 39 Cl

170.3 m 37.24 m 55.6 m

18

109.77 m 14.96 h

1941.9 2167.4 1267.2 1517.5 511 1368.63 2754.03 1342.25 1778.85 477.59 1274.53

100? 42.4 100 73.2 100 100 99.94 52.6 100 10.52 99.94

F Na

24

28

Mg Al 7 Be 22 Na 28

20.9 h 2.241 m 53.29 d 2.602 y

K. Komura et al. / J. Environ. Radioactivity 96 (2007) 103e109

106

Counts/Channel

60

24

Cl

40

20

20

10

0

28

214

Pb

7

Be

214

Mg

Al

15 38

38

S

10

10

5

5

0 1370

0 0 1770 1780 1790 1930 1940 1950 2160 2170 2180

Bi 214 214

Bi

214

Bi

214

Bi

39

Bi

Cl

214

Bi 214

Bi

24 214

10

1

Cl

5

1350

100

Counts/Channel

28

Na

(28Mg)

0 1260 1270 1280 1330

1000

15

10

30 39

Bi

208

Na

TI

7 Be 212

100

Pb

18 F 208

10

24 28

TI

Mg

Na 40

K

28 AI (28Mg)

24 208

Na

TI

1 0

500

1000

1500

2000

2500

3000

Gamma ray energy (keV)

Fig. 1. Gamma-ray spectra of the rain sample collected on October 5, 2004.

the rain sample collected on October 5, 2004. Two full energy spectra are shown in the middle of Fig. 1, and the lower part shows the spectrum obtained for the first 17 ks, and up to 67 ks. The expanded spectra of individual CP nuclides are plotted in the upper part of Fig. 1. Contributions of short-lived 214Pb (T1/2: 26.8 min) and 214Bi (T1/2: 19.9 min), which are daughters of airborne 222Rn, had decreased to negligible levels, as can be seen in the second spectrum. Besides the prominent peak of 7Be, there are also seen small peaks of 24Na and 28Mg (28Al), confirming that their concentrations in the air are extremely low. 3.2. Calculation of concentrations of CP nuclides in a sample Since the half-lives of the CP nuclides are short and corresponding concentrations are very low, it seems preferable to express their concentrations as number of atoms per unit volume (atom L1) instead of Bq L1. Number of atoms of a CP nuclide in a sample can be estimated as Ainfinity ¼

At
; 1  exp  0:693t=T1=2

ð1Þ

where At, Ainfinity and T1/2 denote, respectively, peak areas obtained for the counting time t, the expected area for a measuring time of infinity, and the half-life of a CP nuclide. Using Ainfinity, the number of atoms in the sample can be easily calculated using the detection efficiency (3), the gamma-ray branching ratio (B)

K. Komura et al. / J. Environ. Radioactivity 96 (2007) 103e109

N¼

107

Ainfinity : 3B

ð2Þ

Fig. 2 shows examples of this integral method as applied to 7Be, 24Na, 28Mg and 39Cl. Observed counts and saturation counts estimated by Eq. (1) are shown by open and closed circles, respectively. The saturation activity (counts) converses with counting during the increasing time, confirming the validity of the integral method. Number of atoms of each CP nuclide in the sample calculated using Eq. (2) was decay corrected with the aim of estimating the number of atoms at the time of end of the sampling. This integral method is uncommon, but can be applied independent of half-life such as the case of 238U with half-life of longer than billions of years. Activity can be calculated simply by multiplying the number of atom and the decay constant. 3.3. Number of atoms of CP nuclides per litre of rainwater Among CP nuclides measured in this study, 24Na, 28Mg and 39Cl could be detected in all rain samples because their production rates are higher and their half-lives are long enough. Unfortunately the counting time was not always long enough to detect 22Na, and thus this nuclide could be evaluated only in half of the rain samples. Analysis of 18F data has not been completed as of writing, because of the complexity of the origin of 511 keV annihilation gamma-quanta in the measured spectra. Concentrations of the CP nuclides in rainwater are plotted in Fig. 3. Here, it was assumed that the concentration of each CP nuclide was constant over the sampling time and variations in the intensity of rainfall were not considered. Uncertainties given in Fig. 3 consist only of counting statistics (1s). Systematic uncertainties may occur due to simple assumptions that rainfall and concentrations of CP nuclides were constant during the sampling period. As seen in Fig. 3, the concentrations of CP nuclides in rainwater varied rain-by-rain and even in each rain period. Observed concentrations of 7Be 106e107 atom L1 (n ¼ 16, mean 7  106 [1.05 Bq L1]) are 3 10000000

(a)

150

1000000 100

100000 10000

50

1000

28

Mg (28AI) 1778 keV

37

Be 478 keV

100 0

30000

60000

90000 120000 150000

0 0

2000

200

1500

150

1000

100

500

39 CI 1268 keV

0 0

10000

20000

30000

40000

30000

60000

50

90000 120000 150000

24 Na 2754 keV

0 0

30000

60000

90000 120000 150000

Fig. 2. Estimation of Ainfinity by integral method applied to the rain sample of October 5, 2005. Horizontal and vertical axes are time (s) after the start of counting and integral (open circle) and saturation counts (closed circle) estimated by Eq. (1).

108

K. Komura et al. / J. Environ. Radioactivity 96 (2007) 103e109

Atoms per Liter of Rain Water at the Sampling Time days in 2004 200 1E+08

250

300

350

1E+07 Be-7 1E+06 1E+05 1E+04

Na-22

1E+03 Cl-39 1E+02 Na-24

Mg-28 1E+01

July

Aug

Sep

Oct

Nov

Dec

Fig. 3. Concentrations of CP nuclide during the period from June 2004 to November 2005 observed at Tatsunokuchi, Ishikawa Prefecture, Japan.

to 5 orders of magnitude higher than for 24Na 30e1500 atom L1 (n ¼ 16, mean; 520 [6.7 mBq L1]), for 28Mg 80e600 atom L1 (n ¼ 13, mean; 260 [2.4 mBq L1]), for 39Cl 400e1900 L1 (n ¼ 6, mean; 1200 [250 mBq L1]), and for 22Na 2  103e1  105 L1 (n ¼ 9, mean 2  104 [0.2 mBq L1]). According to Ro¨del (1963), who succeeded in detecting 24Na using 200e250 L of rainwater, the number of 24Na atoms were reported to be 235e455 atom L1, which is little lower but rather well agrees with our value (mean 560 atom L1). On the other hand, our 39Cl values are well in the range of 170e3000 mBq L1 reported by Perkins et al. (1965) who measured not only 39Cl but also 38Cl and 38S using extremely low-background NaI(Tl) scintillation counter. 3.4. Information derived from short-lived CP nuclide concentrations If horizontal and vertical movement or mixing of air-masses occurs in a time scale of the half-lives of short-lived CP nuclides, useful information may be obtained by the measurement of CP nuclides in rainwater. Such measurements can provide information on scavenging process of airborne particles and on pollutants because these are effectively scavenged onto ground surface by rainfall. Measurements of the isotopic ratios such as 38Cl/39Cl and 28Mg/24Na is the most promising tool to analyze such phenomena, not only because half-lives of these pairs are similar, but also because these nuclides can be detected in most of the rain samples. In order to explain observed concentrations of CP nuclides, four-box model composed of stratosphere, upper and lower troposphere and ground surface is now under consideration. In the model, a number of parameters such as production of CP nuclide as a function of altitude, solar activity (sun-spot number), geomagnetic latitude, altitude of cloud and its vertical and horizontal movement, residence time of the cloud (rain drop), and intensity of rainfall must properly be considered.

K. Komura et al. / J. Environ. Radioactivity 96 (2007) 103e109

109

4. Conclusions Concentrations of CP nuclides with half-lives shorter than 1 d could be simultaneously determined by rapid chemistry coupled with ultra-low-background gamma spectrometry. This study is believed to provide useful information for the chemistry and geochemistry of atmosphere, meteorology and hydrology.

References Arnold, J.R., Ali Al-Salih, H., 1955. Beryllium-7 produced by cosmic rays. Science 121, 451e453. Arnold, J.R., 1956. Beryllium-10 produced by cosmic-rays. Science 124, 584e585. Davis Jr., R., Schaeffer, O.A., 1955. Brookhaven National Laboratory, BNL-340. Goel, P.S., 1956. Radioactive sulphur produced by cosmic rays in rain water. Nature 178, 1458e1459. Hamajima, Y., Komura, K., 2004. Background components of Ge detectors in Ogoya undergrond laboratory. Applied Radiation and Isotopes 61, 179e183. Komura, K., 1997. Challenge to detection limit of environmental radioactivity. In: Tsujimoto, T., Ogawa, Y. (Eds.), Proceedings of 1997 International Symposium on Environmental Radiation. Tsuruga, Fukui, Japan, October 20, 1997, pp. 56e75. Komura, K., Hamajima, Y., 2004. Ogoya Underground Laboratory for the measurement of extremely low levels of environmental radioactivity: Review of recent projects carried out at OUL. Applied Radiation and Isotopes 61, 185e189. Lal, D., Narasappay, N., Zutshi, P.K., 1957. Phosphorus isotopes P32 and P33 in rain water. Nuclear Physics 3, 69e75. Lal, D., Goldberg, E.D., Koide, M., 1960. Cosmic-ray-produced silicon-32 in nature. Science 131, 332e337. Marquez, L., Costa, N.L., 1955. The formation of 32P from atmospheric argon by cosmic rays. Nuovo Ciment 2, 1038e1041. Marquez, L., Costa, N.L., Almedia, I.G., 1957. The formation of 22Na from atmospheric argon by cosmic rays. Nuovo Ciment 6, 1292e1295. Perkins, R.W., Thomas, C.W., Hill, M.W., Nielsen, J.M., 1965. Chlorine-38 and sulphur-38 produced by cosmic radiation. Nature 205, 790e791. Ro¨del, W., 1963. Sodium-24 produced by cosmic radiation. Nature 200, 999e1000. Winsberg, L., 1956. The production of chlorine-39 in the lower atmosphere by cosmic radiation. Geochimica et Cosmochimica Acta 9, 183e189.