Introduction

Most plutonium isotopes in the environment today are derived from artificial sources, primarily atmospheric nuclear weapon tests during 1945–1980, though an extremely small amount of 239Pu occurs naturally1. Atmospheric detonations of 502 nuclear devices with a total yield of 440 Mt (TNT equivalent) have occurred at the proving sites shown in Fig. 1, including the two combat uses in Hiroshima and Nagasaki. The former USSR performed nuclear testing mainly at high latitudes in the Northern Hemisphere, whereas the USA performed testing at low latitudes. Additional detonation tests were carried out by People’s Republic of China at middle latitude of Northern hemisphere and by France and the UK in the Southern Hemisphere. Plutonium introduced into the troposphere and stratosphere by these detonations was advected and diffused in the atmosphere before final deposition throughout the world2.

Figure 1
figure 1

Atmospheric nuclear weapon test sites and cumulative total yields at each site (indicated by symbol size) during 1945–196627. Four tests in the Pacific Ocean and three in the Atlantic Ocean are not included because of no description of accurate detonation sites27. Their cumulative yields are approximately 0.11 Mt. Background map is created using Adobe Illustrator CS6 (https://www.adobe.com/).

Full size image

Various industrial and military accidents have dispersed a much smaller amount of Pu compared to nuclear weapon testing. Nuclear accidents at the Chernobyl nuclear power plant (NPP) in the former USSR in 19863 and the Fukushima Dai-ichi NPP in Japan in 2011 dispersed artificial radionuclides including Pu4. A waste storage tank at the Mayak nuclear material production complex in the Chelyabinsk region in the former USSR had major leaks in 1957, and the leakage, containing Pu, flowed into the Techa River5. Furthermore, bomber aircrafts carrying atomic bombs crashed in Palomares, Spain in 19666 and Thule, Greenland in 19687, resulting in local Pu contamination.

Plutonium isotopic compositions differ depending on their origin. For example, the mean 240Pu/239Pu atomic ratio (±a standard deviation) of cumulative global fallout Pu is 0.18 ± 0.01 in surface soils collected from 24 sites in 31–70°N region8. 240Pu/239Pu atomic ratios above 0.18 were observed in soils around the Chernobyl3,9 and Fukushima Dai-ichi NPPs4, whereas the 240Pu/239Pu atomic ratios of spent nuclear fuels from civilian nuclear reactors vary based on the type of reactor and burn-up10,11. Ratios below 0.18 have been measured in soils around aircraft crash-sites12,13. Plutonium isotopic signatures can be used to identify the origin of Pu14,15,16 and as a tracer for studying the behaviour of Pu in the environment17.

Although the 240Pu/239Pu atomic ratio of cumulative global fallout Pu has been widely used to evaluate the origin of Pu, some researchers have reported temporal variations of the ratios in crop archives11, lake sediment cores16,17, and glacial ice core samples11,18,19. The 240Pu/239Pu atomic ratios of aerosol samples collected in the 1960s20 were not in agreement, likely due to differences in sampling locations, the representativeness of the samples, and analytical uncertainties. Despite this discrepancy, peaks in mid-latitude atmospheric concentrations and Pu deposition were observed in 1963, just after the peak of high yield atmospheric testing in 1962. Therefore, 240Pu/239Pu atomic ratios dating to the 1960s are important for the use of Pu atomic ratios as tracers. Although 242Pu/239Pu atomic ratios in atmospheric deposition has been scarcely reported, the ratio is useful for studying on the behaviour of Pu in the environment.

Here we report precise 240Pu/239Pu and 242Pu/239Pu atomic ratios in mid-latitude atmospheric fallout (wet + dry) samples collected in Japan during 1963–1966, as measured by multi-collector inductively coupled plasma mass spectrometry (MC-ICP-MS). Our results are discussed in the context of the atmospheric half-life of Pu transported in the stratosphere.

Results

Atmospheric deposition sample archives

The Meteorological Research Institute (MRI, Japan) has collected monthly atmospheric deposition samples with open surface samplers since April 1957 in Tokyo (35° 42′N, 139° 36′E) and since June 1959 in Akita (39° 43′N, 140° 06′E). Distilled water was put into samplers with effective surface areas of 0.5 or 1 m2. After each monthly collection (wet and dry deposition), the samples were evaporated to dryness, and the residues were dried at 110 °C. An aliquot of each residue was analysed for 239 + 240Pu by α-ray spectrometry, and the results were published as monthly 239 + 240Pu deposition rates21,22,23,24,25. Here, we analysed the 240Pu/239Pu and 242Pu/239Pu atomic ratios of aliquots of the remaining samples collected during March 1963 to May 1966, excluding April 1963 and March 1966 because these samples did not remained. The cumulative 239 + 240Pu deposition in Tokyo during March 1963 to May 1966 was reported to be 19.7 Bq m−2, which is about 40% of that during 1945–2000 (47.7 Bq m−2)24,25. Atmospheric deposition samples from both Tokyo and Akita from seven months (July, August, September, October, and November 1963 and November and December 1964) were analysed to compare between the two locations 450 km apart. For the other months, the analysed samples were collected in either Tokyo or Akita (Table 1).

Table 1 240Pu/239Pu and 242Pu/239Pu atomic ratios of atmospheric deposition samples collected during March 1963 to May 1966 in Tokyo and Akita, Japan.
Full size table

240Pu/239Pu and 242Pu/239Pu atomic ratios of atmospheric deposition samples

The 240Pu/239Pu and 242Pu/239Pu atomic ratios of the atmospheric deposition samples are reported in Table 1. Comparison of the 240Pu/239Pu and 242Pu/239Pu atomic ratios measured in samples collected in Tokyo and Akita during the same month show good correlations between the two locations (Fig. S1; Pearson’s correlation coefficient: r2 = 0.98 for 240Pu/239Pu and r2 = 0.97 for 242Pu/239Pu), i.e. the Pu isotopic compositions of atmospheric deposition samples were almost the same between Tokyo and Akita.

Temporal variations of the 240Pu/239Pu and 242Pu/239Pu atomic ratios deposited in Tokyo and Akita during March 1963 to May 1966 (Fig. 2A,B) are compared to monthly 239 + 240Pu atmospheric deposition fluxes in Tokyo during the same period (Fig. 2C)21. From March 1963 (240Pu/239Pu = 0.186), monthly deposition rapidly increased through June 1963 and then gradually increased to a maximum value of 0.249 in September 1963. The ratio then gradually decreased to 0.180 in August 1964 and remained almost constant with occasional small variations throughout the rest of the study period. The temporal variations of 242Pu/239Pu were similar to those of 240Pu/239Pu.

Figure 2
figure 2

Plutonium atmospheric deposition in Tokyo and Akita, Japan during 1963–1966. (A) 240Pu/239Pu and (B) 242Pu/239Pu atomic ratios of monthly deposition samples collected in Tokyo and Akita. (C) Monthly 239 + 240Pu deposition flux in Tokyo, from Miyake et al.21 240Pu/239Pu atomic ratios (A) are compared to those of aerosol samples collected near ground in Finland26, UK harvest archives11, Alpine11, Arctic18, and Siberian Altai ice-core samples19. Both 240Pu/239Pu (A) and 242Pu/239Pu atomic ratios (B) are compared to those of high-altitude aerosol samples collected at 35° N20. Equations in (A) and (B) present the fitting curves from temporal variations of the ratios of our deposition samples during September 1963 to May 1966.

Full size image

Discussion

The Pu isotopic ratios in global fallout have been measured in various environmental samples. For comparison with our data, we selected Pu isotopic ratios measured in aerosols20,26, plant archives11, and ice core samples11,18,19 (Fig. 2A), because they represent simple, non-remobilized Pu transport pathways and have similar temporal resolutions. The observed rapid increase of the 240Pu/239Pu atomic ratio from the end of 1962 to summer 1963 was also observed in aerosol samples collected near the ground in Finland26, though measurement uncertainties were relatively large, and variations in late 1963 were substantial. Alpine ice core samples showed similar, but narrower, 240Pu/239Pu variations as observed in our data, due to their limited temporal resolution. The peak 240Pu/239Pu atomic ratio observed in 1963 did not appear in Arctic ice core samples18 or aerosols collected at high altitudes (4.6–21 km at 35°N)20. Rothamsted harvest archive samples collected during 1963–196611 showed values similar to our data.

The fission yields of nuclear weapon testing were compiled and partitioned into tropospheric, stratospheric, and local and regional atmospheric inputs shown in Fig. 3 by the United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR)27. Local inputs did not affect Pu deposition in Tokyo and Akita during the study period because of their long distances from proving sites, Hiroshima, and Nagasaki. While there was no atmospheric nuclear weapon testing during 1963, the USA and the former USSR tested their high yield devices during 1961–1962 before the conclusion of the Partial Test Ban Treaty (PTBT) in August 1963. China performed small-scale nuclear testing in 1964 and 1965. In 1966, China and France began testing relatively high yield thermonuclear explosions over 0.1 Mt27.

Figure 3
figure 3

Annual fission yields injected into the stratosphere, troposphere, and local and regional atmospheric regions during 1945–196727.

Full size image

Since the mean residence time of airborne particulates in the troposphere is 30–70 d23,28, the tropospheric Pu input from nuclear debris in 1962 was hard to give large effect to the Pu deposition in Japan in March 1963, i.e. 2–3 months after the final USSR tests at the end of 1962. Thus, most of the Pu deposited in Japan in 1963 was derived from the stratosphere. Nuclear debris is introduced into the stratosphere when the detonation yield exceeds 0.1 Mt29, though its partitioning between the stratosphere and troposphere depends on the latitude of detonation25. Because there were no tests with yields over 0.1 Mt in 1959–196027, and the mean stratospheric residence time of nuclear debris is 1–2 yr30,31,32,33, the observed increase of 239 + 240Pu deposition fluxes and atomic ratios in 1963 must have resulted from high yield nuclear testing conducted during 1961–1962.

Because heavier Pu isotopes are produced in nuclear detonations through successive neutron absorption of 239Pu as fission fuel and 238U as tamper and fission fuel, the 240Pu/239Pu atomic ratio produced in nuclear explosions depends on the neutron flux at detonation14,34. Because higher yield detonations have higher neutron fluxes, higher 240Pu/239Pu and 242Pu/239Pu atomic ratios are expected from higher-yield detonations.

Weapons-grade Pu has very low 240Pu/239Pu atomic ratios of around 0.0735. 240Pu/239Pu atomic ratios in soil and sediment samples around the Nevada (USA) and Semipalatinsk (former USSR) test sites were reported to be 0.03–0.058,36,37,38. These low values are the result of the low mean detonation yields of 12 kt (maximum 74 kt) and 63 kt (maximum 1.6 Mt) at the Nevada and Semipalatinsk sites, respectively, despite the large number of atmospheric nuclear tests performed at each site27. Such low atomic ratios were also found in soil and sediment samples around Nagasaki, Japan39,40,41 (21-kt yield combat detonation)27. In contrast, relatively high 240Pu/239Pu atomic ratios of 0.28–0.34 were found in soil samples from the Pacific Proving Grounds (Bikini, Enewetak, and Rongelap Atolls in the northern Marshall Islands)34,39,42, where 65 atmospheric nuclear tests with a mean detonation yield of 1.7 Mt (maximum 15 Mt) were conducted during 1946–195827. Because the mean detonation yield of 62 tests at the Novaya Zemlya test site (former USSR) during 1961–1962 was 3.5 Mt (maximum 50 Mt)27, it is reasonable that Pu from this area had 240Pu/239Pu atomic ratios similar to or larger than those observed in Marshall Islands soils. Therefore, the high 240Pu/239Pu atomic ratios observed in Japan during 1963–1964 very likely derived from the high-yield testing at Novaya Zemlya in 1961–1962.

According to UNSCEAR27, in 1962, fission-yield partitioned to the upper and lower stratosphere in polar regions (30–90°N including sampling points in this study) were 41.5 and 9.48 Mt, respectively. Thus, significant amounts of Pu with high 240Pu/239Pu atomic ratios were loaded into the upper stratosphere, whereas older debris with relatively low 240Pu/239Pu atomic ratios remained in the lower stratosphere.

In polar regions, the half-life of debris removal from the upper to lower stratosphere is 6–9 months, and that from the lower stratosphere to the troposphere is 3–12 months27. Thus, the observed increase in 240Pu/239Pu atomic ratios in early 1963 was due to debris descending through the polar stratosphere and into the troposphere. The gradual decrease of 240Pu/239Pu atomic ratios after the summer of 1963 resulted from the decreased transport of Pu from the polar stratospheres to the troposphere. We fit the observed 240Pu/239Pu and 242Pu/239Pu atomic ratios during September 1963 to May 1966 as follows:

$${}^{240,242}\,{\rm{P}}{\rm{u}}/{}^{239}\,{\rm{P}}{\rm{u}}={\rm{A}}\cdot {{\rm{e}}}^{{\textstyle \text{-}}{\rm{l}}{\rm{n}}(2)/{\rm{T}}\cdot t}+{\rm{C}}$$
(1)

where T (yr) and t (yr) are a half-life of the atomic ratio and a time from September 1963, respectively, and A and C are fitting constants, and we found the half-life (±a standard error) to be 5.6 ± 1.0 months (Fig. 2A) for 240Pu/239Pu atomic ratio. It is hard to predict the half-life of 240Pu/239Pu atomic ratio in the atmosphere by using an atmospheric model, because the atomic ratio of each detonation is unknown. Quantitative discussion about the half-life will be the research task in future.

The temporal variations of the deposited 242Pu/239Pu atomic ratio were similar to those of the 240Pu/239Pu atomic ratio, i.e. they increased rapidly from March to June 1963, fluctuated throughout the summer of 1963, and gradually decreased until the end of 1964. By fitting the data after September 1963 using Equation (1), we found the half-life of the exponential term to be 5.5 ± 1.1 months (Fig. 2B), very similar to that obtained for 240Pu/239Pu. It is therefore reasonable that the increased 240Pu/239Pu and 242Pu/239Pu atomic ratios from March to June 1963 resulted from high yield nuclear weapon testing during 1961–1962.

Published data on 242Pu/239Pu atomic ratios are scarce in comparison with 240Pu/239Pu because of the low concentrations of 242Pu. Only high-altitude aerosol data were available at sufficient temporal resolution during the 1960s21 to be comparable to our data (Fig. 2B). 242Pu/239Pu atomic ratios in aerosol samples were lower than those observed in our data during 1963, but they were similar to our data during mid-1964 to 1966. This was also the case for 240Pu/239Pu atomic ratios, and these inconsistencies may result from differences in sample type, sampling altitude and location, and duration for one sample.

We compared the correlation between the 242Pu/239Pu and 240Pu/239Pu atomic ratios of our samples to that of surface soil samples collected throughout the Northern Hemisphere in 1970–19718 (Fig. 4). The Pu isotopic ratios of our data were well correlated each other (Pearson’s correlation coefficient: r2 = 0.97, n = 42).

Figure 4
figure 4

240Pu/239Pu and 242Pu/239Pu atomic ratios of Japanese deposition samples during May 1963 to May 1966 compared with those of surface soils collected throughout the Northern Hemisphere8. Solid and dash lines are the regression lines of results from our deposition samples and these surface soils8, respectively.

Full size image

Although the Pu isotopic ratios of the surface soil samples were also well correlated (r2 = 0.90, n = 37), the slope of their regression curve was slightly, but significantly (ANCOVA, p < 0.001) gentler than that of our depositional data. Because the surface soil samples were collected in 1970–718, they were affected by Pu deposition on the ground before 1963 and after 1966. Thus, the gentler slope observed for soil samples may reflect the lower mean yield per detonation during the effective sampling period as compared to that during our sampling period.

Methods

Pu isotopic analyses

We followed previously established methods43 for the decomposition of atmospheric deposition samples and chemical separation of Pu. Briefly, a 1-g aliquot of sample was fused with a mixture of bicarbonate salts and boric acid using a bead sampler (TK-4100, Tokyo Kagaku Co. Ltd., Japan) at 1200 °C for 7 min. The fused sample was then dissolved with 3 M nitric acid, and Pu was separated from the matrix elements using Chelex 100 R resin (Bio-Rad Laboratories, USA). The Pu eluted from the resin was further separated from U using UTEVA resin (Eichrom Industries, USA), because even a trace of U could interfere with ICP-MS Pu measurements. The 240Pu/239Pu and 242Pu/239Pu atomic ratios were measured using a MC-ICP-MS (Nu Plasma HR, Nu Instruments Ltd., UK) equipped with twelve ceramic Faraday cups and three full-sized ion counters that can be utilized simultaneously. The ion counters were used for Pu isotopes, and the Faraday cups for other elements. The configuration of ion counters was optimized for the simultaneous detection for ions with m/z of 239, 240 and 242; therefore we did not measure 241Pu. Plutonium mass bias correction was performed by the external correction technique44 using the U isotopic ratios of a multi-element standard solution (XSTC-829, SPEX CertiPrep, USA). We purified a Pu solution from reference material IAEA Soil-6 (International Atomic Energy Agency, Austria) for use as a working standard during Pu isotopic measurements of the atmospheric deposition samples. The Pu isotopic ratios of the atmospheric deposition samples were measured by the sample-standard bracketing method using the working standard Pu solution. The 240Pu/239Pu and 242Pu/239Pu atomic ratios we measured for IAEA Soil-6 were 0.191 ± 0.001 and 0.0050 ± 0.0001, respectively. Uncertainties of the atmospheric deposition samples were propagated from the uncertainty of the Pu isotopic ratio of the working standard and the standard deviation of results from repeated measurements. 238U can interfere with accurate 240Pu/239Pu analyses through formation of polyatomic 238U1H+ ions (m/z 239). The contribution of 238U1H+ ions was corrected using the count rate of 238U+ ions in the sample and the 238U1H+/238U+ count ratio of the Pu-free U standard solution. Because the mass concentrations of U in the sample solutions were similar to those of Pu, and the 238U1H+/238U+ count ratio was ~1.4 × 10−5 in our system, the corrections for 240Pu/239Pu and 242Pu/239Pu were in the range of 10−6 and 10−8, respectively, i.e. negligibly small compared to the measured sample ratios.

Unfortunately, there is no standard sample with certified values for Pu isotopic ratio. To validate our analytical method, we analysed two international reference sediment samples (IAEA 368, International Atomic Energy Agency, Austria, and NIST 4354, National Institute of Standards and Technology, USA) and a Japanese reference fallout material produced by the MRI from samples collected at 14 locations throughout Japan during 1963–197945. Our Pu isotopic measurements of these standard samples are presented and compared to results from other researchers in Table 2. The obtained 240Pu/239Pu atomic ratios agreed well with previous data9,46,47,48,49,50, whereas the 242Pu/239Pu atomic ratios obtained for IAEA 368 varied considerably because of that sample’s low 242Pu concentration. It is notable that the relative uncertainties of isotopic ratios for NIST 4354 were larger than those for the present samples in Table 1 because of lower concentration of Pu in measuring solution for the former sample.

Table 2 240Pu/239Pu and 242Pu/239Pu atomic ratios measured herein and from literature for IAEA 386, NIST 4354, and Japanese fallout reference material.
Full size table