Journal of Radioanalytical and Nuclear Chemistry

, Volume 307, Issue 3, pp 2117–2122

High-sensitivity determination of radioactive cesium in Japanese foodstuffs: 3 years after the Fukushima accident

  • Katsumi Shozugawa
  • Takuya Saito
  • Mayumi Hori
  • Motoyuki Matsuo

DOI: 10.1007/s10967-015-4407-8

Cite this article as:
Shozugawa, K., Saito, T., Hori, M. et al. J Radioanal Nucl Chem (2016) 307: 2117. doi:10.1007/s10967-015-4407-8


We analyzed 134Cs, 137Cs and 40K in 96 foodstuffs in supermarkets with high sensitivity over 3 years after Fukushima accident. Milk, yoghurt, rice, tea, salmon, cereal, blueberry, miso, and apples had a trace of 134Cs and 137Cs from 10−3 to 100 Bq/kg, however, some mushrooms that were bought in the outer Fukushima prefecture were contaminated by radioactive cesium over the regulatory limit (100 Bq/kg). In view of the 134Cs/137Cs radioactivity ratio, we can conclude that 137Cs detected in remote areas 300 km or more from Fukushima Nuclear power plant contained activity from Pre-Fukushima events such as Chernobyl accident (1986) and atmospheric nuclear explosions (from 1945).


Fukushima accident Radiocesium (137Cs, 134Cs) Foodstuffs in Japan Concentration treatment 


The huge earthquake and the subsequent tsunami on 11 Mar. 2011 caused massive emission of radionuclides from Fukushima Dai-ichi Nuclear Power Plants (FDNPP), TEPCO, Japan. By the catastrophic accident, radionuclides released into the atmosphere were estimated at 520 PBq except Nobel gases such as 133Xe [1]. Within 1 month after the accident, major nuclides related to human exposure mainly by inhalation were short-lived nuclides such as 132Te, 132I, and 131I. 132Te (T1/2 = 7.8 × 102 h) including its daughter 132I (T1/2 = 2.3 h) were estimated at 88–150 PBq [2] into the atmosphere, and 131I (T1/2 = 8.0 days) was also estimated at 150 PBq [3].

After short-lived nuclides decayed, long lived nuclides such as 137Cs (T1/2 = 30.2 years) and 90Sr (T1/2 = 28.8 years) will be in existence for several decades. Regarding 137Cs originates from the Fukushima accident, totally 700 PBq was extant in pressure containers of BWR reactors from No.1 to No.3 as of the accident, and 150 PBq was in contaminated water in each containment vessels [4]. Directive out-flow of 137Cs from containment vessels to the North Pacific Ocean was estimated at 3.6 ± 0.7 PBq [5].

On the other hand, 137Cs released into the atmosphere was estimated at 12–62.5 PBq [6, 7], and mostly fell into the ocean by Prevailing westerlies. Deposited 137Cs on Japanese land was estimated at 1.0–5.0 PBq [8], and measured 137Cs on Japanese land by airborne monitoring was 2.7 PBq [8].

Radiocesium deposited on Japanese land strongly contaminated Fukushima and neighbor prefectures, where there are well-known farmland (rice, vegetables, and fruits) in Japan [9]. Thus, the contamination by long-lived nuclides such as 137Cs and 90Sr will cause the largest concern to both Japanese consumers and producers for several decades.

In April 2012, to achieve further food safety and consumer confidence, the Japanese government revised the regulatory limits of radioactive materials in general foodstuffs to reduce maximum permissible internal exposure dose from 5 mSv/year (500 Bq/kg of 134Cs + 137Cs in general foodstuff) to 1 mSv/year (100 Bq/kg of 134Cs + 137Cs). As the result of a large scale monitoring campaign by the Japanese/Fukushima Government, however, the numbers of general foodstuffs exceeding regulations in Fukushima areas have not been found in supermarkets 3 years after the accident. In the between case of paddy rice harvested in Fukushima, for example, 10,964,774 rice bags (ca. 30 kg/bag) were screened by BGO, CsI(Tl), NaI(Tl), and plastic scintillation gamma detectors with the screening level (50–80 Bq/kg, for radiocesium) from 2014 to 2015 season, and 99.98 % of rice bags were below the screening level. Furthermore, it found that any rice bags were not contaminated over regulatory limit (100 Bq/kg) by HPGe semiconductor detector even if radiocesium was detected over the screening level in Fukushima prefecture [10].

On the other hand, outside the Fukushima prefecture, this study shows that foodstuffs contaminated by radiocesium over the regulatory limits were still in the supermarkets, although the numbers were few. In addition, depending on the area, detected 137Cs might originate from global fallout, not the Fukushima accident. To distinguish the difference between these, detection of 134Cs with high sensitivity must be necessary.

In order to clarify dispersion of radiocesium in Japanese foodstuff, it is necessary to do a concentration treatment before gamma measurements because a large part of radiocesium in Japanese foodstuffs will be at least one order of magnitude lower than 40K.

Here, we show the detail level of 134Cs and 137Cs in general foodstuff harvested outside of Fukushima with high sensitivity analysis by concentration treatment and the HPGe gamma detector, and clarified distribution of radiocesium 3 years after the Fukushima accident.

Materials and methods

Supplied foodstuffs

All foodstuffs for analysis were bought at supermarkets or local shops outside Fukushima prefecture in Japan, that is, mainly from Gunma, Tochigi, Ibaraki prefectures (see Fig. 1). 96 food samples were collected, which were both fresh (meat, green/other vegetables, fish, fruits, milk, bottled/mineral water and so on) and processed (yoghurt, seasonings, soup stock, cereals and so on), and categorized into 18 groups according to some reports [11, 12]. This method is the one of common type of data mining for toxic or harmful matters, such as heavy metals, dioxins, PCBs, and radioactive materials in foodstuffs, and called “market basket analysis” [13].
Fig. 1

Location of Fukushima Dai-ichi Nuclear Power Plant (FDNPP) in Fukushima prefecture and surrounding prefectures in Japan

Concentration of radiocesium

Based on Japanese official methods [14], samples with high water content (>20 % w/w) were concentrated by heat or vacuum freeze drying. In case of liquid samples, such as water, they were heated around 80° for a few days. Liquid samples were concentrated to be at least 5 times smaller than before heating. In case of solid samples, such as rice and meat, they were freeze-dried and heated at 105° for at least 3 days. Furthermore, in order to reduce water contents as possible, the samples were heated at 350° to be carbonized for 1 day. It took 5–7 days on average to concentrateeach sample. Concentration ratios were depending on each water content or density, dispersing from 1.5 to 20.

Gamma measurements and quality control

A high purity germanium detector (HPGe) with 25 % relative efficiency and resolution of 1.80 keV at 1.33 MeV was used to collect gamma rays emitted from concentrated samples. Digital 16 k MCA with integrated HV power supply was used for this system.

Radioactivity of 137Cs and 40K were determined by 661 keV (abundant 85 %) and 1461 keV (10.7 %) gamma ray, respectively. 134Cs, known as having several gamma cascades, was determined by using a weighted average of at least two gamma rays after careful consideration of sum effect and its geometry. Both the 605 keV (97.6 %) and the 796 keV (85.4 %) gamma ray were necessary for determination of 134Cs. The geometry of samples was fixed by plastic vessels, Malinelli vessels (2 L) and U-8 containers (100 mL). Measuring time (live time) of gamma ray was from 3 h to 9.7 days depending on each Compton scattering from that of 40K. Detection limits were set to three sigma of standard deviation (σ) in accordance with Covell’s method. Activity of radiocesium was corrected at the time of purchase of each sample.

Regarding quality control, obtained radioactivity and its standard deviation of Japanese standard material, JSAC 0471 by our measurements, were 86.2 ± 1.8 Bq/kg(134Cs) and 112 ± 1.8 Bq/kg(137Cs), which were in good agreement with certified radioactivities (134Cs: 85.3 ± 5.9 Bq/kg, 137Cs: 115 ± 8 Bq/kg). Uncertainties of 134Cs, 137Cs, and 40K are depending on each activity, but they are <10 % of each activity.


In order to estimate the lost radiocesium due to the concentration process of each sample, recovery was checked by standard addition method. Regarding the liquid samples, 10 mL, 1000 ppm of 133Cs (stable) was added to raw samples before heating. After concentration by heat treatment (80°, from 2 to 3 days), concentrated samples were fractioned and diluted by ultra-pure water (milli-Q®, Merck), and 133Cs in concentrated sample was measured by ICP-MS (ICP-MS 7500, Agilent Technologies).

Results and discussion

Radiocesium in foodstuffs

For the discussion of background excluding radiocesium, Fig. 2 shows radioactivity of 40K in general foodstuffs in the Japansese supermaket. The magnitude of radioactivity ranged 10−1 to 103 Bq/kg. Mineral water had only 0.1 Bq/kg of 40K, while, salt for those with hypertension (K enriched) had the highest radioactivity, which was 8400 Bq/kg. The median radioactivity of all 95 samples was 70.8 Bq/kg.
Fig. 2

Radioactivity of 40K in various foodstuffs (n = 95). 40K was detected in 95 among 96 samples. Uncertainties are less than 10 % of each activity

Figure 3 shows the radioactivity of 134Cs + 137Cs in foodstuffs. 137Cs was not detected in 46 samples with three sigma threshold. Compared with 40K, the median of radiocesium activity was 0.16 Bq/kg, two orders of magnitude lower. Also there was no correlation between 40K and radiocesium (r = 0.089).
Fig. 3

Radioactivity of sum 134Cs and 137Cs in foodstuffs (n = 45). Radiocesium was not detected in other samples (n = 51). Uncertainties are less than 10 % of each activity

We can say the same thing about internal exposure of these nuclides. The dose coefficient (Sv/Bq) [15] both of 137Cs (1.3E−8) and 134Cs (1.9E−8) was about two times higher than that of 40K (6.2E−9), however, the ratio of radiocesium against all radiomaterials was about only 0.48 %. Here, we assume that daily intake is 1 kg food per person, and the effect of 134Cs is negligible due to very small activity (ca. 0.05 Bq/kg). Internal exposure by radiocesium was estimated as 0.76 µSv/year, which was much lower than Japanese regulatory limit equivalent to 1 mSv/year. The results of market basket by the Japanese government was in good agreement with our results, which showed that internal exposure by radiocesium was from 0.7 to 1.9 µSv/year during our research period including Fukushima prefecture [16].

However, Fig. 3 showed that two kinds of mushrooms exceeded the regulatory limit of general foodstuff (100 Bq/kg). These samples sold at local foodshop, such as a rest area (Michi-no-Eki, in Japanese) in Gunma, Tochigi, Ibaraki prefectures, but not Fukushima. Besides mushrooms, simmered fish (Japanese pond smelt, categorized as “fish”) and lotus root (categorised as “other vegetables”) sold at a rest area in Ibaraki had 31.9 and 14.0 Bq/kg of radiocesium with respect. This radioactivity of foods was much higher than that of foods in the general supermarkets. We presumed that the reason why those foods were still in the shops was that the system for measurement of radioactivity was not completely established in those areas at this time.

Lost cesium by concentration

In case of orange juice, added 133Cs was measured before and after concentration treatment. The concentration ratio was 5.8, namely, raw 14.9 kg of orange juice was reduced volume and weight to 2.55 kg. Recovery of cold cesium was 99.6 %, showing that the loss of both cesium and radiocesium by concentration were negligible amounts. Therefore, we assumed that loss by pre-treatment would not affect the radioactivity.

Comparison before and after Fukushima accident

Table 1 shows the median radioactivity (Bq/kg-flesh weight) in foodstuffs categorized in 18 groups. Results of monitoring until 2001 did not report 134Cs in all foodstuffs [11, 17]. In our study (2013–2014), 134Cs was detected in rice, cereals, fruits, other vegetables, mushrooms, seasonings, and fishes, however trace (maximum of median 3.5 Bq/kg, mushrooms), indicating that all of the 134Cs was originated in the Fukushima accident.
Table 1

A comparison of the radioactivity of radiocesium before and after the Fukushima accident

Food categorya

134Cs Bq/kg-flesh weight

137Cs Bq/kg-flesh weight

2013–2014 (this study)

1987–2001 [17]

1994–1995 [11]

2013–2014 (this study)

1987–2001 [17]

1994–1995 [11]













Rice (5)








ND (0.022)


Cereals (2)








ND (0.049)


Nuts and Seeds (4)

ND (0.28)









Potatoes (3)

ND (0.54)




ND (0.63)





Bean & Bean products (3)

ND (0.041)







ND (0.11)


Fruits (1)








ND (0.023)


Green vegetables (2)

ND (0.24)







ND (0.025)


Other vegetables (12)










ND (0.051)


Mushrooms (7)










Seaweeds (2)

ND (0.19)




ND (0.20)




Seasonings (11)








ND (0.043)


Fishes (5)












Meats (3)

ND (0.037)










Eggs (1)

ND (0.010)





ND (0.014)




ND (0.005)


Milk and Milk products (7)










Beverages (11)

ND (0.010)







ND (0.14)


Processed (13)

ND (0.59)





ND (0.56)





Water (4)

ND (0.068)




ND (0.079)




aNumbers of analized samples are shown in parentheses after each category

bND not detected, and detection limits (3 sigma) are shown in parentheses

137Cs with long half-life was detected in 13 groups. Compared with past surveys [11, 17], radioactivity was no different in milk, eggs, and fishes categories before and after the Fukushima accident. On the other hand, there was an increase in radioactivity of meats and mushrooms after the Fukushima accident. Mushrooms, well known as accumulator of cesium [18], were contaminated 10 times higher than before the accident (from 1.7 to 20 Bq/kg). These facts showed that the monitoring system of radiocesium in foodstuffs would be necessary for several decades as long as current regulation is kept.

Contribution of Fukushima accident

134Cs is an activation product in 235U fuel rods, namely 133Cs(n, g)134Cs, whereas 137Cs is one of most typical fission products [19]. Half-life of 134Cs (2.06 years) is much shorter than that of 137Cs (30.2 years) due to its odd–odd nucleus. The latest severe accidents in which radiocesium including 134Cs were emitted into the atmosphere were the Chernobyl accident (1986) and atmospheric nuclear explosions (from 1945) before Fukushima accident (2011). Both 134Cs and 137Cs were omnipresent nuclides at the accident or an event, however, all of the detected 134Cs clearly originated from the Fukushima accident at the time of 2014. In addition, the radioactivity ratio of 134Cs/137Cs in food was 0.98 at the time of the Fukushima accident [20]. Based on this feature, we can distinguish the effect of the Fukushima accident and Pre-Fukushima event quantitatively as long as 134Cs was detected.

Figure 4 shows 134Cs/137Cs ratio corrected as of the accident versus distance from FDNPP. Since the ratio of 134Cs/137Cs indicated about 1 within 300 km radius from FDNPP, the source of radiocesium was clearly Fukushima accident. However, the ratio ranged from 0.4 to 0.6 outside 300 km radius from FDNPP. The sample that was collected at 490 km from FDNPP was “salmon” in North Pacific Ocean, and its 134Cs/137Cs ratio indicated 0.49 ± 0.13. The ratio shows that half of detected 137Cs originated in Fukushima accident, the other half originated from pre-Fukushima events. Since the radioactivity of 134Cs and 137Cs detected in salmon were trace amounts (0.016 ± 0.004 Bq/kg, 0.10 ± 0.005 Bq/kg, respectively) and equivalent to 1/1000 of Japanese regulatory limit, a high level of food safety is maintained regarding this salmon. The 134Cs/137Cs ratio was a useful index to know not only the contribution from the Fukushima accident but also the seasonal migration of fishes as long as 134Cs was detected.
Fig. 4

134Cs/137Cs ratio corrected as the accident (11 Mar. 2011) versus radius distance from FDNPP. The ratio of 1.0 showed that radiocesium in food originated from Fukushima accident


Results of high sensitity analysis of radiocesium in Japanese foodstuffs show that contaminated radiocesium in foodstuffs was small (0.16 Bq/kg, median) 3 years after the Fukushima accident. Almost all foodstuffs showed no difference of radioactivity before and after the accident with the exception of meats and mushrooms. However, this study clarified that some contaminated food exceeded the regulatory limit, and were still found in the areas where monitoring system have not been established. To archive further consumers’ confidence, continuous, discreet and fine monitoring of long half-life nuclides will be necessary for Japanse food safety.


Dr. Akira Ishikawa is acknowledged for providing access to the ICP-MS. This work was supported by JSPS KAKENHI Grant Number 25870158.

Copyright information

© Akadémiai Kiadó, Budapest, Hungary 2015

Authors and Affiliations

  • Katsumi Shozugawa
    • 1
  • Takuya Saito
    • 2
  • Mayumi Hori
    • 1
  • Motoyuki Matsuo
    • 1
  1. 1.Graduate School of Arts and SciencesThe University of TokyoMeguro-kuJapan
  2. 2.Akita Radiation Measuring Station (Beguredenega)KatagamiJapan

Personalised recommendations