Desorption of Intrinsic Cesium from Smectite ... - ACS Publications

Article pubs.acs.org/est

Desorption of Intrinsic Cesium from Smectite: Inhibitive Effects of Clay Particle Organization on Cesium Desorption Keisuke Fukushi,*,† Haruka Sakai,‡ Taeko Itono,§ Akihiro Tamura,§ and Shoji Arai§ †

Institute of Nature and Environmental Technology, ‡School of Natural Systems, College of Science and Engineering, and §Division of Earth and Environmental Sciences, Graduate School of Natural Science and Technology, Kanazawa University, Kakuma, Kanazawa, Ishikawa 920-1192, Japan S Supporting Information *

ABSTRACT: Fine clay particles have functioned as transport media for radiocesium in terrestrial environments after nuclear accidents. Because radiocesium is expected to be retained in clay minerals by a cation-exchange reaction, ascertaining trace cesium desorption behavior in response to changing solution conditions is crucially important. This study systematically investigated the desorption behavior of intrinsic Cs (13 nmol/ g) in well-characterized Na−montmorillonite in electrolyte solutions (NaCl, KCl, CaCl2, and MgCl2) under widely differing cation concentrations (0.2 mM to 0.2 M). Batch desorption experiments demonstrated that Cs+ desorption was inhibited significantly in the presence of the environmental relevant concentrations of Ca2+ and Mg2+ (>0.5 mM) and high concentrations of K+. The order of ability for Cs desorption was Na+ = K+ > Ca2+ = Mg2+ at the highest cation concentration (0.2 M), which is opposite to the theoretical prediction based on the cation-exchange selectivity. Laser diffraction grain-size analyses revealed that the inhibition of Cs+ desorption coincided with the increase of the clay tactoid size. Results suggest that radiocesium in the dispersed fine clay particles adheres on the solid phase when the organization of swelling clay particles occurs because of changes in solution conditions caused by both natural processes and artificial treatments.

■

INTRODUCTION Radioactive cesium (hereinafter radiocesium) dispersed into the environment because of the Fukushima Daiichi Nuclear Plant accident has caused great concern. Some radiocesium is retained in fine particles in surface soils and sediments near the power plant.1−4 The radiocesium migration is related to the transport of fine particles with river and runoff water.2,5 It is generally believed that 2:1 clay minerals such as smectite, vermiculite, and illite are the major sorbents for radiocesium.4,6,7 Actually, Cs sorbs to the clay minerals via a cation-exchange reaction.8 Consequently, the radiocesium retained in fine particles is desorbed to solutions in the presence of high concentrations of foreign cations depending on the cation type. During the transport of fine particles in surface conditions, the surrounding solution conditions must change. Therefore, for the prediction of the migration of radiocesium, elucidating the desorption behavior of radiocesium from clay minerals in response to changes of solution conditions is crucially important. The reported highest level of the radiocesium (134Cs and 137Cs) in soils and sediments in Fukushima, except for areas very close to the power plant, was several tens to hundreds of kBq/kg.1,3,9,10 This level corresponds to the order of 10−12 to 10−10 mol/kg of 134Cs or 137Cs. Soils and sediments usually contain more than 10−5 mol/kg (10 nmol/g) of intrinsic 133Cs.11,12 Assuming that the © XXXX American Chemical Society

chemical behavior of radiocesium can be regarded as identical to that of 133Cs, it is necessary to ascertain the intrinsic Cs desorption from clay minerals to understand radiocesium desorption behavior. The clay mineral composition of soils in Fukushima is dominated by smectites,6 which are common clay mineral species in soils and sediments and which possess the ability to retain Cs.13,14 Many investigations have assessed the adsorption of Cs to smectite,13−20 but few reports describe Cs desorption from smectite.16,17,20 Iijima et al. investigated the adsorption and desorption of Cs from colloidal smectite in diluted solutions.16 They reported that the Cs uptake in the purified smectite is reversible even in very low Cs loadings. However, in suspensions of smectite, the spatial organization of clay platelet, i.e., quasicrystal or tactoid formation, usually occurs in the presence of electrolyte cations depending on the type and concentration.21−23 The large tactoid formation of smectite leads to the hysteresis of ion exchange reactions,18,24,25 but the reversibility of Cs desorption from smectite has not been assessed in the presence of various types and wide ranges of Received: June 6, 2014 Revised: August 21, 2014 Accepted: August 21, 2014

A

dx.doi.org/10.1021/es502758s | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

Desorption Experiments. The stock smectite suspensions were diluted 10 times with double-distilled demineralized water. The diluted suspensions were mixed at least one night on the mix-rotor at 25 ± 0.5 °C. A small volume of the stock solutions of each cation was added to the suspension to adjust 0.0002 to 0.2 M of cation concentrations in the system. Experiments using the mixed cation solution of 0.01 M Na+ and 0.001 M Mg + were also conducted. The total solid concentration in each batch was 0.96 g/L. The suspensions were placed on the mix-rotor at 25 ± 0.5 °C for 24 h. Preliminary results showed that the suspension pH was 6.4−7.3 (Figure S1, Supporting Information). Quadruplicate batches were conducted for a single condition. After the reaction times, separation of the supernatants from solids was conducted using high-speed centrifugation (22000g) for 20 min. The supernatants were filtered with a 0.2 μm membrane. They were then acidified by the addition of a small amount of concentrated HNO3. The acidified supernatants were used for Cs measurements by ICP-MS (X7; Thermo Elemental) conducted using internal standard method (10 ppt detection limit) and Si and Al measurements using an ICP-optical emission spectrometer (OES; 710-ES; Varian, Inc.). Kinetic studies of each cation (0.01 M) of 1−50 h were performed using the same procedure. Grain Size Distribution. The volumetric basis grain size distributions of clay particles in the suspension in the presence or absence of cations were measured using a laser diffraction grain size analyzer (SALD-2200; Shimadzu Corp.). The measurable range of the grain size is 10 nm to 1 mm. The suspensions were mixed at least one night on the mix-rotor at 25 ± 0.5 °C before each cation was added. A small volume of the stock solution of each cation was added to the suspension to adjust 0.0002 to 0.2 M of the cation concentrations in the system. The suspensions were placed on the mix-rotor at 25 ± 0.5 °C for 24 h and were used for grain size distribution measurements. Solid concentrations in the presence of Na+ and K+ were 0.46 g/L. Those in the presence of Ca2+, Mg2+ and mixed cation (Mg 2+ + Na + ) were 0.096 g/L. Solid concentrations with 0.96 g/L (same solid concentrations as batch desorption experiments) for most conditions were too high to obtain reliable results because of the occurrences of the diffused reflection in the equipment. No significant difference was found in the grain size distribution among the different solid concentrations (Figure S2, Supporting Information). Modeling. Cation-exchange modeling was performed using REACT from the Geochemist’s Workbench.30 The activity coefficient was calculated using an extended Debye−Hückel equation. Thermodynamic data for aqueous species were from “thermo.dat” of the Geochemist’s Workbench. The following exchange reactions and the corresponding selectivity coefficients were considered in the modeling.

concentrations of electrolyte cations. Therefore, this study systematically examined the desorption behavior of intrinsic Cs+ in well characterized Na-montmorillonite in electrolyte solutions (NaCl, KCl, CaCl2, and MgCl2) under widely different cation concentrations.

■

MATERIALS AND METHODS Materials. A smectite sample (Kunipia-F) was supplied by Kunimine Industries Co. Ltd. Kunipia-F is a Na-type smectite prepared by the elutriation of crude Tsukinuno bentonite (Kunigel-V1). The mineralogical, physical, and chemical properties of Kunipia-F are presented in Table 1. The reported Table 1. Mineralogical, Chemical, and Physical Properties of Kunipia-F properties minerals (wt %) montmorillonite quartz calcite soluble impurities (mol/g) Na2SO4 NaCl exchangeable cations (mequiv/g) cation exchange capacity, CEC exchangeable Na+ exchangeable K+ exchangeable Mg2+ exchangeable Ca2+ intrinsic Cs+ (mol/g) Others median grain size (μm) in diluted solution specific surface area (m2/g) specific density (kg/m3)

sources 99 traces (0.5) traces (0.5)

ref 14 ref 14 ref 14

4.17 × 10−5 1.54 × 10−5

ref 14 ref 14

1.08 0.975 0.0119 0.00011 0.0918 1.3 × 10−8

ref 14 ref 14 ref 14 ref 14 ref 14 this study

1.7 810 2880

this study ref 14 ref 14

chemical compositions of Kunipia-F are given in Table S1 (Supporting Information). In this study, the intrinsic Cs content in the smectite was measured using laser-ablation inductive coupled plasma−mass spectrometry (LA-ICP−MS) at Kanazawa University (Table S2, Supporting Information).26 Sample pellets for LA-ICP-MS measurements were prepared according to the process described by Ito et al.27 The NIST SRM612, used as the calibration standard, was analyzed at the beginning of the measurements. The element concentration of NIST SRM612 for the calibration was selected from the preferred values reported by Pearce et al.28 Single-spot analysis was conducted randomly for five pellets, yielding 10 spot data. Data reduction was facilitated by 29Si as internal standard elements, based on the average of reported SiO2 contents of Kunipia-F (56.8 ± 2.5 wt % from Table S1, Supporting Information), following a protocol that was fundamentally identical to that outlined by Longerich et al.29 The 133Cs content in Kunipia-F was measured as 13 nmol/g. Details of the measurement results are presented in Table S3 (Supporting Information). A single stock smectite suspension was used for all experiments in this study. The solid concentration of the stock smectite suspension was 9.6 g/L. The pH of the original stock suspension was slightly alkaline (pH 9−10). A small amount of concentrated HCl was added to the suspension to adjust the pH to neutral (pH 6.7).

>X:Na + + Cs+ = >X:Cs+ + Na +, KNa_Cs =

β> X:Cs+a Na+ β> X:Na+aCs+ (1)

>X:Na + + K+ = >X:K+ + Na +, KNa_K =

β> X:K+a Na+ β> X:Na+a K+ (2)

>X:Na + + H+ = >X:H+ + Na +, KNa_H =

β> X:H+a Na+ β> X:Na+a H+ (3)

B



Article

TOTNa + = [Na +]added + [Na +]initial + [Na +]impurity

2>X:Na + + Ca 2 + = >X 2:Ca 2 + + 2Na +, KNa_Ca =

(13)

2 + β> X:Ca 2+a Na

+

β>2X:Na+aCa 2+

TOTi = [i]added + [i]initial where i : K , Ca , and Mg

(5)

Therein, [i]added denotes the added cations concentrations for monovalent and divalent cations; [i] initial denotes the exchangeable cations in initial smectite per volume of solutions given as shown below.

(14)

2 + β> X:Mg 2+a Na

β>2X:Na+a Mg 2+

[i]initial = initial exchangeable i (mol/g)

In those equations, >X denotes the exchangeable site, a represents the aqueous activity, and β stands for the activity of a species complexed with an exchange site. Surface activity coefficients are not well-known quantities. Here, we assume that the activity coefficients of the species with exchange site are unity, as in previous studies.31 By the Gaines−Thomas convention,32 βj is given in terms of the fraction of the total electrical equivalents of exchange capacity occupied by cation j, with charge k. βj =

mequiv > xk : j per g solid CEC

2+

(4)

2>X:Na + + Mg 2 + = >X 2:Mg 2 + + 2Na +, KNa_Mg =

2+

× solid concentration (g/L)

(15)

The values of initial exchangeable cations are shown in Table 1. [Na+]impurity stands for the Na+ concentrations derived from the soluble impurity (Na2SO4 and NaCl in Table 1). It is noteworthy that the [Na+]initial + [Na+]impurity was almost identical to that of CEC. In this study, [Cs+]sol in eq 7 were calculated as a function of added cation concentrations. pH in the solution in the modeling was assumed to be 6.7, which was the average of all suspension pH after the reactions. The pH of reacted suspensions was 6.4−7.3 (Figure S1, Supporting Information). No difference was observed in the calculated [Cs+]sol within the pH range.

(6)

Therein, CEC is the cation-exchange capacity given in units of mequiv/g. The value of 1.08 mequiv/g was taken from the that of Kunipia-F given after Tachi and Yotsuji (Table 1).14 The value of KNa_Cs was estimated from this study as described later. The values of other selectivity coefficients were obtained from a report by Tachi and Yotsuji (Table 2).14

■

RESULTS AND DISCUSSION Desorption of Intrinsic Cs from Smectite in the Presence of Major Cations. Figure 1 presents the aqueous

Table 2. Selectivity Coefficients for Cation Exchange of Kunipia-F Used for This Study reaction

log K

source

>X:Na+ + H+ = >X:H+ + Na+ >X:Na+ + K+ = >X:K+ + Na+ 2>X:Na+ + Ca2+ = >X2:Ca2+ + 2Na+ 2>X:Na+ + Mg2+ = >X2:Mg2+ + 2Na+ >X:Na+ + Cs+ = >X:Cs+ + Na+

0.10 0.42 0.69 0.67 1.60 1.82

ref 14 ref 14 ref 14 ref 14 ref 14 this study

The mass balance equations, except for that for H+, are presented below. TOTCs+ = [Cs+]sol + [>x:Cs+]

(7)

TOTNa + = [Na +]sol + [>x:Na +]

(8)

TOTK+ = [K+]sol + [>x:K+]

(9)

TOTCa 2 + = [Ca 2 +]sol + [>x2:Ca 2 +]

TOTMg

2+

2+

2+

= [Mg ]sol + [>x2:Mg ]

Figure 1. Desorption kinetics of Cs from smectite. Cation concentrations in the kinetic studies were 0.01 M. The reaction times were 1, 2, 6, 12, 24, and 48 (K+, Ca2+ and Mg2+) or 56 (Na+) h.

(10) (11)

Cs concentrations in the presence of 0.01 M cations with time. Addition of all cations led to the desorption of Cs in solutions. The amounts of the desorption depend on cation types. Within the experimental errors obtained from the standard deviation of the quadruplicate experiments, the Cs concentration did not change with the reaction times. Figure 2 shows the Cs concentrations in solutions after the desorption experiments as a function of cation concentrations. Incomplete separation of the supernatant with clays was observed by centrifugation with low cation concentrations ([Na+] < 0.01 M, [K+] < 0.005 M, [Ca2+] and [Mg2+] < 0.0005

In those equations, TOTi denotes the total concentrations of ith cation in the system. [i]sol denotes aqueous concentrations of i the cations. [>x:i] or [>x2:i] stands for the concentrations of ith cation in exchangeable site in solution given in mol/L. TOTi is the total cation concentration in a system and calculated from the following equations. TOTCs+ = initial intrinsic Cs+ (mol/g) × solid concentration (g/L)

(12) C



Article

Figure 2. Cs concentrations in solution after the desorption experiments as functions of added Na+ (a), K+ (b), Ca2+ (c), and Mg2+ (c) concentrations. Solid lines in the figures represent the modeled Cs concentration based on the cation exchange model.

(approximately 10−8 M) in a 1 mM NaCl solution.16 To assess whether the observed desorption behavior by each cation is explainable by the reversible cation exchange, classical cationexchange modeling was applied. The selectivity coefficients used for this study are presented in Table 2. The selectivity coefficient for Cs+ to Na+ as log KNa_Cs = 1.82 gave a reasonable fit (Figure 1a), indicating that desorption of Cs in the presence of Na+ is explainable by a reversible cation-exchange reaction. Tachi and Yotsuji compiled the selectivity coefficient as log KNa_Cs =1.60 for Kunipia-F,14 which is close to the value obtained from the present study. Trace impurities of micaceous minerals (or illite/smectite interstratified minerals) in smectite are known to be very selective for Cs adsorption and fixation at low Cs loading.13,33,34 Several authors proposed multisite model considering the “frayed edge site” to represent the effect of micaceous minerals on trace Cs sorption on bentonite.13,34 Missana et al. examined the sorption behavior of Cs on natural Spanish FEBEX bentonite under widely diverse conditions of pH, ionic strength, and Cs concentrations.13 A two-site exchange model approach was adopted to interpret and model sorption data. The selectivity coefficient for Na+ of highly selective sites for Cs sorption, resembling those present in micaceous minerals with very low capacity but controlling

M; Figure S3, Supporting Information). Results of Cs desorption obtained from these experimental conditions were omitted from this figure. The Cs concentration increased with the Na+ concentration (Figure 2a). More than half of the total Cs in smectite was released to the solution at the highest Na+ concentration (0.2 M). The Cs concentration also increased with K+ concentrations (Figure 2b). However, it became almost constant around 7 nM at K+ concentrations higher than 0.05 M. The Cs desorption behaviors in the presence of Ca2+ differed from those in monovalent cations (Figure 2c). The Cs concentration at the lowest concentration (0.0005 M) was 2 nM. The concentrations increased slightly with the Ca2+ concentrations. The Cs desorption behavior in the presence of Mg2+ was almost identical to that in Ca2+ (Figure 2d). These results show qualitatively that Na+ and K+ in solutions acts as the desorption agent for the intrinsic Cs in smectite, although the desorption is inhibited at higher K+ concentrations (>0.05 M). However, the divalent cations in solution act as desorption agents only to a slight degree. Intrinsic Cs+ in smectite is expected to be retained by the structural negative charge because of isomorphic substitution in the octahedral layer. Iijima et al. described that the adsorption of Cs on Kunipia-F was reversible even at very low Cs loading D



Article

presence of K+ less than 0.05 M was 2−3 μm. However, it increased exponentially with K+ concentrations higher than 0.05 M. The grain size at the highest K+ concentration (0.2 M) was higher than 100 μm. The grain size in the presence of Ca2+ at the lowest concentration (0.0002 M) was 3 μm. Those sizes increased abruptly at 0.0005 M and increased gradually with the Ca2+ concentrations at [Ca2+] > 0.005M. Above 0.005 M of Ca2+, the median grain sizes were 10−30 μm. Grain sizes in the presence of Mg2+ were generally similar to those of Ca2+ but were systematically higher than those of Ca2+. Ahmad and Karube investigated the critical coagulation concentration (CCC) of Kunipia-F for different cation types and concentrations.35 They ascertained the CCC from visual inspection of the colloidal stability after 24 h settling of the suspension. They found that coagulation of smectite occurred at 3−30 mmol/L of NaCl solution, depending on the pH. They found no significant differences in CCC in monovalent cation salt solutions (NaCl, KCl, and NH4Cl). Their observations of the aggregation behavior in monovalent cation salt solutions were generally inconsistent with those of the present study, which showed that the grain size in NaCl media did not change in the entire range of experimental conditions (2 mM to 0.2 M). Moreover, we found significant differences between grain sizes in Na+ and K+ media at higher cation concentrations (>0.05 M). Three association modes of the smectite platelets exist:36 edge-to-edge, edge-to-face and face-to-face. Heller et al. demonstrated edge-to-edge and edge-to-face associations of smectite dominate at lower concentrations in monovalent electrolyte solutions.36 The face-to-face associations are associated with desorption water from the clay interlayer to form tactoid or quasi-crystal24,37 and are expected to have strong bonding, which cannot be dissociated easily in the shaking conditions. Basal spacing of smectites saturated with Na+ in dilute solution ( Na+ at any given cation concentration, which is consistent with the selectivity of cations to the interlayer of clay minerals.32 However, observations showed that the order depends on cation concentrations and that it is Na+ = K+ > Ca2+ = Mg2+ at the highest cation concentrations (0.2 M), which is opposite to the theoretical prediction. Effects of Spatial Organization on the Inhibition of the Intrinsic Cs Desorption. The median grain sizes calculated on a volumetric basis (hereinafter, grain sizes) in the presence of Na+ were around 2 μm in the full range of experimental conditions (Figure 3). The grain size in the

Figure 3. Median grain size measured using a laser diffraction grain size analyzer with no other treatment such as ultrasonic dispersion. E



■

exchange with the internal surface sites because of the large activation energy. Therefore, large tactoid formation associated with the increase of the proportion of interlayer site to the external site results in the changes of apparent chemical affinity of cations. Tournassat et al. considered that the external surface site and interlayer site have different affinities for cations.22 The overall affinity of cations on the smectite depends on the tactoid size. Three physicochemical reactions must exist in systems with K+, Ca2+, and Mg2+. The first is cation exchange of the foreign cations (K+, Ca2+, and Mg2+) with interlayer cations (mainly Na+). The second is associations of smectite platelets. The third is desorption of the intrinsic Cs+ by the foreign cations. Spectroscopic and molecular simulation reports have described that Cs primarily forms an inner sphere surface complex on smectite.40,41 Results of previous reports have described that Cs+ diffuses very slowly in the Na−smecite interlayer,42 which indicates that the Cs+ desorption is a kinetically slow process and that the associations of smectite particles possibly occur before the entire Cs+ desorption. The Cs+ inside the tactoids cannot access the external solution directly.24 In a single 0.001 M Mg2+ solution, the smectite particles were aggregated (Figure 3) and the Cs desorption was inhibited (Figure 2d). However, in the presence of a high concentration of Na+, the grain size decreased markedly to less than 5 μm (Figure 3). The Cs desorption was not inhibited. This result from the mixed cation solutions supports the mechanism of the inhibition of Cs desorption because of the increase of the smectite tactoid size. This study demonstrated that the desorption of intrinsic Cs from clay particles is influenced not only by the cationexchange reaction but also by the organization of the clay particles with response to the changes of solution chemistry. The ionic strength and hardness of terrestrial water in Japan are generally low. For example, the concentrations of Na and Ca in a reservoir in Iitate village, Fukushima prefecture, which was heavily contaminated by the accident,1 were both less than 0.2 mM (Table S4, Supporting Information). Fine particles in the terrestrial water in the area are most likely in a dispersed state. Therefore, radiocesium emitted from the accident adsorbed to dispersed clay particles in terrestrial conditions. During the transport of fine particles in surface conditions, the surrounding solution conditions must change. According to a prediction based solely on the cation exchange reaction, most of the retained Cs must be released in the solution with high ionic strength, shown as modeled lines in Figure 2. However, the present study demonstrated that a solution with high ionic strength inhibits Cs desorption especially when the divalent cations are important species. In Fukushima prefecture, limestone is distributed in downstream areas of some rivers43 that flow from areas that are severely contaminated by radiocesium. The addition of calcium from the calcium carbonate to river water might influence radiocesium migration. However, highly concentrated Na+ (

Desorption of Intrinsic Cesium from Smectite ... - ACS Publications

Recommend Documents