Introduction

Following the recent accident at the Fukushima Dai-ichi nuclear power plant in Japan, radioactive contamination has been observed near the reactor site. In order to search for more widespread distribution of radioactive material, we collected rainwater samples in Berkeley, Oakland, and Albany, California and examined them for the presence of above normal amounts of radioactivity.

Methods

Rainwater samples were collected from March 16–26, 2011 by placing plastic containers outside in the Oakland and Berkeley hills and in Albany. Collection periods varied from a few to approximately 12 hours. Following collection, each sample was placed directly into a Marinelli beaker for gamma-ray counting. This type of re-entrant cylindrical container fits directly over a gamma-ray detector and allows liquid samples to be counted with relatively high detection efficiency. No chemical or physical processing of any kind was done to the rainwater samples before counting.

Each sample was gamma-ray counted using a 60% relative efficiency high-purity germanium detector. The Marinelli beaker was placed directly over the end-cap of the detector. The detector was shielded with 10–15 cm of lead in order to reduce the gamma-ray background from natural sources. Gamma-ray energy spectra from 0.02–1.58 MeV were collected in 16384 channels using the ORTEC Maestro data acquisition system. Counting periods varied from 1–24 hours. Energy calibrations were performed using standard gamma ray sources. The efficiency of the detector for the Marinelli geometry was calibrated using the technique described by Perillo Isaac et al. [1]. Measured masses of high-purity LuCl₃, LaCl₃, and KCl obtained from Alfa-Aesar were dissolved in 1 liter of water and placed into a Marinelli beaker. This mixture of chemicals provides known emission rates of gamma rays at 88, 202, 307, 789, 1436, and 1461 keV from the decays of ¹⁷⁶Lu, ¹³⁸La, and ⁴⁰K, respectively [2]. This source was then counted in the same manner as the rainwater samples. This source thus provided an efficiency curve that spans the energy range of interest for our measurements. Based on these measurements, our photo-peak detection efficiency at 364 keV was determined to be 0.026 and at 723 keV to be 0.018.

Results

Collection of rainwater started on the night of March 15/16. This sample showed no evidence of fission-fragment gamma-rays. From this spectrum, a 1-sigma upper limit on the ¹³¹I activity concentration in this sample was established to be <0.016 Bq L⁻¹ (<0.43 pCi L⁻¹). The first sample which showed activity above background was collected on March 18. Figure 1(a) illustrates the gamma-ray spectrum we observed from a 24-hour count of a 1-liter sample of rainwater collected in Oakland on March 18. For comparison, we show in Figure 1(b) the spectrum we observed from a 24-hour count of a 1-liter sample of Berkeley tap water. Note that the gamma-ray lines seen in figure 1(b) do not originate from the tap water, but are the part of the natural background present in our laboratory environment. The rainwater spectrum shown in Figure 1(a) clearly shows the presence of ^131,132I, ¹³²Te, and ^134,137Cs. These are relatively volatile fission fragments produced with large cumulative yields from the fission of ²³⁵U and ²³⁹Pu [3]. The half-lives (2) of ¹³²Te (3.26 days) and ¹³¹I (8.04 days) are sufficiently short in order for them to present now, that they must have been released from the reactor core(s) rather than from spent fuel repositories. The short-lived ¹³²I, t_1/2 = 2 hours [2], is only present in our water samples because of the in-situ beta decay of ¹³²Te. We see no evidence for the presence of more refractory fission fragments such as ⁹⁵Zr, ⁹⁹Mo, ¹⁴⁰Ba, or ¹⁴⁴Ce that are also produced with large yields.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Gamma-ray spectra observed from rainwater and from tap water.

(a) a portion of the gamma-ray spectrum observed from counting 1 liter of rainwater collected in Oakland on March 18, 2011. Fission-product gamma rays are indentified by their energies in keV and by their beta-decay precursor. (b) the same region of the spectrum collected from counting a 1-liter sample of Berkeley tap water. Both spectra represent the same counting period of 24 hours. Note that the gamma-ray peaks seen in the bottom spectrum do not originate in the tap water but are part of the natural background present in our laboratory environment.

https://doi.org/10.1371/journal.pone.0024330.g001

From each of our measured spectra, the observed counting rates were determined for the major gamma-ray lines seen from each of the above-mentioned fission fragments. Our measured efficiency curve was then used to convert these counting rates into activities per liter of water. The activities were corrected for decay back to the time of collection. The results of our measurements as a function of time of sample collection for ^131,132I, ¹³²Te, and ^134,137Cs are shown in Figure 2. Note that the activities of all the other observed fission fragments are substantially smaller than that of ¹³¹I and all seem to track each other from day to day. The time dependence of the activities we observe in San Francisco Bay area rainwater is undoubtedly a complicated function of many variables including release rates at the Fukushima Dai-ichi reactor site, wind patterns and velocities, as well as local weather conditions.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Activity concentrations of ^131,132I, ¹³²Te, and ^134,137Cs in Bq L⁻¹ measured in San Francisco Bay area rain water as a function of time.

https://doi.org/10.1371/journal.pone.0024330.g002

Discussion

Because of their short half lives, ¹³²Te and ¹³¹I quickly reach their saturation activity levels in the course of normal reactor operation. The cumulative yields of ¹³²Te and ¹³¹I from the thermal fission of ²³⁵U are 4.30% and 2.89%, respectively [3]. For ²³⁹Pu thermal fission, these same cumulative yields are 5.14% and 3.86%, respectively [3]. Thus at the cessation of normal reactor operation, the ratio of activities of ¹³²Te/¹³¹I should be 4.30/2.89 = 1.49 (from ²³⁵U fission) or 5.14/3.86 = 1.33 (from ²³⁹Pu fission). Once normal reactor operations cease, these two activities decrease in time with their respective half lives. Our measured ratios of these two activities can be compared to the expectations from free decay to search for chemical fractionation effects in the processes that led to the release of the isotopes from the reactor(s) and their subsequent transport from Japan to California. For example, 10 days after the cessation of fission production, the ratio of ¹³²Te/¹³¹I activity should be approximately 0.61 (0.54) from a ²³⁵U (²³⁹Pu) fission source. From the data shown in Figure 2, it can be seen that the ratios of ¹³²Te/¹³¹I activities we observe are substantially less than what would be expected if Te and I were released and transported with equal efficiencies. Previous studies of the relative volatilities of Te and I [4] show that Te is much less volatile than I and thus much less tellurium should have been released than iodine. The results we obtained are consistent with this expectation.

The maximum concentration of ¹³¹I we observed in the rainwater was 16 Bq L⁻¹ (430 pCi L⁻¹ from the sample collected on March 24. This maximum activity concentration can be compared to the US EPA limit on ¹³¹I allowed in drinking water of 3 Bq L⁻¹ (81 pCi L⁻¹) [5]. If a person were to drink a typical amount of water per day containing the EPA limit of ¹³¹I, then in one year he or she would receive a whole body dose of <0.04 mSv (4 mrem). This dose should be compared to the US average annual radiation dose of 6.2 mSv (620 mrem) [6]. Due to the short half life of ¹³¹I, it is extremely unlikely that the public will be exposed to anywhere near these levels in drinking water. Thus the levels of fallout we have observed in San Francisco Bay area rain water pose no health risk to the public. A preliminary version of this manuscript was posted online on March 30, 2011 [7].

Acknowledgments

We wish to thank Ren P. Angell for his assistance in collecting the Albany water samples.