Introduction

Since the accident at the Fukushima-Daiichi nuclear power plant (FNPP) in 2011, there has been an immense need for novel ways to remove radioisotopes from contaminated water. A substantial number of radioisotopes such as strontium-90 (90Sr), cesium-134 (134Cs), and cesium-137 (137Cs) were released from FNPP into the environment, thereby contaminating the seawater, groundwater, and land surface1,2,3. The radioactivity of the Sr and Cs released from FNPP onto the land surface was estimated to be around 8.56 PBq for 90Sr and 288 PBq for 134Cs and 137Cs, whereas that of the strontium and cesium directly released into the ocean was estimated to be 52 GBq for 90Sr and 3.5 PBq for 134Cs and 137Cs4, 5. Of these radioisotopes, 90Sr poses a major threat to human health and the environment because it can replace calcium (Ca) in bones and in plants, leading to genetic mutations6. Thus, studying the migration of 90Sr in solution and determining suitable decontamination procedures are important issues to be researched.

Sorbent materials such as minerals5, 7,8,9,10,11,12,13 and polymeric materials13 have been used for the removal of 90Sr so far, however, development of a sorbent with both high sorption capacity and high ion selectivity is essential. Usually, several types of ion species such as magnesium ion (Mg2+) and calcium ion (Ca2+) are present in contaminated seawater and groundwater5, 14, 15, making it difficult for sorbents to maintain their targeted ion sorption capacity. Sorbents with Cs+ selectivity have been developed16, 17, but materials with Sr selectivity have hardly been reported.

Hydroxyapatites (HAPs), which are calcium phosphate crystals, are the main components of our teeth and bones. They are of considerable interest in environmental and biomedical fields due to their ion-exchange ability and sorption capacity5, 12, 18, 19. In HAP crystals, Ca2+ ions are located at two distinct crystallographic sites, which provide exchange sites for a wide range of divalent cations20,21,22. As Sr2+ and Ca2+ have the same charge and a similar ionic radii (1.2 Å23 and 1.0 Å23, respectively), Sr2+ can be adsorbed onto HAP with high efficiency4, 12, 24. This preferential selectivity for Sr would make HAP an effective absorbent for the removal of Sr2+. In fact, a HAP reactive barrier has already been successfully trialed at the Hanford nuclear site in the USA to prevent the migration of 90Sr into the Columbia River25. Recently, HAPs are also being tested as a potential remediation material at the FNPP site15.

The chemical composition of HAPs can be modified from the stoichiometric form (Ca10(PO4)6(OH)2) to the Ca-deficient form (Ca10−x(HPO4)x(PO4)6−x(OH)2−x) by selecting appropriate Ca/P molar ratios18, 26,27,28. Natural HAPs in teeth and bones also have a nonstoichiometric form18. Moreover, HAPs can maintain their crystal structure regardless of the Ca/P molar ratio; thus, Ca-deficient HAP has vacant Ca2+ sites within its crystal structure18, 27, 28. It is known that the targeted ion-sorption properties of HAPs strongly depend on the Ca/P molar ratio29,30,31,32. In this study, we investigated the potential of Ca-deficient HAP for cleaning up contaminated water via Sr2+ sorption.

Results

Material characterization

The physical–chemical characterization of Ca-deficient hydroxyapatite (DEF-HAP, Ca/P = 1.38) and stoichiometric hydroxyapatite (ST-HAP, Ca/P = 1.68) is summarized in Table 1. Transmission electron microscopy (TEM) images of the materials (Fig. 1(a) and (b)) show that both DEF-HAP and ST-HAP form rod-like crystals and that the crystals of DEF-HAP are smaller than those of ST-HAP. The size of the crystals was estimated by analyzing these images; the average length (L) of the crystals of DEF-HAP (43 ± 5.7 nm) was approximately 46 nm shorter than that of ST-HAP (89 ± 17 nm), whereas their average widths (W) were almost identical (~17 nm) (Table 1). The X-ray powder diffraction (XRD) profiles of DEF-HAP and ST-HAP (Fig. 1(c)) have peaks in the 2θ range of 10–70°, and the positions of the peaks are similar to those of hydroxyapatites having a hexagonal structure (P63/m)24. The broader peak patterns observed for DEF-HAP might be due to its smaller crystal sizes, which agrees with the TEM observations. A broad band observed at a 2θ angle in the range 15°–28° might have originated from an amorphous component in the HAPs. Notably, the integrated intensities of the broad bands for DEF-HAP and ST-HAP are almost identical. At pH 7, the zeta potentials of DEF-HAP and ST-HAP were −20.1 and −5.46 mV, respectively.

Table 1 Physical–chemical characterization of DEF-HAP and ST-HAP.
Figure 1
figure 1

HAP crystal structures. Panels (a,b) show the TEM images of DEF-HAP and ST-HAP, respectively. The scale bars represent 200 nm. (c) X-ray powder diffraction profiles of DEF-HAP and ST-HAP. Miller indices corresponding to an hydroxyapatite phase ((P63/m)24 are indicated.

Effect of competing cations

The effect of competing cations on the Sr2+ sorption capacity of DEF-HAP and ST-HAP was evaluated using an aqueous solution containing HAPs (35 mg), Sr2+ (0.05 mmol/L), and either Mg2+ or Ca2+ (0–50 mmol/L) in MilliQ water (7 mL). The solution was stirred for 5 h at pH 7. The supernatant was obtained by centrifugation and filtration, and the concentration of Sr2+ was determined using inductively coupled plasma-mass spectrometry (ICP-MS). Notably, the equilibrium time for the sorption was confirmed to be 1 h; hence, five hours were sufficient to evaluate the sorption properties of the HAP materials (refer to Fig. S1 in the Supplementary Information). Based on the concentration of Sr2+ in the supernatant, the sorption efficiency (A eff ) of DEF-HAP and ST-HAP was calculated as follows:

$$\,{A}_{eff}( \% )=100\times (1-\frac{C}{{C}_{0}}),$$
(1)

where C 0 and C are the initial and final (after 5 h) concentrations (mg/L) of Sr2+, respectively.

As shown in Fig. 2(a), DEF-HAP demonstrates a high A eff in the absence of Mg2+ and Ca2+ (83%), and the values remain almost constant in the presence of Mg2+ and Ca2+ at concentrations between 0.1 and 1.0 mmol/L. The values decrease as the concentration increases above 1.0 mmol/L, decreasing to 42% for Mg2+ and 14% for Ca2+ at 50 mmol/L. Conversely, ST-HAP exhibits a lower A eff than DEF-HAP in the absence of Mg2+ and Ca2+ (74%) and the values decrease even in the presence of small amounts of Mg2+ and Ca2+ (72% for Mg2+ and 51% for Ca2+ at 0.1 mmol/L). These values significantly decrease to 14% for Mg2+ and 0.5% for Ca2+ at 50 mmol/L. The dashed and solid lines in Fig. 2(a) and (b) denote the Mg2+ and Ca2+ concentrations in seawater and groundwater at FNPP, respectively. Our results demonstrate that DEF-HAP can maintain its Sr2+ sorption capacity under the existing groundwater conditions at FNPP.

Figure 2
figure 2

Effect of competing ions on the Sr2+ sorption capacity of HAPs. The change in A eff for DEF-HAP (dark gray) and ST-HAP (light gray) as a function of the concentration (mmol/L) of (a) Mg2+ ([Mg2+]) and (b) Ca2+ ([Ca2+]) is shown. The dashed and solid lines represent the approximate average [Mg2+] and [Ca2+] in seawater14 (52 mmol/L for Mg2+ and 9.4 mmol/L for Ca2+) and in FNPP ground water15 (0.86 mmol/L for Mg2+ and 0.70 mmol/L for Ca2+), respectively. The hydrated ionic radii of Mg2+ and Ca2+ are 0.6 Å and 1.0 Å, respectively. α is the ionic hydrated radius (Å).

Dependence on initial concentration

We investigated the dependence of the Sr2+ concentration on the sorption capacity of DEF-HAP and ST-HAP using aqueous solutions initially containing 0.01, 0.05, 0.1, 1, 5, and 10 mmol/L of Sr2+ at pH 7. Their supernatants were obtained after 5 h, and A eff was estimated using Eq. (1). As shown in Fig. 3(a), the A eff of DEF-HAP is 82% at 0.01 mmol/L; this value remains almost constant as the concentration is increased to 0.1 mmol/L. For higher Sr2+ concentrations, A eff decreases with an increase in the concentration and decreasing to 10% at 10 mmol/L. For ST-HAP, the A eff at 0.01 mmol/L is 70%, which significantly decreases above 0.05 mmol/L. The values of A eff for DEF-HAP are higher than those of ST-HAP at all measured concentrations. The most significant difference was observed at 1 mmol/L where the value of A eff for DEF-HAP was 60%, which is 75% higher than that of ST-HAP at the same concentration, that is, 15%.

Figure 3
figure 3

Dependence of initial concentration of Sr2+ on the sorption capacity of HAPs. (a) A eff of DEF-HAP (closed circles) and ST-HAP (open circles) as a function of Sr2+ concentration ([Sr2+]) between 0.01 and 10 mmol/L. The dashed lines are guides to the eye. The hydrated ionic radius of Sr2+ is 1.2 Å. α is the ionic hydrated radius (Å). (b) Sorption isotherm of DEF-HAP (closed circles) and ST-HAP (open circles) fitted to a Langmuir isotherm model using the values of q max and b (solid lines) (refer to Fig. S2 in the Supplementary Information).

Figure 3(b) shows the sorption isotherms of DEF-HAP and ST-HAP fitted to a Langmuir isotherm model33. The model assumes monolayer coverage on a surface with a finite number of identical sites and is represented as follows33:

$$\frac{{C}_{e}}{{q}_{e}}=\frac{1}{{q}_{{ma}{{x}}^{b}}}+\frac{{C}_{e}}{{q}_{\max }},$$
(2)

where C e is the equilibrium concentration (mg/L) of Sr2+, q e is the amount of adsorbed Sr2+ at equilibrium (mg/g), and q max (mg/g; maximum sorption capacity) and b (affinity) are Langmuir constants, respectively. The parameters q max and b, evaluated via Eq. (2) (refer to Fig. S2 in the Supplementary Information), and the coefficients of determination (R 2) are summarized in Table 2. The result shows that DEF-HAP (15.4 mg/g) has a larger maximum sorption capacity for Sr2+ than ST-HAP (3.11 mg/g).

Table 2 Langmuir isotherm parameters for DEF-HAP and ST-HAP.

Extended X-ray absorption fine structure (EXAFS) measurements

The Sr2+ sorption mechanism on DEF-HAP and ST-HAP was investigated via EXAFS measurements. Figure 4(a) and (b) shows the EXAFS spectra of DEF-HAP and ST-HAP around the K-edge of Sr and the Fourier transforms of DEF-HAP and ST-HAP at concentrations of 10 and 100 mmol/L of Sr2+ (these samples are referred to as DEF-HAP-Sr10, DEF-HAP-Sr100, ST-HAP-Sr10, and ST-HAP-Sr100). The features of the spectra are similar to those of a previously reported Sr-doped hydroxyapatite22. As shown in Fig. 4(b), three peaks (P1, P2, and P3) are observed around r (interatomic distance) = 1.9, 2.8, and 3.8 Å for all samples. These peaks are attributed to two O shells and one P shell nearest to the Sr22, 34, 35. The EXAFS spectrum obtained for ST-HAP-Sr100 differs from that of the other samples as the P1 peak shifts to a longer r, whereas P2 shifts to a shorter r when ST-HAP is exposed to 100 mmol/L of Sr2+; moreover, the peak intensity of P3 decreases. The spectral line shapes of DEF-HAP are similar irrespective of the Sr2+ concentration.

Figure 4
figure 4

Sr K-edge EXAFS analysis of the HAP samples. (a) EXAFS spectra of DEF-HAP and ST-HAP exposed to 10 and 100 mmol/L of Sr2+, where k is the wavenumber. (b) Fourier transforms of the EXAFS spectra where r is the interatomic distance. Notably, the r does not take phase shifts into account. Transforms were conducted with k 2 weighting. Three peaks (P1, P2, and P3) are observed at around r = 1.9, 2.8, and 3.8 Å in all samples.

Discussion

Sorption tests for HAPs in the presence of competing ions showed that the A eff of both DEF-HAP and ST-HAP is affected less by Mg2+ than by Ca2+ (Fig. 2). Divalent cations can be exchanged with Ca2+ in crystals; moreover, ions with hydrated radii in the range of 0.9–1.3 Å are easily exchanged as compared with other ions36. Since the hydrated ionic radius of Sr2+ is 1.2 Å23, sorption of Sr2+ might occur prior to that of Mg2+, which has a hydrated ionic radius of 0.6 Å23. With Ca2+ as the competing ion, DEF-HAP maintains its A eff for a specific range of concentrations. This means the Sr2+ sorption capacity of DEF-HAP is equal to or greater than its Ca2+ sorption capacity. The result implies that the vacant Ca2+ sites in DEF-HAP are slightly larger than the ionic radius of Ca2+, making them more suitable for Sr2+ than for Ca2+.

The A eff of DEF-HAP is higher than that of ST-HAP at all the tested concentrations, as shown in Fig. 3. The sorption of divalent cations onto HAPs occurs both at the surface and within the crystals21. In aqueous solutions, strontium mainly exists as Sr2+ at pH 737, and these positively charged species are actively adsorbed at the negatively charged sites in the HAPs by electrostatic attraction. Thus, the higher negative charge on the surface of DEF-HAP (zeta potential of −20.1 mV) compared with that on the surface of ST-HAP (zeta potential of −5.46 mV) contributes to its higher absorptivity of Sr2+. The higher negative zeta potential of DEF-HAP might be due to the existence of PO4 3−, OH, or HPO4 surface ions; similar observations have also been made for other Ca-deficient hydroxyapatites38, 39. These results suggest that DEF-HAP has a higher Sr2+-selective sorption capacity owing to its negatively charged surface and the vacant Ca2+ sites within its crystal structure.

In Fig. 4(b), EXAFS spectra of DEF-HAP for different Sr2+ concentrations appear to be independent of concentration, suggesting that the sorption sites of DEF-HAP remain unchanged for all the studied conditions; however, the ST-HAP spectra vary with the Sr2+ concentration. In the spectrum of ST-HAP-Sr100, P1, which is related to the O shell nearest to the Sr, shifts to a larger r, indicating that there are two types of sites providing shorter and longer Sr–O distances in ST-HAP. This implies that one site providing a short Sr–O distance is occupied at a certain Sr2+ concentration below 100 mmol/L; moreover, the other site providing long Sr–O distance becomes more dominant in ST-HAP. The spectra for DEF-HAP suggest that DEF-HAP only contains sites that provide a shorter Sr–O distance, in which O strongly interacts with Sr2+. This might contribute to the higher value of q max observed in Fig. 3.

In conclusion, Sr2+ sorption tests in the presence of competing cations demonstrate that DEF-HAP has a higher Sr2+ sorption selectivity than ST-HAP even for conditions similar to those in the groundwater at FNPP. The difference in sorptivity for the competing cations indicates that the ionic radius is one of the important factors affecting the sorption of ions onto HAPs and that HAPs are efficient sorbents for Sr2+, which has a hydrated ionic radius of 1.2 Å. Furthermore, DEF-HAP could maintain its Sr2+ sorption capacity even in the presence of Ca2+. Compared with ST-HAP, DEF-HAP showed a higher A eff for solutions with various Sr2+ concentrations. The EXAFS results demonstrate that compared with ST-HAP, DEF-HAP has significantly more sorption sites where Sr2+ can be stably and preferentially sorbed. These results provide useful insights that may assist the future development of HAP-based sorption materials for 90Sr by controlling their chemical composition.

Methods

Materials

Stoichiometric hydroxyapatite (Ca10(PO4)6(OH)2, Ca/P = 1.68, ST-HAP) and Ca-deficient hydroxyapatite (Ca10−x(HPO4)x(PO4)6−x(OH)2−x, Ca/P = 1.38, DEF-HAP) were supplied by Taihei Chemical Industrial Co., Ltd., Japan. ST-HAP was synthesized by mixing calcium hydroxide and phosphoric acid solutions40, 41. The precipitate was dried at 100 °C. DEF-HAP was synthesized by mixing slurries of calcium hydrogen phosphate dihydrate with an aqueous solution of sodium hydroxide31, 41. The precipitate was dried at 150 °C. Strontium chloride (SrCl2), magnesium chloride (MgCl2), and calcium chloride (CaCl2) were purchased from Wako Pure Chemical Industries Co., Ltd., Japan. Sr standard solutions (1 mg/mL) for ICP-MS (PerkinElmer Japan Co., Ltd., United States) were also purchased from Wako Pure Chemical Industries. All materials were used without further purification.

Characterizations

XRD measurements were performed using an X-ray diffractometer (Ultima IV, Rigaku Co., Ltd. Japan) with Cu-K α radiation (λ = 0.15418 nm) at room temperature. The sample was mounted on a glass plate and optically centered on the diffractometer. The diffraction data was collected in steps of 0.02° in for 2θ angles between 10° and 70°. TEM images of the HAP samples were obtained using a JEOL TEM (HT-7700, Hitachi, Tokyo, Japan) operated at an electron beam accelerating voltage of 100 kV. The zeta potentials of the HAP samples were measured at room temperature using a Malvern Zetasizer (Marvern Instruments Ltd., Worcestershire, UK). The unit automatically calculates the electrophoretic mobility of the particle and converts the electrophoretic mobility into a zeta potential using the Smoluchowski equation. One milligram of each HAP sample was dispersed in 1 mL of MilliQ water at pH 7 and stirred for 5 h. Each data point in the A eff plots is an average of approximately three measurements. The pH of the suspension was adjusted using an aqueous solution of 0.1 mmol/L hydrochloric acid (HCl). The pH of the suspension was measured using a digital pH meter (Mettler Toledo Co., Ltd., Switzerland).

Sorption tests

The effect of competing ions on the Sr2+ sorption capacities of DEF-HAP and ST-HAP was investigated using SrCl2, CaCl2, and MgCl2 solutions. We stirred 7 mL of aqueous solutions containing 35 mg of DEF-HAP or ST-HAP and Sr2+ (0.05 mmol/L) in MilliQ water for 5 h at room temperature in the presence of Ca2+ or Mg2+ at concentrations of 0.1, 0, 1, 10, or 50 mmol/L. The initial pH was adjusted to 7 using 0.1 mmol/L of HCl. Aliquots (1 mL) of the suspension were taken after 5 h and were centrifuged at 10,000 g for 5 min. The pH was maintained at 7 for this study since the groundwater at FNPP generally has a pH of 7. The supernatant was filtered through a 0.1-mm polyvinylidene difluoride (PVDF) filter (Millex-GV, Millipore Co., United States). The concentrations of Sr2+ in the supernatant were measured via ICP-MS (PerkinElmer Japan Co., Ltd., Japan).

To investigate the dependence of Sr2+ concentration on the sorption efficiency, 35 mg of DEF-HAP or ST-HAP was mixed with Sr2+ at concentrations of 0.01, 0.05, 0.1, 1, 5, 10, or 100 mmol/L for 5 h at room temperature. The supernatant was obtained after 5 h, and the sorption efficiency was estimated as aforementioned.

EXAFS measurements

EXAFS measurements were performed at the Sr K-edge on the BL14B1 line of SPring-842, 43, where X-rays (~16,100 eV) were continuously generated from a bending magnet source. The incident X-rays were monochromatized by two silicon (311) crystals. For the measurements, powder samples were pressed into pellets and placed in the X-ray path. The absorption data were analyzed using the ATENA software package44.