135Cs activity and 135Cs/137Cs atom ratio in environmental samples before and after the Fukushima Daiichi Nuclear Power Plant accident

135Cs/137Cs is a potential tracer for radiocesium source identification. However, due to the challenge to measure 135Cs, there were no 135Cs data available for Japanese environmental samples before the Fukushima Daiichi Nuclear Power Plant (FDNPP) accident. It was only 3 years after the accident that limited 135Cs values could be measured in heavily contaminated environmental samples. In the present study, activities of 134Cs, 135Cs, and 137Cs, along with their ratios in 67 soil and plant samples heavily and lightly contaminated by the FDNPP accident were measured by combining γ spectrometry with ICP-MS/MS. The arithmetic means of the 134Cs/137Cs activity ratio (1.033 ± 0.006) and 135Cs/137Cs atom ratio (0.334 ± 0.005) (decay corrected to March 11, 2011), from old leaves of plants collected immediately after the FDNPP accident, were confirmed to represent the FDNPP derived radiocesium signature. Subsequently, for the first time, trace 135Cs amounts before the FDNPP accident were deduced according to the contribution of global and FDNPP accident-derived fallout. Apart from two soil samples with a tiny global fallout contribution, contributions of global fallout radiocesium in other soil samples were observed to be 0.338%–52.6%. The obtained 135Cs/137Cs database will be useful for its application as a geochemical tracer in the future.

is direct for soil samples and indirect via soil-to-plant transfer for newly grown plant samples. Therefore, these data would not reflect the specific FDNPP derived radiocesium signature accurately. For example, two recent studies reported relatively higher 135 Cs/ 137 Cs atom ratios for the standard reference materials released by the Japan Society for Analytical Chemistry (JSAC) 14,26 . These values were somewhat higher than the data (0.32-0.41) from heavily contaminated samples 6,25,27,28 . This indicated a rising proportion of the contribution from global fallout and other background sources in lightly contaminated samples due to the FDNPP accident, since the 134 Cs activities (< 360 Bq kg −1 , decay corrected to March 11, 2011) were relatively low for these standard reference materials. Therefore, increasing the numbers of data for lightly contaminated samples is also highly desired for the analysis of low 135 Cs activities, especially for obtaining a database of 135 Cs/ 137 Cs atom ratios to illustrate the influence of the FDNPP accident.
However, measurement of 135 Cs encounters a great challenge, and there were no 135 Cs data available in Japanese environmental samples before the FDNPP accident. It was only 3 years after the FDNPP accident that limited 135 Cs values could be measured and therefore reported in heavily contaminated environmental samples 6,25,27,28 . The challenge to measure 135 Cs/ 137 Cs in lightly contaminated samples is due to the isobaric ( 135 Ba and 137 Ba) and polyatomic ( 95 Mo 40 Ar + , 97 Mo 40 Ar + , 119 Sn 16 O + , and 121 Sb 16 O + ) interferences, along with the peak tailing effect from 133 Cs during ICP-MS analysis. These will result in higher experimental uncertainties in the measured 135 Cs/ 137 Cs, which may be insufficient to discriminate different sources. Recently, an almost interference-free and non-peak tailing spectrum has been achieved by combining ammonium molybdophosphate adsorption, cation-exchange chromatography, and triple-quadrupole inductively coupled plasma-mass spectrometry (ICP-MS/MS) analysis 26 . The low detection limits of 2.59 × 10 −5 and 3.28 Bq kg −1 for 135 Cs and 137 Cs make it possible to carry out precise 135 Cs/ 137 Cs ratio analysis in lightly contaminated samples.
In the present study, soil and plant samples, both heavily and lightly contaminated by the fallout from the FDNPP accident, were collected immediately after the accident. Subsequently, conventional γ spectrometry was combined with the most advanced ICP-MS/MS available to measure the activities of 134 Cs, 135 Cs, and 137 Cs for studying their distribution and risk assessment. Furthermore, the activity ratios of 134 Cs/ 137 Cs and the atom ratios of 135 Cs/ 137 Cs were also obtained for (1) illustrating the precise radiocesium signature due to the FDNPP accident from almost global fallout-free samples; and (2) gaining new knowledge on the contributions of global fallout and FDNPP accident derived fallout for the lightly contaminated samples. A preliminary database of 135 Cs/ 137 Cs atom ratios due to the FDNPP accident was built in order to widely apply the 135 Cs/ 137 Cs ratio as a new tracer of radiocesium in the future. Finally, the background values of 135 Cs and 135 Cs/ 137 Cs before the FDNPP accident were presented for the first time.

Results
The activities of 134 Cs, 135 Cs and 137 Cs, the activity ratios of 134 Cs/ 137 Cs, and the atom ratios of 135 Cs/ 137 Cs are shown in Table S1. When decay-corrected to March 11, 2011, the 134 Cs and 137 Cs activities in the soil covered wide ranges, from 12.9 to 113 kBq kg −1 and from 14.2 to 110 kBq kg −1 (in dry weight), respectively. Similarly, the highest 135 Cs activity (0.500 ± 0.014 Bq kg −1 ) was 2 orders of magnitude higher than that of the lowest available value (0.005 ± 0.001 Bq kg −1 ), except for three paddy soil samples below the detection limit. Among these three samples, one was a paddy field that was covered by a greenhouse plastic cover and it had the lowest 134 Cs and 137 Cs activities (12.9 ± 4.3 and 14.2 ± 3.8 Bq kg −1 , respectively). The arithmetic means of the 134 Cs/ 137 Cs activity ratio (1.033 ± 0.006) and 135 Cs/ 137 Cs atom ratio (0.334 ± 0.005), from the old leaves of plants collected immediately after the FDNPP accident, were confirmed to represent the FDNPP derived radiocesium signature. Subsequently, trace amount of 135 Cs with the highest activity of 0.0212 ± 0.0024 Bq kg −1 and 135 Cs/ 137 Cs atom ratios up to 4.02 before the FDNPP accident were deduced according to the contribution of the global fallout and the FDNPP accident-derived fallout.

Discussion
Higher radiocesium activities were observed in the northwest direction from the FDNPP, in agreement with the observation that the radionuclides were mainly deposited northwest of the site in a strip approximately 40 km in length 1,4,29-31 . Radiocesium was not released simultaneously from the fuel in the reactors and the spent fuel pools (SFPs) of Units 1-3, and the SFP of Unit 4, but sequentially over the time-span of several days 32 . Therefore, heterogeneous deposition of 134 Cs, 135 Cs, and 137 Cs occurred on the ground, which was consistent with the deposition of other radionuclides 33 . These heterogeneous distributions were also influenced by the initial deposition parameters, e.g., local wind direction and precipitation 1 , and the post-depositional redistribution, e.g., transportation of the soil particle-bound radiocesium by the surface runoff 34 . It can be concluded that certain 134 Cs/ 137 Cs and 135 Cs/ 137 Cs atom ratios, other than the distribution of radiocesium activities, may provide valuable information about sources of radiocesium, by comparing them with the signatures of Units 1-4.
In order to reduce the influence of global fallout on obtaining more accurate 134 Cs/ 137 Cs and 135 Cs/ 137 Cs signatures for the FDNPP accident, extremely heavily contaminated plant leaves with higher 134 Cs (0.076-0.205 MBq kg −1 ) and 137 Cs (0.073-0.199 MBq kg −1 ) were also selected for the 135 Cs study. As shown in Table S1, relatively higher 135 Cs concentrations, varying from 0.320 to 0.881 Bq kg −1 -wet weight, were observed as expected.
Because of the easiness and importance of its measurement, the 134 Cs/ 137 Cs activity ratio was immediately applied as a distinguishing indicator between the Chernobyl accident and the FDNPP accident fallout 1,24 , and as an index for evaluating the contamination from each FDNPP reactor unit 11,25 . Compared with soil samples, it is more plausible to employ old leaves of plants collected immediately after the FDNPP accident to illustrate the radiocesium signature of FDNPP accident; since old leaves of plants have less deposition of global fallout than that of soil, and plant leaves generally have non-accumulating characteristics for Cs via soil-to-plant transfer 35 . The 134 Cs/ 137 Cs activity ratios for the plant leave samples covered a narrow range of 1.02-1.04. Moreover, the 135 Cs/ 137 Cs atom ratios for the plant leaves were also in a narrow range of 0.326-0.338, which was comparable with the range of values from leaf litter samples (0.333-0.341) collected in May 2011 6,27 . On the other hand, wider ranges of 0.907-1.05 and 0.315-0.419 were observed for the 134 Cs/ 137 Cs activity ratios and 135 Cs/ 137 Cs atom ratios in soil samples, respectively, as reported by the Ministry of Economy Trade and Industry (METI) 36 and other authors 11,23,24 .
In terms of radionuclide signatures, radiocesium ratios released due to the Chernobyl accident, the FDNPP accident, and the global fallout from atmospheric nuclear weapon explosions were clearly distinct in Japan. Chernobyl-derived radiocesium had a lower 134 Cs/ 137 Cs activity ratio of 0.50-0.52 Bq/Bq (decay corrected to April 26, 1986) 37 , and presented a higher 135 Cs/ 137 Cs atom ratio of 0.480-0.589 (decay corrected to March 11, 2011) 6,12,13,18,26,27,38 . The global fallout radiocesium after the Chernobyl accident that fell in Japan had a 134 Cs/ 137 Cs activity ratio of 0.48-0.63 Bq/Bq (decay corrected to April 26, 1986) 39 . Unfortunately, the 135 Cs/ 137 Cs atom ratio in the global fallout of Japan is not available, and even values in other places were limited due to the challenge in measuring 135 Cs 19,21,22 . Figure 1 presents the variations of 134 Cs/ 137 Cs activity ratios (in green) and 135 Cs/ 137 Cs atom ratios (in blue) compared with corresponding 134 Cs activities in order to illustrate the effects of both global fallout and FDNPP accident fallout. Since the half-life of 134 Cs is short (t 1/2 = 2.06 y), the environmental 134 Cs contamination before the FDNPP accident was basically zero. Therefore, the environmental 134 Cs activities after the FDNPP accident indicated the contamination level due to this accident, that is, lower environmental 134 Cs activities indicated a smaller proportion of radiocesium contamination from the FDNPP accident and a larger proportion of radiocesium contamination from the global fallout. The global fallout contribution to environmental samples would increase and it would become obvious in samples lightly contaminated by the FDNPP accident; that is, the 134 Cs/ 137 Cs activity ratio would present a decreasing trend and 135 Cs/ 137 Cs atom ratio would show an increasing trend for these samples. The reported higher 135 Cs/ 137 Cs atom ratios (0.3808-0.3896) from grass, bark, and moss, collected in August 2011, may have some influence from the soil-to-plant transfer of global fallout radiocesium 25 . Sample S40 from a paddy field in Namie Town, with low 134 Cs (84 ± 5 Bq kg −1 ; Table S1) and 135 Cs below the detection limit, presented abnormally low 134 Cs/ 137 Cs activity ratios of 0.489 ± 0.045, indicating a larger contribution of 137 Cs from global fallout for that collection site. Furthermore, the higher 135 Cs/ 137 Cs atom ratio of JSAC standard reference materials also indicated a larger contribution of global fallout and other background sources given their lower 134 Cs activity (< 360 Bq kg −1 , decay corrected to March 11, 2011) 14,26 . In brief, radiocesium isotopic ratios, from the old leaves of plants collected immediately after the FDNPP accident, have great potential to represent the specific radiocesium signatures derived from the FDNPP accident.
For the source term identification, the evaluated isotopic compositions in the reactor cores of damaged Units 1-3 and in the SFPs have been frequently employed ( Table 1). The results are given for irradiated uranium pellets and activated cladding tubes of zirconium alloy in the core and the SFPs of the respective reactors 40 . Radionuclides released into the stagnant water in the turbine buildings and their surroundings on the FDNPP site have been measured, providing a more accurate radioactive signature of the FDNPP accident 41 . The 134 Cs/ 137 Cs activity ratios and 135 Cs/ 137 Cs atom ratios in these two kinds of data were employed in the present study to illustrate radiocesium source for the FDNPP accident. Figure 2 compares the 134 Cs/ 137 Cs activity ratios and 135 Cs/ 137 Cs atom ratios in the old leaves of plants collected immediately after the FDNPP accident, nuclear fuel in the damaged reactor cores, and nuclear fuel in the SFPs. The radiocesium isotopic compositions of all four SFPs were clearly different from those observed in the heavily contaminated leaves, indicating the potential release of radiocesium from SFPs was negligible, if any occurred. As shown in Table S2, the METI has estimated the released amounts of 137 Cs from Unit 1, Unit 2, and Unit 3 were 3.86%, 91.5%, and 4.64%, respectively 36 . In a cluster analysis, the radiocesium isotopic compositions of sampled leaves belonged to the cluster of Unit 2 and Unit 3 ( Figure 2). Because the 135 Cs/ 137 Cs atom ratios were not available in the stagnant water samples from the FDNPP, only 134 Cs/ 137 Cs activity ratios in the highly contaminated leaves were compared to that from the stagnant water in the basement of the turbine building (Tb), basement of the reactor building (Rb), and a trench (Tr), as shown in Figure 3. It was observed clearly that the 134 Cs/ 137 Cs activity ratio in the basement of Tb of Unit 2 was the closest value to that of the leaves.
All these findings further illustrated that the arithmetic means of the 134 Cs/ 137 Cs activity ratio (1.033 ± 0.006) and 135 Cs/ 137 Cs atom ratio (0.334 ± 0.005), from the old leaves of plants collected immediately after the FDNPP accident, could be applied as the specific FDNPP accident-derived radiocesium signatures to study the effect of the FDNPP accident on other lightly contaminated environmental samples.
Obtaining background data in the environment on radiocesium isotopic ratios is essential to widen their application as a geochemical tracer in the future and for assessment of the environmental impact of Cs released from the FDNPP accident. However, to the best of authors' knowledge, there were no reported 135 Cs/ 137 Cs atom ratios in environmental samples before the FDNPP accident in Japan, and the first sets of 135 Cs/ 137 Cs atomic ratios in plants and soil 6,25,27 and in rainwater 28 were not reported until 3 years after the FDNPP accident, due to the challenge in measuring 135 Cs. Before the FDNPP accident, the most recent major radiocesium contribution was from the Chernobyl accident. After the Chernobyl accident, the Meteorological Research Institute at Tsukuba City monitored 16 radionuclides, including 134 Cs and 137 Cs, in the fallout deposited in Japan during a monthly sampling program of the total fallout at eleven stations to obtain the temporal and spatial distributions of the radionuclides 39 . In the present study, the 135 Cs background in the Fukushima area was obtained based on the following assumptions: (1) The radiocesium in Japanese environmental samples were from global fallout due to the combination of atmospheric nuclear weapon explosions and the Chernobyl accident, and fallout from the FDNPP accident; (2) The mean 134 Cs/ 137 Cs activity ratio of (2.13 ± 0.17) × 10 −4 (Table S3) (decay corrected to March 11, 2011) in the wet and dry precipitation depositions from the eleven Japanese stations in 1986 was employed as the 134 Cs/ 137 Cs global fallout background activity ratio before the FDNPP accident;  40 . b Activity ratio reported by Nishihara et al. 41 . c Atom ratio reported by Nishihara et al. 41 .   Subsequently, a simple two-end member mixing model was employed to calculate the relative contributions of the background radiocesium: where R is the 134 Cs/ 137 Cs activity ratio or 135 Cs/ 137 Cs atom ratio; and subscripts F, G, and S refer to the FDNPP accident fallout, global fallout, and the soil sample, respectively. These relative contributions of the background radiocesium are shown in Table S4. Apart from two soil samples with a tiny contribution from global fallout, the observed contribution of global fallout radiocesium ranged from 0.338% to 52.6% in all other samples. Sample S40 from a paddy field in Namie Town, with low 134 Cs (84 ± 5 Bq kg −1 ; Table S1), had the highest contribution of global fallout radiocesium and it was abnormally high. The 135 Cs background before the FDNPP accident was a trace amount, with data of five samples below the limit of detection. For other samples, a larger variation was observed, with the maximum value (0.0212 ± 0.0024 Bq kg −1 ) two orders of magnitude higher than the lowest value. As shown in Fig. 4, the soil samples collected close to the FDNPP had relatively higher 135 Cs concentrations, which may have been due to the operation of the FDNPP from 1971 to 2011. Regarding the 82.1% availability of the 135 Cs/ 137 Cs atom ratios in the global fallout before the FDNPP accident, the range of 0.028-4.02 (decay corrected to March 11, 2011) was observed in the present study. These are the first batch of deduced values of 135 Cs contents and 135 Cs/ 137 Cs atom ratios in global fallout before the FDNPP accident in Japan, and they can be employed in the future to illustrate the exact contribution of the FDNPP accident to the environment. It should be noted that more factors, such as the operation of the FDNPP, should be considered in the model to get a more accurate 135 Cs background value in the future.

Methods
Soil and Plant Sampling. The details of procedures for sampling and pretreating soil and plant samples have been described elsewhere 33 . Surface soils (0-5 cm) and plant leaves were collected from 67 sites in Fukushima Prefecture (Fig. 5) on four sampling expeditions in 2011, from April 12 to 16, April 26 to 28, June 6 to 10, and June 15 to 16. For the soil analysis, stones and plant roots were removed by handpicking and soil was transferred into a 100-mL polystyrene container. Leaves were collected from the upper part of the plant, so as not to be newly grown after the radionuclide contamination, to minimize soil-to-plant transfer from the contaminated soil. Then, they were cut into 1 × 1 cm pieces with scissors prior to further treatment.

Measurement of Isotopes.
The concentrations of 134 Cs and 137 Cs were determined by γ -ray spectroscopy (ORTEC GEM-40190, Seiko-EG&G, Tokyo, Japan) at energies of 604 keV and 662 keV, respectively. Mixed gamma standard sources distributed from the Japan Radioisotope Association were employed for efficiency correction. After organic matter decomposition in a muffle oven at 450 °C for 2 h, leaching with 20 mL of concentrated HNO 3 was performed in PFA jars with lids (Savillex, Eden Prairie, MN, USA) on a hot plate at 180 °C for 2 h. During acid leaching, for those samples with high organic contents, 5 mL H 2 O 2 was added. After filtration, the sample solutions were adjusted to 4 M HNO 3 for Cs separation. The purification of Cs was conducted as described in Figure S1 following the method of Yang et al. 26 . Agilent 8800 (ICP-MS/MS, Agilent Technologies, Santa Clara, CA, USA) featuring an octopole collision/reaction cell situated between quadrupole mass filters (first, Q1; second, Q2) was employed for analysis of 134 Cs/ 137 Cs and 135 Cs/ 137 Cs ratios. The optimized operation parameters are summarized in Table S5. Finally, 135 Cs activity could be obtained by combining the data from γ spectrometry and ICP-MS/MS.