ww.sciencedirect.com

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 4 3 9 7e4 4 0 4

Available online at w

journal homepage: www.elsevier .com/locate/watres

Removal of radioactive iodine and cesium in waterpurification processes after an explosion at a nuclear powerplant due to the Great East Japan Earthquake

Koji Kosaka a,*, Mari Asami a, Naoya Kobashigawa a, Keiko Ohkubo a, Hiroshi Terada a,Naohiro Kishida a, Michihiro Akiba b

aDepartment of Environmental Health, National Institute of Public Health, 2-3-6 Minami, Wako, Saitama 351-0197, JapanbNational Institute of Public Health, 2-3-6 Minami, Wako, Saitama 351-0197, Japan

a r t i c l e i n f o

Article history:

Received 7 February 2012

Received in revised form

6 May 2012

Accepted 27 May 2012

Available online 7 June 2012

Keywords:

Radioactive iodine

Radioactive cesium

Powdered activated carbon

Chlorination

Coagulation

Water purification process

* Corresponding author. Tel.: þ81 48 458 630E-mail address: kosaka@niph.go.jp (K. Ko

0043-1354/$ e see front matter ª 2012 Elsev10.1016/j.watres.2012.05.055

a b s t r a c t

The presence of radionuclides at five water purification plants was investigated after an

explosion at a nuclear power plant hit by the Great East Japan Earthquake on 11 March

2011. Radioactive iodine (131I) and cesium (134Cs and 137Cs) were detected in raw water in

Fukushima and neighboring prefectures. 131I was not removed by coagulationeflocculation

esedimentation. 131I was removed by granular activated carbon (GAC) and powdered

activated carbon (PAC) at a level of about 30%e40%, although 131I was not removed in some

cases. This was also confirmed by laboratory-scale experiments using PAC. The removal

percentages of 131I in river and pond waters by 25 mg dry/L of PAC increased from 36% to

59% and from 41% to 48%, respectively, with chlorine dosing before PAC. 134Cs and 137Cs

were effectively removed by coagulation at both a water purification plant and in

laboratory-scale experiments when turbidity was relatively high. In contrast, 134Cs and137Cs in pond water with low turbidity were not removed by coagulation. This was because134Cs and 137Cs in river water were present mainly in particulate form, while in pond water

they were present mainly as cesium ions (134Csþ and 137Csþ). However, the removal of 134Cs

and 137Cs in pond water by coagulation increased markedly when 134Cs and 137Cs were

mixed with sediment 24 h before coagulation.

ª 2012 Elsevier Ltd. All rights reserved.

1. Introduction Several groups have investigated the occurrence of 131I,

The Great East Japan Earthquake occurred on 11 March 2011,

and caused an explosion at the Tokyo Electric Power Company

(TEPCO) Fukushima Daiichi Nuclear Power Plant, which

resulted in the release of large amounts of radionuclides into

the environment (Water Supply Division (WSD), Health

Service Bureau (HSB), Ministry of Health, Labour, and Welfare

(MHLW), 2011). Radioactive iodine (131I) and cesium (134Cs and137Cs) were detected in drinking water after the explosion

(WSD, HSB, MHLW, 2011; Ikemoto and Magara, 2011).

6; fax: þ81 48 458 6305.saka).ier Ltd. All rights reserved

134Cs, and 137Cs at water purification plants (Goossens et al.,

1989; Esumi et al., 1986; Gafvert et al., 2002; Kamata et al.,

1973) although data after nuclear accidents were limited. It

was considered that radionuclides existed in particulate and

dissolved forms in environmental water considering their

forms in the air released after nuclear power plant accidents

(Noguchi and Murata, 1988). They may also exist in colloidal

form. Dissolved 131I is separated into various forms, such as

radioactive iodide ion (131I�), iodine molecule (131I2), hypo-

iodous acid (HO131I), iodate ion (131IO�3 ), and organic 131I. The

.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 4 3 9 7e4 4 0 44398

proportions of the forms of 131I were considered to be different

among environmental waters and to change due to reactions

in the environmental water. For example, 131I� slowly trans-

forms into 131IO�3 in environmental water under the action of

sunlight (Lettinga, 1972). 131I2 and HO131I react with natural

organic matter (NOM) in the environmental water. Thus, it

was assumed that 131I removal at water purification plants

would depend on the environmental water. Information

regarding the occurrence of radionuclides at water purifica-

tion plants is valuable. However, their levels in the environ-

mental water are usually very low, and thus such data are

limited.

There have also been experimental investigations of 131I,134Cs, and 137Cs removal by water purification processes, such

as coagulation and powdered activated carbon (PAC) (Morton

and Straub, 1955; Lettinga, 1972; Brown et al., 2008a, b).

Brown et al. (2008a and 2008b) evaluated the treatability of

radionuclides by each unit process applied to water purifica-

tion plants based on data obtained from the literature. These

data are useful, but in many cases the experiments were

performed using only certain forms of radionuclide; e.g., 131I�

and radioactive cesium ion (134Csþ and 137Csþ). As described

above, radionuclides seem to exist in particulate and dis-

solved forms in environmental water. Thus, the removal of131I, 134Cs, and 137Cs from environmental water may occa-

sionally be different from those of 131I�, 134Csþ, and 137Csþ,respectively. It was reported that 131I� removal by PAC was

low (Lettinga, 1972); however, some other forms of 131I (e.g.,

organic 131I) were removed by PAC (Summers et al., 1988). In

addition, it was reported that ionic forms of 137Cs-radioactive

barium (137Ba) were not removed by coagulation (Morton and

Straub, 1955). However, Csþ is known to be sorbed by soils

and minerals (Cornell, 1993; Staunton, 1994). Thus, it was

considered that particles of 134Cs and 137Cs could be removed

by coagulation. In fact, it was reported that removal of ionic

forms of 137Cse137Ba by coagulation was increased in the

presence of clay (Morton and Straub, 1955). As in the case

regarding the occurrence of radionuclides at water purifica-

tion plants, only limited experimental data obtained using

environmental water containing radionuclides are available.

In the present study, we investigated the removal of 131I,134Cs, and 137Cs at water purification plants after the explosion

at a nuclear power plant caused by the Great East Japan

Earthquake. We also investigated 131I removal by PAC and

removal of 134Cs and 137Cs by coagulation taking their forms in

environmental water into consideration.

2. Material and methods

2.1. Reagents and solutions

The reagents and solutions used are described in Supple-

mentary Materials. Three types of PAC (PAC-1 to PAC-3) were

used in the present study.

2.2. Sampling

Raw, process, and finished water samples at five water puri-

fication plants (WPP-1 to WPP-5) were collected in April 2011.

The details of these water purification plants and sampling

procedures are described in Supplementary Materials. In

cases where levels of 131I, 134Cs, and 137Cs levels in raw water

were higher than 1 Bq/L, the data were used for analysis. Pond

water was also collected from Iitate Village in Fukushima

Prefecture on 7 May 2011. Water quality data of raw waters of

WPP-1 to WPP-5 and pond water are listed in Table S1. Gran-

ular activated carbon (GAC) and sand from the rapid sand filter

at WPP-1 were collected from April to September 2011. The

GAC was that at depths of up to 5 cm from the surface.

Moreover, sediments were collected from the Yodo and Ara

Rivers, and were designated as sediments A and B,

respectively.

2.3. Removal test of radioactive iodine and cesium

Removal of 131I in river and pond waters by PACs was inves-

tigated. To investigate the form of 131I, its removal by silver ion

(Agþ) was also performed. Removal of 134Cs and 137Cs in river

and pond waters by coagulation with polyaluminum chloride

(PACl) was investigated. Removal of 134Cs and 137Cs in pond

waters by PAC (PAC-1) and cation exchange resin was also

investigated. The procedures of the removal tests of 131I, 134Cs,

and 137Cs are described in Supplementary Materials.

2.4. Analytical methods

Gamma-emitting radionuclides (131I, 134Cs, and 137Cs) in the

samples were determined using high-purity germanium

semiconductor detectors (GX2518; Canberra Co., Meriden, CT

and EGPC20-190-R; EURYSIS Co., Cedex, France). In the case of

occurrence of radionuclides at water purification plants, their

levels were expressed as those on sampling days and time

using their half-lives. In the radionuclide removal experi-

ments, their levels on each experimental daywere used. Other

analytical methods are described in Supplementary Materials.

3. Results and discussion

3.1. Levels of radioactive iodine and cesium in drinkingwater after explosion at a nuclear power plant

After the explosion at a nuclear power plant, the levels of 131I,134Cs, and 137Cs in drinking water have been investigated at

Fukushima Prefecture and its 10 neighboring prefectures

(Fig. S1) (Ikemoto and Magara, 2011). Figure S3 shows the

results of monitoring of 131I in drinking water at Fukushima

Prefecture and its 10 neighboring prefectures in March to May

2011 (WSD, HSB, MHLW, 2011; Asami and Akiba, 2011). During

the monitoring, results indicating > 100 Bq/kg of 131I in

drinking water were only reported in March 2011. 131I in

drinking water exceeded 300 Bq/kg and drinking water for the

general public was restricted at 1 water utility in Iitate Village,

Fukushima Prefecture (maximum, 965 Bq/kg). 131I in drinking

water exceeded 100 Bq/kg and drinking water for infants was

restricted at 20 water utilities in 5 prefectures (Fukushima,

Ibaragi, Tochigi, and Chiba Prefectures, and Tokyo Metropol-

itan Government) (WSD, HSB, MHLW, 2011). Figure S4 shows

the profiles of 131I in tap water in Iitate Village, Fukushima

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 4 3 9 7e4 4 0 4 4399

Prefecture, and in finished water at a water purification plant

in Tokyo Metropolitan Government (WSD, HSB, MHLW, 2011).

The initially high levels of 131I in drinkingwaterwere shown to

have dropped considerably.

Figure S5 shows the results of monitoring of the sum of134Cs and 137Cs in drinking water in Fukushima Prefecture and

its 10 neighboring prefectures from March 2011 to January

2012 (WSD, HSB, MHLW, 2012). The maximum of the sum of134Cs and 137Cs was 180.5 Bq/kg in tap water in Tamura City,

Fukushima Prefecture, which was lower than 200 Bq/kg, an

index of restriction of drinking water for the general public.

Figure S6 shows the profiles of the sum of 134Cs and 137Cs in

tap water in Iitate Village, Fukushima Prefecture (WSD, HSB,

MHLW, 2011). The trend of the sum of 134Cs and 137Cs was

different from that of 131I.

Compared to the data of radionuclides in drinking water,

those in raw water were limited and therefore their intro-

duction is not described here. One reason for this was that the

number of inspection institutes with monitoring instruments

was small just after the nuclear power plant accident and it

was difficult to monitor the levels of radionuclides in samples

other than drinking water. Similarly, in the present study,

samples with high levels of radionuclides could not be

collected because of the confusion immediately following the

accident.

3.2. Removal of radioactive iodine at water purificationplants

Fig. 1 and Fig. S7 show the profiles of 131I at WPP-1 and WPP-2

to WPP-5, respectively. As shown in Fig. S2 and Table S1, their

treatment flows and sampling days were different. 131I

removal by unit purification processes was evaluated based

on the results of WPP-1 to WPP-5.

3.2.1. CoagulationFig. 1 shows the profiles of 131I atWPP-1. 131I level in rawwater

was 2.9 Bq/L and that in rapid sand filtration was 2.8 Bq/L. At

WPP-2, 131I levels in raw water and sand filtration on 20 April

were 5.4 and 5.7 Bq/L, respectively (Fig. S7(a)). Thus, 131I was

not removed by coagulationeflocculationesedimentation. A

similar tendency was also seen at WPP-2 on 28 April. In

0.0

2.0

4.0

6.0

Raw water Sedimentation Rapid sand f iltration

GAC

131 I

(Bq/

L)

Fig. 1 e Profiles of 131I at WPP-1.

addition, 131I in raw water at WPP-1 and WPP-2 was consid-

ered to be present not in particulate form but in dissolved

form.

3.2.2. Activated carbon treatmentGAC treatment is applied at WPP-1. 131I levels before and after

GAC treatment were 2.8 and 1.9 Bq/L, respectively, and

therefore 34% of 131I was removed (Fig. 1). 131I was detected in

GAC at 1500 Bq/kg on 22 April and at 94 Bq/kgwet weight on 30

May (Table S2). These results supported the removal of 131I by

GAC. Biological activated carbon (BAC) treatment followed by

rapid sand filtration is applied at WPP-2 (Fig. S2). The removal

percentages of 131I on 20 and 28 April were 4% and >39%,

respectively, and were different between sampling days. The

results on 20 April may be more reliable because 131I levels on

28 April were lower. However, on 28 April, 131I was not

detected either after BAC treatment or rapid sand filtration

although 131I levels were similar in other samples. Thus, the

treatability of 131I by BAC was not clarified in the present

study. If 131I was removed by BAC to some degree, some 131I

was considered to exist in forms that could be better adsorbed

by AC or biologically removed, such as organic 131I.

PAC treatment was applied at WPP-3 to WPP-5 during the

study (Fig. S2). 131I levels in rawwater at WPP-3 toWPP-5 were

1.0e2.9, 1.5, and 1.5 Bq/L, respectively (Fig. S7). PAC concen-

trations at WPP-3 to WPP-5 were 5, 20, and 30 mg dry/L,

respectively. From the results between raw water and rapid

sand filtration, the removal percentages of 131I at WPP-3 to

WPP-5 were 0% e >27%, 13%, and 39%, respectively. It was

notable that the removal percentage of 131I atWPP-3 on 5 April

was regarded as 0%. From the results at WPP-1 to WPP-5, it

was considered that 131I was removed by GAC and PAC

ranging from about 30% to 40%, although 131I was not removed

in some cases.

It was reported that 131I� removal by PAC was low and

increased by a combination of chlorination and PAC (Lettinga,

1972). On the other hand, in the case of 131I-containing water,

it was reported that the removal percentages of 131I in raw

water and process water after sand filtration by PAC (30 mg/L)

were 100% and 39%, respectively, and that in rain water by

PAC (50 mg/L) was 37% (Honma et al., 1988). Prechlorination

was applied at the water purification plant. It was notable that

in this report reference data were used as the results for raw

water because 131I level in raw water was below the limit of

quantification. In the present study, prechlorination was

applied at WPP-1 andWPP-4 on the sampling days. Therefore,

similar to the previous study by Honma et al. (1988), it was

considered that 131I was partially removed by AC (i.e., PAC or

BAC) without prechlorination at WPP-2, WPP-3, and WPP-5.

This was presumed to be because not only 131I� but also

other forms of 131I were present in raw water. No clear effects

of prechlorination were observed in the present study.

3.2.3. Slow sand filtrationSlow sand filtration is applied at WPP-2 (Fig. S2). The removal

percentages of 131I by slow sand filtration on 20 and 28 April

were 67% and 3%, respectively. The removal efficiencies of 131I

were significantly different between sampling days. On 20

April, 131I level dropped after slow sand filtration. Therefore,131I may be removed by slow sand filtration; however, the

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 4 3 9 7e4 4 0 44400

results were not conclusive in the present study. If 131I was

removed by slow sand filtration, it was considered that some131I existed in forms that could be biologically removed, such

as organic 131I.

3.3. Removal test of radioactive iodine

Fig. 2 shows removal percentages of 131I in river and pond

waters by PAC (PAC-1 was used). 131I levels in river and pond

waters before treatment were 3.4e3.5 and 7.4e9.5 Bq/L,

respectively. Reaction time was 30 min. Removal percentages

of 131I in river water at 2.5, 5, 10, 25, and 50 mg dry/L of PAC

were 11%, 11%, 9%, 36%, and 71%, respectively. Those in pond

water at 10 and 25 mg dry/L of PAC were 13% and 41%,

respectively. It was reported that the removal percentages of131I� by 1000 mg/L of PAC were 4.1%e6.1% at pH 8.5 (contact

time, 30 min) (Lettinga, 1972). Thus, the removal efficiency of131I by PAC in the present studywas higher than that of 131I� in

previous studies.

The environmental water associated with the explosion at

a nuclear power plant was considered to contain various

forms of 131I, including 131I� and various types of organic 131I. It

was reported that the removal percentage of radioactive

iodomethane (CH3131I), one of the organic 131I forms detected in

air after the accident at the Chernobyl nuclear power plant

(Noguchi and Murata, 1988), in humic substance solution

(5.98 mg C/L, pH 6.5) by 100 mg/L of AC with a reaction time of

2 dayswas 92% (Summers et al., 1988). Thus, the differences in

results between the present study and that of Lettinga (1972)

were considered to be at least partly due to differences in

the form of 131I. This agreed with the results presented in the

previous section.

To estimate the proportions of 131I� to total 131I, the

removal of 131I in pond waters by Agþ was investigated after

acidification of the samples (Fig. S8). The removal percentages

of 131I by 0.5 and 1.0 g/L were 34% and 28%, respectively. Agþ is

known to react with halide ions, such as chloride, bromide,

and iodide ions, with resultant deposition of silver halides.

0

20

40

60

80

100

0 10 20 30 40 50

Rem

oval

per

cent

age

of 13

1 I%

PAC (mg-dry/L

River (PAC)River (Cl2 + PAC)Pond (PAC)Pond (Cl2 + PAC)

Fig. 2 e 131I removal in river and pond waters by PAC (131I

in river water, 3.4e3.5 Bq/L; 131I in pond water, 7.4e9.5 Bq/

L; PAC, PAC-1; reaction time of PAC, 30 min).

Thus, it was considered that the proportion of 131I� was

around 30% under the assumption that among the 131I species

only 131I� was removed by Agþ. This result supported the

presence of various forms of 131I in environmental water.

The effects of chlorination on the removal percentages of131I by PAC were investigated (Fig. 2). The order of treatment

was chlorination for 10 min followed by PAC for 30 min. Free

chlorine concentration in river and pond waters after 10 min

of chlorination was around 0.10e0.15 mg/L. Removal

percentages of 131I in river water at 10, 25, and 50 mg dry/L of

PAC were 41%, 59%, and 71%, respectively. Those in pond

water at 5, 10, 25, and 50 mg dry/L of PAC were 20%, 33%, 48%,

and 62%, respectively. Fig. S9 also shows the effects of chlo-

rination on the removal percentages of 131I in pond water by

PACs (PAC-1, PAC-2, and PAC-3). PAC concentrationwas 25mg

dry/L and other reaction conditions were the same as those in

Fig. 2. Removal percentages of 131I by PAC-1, PAC-2, and PAC-3

without chlorination were 43%, 11%, and 21%, respectively.

Those with chlorinationwere 48%, 27%, and 36%, respectively.

These results indicated that 131I removal by PAC increased

with chlorination, while the effects of chlorination were

higher at lower PAC concentration or lower removal by PAC

only.

Lettinga (1972) reported that 131I� removal by PAC

increased with chlorine dosing before PAC treatment. The

tendency in this previous study agreed with our results

presented here. In addition, Summers et al. (1988) reported

that the removal percentage of 131I in river water by 100 mg/L

AC at a reaction time of 2 days was 57% when 131I� was used

as 131I and 1 mg/L of chlorine was initially added. Addition of

chlorine simulated prechlorination in the actual purification

process. I� reacts rapidly with chlorine and transforms

mainly into HOI at neutral pH (Lettinga, 1972; Bichsel and von

Gunten, 1999). The sorption capacities of HO131I and 131I2with AC were reported to be higher than that of 131I� by 15-

and 60-fold, respectively (Mandi�c et al., 1996). In addition, it

was also reported that I� or organic I was formed when HOI

reacted with NOM in environmental water (Bichsel and von

Gunten, 1999). It was presumed that removal of organic 131I

by PAC was higher than 131I� as in the case of CH3131I

(Summers et al., 1988). Therefore, it was considered that

a combination of chlorination and PAC was effective for 131I�

removal among 131I species because of the formation of

HO131I. 131I removal in river water by PAC only was lower

than that in pondwater (Fig. 2). However, 131I removal in river

water by combination of chlorination and PAC was higher

than that in pond water. Thus, it was presumed that the

proportion of 131I� in river water was greater than that in

pond water.

HOI transformed by chlorination of I� is further trans-

formed into IO�3 by reaction with chlorine (Lettinga, 1972;

Bichsel and von Gunten, 1999). The transformation yield of

IO�3 depends on chlorination conditions, such as pH, chlorine

concentration, and reaction time. It was reported that the

distribution coefficients of 131IO�3 for active charcoal at pH 6e8

were lower than those of 131I� by one or two orders of

magnitude (Ikeda and Tanaka, 1975). Lettinga (1972) reported

that 131I� removal decreased by combination of chlorination

and PAC when chlorination reaction time was longer or

chlorine concentration was much higher. Fig. S10 shows the

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 4 3 9 7e4 4 0 4 4401

effects of chlorination conditions on 131I removal in pond

waters by PAC (PAC-1). In the case of chlorination followed by

PAC, the removal percentages of 131I at chlorination reaction

times of 30 s and 10minwere 32% and 33%, respectively. In the

case of PAC followed by chlorination, the value was 33%. That

is, 131I removal by the combination of chlorination and PAC

was unaffected by chlorination conditions. Therefore, it was

considered that in the present study, chlorine concentration

was not excessive and the transformation of 131I� into 131IO�3

was small.

Fig. S11 shows the effects of reaction time of PAC on the

removal of 131I in pond water by combination of chlorination

and PAC (PAC-1). The reaction timeswere 5, 10, and 30min 131I

levels in pond water before treatment were 7.4e9.5 Bq/L. PAC

concentration was 25mg dry/L and chlorination reaction time

was 10 min. The removal percentages of 131I at reaction times

of 5, 10, and 30 min with PAC were 36%, 48%, and 48%,

respectively. That is, the removal percentages at reaction

times of 10 and 30minwere similar under the conditions used

in the present study.

3.4. Removal of radioactive cesium at water purificationplants

Among the five water purification plants, the levels of 134Cs

and 137Cs at WPP-2 were sufficiently high to investigate their

removal during water purification processes. Fig. 3 shows the

profiles of 134Cs and 137Cs at WPP-2. On 20 April, 134Cs and137Cs levels in raw water were 2.5 and 3.0 Bq/L, respectively.

Those on 28 April were 5.6 and 6.4 Bq/L, respectively. Although

low levels of 137Cs and 134Cs were detected after sand filtration

and BAC, respectively, on 20 April, 134Cs and 137Cs were not

detected in other process water samples on 20 or 28 April.

Turbidities in raw water on 20 and 28 April were 15 and

51 degrees, respectively. The proportions of 134Cs and 137Cs

particles in raw water on 28 April were 86% and 87%, respec-

tively (details are presented in the next section). These results

suggested that 134Cs and 137Cs at WPP-2 existed mainly as

particles, and they were therefore effectively removed by

coagulationeflocculationesedimentation.

0

3

6

9

12

15

20 Apr. 28 Apr. 20 Apr. 28 Apr. 20 Ap

Raw water Sedimentation

137 C

s an

d 13

7 Cs

(Bq/

L)

< 1.0 < 0.59

< 0.50 < 0.83

< 0.4

Fig. 3 e Profiles of 134Cs

It was reported that 134Cs and 137Cs were detected in

sludge from water purification plants in Fukushima and its

neighboring prefectures after the explosion at the nuclear

power plant (WSD, HSB, MHLW, 2012). These observations

suggested that particles of 134Cs and 137Cs were removed

during coagulationeflocculationesedimentation. On the

other hand, 134Cs and 137Cs were occasionally detected in

drinking water after the explosion (WSD, HSB, MHLW, 2011).

As 134Csþ and 137Csþ (ionic form of 137Cse137Ba) were not

removed by coagulation in laboratory-scale experiments

(Morton and Straub, 1955), this was presumed to be because

particles were not sufficiently removed during water purifi-

cation processes or the proportions of dissolved forms of134Cs and 137Cs (e.g., 134Csþ and 137Csþ) were high. It was

reported that removal efficiency of 137Cs was low during

coagulationeflocculationesedimentation followed by rapid

and slow sand filtration when the turbidity in raw water was

0.80 FNU (Gafvert et al., 2002). At a pilot-scale, the maximum

removal percentages of 134Cs by coagulatione-

flocculationesedimentation and filtration were 31% and 56%,

respectively (Goossens et al., 1989). These reports also sug-

gested that the proportions of particulate and dissolved

forms of 134Cs and 137Cs were dependent on environmental

water.

Although 134Cs and 137Cs were not detected in raw or

process water at WPP-1 in the present study, they were

detected in some samples of GAC and sand (Table S2). In the

case of GAC, 134Cs and 137Cs on 22 April were 7.4 and 13 Bq/kg

wet, respectively. It was reported that removal of ionic forms

of 137Cse137Ba by ACwas low (Rivera-Utrilla et al., 1984). Thus,

the reason for the detection of 134Cs and 137Cs in GAC was

presumed to be because flocks containing 134Cs and 137Cs

passing the sand filter were captured at GAC. In the case of

sand, 134Cs and 137Cs on 31 May were 40 and 51 Bq/kg wet,

respectively. Those on 25 August were 13 and 11 Bq/kg wet,

respectively. One reason for the detection of 134Cs and 137Cs in

sand was presumed to be because flocks containing 134Cs and137Cs passing the sedimentation tank were captured at the

rapid sand filters. It was reported that 137Csþ in purewaterwas

removed by sand filters despite the low ion exchange

r. 28 Apr. 20 Apr. 28 Apr. 20 Apr. 28 Apr.

BAC Rapid sand f iltration

Slow sand f iltration

137Cs134Cs

< 0.58 < 0.76

< 0.57 < 0.58

< 0.69 < 0.63

< 0.60 < 0.65

7 < 0.67

and 137Cs at WPP-2.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 4 3 9 7e4 4 0 44402

capacities (Ohshio, 1960). The average removal percentages of137Csþ by 13 types of sand filter were from47.7% to 99.7%when

the ratio of passingwater volume to sand filter volumewas 15.

In this previous study, the sand was prepared for rapid/slow

sand filters at water purification plants or already used at

water purification plants. As shown in the next section, some

proportions of 134Cs and 137Cs existed as 134Csþ and 137Csþ,respectively, in river and pond waters, although the propor-

tions of 134Csþ and 137Csþ were very different between the

environmental waters. Thus, another possibility for the

detection of 134Cs and 137Cs in sand was that 134Csþ and 137Csþ

were removed by the ion exchange ability of sand and were

accumulated although their levels in process water were low.

3.5. Removal test of radioactive cesium

Fig. 4 shows the removal percentages of 134Cs and 137Cs in river

and pond waters by coagulation. 134Cs and 137Cs levels in river

water were 5.6 and 6.4 Bq/L, respectively. 134Cs and 137Cs levels

in pondwater were 9.8 and 11 Bq/L, respectively. Turbidities in

river and pond waters were 51 and 0.3 degrees, respectively. In

the case of river water, removal percentages of 134Cs and 137Cs

were 94% and 95%, respectively. In the case of pond water,

removal percentages of 134Cs and 137Cs were 5% and 6%,

respectively. In addition, the removal percentages of 134Cs and137Cs in river waters by filtration only were 86% and 87%,

respectively, and those in pond waters were 0%. These results

showed that 134Cs and 137Cs in river waterwere presentmainly

in particulate form, and were therefore effectively removed by

coagulation. These results supported those shown in Fig. 3. On

the other hand, 134Cs and 137Cs in pond water were present

mainly in dissolved form, and therefore were not removed by

coagulation. In addition, the removal percentages of 134Cs by

coagulation in both river and pondwaterswere similar to those

of 137Cs. Thus, it was considered that the proportions of the

forms of 134Cs and 137Cs in both river and pond waters were

similar. Moreover, in both the river and pond waters, the

removal percentages of 134Cs and 137Cs by coagulation were

slightly higher than those by filtration. In the present study, it

0

20

40

60

80

100

Filtration (River) Coagulation (River) Filtration (Pond) Coagulation (Pond)

Rem

oval

per

cent

ages

of 13

4 Cs

and

137 C

s (%

)

134Cs137Cs

Fig. 4 e Removal of 134Cs and 137Cs in river and pond

waters by filtration and coagulation (134Cs, 137Cs in river

water, 5.6 and 6.4 Bq/L; 134Cs, 137Cs in pond water, 9.8 and

11 Bq/L; turbidity in river and pond water, 51 and 0.3

degrees).

was operationally defined that components passing 0.5-mm

filters were in the dissolved form. Thus, it was presumed that

some percentages of 134Cs and 137Cs may be sorbed onto small

particles <0.5 mm in diameter, and such 134Cs and 137Cs were

removed by coagulation with these small particles. Alterna-

tively, some percentages of 134Cs and 137Cs may have been

present in dissolved form, which could be removed by coagu-

lation (e.g., colloids).

To investigate the types of dissolved forms of 134Cs and137Cs in pond water, we examined their removal by PAC and

cation exchange resin (Fig. S12). Removal percentages of 134Cs

and 137Cs by cation exchange resin were 95% and 92%,

respectively. On the other hand, removal percentages of 134Cs

and 137Cs by PACwere 7% and 0%, respectively. It was reported

that cation exchange resin showed strong activity for removal

of the ionic form of 137Cse137Ba (Morton and Straub, 1955). It

was also reported that the removal levels of the ionic form of137Cse137Ba by AC (Rivera-Utrilla et al., 1984) and coagulation

were low (Morton and Straub, 1955). These results showed

that 134Cs and 137Cs in pond water exist mainly as Csþ.Moreover, the sample of river water after coagulation was

further treated by cation-exchange resin. 134Cs and 137Cs were

not detected in the sample after cation-exchange (data not

shown). Thus, it was presumed that the forms of 134Cs and137Cs remaining after coagulation in the river water were134Csþ and 137Csþ, respectively.

The effects of suspended solids on the removal of 134Cs and137Cs in pond water by coagulation were investigated. As

suspended solids, sediment A or B was initially added to pond

water and mixed at 24 h. Subsequently, coagulation of the

sample was performed. When the amounts of sediment A

added were 0, 5, 10, and 30 degrees as turbidity, the removal

percentages of 134Cs were 5%, 56%, 67%, and 92%, respectively

(Fig. 5). Those of 137Cs were 6%, 60%, 73%, and 91%, respec-

tively. When the amounts of sediment B added were 10

degrees as turbidity, the removal percentages of 134Cs and137Cs were 94% and 90%, respectively. These results indicated

that removal of 134Cs and 137Cs by coagulation increased with

0

20

40

60

80

100

0 10 20 30

Rem

oval

per

cent

ages

of 13

4 Cs

and

137 C

s (%

)

Turbidity

134Cs (Sediment A)134Cs (Sediment B)137Cs (Sediment A)137Cs (Sediment B)

Fig. 5 e Effects of suspended solids on removal of 134Cs and137Cs in pondwater by coagulation (134Cs and 137Cs, 9.8 and

11 Bq/L; turbidity in pond water, 0.3 degrees).

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 4 3 9 7e4 4 0 4 4403

the addition of sediment. Csþ is known to be sorbed by soils

and minerals (Cornell, 1993; Staunton, 1994). Thus, in the

present study it was considered that 134Csþ and 137Csþ in pond

waterwere sorbed on the sedimentswithin 24 h of contact and

were removed by coagulation with sediments. Results similar

to those of the experiments using 137Csþ (or ionic forms of137Cse137Ba) were reported previously (Morton and Straub,

1955; Kobayashi et al., 1969). However, the degree of increase

in removal of 134Cs and 137Cs by addition of suspended solids

was higher in the present study than in these previous studies.

That is, in the study by Morton and Straub (1955), the removal

percentages of ionic forms of 137Cse137Ba by coagulation in the

absence and presence of clay (100 mg/L) were 0%e6% and

35%e65%, respectively. In the study of Kobayashi et al. (1969),

the removal percentages of 137Csþ in river water (turbidity, 10

degrees) by coagulation were 10.9%e31.0% and those after

addition of sediments to 95 degrees as turbidity were 71%e

76%. The reaction times of 137Csþ (or ionic forms of137Cse137Ba) with suspended solids before coagulation were

not described in either of these previous reports. Thus, it was

presumed that the degree of improvement in the removal of134Cs and 137Cs by addition of suspended solidswas dependent

on the types of suspended solids and experimental conditions.

4. Conclusions

From the occurrence of 131I at water purification plants, it was

shown that 131I was not removed by coagulatione-

flocculationesedimentation. The removal of 131I by GAC and

PAC ranged from 30% to 40%, although 131I was not removed in

some cases.

Laboratory-scale experiments indicated that the removal

percentages of 131I in river and pond waters by 25 mg dry/L of

PAC (PAC-1) were 36% and 41%, respectively. These levels

were increased by the combination of chlorination and PAC

(59% for river water and 48% for pond water).

The occurrence of 134Cs and 137Cs at water purification

plants indicated that they were effectively removed by coag-

ulationeflocculationesedimentation when turbidity in raw

water was relatively high.

The results of laboratory-scale experiments showed that134Cs and 137Cs in river water with relatively high turbidity (51

degrees) were removed by coagulation, but those in pond

water with low turbidity (0.3 degrees) were not. The propor-

tions of particulate and dissolved forms of 134Cs and 137Cs

were high in river and pond waters, respectively. The removal

of 134Cs and 137Cs in pond water by coagulation increased

markedly when 134Cs and 137Cs were mixed with sediment at

24 h before coagulation.

Acknowledgments

The authors thank Mr. Masami Oya, Mr. Yasuo Tanaka, Mr.

Nobuaki Munakata, and Mr. Yusei Kobayashi (Hanshin Water

Authority, Japan) for their help with the experiments. The

authors also thank the officials of water utilities for sample

collection.

Appendix A. Supplementary material

Supplementary data related to this article can be found online

at doi:10.1016/j.watres.2012.05.055.

r e f e r e n c e s

Asami, M., Akiba, M., 2011. Radionuclides in drinking water: factsand response. Journal of National Institute of Public Health 60,306e313 (in Japanese).

Bichsel, Y., von Gunten, U., 1999. Oxidation of iodide andhypoiodous acid in the disinfection of natural waters.Environmental Science & Technology 33, 4040e4045.

Brown, J., Hammond, D., Wilkins, B.T., 2008a. Handbook forAssessing the Impact of a Radiological Incident on Levels ofRadioactivity in Drinking Water and Risks to Operatives atWater Treatment Works (HPA-RPD-040). Health ProtectionAgency, Chilton, Oxfordshire.

Brown, J., Hammond, D., Wilkins, B.T., 2008b. Handbook forAssessing the Impact of a Radiological Incident on Levels ofRadioactivity in Drinking Water and Risks to Operatives atWater Treatment Works: Supporting Scientific Report (HPA-RPD-041). Health Protection Agency, Chilton, Oxfordshire.

Cornell, R.M., 1993. Adsorption of cesium on minerals: a review.Journal of Radioanalytical and Nuclear Chemistry 171,483e500.

Esumi, S., Fuzii, K., Terai, K., Yamamoto, H., 1986. Effect ofaccident of Chernobyl nuclear power plant in the Soviet Unionobserved in raw water and drinking water. Report of theShimane Prefectural Institute of Public Health andEnvironmental Science 28, 48e52 (in Japanese).

Gafvert, T., Ellmark, C., Holm, E., 2002. Removal of radionuclidesat a waterworks. Journal of Environmental Radioactivity 63,105e115.

Goossens, R., Delville, A., Genot, J., Halleux, R., Masschelein, W.J.,1989. Removal of the typical isotopes of the Chernobyl fall-outby conventional water treatment. Water Research 23,693e697.

Honma, S., Takahashi, T., Yamagaki, K., 1988. Contamination ofradionuclides in water supply in Niigata city by the accident ofthe Chernobyl nuclear power plant and their removalexperiments by activated carbon. Journal of Japan WaterWorks Association 640, 24e29 (in Japanese).

Ikeda, N., Tanaka, K., 1975. Column chromatographicseparation of radioactive tellurate, tellurite, iodide andiodate by active charcoal. Journal of Chromatography 114,389e395.

Ikemoto, T., Magara, Y., 2011. Measures against impacts ofnuclear disaster on drinking water supply systems inJapan. Water Practice and Technology 6. doi:10.2166/wpt.2011.078.

Kamata, H., Yukawa, M., Saeki, M., 1973. Studies on movement of90Sr, 106Ru, 137Cs, 144Ce in environment, connected with thecity water supply. Japanese Journal of Toxicology andEnvironmental Health 19, 334e339 (in Japanese).

Kobayashi, U., Ohzeki, S., Souda, H., Yamamoto, H., 1969. Basicexperiment on removal of radionuclides in water. Journal ofJapan Water Works Association 417, 389e395 (in Japanese).

Lettinga, G., 1972. Radioactivity and Water Supplies.Mandi�c, M., Vukovi�c, �Z., Lazi�c, S., Rai�cevi�c, S., 1996. Sorption of

hypoiodous acid on activated carbon. Journal ofRadioanalytical and Nuclear Chemistry 208, 453e460.

Morton, R.J., Straub, C.P., 1955. Removal of radionuclides fromwater by water treatment processes. Journal of the AmericanWater Works Association 12, 545e557.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 4 3 9 7e4 4 0 44404

Noguchi, H., Murata, M., 1988. Physicochemical speciation ofairborne I-131 in Japan from Chernobyl. Journal ofEnvironmental Radioactivity 7, 65e74.

Ohshio, T., 1960. Removal of 137Cs and 90Sr by sandfiltration. Journalof JapanWater Works Association 306, 42e47 (in Japanese).

Rivera-Utrilla, J., Ferro-Garcia, M.A., Mata-Arjona, A., Gonzalez-Gomez, C., 1984. Studies on the adsorption of Caesium,Thallium, Strontium and Cobalt radionuclides on activatedcarbons from aqueous solutions. Journal of ChemicalTechnology and Biotechnology 34A, 243e250.

Summers, R.S., Fuchs, F., Sontheimer, H., 1988. The fate andremoval of radioactive iodine in the aquatic environment. In:Aquatic Humic Substances, pp. 623e636. Chapter 35.

Staunton, S., 1994. Adsorption of radiocaesium on varioussoils: interpretation and consequences of the effects ofsoil: solution ratio and solution composition on thedistribution coefficient. European Journal of Soil Science45, 409e418.

WSD, HSB, MHLW, 2011. Interim Report For Measures AgainstRadioactive Materials in Drinking Water (in Japanese).accessed on 28.04.12. http://www.mhlw.go.jp/stf/shingi/2r9852000001g9fq-att/2r9852000001g9jp.pdf.

WSD, HSB, MHLW, 2012. Documents of Sixth Meeting forMeasures against Radioactive Materials in Drinking Water (inJapanese). accessed on 28.04.12. http://www.mhlw.go.jp/stf/shingi/2r98520000022wul.html.