Fukushima-derived radionuclides in sediments of the Japanese Pacific Ocean coast and various Japanese water samples (seawater, tap water, and coolant water of Fukushima Daiichi reactor unit 5)

We investigated Ocean sediments and seawater from inside the Fukushima exclusion zone and found radiocesium (134Cs and 137Cs) up to 800 Bq kg−1 as well as 90Sr up to 5.6 Bq kg−1. This is one of the first reports on radiostrontium in sea sediments from the Fukushima exclusion zone. Seawater exhibited contamination levels up to 5.3 Bq kg−1 radiocesium. Tap water from Tokyo from weeks after the accident exhibited detectable but harmless activities of radiocesium (well below the regulatory limit). Analysis of the Unit 5 reactor coolant (finding only 3H and even low 129I) leads to the conclusion that the purification techniques for reactor coolant employed at Fukushima Daiichi are very effective.


Introduction
Following the earthquake on March 11, 2011, a gigantic tsunami destroyed the cooling systems of Fukushima Daiichi nuclear power plant (NPP)(Japan) and caused partial melt-down of 3 reactor cores. In the course of the nuclear accident a total of 520 PBq (excl. noble gases) were released to the atmosphere [1]. More than 99 % of the released substances were radionuclides of Kr, Te, I, Xe, and Cs. These nuclides have been monitored globally in air [2,3]. Less volatile radionuclides such as radiostrontium [4,5], plutonium [6][7][8] or radionuclides that are difficult to measure, such as 3 H [9,10], 135 Cs [11] or 35 S [12] have been monitored much less frequently [13]. In any case, at least 80 % of the airborne radionuclides were transported offshore by the wind [14,15], and a yet unknown amount has leaked directly into the Pacific Ocean. Although several studies have addressed the impact of the accident on the Pacific Ocean and its organisms [16][17][18], our knowledge on the impact of the Fukushima nuclear accident on the marine environment is yet far from complete.
In the current study, we investigated several interesting and unique sample materials, many of which were taken directly after the accident and/or inside the ''exclusion zone'' around the crippled reactors. Target nuclides were radiocesium [ 134 Cs (T 1/2 = 2.1 years) and 137 Cs (T 1/2 = 30.2 years)] as well as the understudied nuclides 90 Sr (T 1/2 = 28.6 years) and 3 H (T 1/ 2 = 12.3 years). For two samples, also 129 I analysis was performed. It is a long-lived (T 1/2 = 15.7 million years) fission product that can be used as an (environmental) tracer nuclide [19] that is also useful for retrospective dosimetry after the decay of short-lived 131 I (T 1/2 = 8 days) [20] which accumulates in the thyroid [21] and is mostly responsible for thyroid cancer cases after Chernobyl.

Materials and methods
Environmental materials analyzed in this study included sediments and seawater from the Japanese Pacific Ocean coast inside the exclusion zone, tap water from Tokyo,  The coolant was investigated also for 129 I to  scrutinize the overall efficiency of the techniques currently  employed for radionuclide removal from water at the  Fukushima Daiichi NPP; most fission and activation products are cations that can be removed using natural or artificial cation exchanger. Iodine, however, is usually anionic (iodide or iodate) and thus more challenging to remove. A summary of all samples, the exact location and date of sampling, distance to the Fukushima Daiichi NPP as well as target nuclides is given in Table 1.
A map of the sample locations (excluding the seawater sample from Hawaii), is given in Fig. 1.
Gamma measurements for 134 Cs and 137 Cs were conducted with an ORTEC TM 364 cm 3 HPGe detector with a 0.76 mm Be window (2.32 keV resolution at the 1332 keV 60 Co peak; 87.4 % relative efficiency). The samples were weighed into 60 or 125 mL polypropylene containers (typically 100 mL for water samples; 70 g for sediment samples) that were placed on top of the detector. The samples were measured for at least 24 h. Detector efficiency calibration was done with an Eckert and Ziegler Ò multinuclide standard solution that was diluted in 2 M HCl (aqueous samples) or dispersed in quartz sand (sediment samples) (see [22]). Tritium ( 3 H) and 90 Sr was conducted with liquid scintillation counting (LSC). The LSC measurements were performed using a LabLogic 300SL Liquid Scintillation Counter with TDCR Technology (Hidex, Mustionkatu 2, FIN-20750 Turku, Finland). The user interface software is provided by MikroWin TM (Mikrotek Laborsysteme GmbH, Olper Strasse 35, D-51491 Overath, Germany). Instrument output was transferred to a vendorprovided PC-application spreadsheet with macros where the raw data could be analyzed and a graphic routine provided for convenient data visualization [23].
For radiocesium measurements, the sediment samples were dried at 100°C, weighed and filled into calibrated geometries at the gamma detector station at Colorado State University. Water samples were weighed and measured for gamma emitters without further pre-treatment. Pure beta emitters such as 90 Sr or 3 H require chemical separation and purification prior to beta counting in the LSC. Water samples were carefully distilled to dryness; the distillate was mixed with Ultima Gold TM low level tritium (LLT) scintillation cocktail for measurement. Ocean sediments were treated with nitric acid for Sr leaching and purified using the Eichrom TM SR resin (2 mL cartridges), as described elsewhere [4]. In brief, the separation process involved weighing of the dried sand into a flask, addition of 8 mL HNO 3 (8 M), 2 mL H 2 O 2 (30 %), 1 ml Sr carrier solution (c Sr = 1.2 mg mL -1 ), and 3 mL of concentrated HNO 3 (70 %). This solution was boiled under reflux for 30 min and then filtered through a paper filter. The filtrate was loaded onto preconditioned (with 8 M HNO 3 ) SR resin cartridges. The flask was rinsed with 3 9 2 mL HNO 3 (8 M). The resin was then rinsed with 10 9 1 mL of a mixture of HNO 3 (3 M) and oxalic acid (0.05 M). Then the Sr fraction was eluted with 0.01 M HNO 3 , which we found to be sufficiently acidic to not eluate any radiolead from the column. Water samples (50 mL) were either brought to 8 M HNO 3 by addition of equal amounts of 16 m HNO 3 , or the solid residue after tritium distillation was taken up in 8 M HNO 3 . Since the natural Sr content in water is much lower than from leached sediments, the amount of Sr carrier was increased to 1.7 mg. The loading, rinsing and eluting processes were conducted as described above.
We found that the acidic eluate (0.01 M HNO 3 ) does not mix well with the Ultima Gold TM LSC cocktail. In order to remove the acid, the eluate was evaporated to almost dryness and then taken up in 1 mL H 2 O again. This step was repeated 10 times. Then, the Sr fraction was transferred to LSC vials and the flask was rinsed with 4 9 0.5 mL H 2 O. Finally 18 mL of scintillation cocktail were added.
Addition of HNO 3 to sediments should be performed carefully as the carbonate fraction of the sand will vigorously decompose under formation of CO 2 gas which could cause the overflow of the flask with acidic foam. In our case, the sediments proved to have a very low carbonate content. Also organic substances may react violently upon exposure to the highly oxidizing HNO 3 /H 2 O 2 mixture, especially when heated. If in doubt, the use of protective measures (such as gloves, face shields) is recommended [24][25][26].
For the accelerator mass spectrometry (AMS) measurements of 129 I, iodine was separated from the matrix in two steps. First, Woodward iodine carrier was added and all the iodine was transferred into iodide. All iodine species were oxidized with Ca(OCl) 2 to iodate and afterwards reduced with NH 3 OHCl and NaHSO 3 to iodide. The second step consisted in an ion exchange separation using a DOWEX Ò 1 9 8 analytical grade ion exchange resin, which was preconditioned with KNO 3 . After rinsing the ion exchange columns with high purity water and a 0.5 mol L -1 KNO 3 solution, the iodine was eluted with concentrated potassium nitrate solution (2.25 mol L -1 ). The iodine was precipitated as AgI, which was dried, mixed with silver powder (AgI:Ag 1:4 by weight), and pressed into titanium targets for AMS measurement. Measurements were performed with low-energy accelerator mass spectrometry (AMS) at the 0.5 MV TANDY facility at ETH Zurich, Switzerland.
For the ICP-MS measurement of 127 I no iodine matrix separation was necessary. The samples were diluted and tetramethylammoniumhydroxide (TMAH) was added, which causes hydrolysis of any organic compounds and reduces the redox potential. The stability of Iincreases in the sample relative to I 2 and IO 3 -. Moreover, iodide and iodate have not significantly different sensitivities in the ICP-MS measurements. A Thermo X7 (Thermo Fisher Scientific) was employed for the detection and quantification of stable 127 I [27].

Results and discussion
Results of the radionuclide analysis are summarized in Table 2. Tritium, 90 Sr, 134 Cs, and 137 Cs were obtained by radiometric means, 129 I by AMS. Please note that the seawater sample from Hawaii was only analyzed for 129 I, as radiometric methods were not deemed successful for the presumably very low contamination levels.

Ocean sediments
As expected, Pacific Ocean sediments from inside the exclusion zone exhibited the highest contamination levels. The correlation between radiostrontium and 134 Cs as well as 137 Cs is illustrated in Fig. 2.
Although a slight distance dependency can be observed (at least sample Sed 1 that was taken closest to the Fukushima Daiichi NPP exhibited the highest activity concentration), it was surprising to note that sample Sed 3 (taken 16.1 km north of the NPP) exhibited the lowest activities. It is also interesting that the radiocesium activity concentrations fluctuate by more than an order of magnitude within the samples, the 90 Sr activity concentrations, however, only by a factor of 2. The 90 Sr/ 137 Cs activity ratios (where calculable) were 0.012 ± 0.002 (Sed 1), 0.036 ± 0.005 (Sed 2), 0.15 ± 0.03 (Sed 3), 0.012 ± 0.002 (Sed 4), and 0.024 ± 0.004 (Sed 5). It is unclear why Sed 3 has a ten times higher 90 Sr/ 137 Cs ratio than the other sediments (which fluctuate within a factor of 3). This odd ratio is not so much due to a high 90 Sr concentration, which is in the range of the other samples, but rather due to an exceptionally low 137 Cs concentration. The reason for this anomaly (which has been observed in similar manner in terrestrial environmental samples from north of the NPP [4] ) is yet unclear. Although, from an radioecological point of view, sediment is not comparable to food, the high 90 Sr/ 137 Cs ratio of 0.15 challenges the governmental assumption for the regulatory limits in food, namely a constant 90 Sr/ 137 Cs ratio in food of B0.1 [28]. Seafood organisms exposed to high-90 Sr effluents may exhibit higher 90 Sr concentrations than covered by this governmental assumption.
The 134 Cs/ 137 Cs activity ratios are reliable source identifiers [29]. The average 134 Cs/ 137 Cs ratio at the time of the accident (March 11, 2011) was found to be 0.98 [28]. Due to the shorter half-life of 134 Cs, this average ratio has been calculated to have gone down to 0.77 by the time of sampling (Dec 20, 2011). The ratios found were 0.81 ± 0.05 (Sed 1), 0.91 ± 0.08 (Sed 2), 0.75 ± 0.25 (Sed 3), 0.77 ± 0.01 (Sed 4), and 0.76 ± 0.02 (Sed 5). Sed 2 seems to exhibit a somewhat higher 134 Cs/ 137 Cs ratio than the other sediments, however, with the given uncertainty margins, this outlier seems not significant. It would be interesting to study the 135 Cs/ 137 Cs ratio [11] to possibly identify the source of these contaminations.
Other studies [30] for radiocesium in sea sediments found up to 250 Bq kg -1 -dry in Sendai (north of the Fukushima Daiichi NPP), which is naturally somewhat lower than the levels reported in this study, but in good agreement with our results. Our findings are also in good agreement with the activity concentrations reported by Yamamoto et al. [31]. A study on sediments from the Sea of Japan naturally found much lower radiocesium activity concentrations (0.25 Bq kg -1 137 Cs and 0.15 Bq kg -1 134 Cs in the Sado Basin) [32].  For comparison, it may be interesting to note that the activity levels found in these Ocean Sediments are roughly comparable to the radiocesium inventory in soil in Central Europe, e.g., in Austria [22], which is still significantly polluted from the Chernobyl accident and the fallout from 20 th century's atmospheric explosions.

Seawater
With only small volumes available for analysis, it was not expected to detect any radionuclides. However, low activities of radiocesium were found (with high uncertainties), as summarized in Table 2. Given the high uncertainties 134 Cs/ 137 Cs ratios are less meaningful; they are 0.89 ± 0.3 (OW 1),\1.08 (OW2), [4.5 (OW4), and 1.5 ± 0.7 (OW5). In any case, our findings for seawater (up to 2.8 Bq kg -1 137 Cs and 2.5 Bq kg -1 134 Cs) are in good agreement with previously reported data [30] for seawater.
No tritium was detected in any of the seawater samples. The 129 I activity concentration in the seawater sample from Hawaii from 2012 (2.37 9 10 -8 ± 1.9 9 10 -9 ; 129 I/ 127 I ratio 1.10 9 10 -10 ) are in the range of the pre-Fukushima background reported by Stan-Sion et al. [33].

Tap water
Tap water from Tokyo exhibited contamination levels that were of no health concern (up to 4.3 Bq kg -1 137 Cs). The activity concentrations found in this study are in good agreement with previously reported tap water results [28,29]. The radiocesium activity concentrations were clearly below the regulatory limit for radiocesium in tap water of 200 Bq kg -1 . Previous studies [28] showed that 131 I was probably also present in Tokyo's tap water at the time of sampling; however, the measurements of this study were performed too late for the detection of this short-lived radionuclide. Strontium-90 and tritium could not be detected in tap water because of too low sample volumes and the limited detection limit of the LSC, respectively.
The 134 Cs/ 137 Cs activity ratios slightly deviated from the average ratio [28] of 0.98 at the time of the accident (0.97-0.96 at the time of sampling would have suggested): we found ratios of \0.2 (TW6), 0.90 ± 0.25 (TW7), 0.87 ± 0.09 (TW8), and 1.7 ± 0.6 (TW9). Like TW6, TW7, and TW8, also the ratios found in Tokyo tap water in a previous study [28] were consistently lower than the average; only TW9 exhibited a higher ratio, which raises the question of the source of these contaminations.

Reactor coolant
Without a doubt, the sample of reactor coolant was the most ''unusual'' sample of this study. The analysis of this sample allows the investigation of the performance of the water purification techniques currently employed at the Fukushima Daiichi NPP. Purification is performed in multiple steps, including the application of charcoal and zeolithe as adsorbent materials as well as reverse osmosis techniques. This purification method proves to be extremely efficient: No radiocesium and no radiostrontium could be detected. Also, the content of 129 I is surprisingly low; it is just two orders of magnitude higher than 2012 Pacific Ocean water from Hawaii (see Table 2). By far the dominant activity in the reactor coolant is, as expected, tritium (HTO) that cannot be removed from water by chemical means.

Conclusions
The sea sediments investigated in this study were found to be contaminated with radiocesium (up to almost 800 Bq kg -1 ) and radiostrontium (up to 5.6 Bq kg -1 ). This is one of the first reports on radiostrontium in sea sediments from the Fukushima exclusion zone. Seawater exhibited radiocesium contamination levels up to 5.3 Bq kg -1 , which is also in agreement with previous findings. Tap water from Tokyo weeks after the accident exhibited detectable but harmless activities of radiocesium (well below the regulatory limit). The investigation of the reactor coolant (finding only 3 H and even low 129 I) leads to the conclusion that the purification techniques for reactor coolant employed at Fukushima Daiichi are very effective.