Introduction

Under the international situation of energy crisis, nuclear power as a sustainable energy source has become one of the most major energy sources in the world. With the peaceful utilization of nuclear energy, it has been an essential task to dispose safely various radionuclides [1, 2].

Uranium, a chemical homolog of hexavalent actinide, and thorium, a tetravalent actinide, are rather important materials in nuclear industry. These dangerous radionuclides are easily mobile into the water, soil and air systems, and then continuously accumulate because of extreme long half-life. Once inside the living bodies through the transmission of food chain, it will give rise to many irradiation damages, including skin corrosion, toxic hepatitis, kidney injury and even cancer. For the sake of environmental protection and human health, it is of paramount importance to remove of U(VI) and Th(IV) from radioactive waste.

Up to now, various techniques including ion-exchange, chemical precipitation, adsorption, filtration and solvent extraction, have been used to remove radionuclides [3,4,5,6]. Among all these methods, adsorption has been proved to more excellent than others owing to its ease of operation, low cost, high efficiency and wide applications [7].

Recently, it was found that the great effort was made in the adsorption of radionuclides on clay minerals because of widespread in many countries, high adsorption capacity and high selectivity of radionuclides [8]. A number of studies about the adsorption of radionuclides on montmorillonite [9], sepiolite [10] and illite [11] have been carried out. Apart from what is mentioned above, as a type of natural aluminosilicate clay mineral, halloysite has received much attention due to its tubular porous structure and high surface area. Meanwhile, halloysite is very widespread in china, rich in Jiangxi, Sichuan, Hunan and Guangdong. Similar to kaolinite, halloysite is also a superposition of one silicon-oxygen tetrahedron sheet and one aluminium-oxygen octahedron sheet with 1:1 stoichiometric ratio [12]. The distinction between kaolinite and halloysite is the existence of more interlayer water in halloysite. Tubular halloysite has a highly unusual meso/macroscopic superstructure and cylindrical structure rolled into by 10–15 layers of alumina and silica bimolecular layers. The chemical formula can be shown as Al2Si2O5(OH)4·nH2O [13, 14]. In each halloysite nanotube, the internal lumen surface consists of aluminol (Al–OH) groups, whereas the external surface is mainly composed of siloxane Si–O–Si groups. The unit layers are separated by amount of water molecules. One side of the interlayer are Si–O–Si groups, the other side are Al–OH groups. There are also some Al–OH and Si–OH groups on the edges of the nanotube [15]. Its schematic diagram and chemical structure are shown in Fig. 1 [16, 17]. An outstanding feature is a layered and hollow tubular structure, similar to the carbon nanotube. Thanks to some excellent characteristics, e.g. abundant resources, low cost, hollow and layered structure, high surface area and high ions exchange capacity, halloysite nanotubes have been studied widely as an adsorbent. Previous studies reported that halloysite nanotubes were used to remove Zn(II) [18], Co(II) [4] and U(VI) [19]. Therefore, halloysite nanotubes also have been considered as a promising adsorbent. However, little attention has been devoted to studying the adsorption of Th(IV) and U(VI) on halloysite in detail, especially the simultaneous removal of U(VI) and Th(IV) ions by halloysite nanotubes from aqueous solution.

Fig. 1
figure 1

Scheme of the halloysite nanotube (left) and its chemical structure (right)

Full size image

This paper focused on study the feasibility of the individual and simultaneous removal of U(VI) and Th(IV) ions by halloysite nanotubes from aqueous solution. The halloysite nanotubes were characterized by potentiometric titration, N2-BET, TEM, XRD, XRF, FT-IR and XPS. The effect of different conditions on adsorption was studied, such as contact time, adsorbent dosage, pH, ionic strength and temperature, and the adsorption kinetic and thermodynamics models were determined. Furthermore, XRD, FT-IR and XPS were employed to discuss the adsorption mechanisms of Th(IV) or U(VI) or Th(IV)/U(VI) onto halloysite nanotubes. In addition, the selective desorption of Th(IV)/U(VI) from halloysite nanotubes was studied in detail.

Experimental

Materials

The halloysite nanotubes were obtained from Henan province (China). The halloysite nanotubes were treated with dilute HCl, followed by rinsing with deionized water and drying at 40 °C. Then small-size particles with larger surface area were gained after milling and sieving through 200-mesh screen. The stock solution of Th(NO3)4 and UO2(NO3)2 were prepared by dissolving Th(NO3)4·5H2O and UO2(NO3)2·6H2O in deionized water. All other reagents were analytic grade.

Characterization

The phase composition of halloysite nanotubes was identified by powder X-ray diffraction (XRD, Philips XTert PRO) with a Cu/Kα radiation operated at 40 kV and 60 mA. X-ray Fluorescence (XRF, Xenemetrix X-Calibur) also was used to measure the chemical composition of halloysite nanotubes. Fourier transform infrared spectra (FT-IR; Bruker Vertex70v) of halloysite nanotubes was recorded between 4000 and 400 cm−1 using KBr pellet. The BET (Bechman Coulter SA3100) nitrogen surface area was measured at 393 K for 4 h. The morphology of halloysite nanotubes was studied using transmission electron microscope (TEM; Hitachi S-4802). The point of zero charge (pHpzc) was obtained by the potentiometric titration [20].

Batch adsorption and desorption experiments

The experimental procedures of batch sorption experiments were identical to those already applied [21]. Batch adsorption experiment was carried out using a temperature controlled orbital shaker in polyethylene test tubes sealed with screwcaps. The pH value of initial solution was adjusted by adding 0.1 or 0.01 mol/L HNO3 or NaOH solution.

There are two approaches for desorption experiment. (I). At the end of adsorption experiments, half of the supernatant was pipetted and the equal volume of background electrolyte solution (0.1 mol/L) with same pH was added, then the mixture was shaken and centrifuged at the same conditions as in adsorption experiments; (II). The desorption agent solution was mixed with the dried halloysite nanotubes loaded with Th(IV) and U(VI), then the mixture was shaken for 24 h.

The concentration of Th(IV) or U(VI) was measured by spectrophotometry using arsenazo(III) complexing agent. All experimental data were averages of duplicate measurements with relative errors of adsorption percentage less than 5%. The adsorption percentage (% = (C0 − Ce)/C0 × 100%) and solid phase concentration (q = (C0 − Ce) ×  V/m) were calculated on the basis of the difference between the initial concentration (C0) and equilibrium concentration (Ce), adsorbent mass (m) and the solution volume (V) [21].

Results and discussion

Characterization of the halloysite nanotubes

The TEM image (Fig. 2a) showed the morphological features of halloysite nanotubes: it has a hollow and open-ended tubular morphology. The diffraction peaks in XRD pattern (Fig. 2b) indicated that the phase composition was primarily halloysite nanotubes (2θ = 11.8°, 20.0°, 24.7°, 35.0°, 54.5°, and 62.6° indexed by JCPDS No.29-1487), meanwhile contains minor quartz (26.7°) as impurities [22], which can be further supported by the XRF results (Table 1). It was in good agreement with previously published pattern for halloysite [23]. The interlayer distances of halloysite nanotubes was calculated from the Bragg’s law: nλ = 2dsinθ, where n is an integer, λ is the wavelength of incident wave, d is the interlayer distance, and θ is the diffraction angle [24]. The interlayer distances d is 7.40 Å by calculation, therefore n equals 1 in Al2Si2O5(OH)4.nH2O [15].

Fig. 2
figure 2

a TEM image of halloysite nanotubes, b XRD of halloysite nanotubes, c FTIR of halloysite nanotubes, d pore size distribution of halloysite, e N2 adsorption–desorption isotherms of halloysite nanotubes, f acid–base potentiometric titration plots of halloysite nanotubes

Full size image
Table 1 Chemical composition of the halloysite nanotubes obtained by XRF
Full size table

FT-IR spectroscopy can be used to characterize the functional groups of halloysite nanotubes, as shown in Fig. 2c. The band at 3697 cm−1 was attributed to the outer-surface hydroxyl groups of Si–OH stretching vibrations. The characteristic band at 3621 cm−1 was ascribed to the inner-surface hydroxyl groups of Al–OH stretching vibrations [25]. The broad band at 3443 cm−1 arose from O–H stretching of adsorbed water [26]. 1638 cm−1 was the bending vibration of water O–H [27]. The band at 1094 cm−1 was assigned to Mg-O stretching vibration. The strong band at 1032 cm−1 corresponded to the in-plane Si–O stretching vibration (Si–O–Si and O–Si–O), and Si–O perpendicular stretching appeared at 750, 690 cm−1 [26]. The strong band at 910 cm−1 was ascribed to Al–O groups. The peaks at 537, 470 cm−1 were ascribed to band of Al–O–Si [25]. The results were consistent with chemical structure of halloysite nanotubes [17].

Figure 2d showed the distribution of the pore diameter of halloysite nanotubes and average pore size was 20.25 nm. The N2 adsorption–desorption isotherms of halloysite nanotubes was shown in Fig. 2e. The adsorption–desorption isotherm of halloysite nanotubes exhibited a type IV shape based on IUPAC classification [28] with an obvious hysteresis loops, and a drastic upturn at high pressure region and no saturation plateau, which explained the presence of mesopores(2–50 nm) in the structure, even a few macrospores(> 50 nm). The BET nitrogen surface area of halloysite nanotubes is 55.65 m2/g. Obviously, it is beneficial to the adsorption and storage of certain ions [29].

Potentiometric titration was employed to determine the surface charge of halloysite nanotubes. The surface charge will greatly influence the sorption affinity of halloysite halloysite at various pH values. The value of ΔQH(the proton excess on solid surface) as a function of pH was presented in Fig. 1f. The pHpzc (point of zero charge) was measured to be 4.18 for the natural halloysite. It indicated that the surface is positively charged at pH < 4.18 and negatively charged at pH > 4.18.

Effect of contact time (adsorption kinetics)

Figure 3a showed the adsorption of U (VI) and Th(IV) as a function of time on the halloysite nanotubes. Th(IV) and U(VI) ions removal was faster during the initial stage due to a large number of vacant surface sorption sites and hollow mesopores. With the sites being gradually occupied, the repulsive forces resulted in slowing down of adsorption. To obtain complete equilibrium state, contact time of 40 h was selected in the following experiments.

Fig. 3
figure 3

a Influence of the contact time on the adsorption quantities and adsorption percent of Th(IV) (black and blue diamonds) and U(VI) (black and blue circles) adsorbed by the halloysite nanotubes, pH = 4.0 ± 0.1, m/V = 5 g/L, \(C_{\text{Th(IV)}}^{0} = C_{\text{U(VI)}}^{0} = 2.5 \times 10^{ - 4} \,{\text{mol}}/{\text{L}}\), T = 298 ± 1 K; b Pseudo-second-order plot of Th(IV) and U(VI) ions adsorption onto the halloysite nanotubes; c Pseudo-first-order plot of Th(IV) and U(VI) adsorption on the halloysite nanotubes. (Color figure online)

Full size image

It was clear from Fig. 3a that the amount and percent of adsorbed Th(IV) was much larger than that of U(VI). The application of pseudo-first-order kinetic model [Eq. (1)] and pseudo-second-order model [Eq. (2)] kinetic models was useful to determine the mechanisms during the adsorption process.

$$\ln \left( {q_{\text{e}} - q_{t} } \right) = \ln q_{\text{e}} - k_{1} t$$
(1)
$$t/q_{t} = 1/\left( {k_{2} \times q_{\text{e}}^{2} } \right) + t/q_{\text{e}}$$
(2)

where qt and qe (mg/g) were adsorption capacity of on natural halloysite at t and equilibrium time, respectively. k1 (h−1) and k2 (g/(mg h)) represented the rate constant of pseudo-first-order and pseudo-second-order adsorption model.

The kinetic adsorption process of U(VI) and Th(IV) on natural halloysite can be well described by the pseudo-second-order model by correlation coefficient (R2 = 0.99 or 1.00) (shown in Table 2). This also suggested that the adsorption process of Th(IV) and U(VI) on the halloysite nanotubes was chemisorption [30].

Table 2 Parameters of pseudo-first-order and pseudo-second-order kinetic model for U(VI) and Th(IV) adsorption on the halloysite nanotubes
Full size table

Effect of adsorbent dosage

The effect of adsorbent dosage on the removal of U(VI) and Th(IV) ions was displayed in Fig. 4. It showed that the adsorption percent of uranyl and thorium ions increased with the increase of adsorbent dosage from 1 to 5 g/L. Taking the treatment cost into consideration, the optimal adsorbent dosage to remove U(VI) and Th(IV) ions was 5 g/L. The equilibrium data indicated that the capacity of the adsorption on the halloysite nanotubes was Th(IV) > U(VI). The result was explained by covalent index of metal and free energy of hydration, which can be calculated from the following equation [31, 32]:

Fig. 4
figure 4

Effect of halloysite nanotubes dosage on the removal of U(VI) and Th(IV). pH = 4.0 ± 0.1, I = 0.1 mo/L, T = 298 ± 1 K, \(C_{\text{Th(IV)}}^{0} = C_{\text{U(VI)}}^{0} = 1.5 \times 10^{ - 4} \,{\text{mol}}/{\text{L}}\), V = 6.0 ml

Full size image
$${\text{Covalent}}\,{\text{index}} = X_{\text{m}}^{2} r$$
(3)
$${\text{Free}}\,{\text{energy}}\,{\text{of}}\,{\text{hydration}} = - 69500Z_{\text{c}}^{2} /r$$
(4)

where Xm, Zc and r is electronegativity, cation charge and atomic radius, respectively. It is well known that the adsorptive capacity of adsorbent for metal with high covalent index is greater than that for metal with lower covalent index [31]. Covalent index for Th(IV) and U(VI) sorption was 303.86 and 297.09, respectively(shown in Table 3). Meanwhile, the free energy of hydration is another parameter for explanation of sorption process. The higher free energy of hydration for metal, the tendency to remain in liquid phase was greater [32]. Free energy of hydration values were − 6184.6 kJ mol−1 for Th(IV) and − 16,038.5 kJ mol−1 for U(VI) in Table 3. These values indicated that Th(IV) complexes tended to adsorb on halloysite nanotubes, while U(VI) complexes tended to remain in liquid phase. Therefore, the order of metal ion adsorbed on halloysite was: Th(IV) > U(VI). Based on the above reason, the adsorption capacity of Th(IV) on some adsorbents is stronger than that of U(VI) [33].

Table 3 Covalent index and Free energy of hydration for Th(IV) and U(VI) ions
Full size table

Effect of pH and ionic strength

The solution pH is an important factor for adsorption behavior of U(VI) and Th(IV) ions onto halloysite nanotubes. Therefore, the pH-dependent adsorption of U(VI) and Th(IV) was studied. The adsorption capacity or percentage of U(VI) and Th(IV) onto halloysite nanotubes as a function of pH at ionic strength 0.01, 0.10 and 0.50 mol/L NaNO3 was shown in Fig. 5a, b. For U(VI), it can be seen that there were three regions: at initial region, the adsorption capacity (or percent) increased gradually with pH increasing from approximately 2.0–4.2; the second region, adsorption increased steeply as a function of pH from approximately 4.2–5.2; and a final region, the adsorption slightly decreased in the pH rang of approximately 5.2–7.0. The observed adsorption trend was in relation to the surface charge of halloysite nanotubes, the dissociation of functional groups (Si–OH and Al–OH) and the distribution of hydrolyzed uranyl species in solution. As mentioned above, halloysite nanotubes surface had positive charges at pH < 4.18 (pHpzc) and negative charges at pH > 4.18. Uranium exists primarily as UO22+ ions (pH < 4), and uranium complexes such as UO2(OH)+, (UO2)2(OH) 2+2 , UO2(OH)2 and (UO2)3(OH) +5 are formed at pH > 4.0, while (UO2)2CO3(OH) −3 , (UO2)2(CO3) 2−2 and UO2(CO3) 4−3 can’t be neglected at pH > 5.5 [3]. Therefore, the surface positive charge of halloysite nanotubes was not favorable to adsorption of these positively-charged species because of electrostatic repulsion at pH < 4.18. In contrast, at pH > 4.18, the surface electronegativity increases and surface sites became more dissociated with increasing pH values, which was in favor of adsorption owing to electrostatic attraction and the enhancement of surface complexation ability. At pH > 5.5, (UO2)2CO3(OH) −3 , (UO2)2(CO3) 2−2 and UO2(CO3) 4−3 were difficult to be absorbed on the negatively-charged surface due to electrostatic repulsion, resulting in the slight decrease of uranium adsorption capacity (or percent). For Th(IV), it was also found the adsorption capacity (or percent) increased with pH increasing. Up to pH 3.5, the adsorption was gradually increased. At higher pH > 4.2, the Th(IV) adsorption become nearly constant. At pH < 4.2, thorium existed as positively-charged Th4+, Th(OH)3+, Th(OH) 2+2 and Th(OH) +3 in the aqueous solution [3, 21]. and the percent of hydrolysis species increased with pH increasing. However, the surface of halloysite nanotubes carried positive charges pH from 2.0 to 4.18. Therefore, it was unfavorable for thorium adsorption because of the repulsive electrostatic forces between the positively charged surface and the positively charged thorium species. In addition, according to the thermodynamic model of adsorption [34], the Gibbs free energy of adsorption (ΔGads) involved not only the coulombic electrostatic energy (ΔGcoul), but also the hydration energy (ΔGhyd) and the chemical bond energy (ΔGchem). As mentioned above, the coulombic electrostatic energy was adverse for thorium adsorption, and the dehydration energy was often positive and unfavorable to the adsorption. Therefore, the chemical bond energy was the sole driving force for thorium adsorption. That is to say, the strong chemical bonds between the surface of halloysite nanotubes and the hydrolysis thorium species were formed, and it overcame the sum of repulsive electrostatic force and dehydration energy. As also shown in Fig. 4a, b, for U(VI) adsorption on the halloysite nanotubes, three adsorption edges merged into a single one at three ionic strengths 0.01, 0.10 and 0.50 mol/L NaNO3, which suggested that the adsorption of U(VI) onto halloysite nanotubes was insensitive to ionic strength, while the adsorption of Th(IV) was greatly influenced by ionic strength at lower pH. Specifically, Th(IV) adsorption decreased with the increase of NaNO3 concentration. When Th(IV) was adsorbed by halloysite nanotubes, the presence of the competing cations (Na+) will reduce the adsorption, suggesting ionic interaction nature of the adsorption [35]. In addition, the activity coefficient of thorium ions was greatly influenced at higher ionic strength, which limited their migration to sorbent surfaces [1]. Furthermore, ionic strength increased has been testified to weaken repulsion among halloysite nanotubes, resulting in aggregation of halloysite nanotubes, decrease of available adsorption sites and decrease of adsorption percent [4]. Generally, outer-sphere surface complexation and cation exchange is impacted by ionic strength, while inner-sphere surface complexation is influenced by pH values obviously [18, 36]. Therefore, the adsorption of Th(IV) on halloysite nanotubes was the result of cation exchange combined with out-sphere surface complexation, while the adsorption mechanism of U(VI) was inner-sphere surface complexation.

Fig. 5
figure 5

a, b Effect of pH and ionic strength on Th(IV)/U(VI) adsorption capacity(percentage) onto the halloysite nanotubes, \(C_{\text{U(VI)}}^{0} = C_{\text{Th(IV)}}^{0} = 1.5 \times 10^{ - 4} \,{\text{mol}}/{\text{L}}\), T = 298 ± 1 K, m/V = 5 g/L. All error bars represent standard deviation from triplicate experiments

Full size image

In addition, as can be seen from Fig. 5a, the optimum pH for Th(IV) and U(VI) separation is 4.1–4.3.

The effect of temperature on adsorption

The temperature is an important parameter for the adsorption of radionuclides on adsorbent. The adsorption experiments of U(VI) or Th(IV) on halloysite nanotubes were carried out at 298, 318 and 338 K. The results were displayed in Fig. 6a, b. For U(VI) and Th(IV), the adsorption capacity increased with the temperature raising, indicating that high temperature was favored for the adsorption process. This is accountable by the increase of ions activity and the acceleration of ions diffusion [18, 36]. In addition, the increase of initial metal concentration enhanced the equilibrium adsorption. The initial concentration provided enough driving force for the transfer of metal ions from the solution to adsorbent. The increase of initial metal concentration also increased the interaction between the metal ions in the solution and absorbent. Therefore, the increase of initial metal concentration enhanced the equilibrium adsorption of U(VI)and Th(IV) ions [37]. It was also observed that the adsorption capacity of halloysite nanotubes for Th(IV) was greater than that for U(VI).

Fig. 6
figure 6

Adsorption of a Th(IV) and b U(VI) on the halloysite nanotubes at different temperatures, T = 40 h, pH = 3.5 ± 0.1, I = 0.1 mol/L and m/V = 5 g/L; c, d Freundlich isotherm of Th(IV) and U(VI) on the halloysite nanotubes, respectively; e, f D–R of Th(IV) and U(VI) on the halloysite nanotubes, respectively

Full size image

In order to understand the adsorption process and adsorption mechanism of U(VI) and Th(IV) on halloysite nanotubes, adsorption data were analyzed by Langmuir, Freundlich and D–R models by Fig. 6c–f) and the results was shown in Table 4. From the correlation coefficients R2 obtained, Freundlich and D–R models were more suitable for describing the adsorption isotherms of U(VI) and Th(IV) on halloysite nanotubes. It can be seen from n value that the adsorption of thorium on halloysite nanotubes was favorable (n > 1), while that of uranium on halloysite nanotubes was unfavorable (n < 1). In the meantime, the value for the bonding energy was E > 8kJmol−1 (shown in Table 4), thus the mechanism involved was mainly chemical adsorption with physical adsorption which can be explained by the later adsorption–desorption experiments.

Table 4 Parameters of Langmuir, Freundlich and D–R model for U(VI) and Th(IV) of adsorption on the halloysite nanotubes
Full size table

The thermodynamic parameters, such as Gibbs free energy (ΔG0), enthalpy (ΔH0) and entropy (ΔS0), were calculated from the following equation:

$$\ln K_{\text{d}} = \frac{{\Delta S^{0} }}{R} - \frac{{\Delta H^{0} }}{RT}$$
(5)
$$\Delta G^{0} = \Delta H^{0} - T\Delta S^{0}$$
(6)

Relevant data calculated from Eqs. (5, 6) were listed in Table 5. The results showed that the adsorption process of U(VI) and Th(IV) on halloysite nanotubes was an endothermic(ΔH0 > 0) and spontaneous process (ΔG0 < 0) under the experimental conditions. Meanwhile, positive values of ΔS0 suggested the increase in randomness at the solid–liquid interface during adsorption. In short, the adsorption of Th(IV) and U(VI) on halloysite nanotubes was an endothermic and spontaneous process.

Table 5 Thermodynamic parameter of Th(IV) and U(VI) adsorption onto the halloysite nanotubes
Full size table

Competitive adsorption of uranium (VI) and thorium (IV) ions

Generally, the competitive adsorption behavior of binary-component on the adsorbent may exhibit three types: synergism (the adsorption amount increases when the other component exists), antagonism (the adsorption amount decreases with the addition of the other component) and non-interaction (the binary-component adsorbates have no mutual effect on the adsorption) [3, 38]. The competitive adsorption was carried out by the form of Th(IV)/U(VI) and U(VI)/Th(IV), respectively. This is to say the former metal with a constant concentration and the latter metal with varied concentration. The results of the experiments were shown Fig. 7. Figure 7a showed that the initial Th(IV) concentration remained constant and U(VI) concentration was varied. The amounts of Th(VI) adsorbed on halloysite nanotubes exhibited slight decrease when U(VI) concentration was increased. Such small decrease indicated that U(VI) competed weakly with Th(IV); Fig. 7b showed that initial U(VI) concentration was constant and Th(IV) concentration was increased. The amounts of Th(IV) adsorbed on halloysite nanotubes increased while the amount of uranium on halloysite nanotubes decreased when the Th(IV) concentration was increased. The trend suggested that Th(IV) competed strongly with U(VI). Especially, the amounts of U(VI) adsorbed on halloysite nanotubes was not zero as Th(IV) concentration increased. These results can also be explained by the previous covalent index of metal free energy of hydration The phenomenon also showed that halloysite nanotubes may possess some adsorption sites selectively adsorbed U(VI) even in higher Th(IV) concentration. In addition, previous experiments have shown that the adsorption amount of Th(IV) in single Th(IV) adsorption system on halloysite nanotubes was much more than that the adsorption amount of U(VI) in single U(VI) adsorption system on halloysite nanotubes. Based on above, it is believed that the mutual effect of the binary-component of uranium (VI) and thorium (IV) ions was antagonistic. In the meanwhile, U(VI) and Th(IV) systems with the same concentration was studied. The equilibrium adsorption data obtained from the binary-component system also suggested that the adsorption capacity of halloysite nanotubes for Th(IV) is better than that for U(VI), which is in good agreement with the single-component system.

Fig. 7
figure 7

Adsorption amount of Th(IV) (blue circles) and U(VI)(red circles) on the halloysite nanotubes at the constant initial concentration of Th(IV) (a) and adsorption amount of Th(IV) (blue circles) and U(VI)(red circles) on the halloysite nanotubes at the constant initial concentration of U(VI), (b) m/V = 5 g/L, T = 298 ± 1 K, pH = 3.5 ± 0.1, I = 0.1 mol/L NaNO3. (Color figure online)

Full size image

Desorption of U(VI) and Th(IV) study

The desorption behaviors help to understand the interaction mechanism between U(VI) or Th(IV) and halloysite nanotubes. Desorption experiments can be divided into two categories: reversible desorption and specific desorption. Reversible desorption is mainly used to verify whether the adsorption reaction is reversible. Specific desorption is a process in which reagents are used to elute the adsorbed substances. Batch desorption experiments were performed and displayed in Fig. 8. As shown in Fig. 8a, the desorption isotherms of Th(IV) and U(VI) were obviously higher than that adsorption isotherms, respectively, which displayed that adsorption -desorption hysteresis has occurred and the adsorption was irreversible. The irreversible adsorption indicated that complete desorption cannot occur without intervention. Meanwhile, the irreversible adsorption also suggested that strong surface complexation of Th(IV) and U(VI) with halloysite nanotubes occurred. Figure 8b showed the desorption of Th(IV)/U(VI) from natural halloysite nanotubes by different desorption agents. The results displayed that these agents were suitable for desorption of U(VI) on halloysite nanotubes, only EDTA was suitable for desorption of Th(IV) on halloysite nanotubes (LogKU-EDTA = 13.7, LogKTh-EDTA = 26.6) [39]. The separation of U(VI) and Th(IV) was achieved by a two-step continuous desorption process (desorption agent may be Na2CO3/EDTA).

Fig. 8
figure 8

a Adsorption and desorption isotherms of U(VI) and Th(IV) ions on the halloysite nanotubes at T = 298 ± 1 K, m/V = 5 g/L, I = 0.1 mol/L. b Desorption of U(VI) and Th(IV) ions on the halloysite nanotubes by different desorption agents, Cdesorption agenT = 0.1 mol/L. Error bars represent standard deviation from 3 experiments

Full size image

XRD, FTIR and XPS

The spectroscopic techniques, such as XRD, FTIR, and XPS can help us to obtain reliable interaction mechanisms between clay and metal. XRD spectra were employed to obtain the interlayer distance data of halloysite nanotubes and halloysite nanotubes adsorbed U(VI) or (and) Th(IV) (shown in Fig. 9). The interlayer distances calculated from XRD spectra were 7.44 Å, 7.45 Å 7.49 Å and 7.50 Å for halloysite, U(VI)-halloysite, Th(IV)-halloysite, and Th(IV)-U(VI)-halloysite, respectively. The addition of U(VI) or (and) Th(IV) induced different increase of halloysite nanotubes interlayer distance and the addition of Th(IV) led to a more significant increase. These results indicated that the interlayer sites of halloysite nanotubes mainly fit for the adsorption of Th(IV), while U(VI) formed surface complexes by edge sharing with [Al(O,OH)6] octahedra and/or [Si(O,OH)4] tetrahedral sites of halloysite. And this phenomenon can explain the higher Gibbs free energy of U(VI) adsorbed on halloysite nanotubes than that of Th(IV) adsorbed on halloysite nanotubes (shown in Table 5). The results are similar to Th(IV)/U(VI) adsorbed on graphene oxide [40].

Fig. 9
figure 9

X-Ray diffraction patterns of halloysite nanotubes before and after U(VI) and Th(IV) adsorption. C0 = 1.5 × 10−4 mol/L, I = 0.1 mol/L, m/V = 5 g/L, pH = 4.0 ± 0.1, T = 298 ± 1 K

Full size image

Figure 10 showed the FTIR spectrum of halloysite nanotubes before and after U(VI) or Th(IV) adsorption. As can be seen from Fig. 10, the weak peaks at 1383 cm−1 showed that NO3− groups were co-adsorbed as the counterions or coordination ligands on halloysite nanotubes [39]. The intensity of 3621 cm−1 (Al–OH) and 1638 cm−1 (H–OH) peaks weakened, which indicated that Th(IV) or U(VI) may mainly interact with interlayer Al–OH by coordination or electrostatic interaction.

Fig. 10
figure 10

FTIR spectra of halloysite nanotubes before and after Th(IV) and U(VI) adsorption. C0 = 1.5 × 10−4 mol/L, I = 0.1 mol/L, m/V = 5 g/L, pH = 4.0 ± 0.1, T = 298 ± 1 K

Full size image

XPS can be used to analyze the change in the chemical composition and structure of halloysite after adsorption of U(VI) and Th(IV). Figure 11 showed the XPS spectra of halloysite nanotubes (including adsorption of U(VI) and Th(IV)). Peaks corresponded to signals in Table 6. The binding energy for Al2p and Si2p of halloysite nanotubes increased after adsorption U(VI) and Th(IV), which was attributed to decrease of extranuclear electron cloud density surrounding the Al(Si) atom [39]. Therefore, The result of adsorption may be the formation of covalent Si–O–Th(U) or Al–O–Th(U) bonds. Figure 12 showed high-resolution XPS spectra of O1s peaks before and after U(VI) and Th(IV) adsorption on halloysite nanotubes. Meanwhile, Table 7 listed the fitting results of curve fitting of O1s XPS spectra before and after U(VI) and Th(IV) adsorption on halloysite nanotubes. From Fig. 12 and Table 7, changes of peak position and ratio (%) of O–U or (and) O–Th were observed in the XPS spectra of halloysite-U(VI), halloysite-Th(IV), and halloysite-U(VI)-Th(IV). The peak position of O–H, O–U(VI) and O–Th(IV) didn’t change after adsorption, while O–H, O–U(VI) and O-Th(IV) peak ratio has changed after adsorption. Especially, simultaneous adsorbing U(VI) and Th(IV), O–Th of XPS peak ratio slightly increased compared with only adsorbing Th(IV), while O–U of XPS peak ratio was slightly less compared with only adsorbing U(VI), which illustrated that Th(IV) has a much stronger affinity for halloysite nanotubes than U(VI).

Fig. 11
figure 11

XPS wide scan spectra of the halloysite nanotubes the before and after U(VI) and Th(IV) adsorption. C0 = 1.5 × 10−4 mol/L, I = 0.1 mol/L, m/V = 5 g/L, pH = 4.0 ± 0.1, T = 298 ± 1 K

Full size image
Table 6 Electron binding energy of elements on the halloysite nanotubes of XPS spectra
Full size table
Fig. 12
figure 12

High-resolution XPS spectra of O1s peaks before and after U(VI) and Th(IV) adsorption on the halloysite nanotubes. a Halloysite, b U(VI)-halloysite, c Th(IV)-halloysite, d Th(VI)-U(VI)-halloysite, C0 = 1.5 × 10−4 mol/L, m/V = 5 g/L, pH = 4.0 ± 0.1, T = 298 ± 1 K

Full size image
Table 7 Curve fitting results of XPS spectra of O1s peaks before and after U(VI) and Th(IV) adsorption on the halloysite nanotubes
Full size table

Figure 13 showed high-resolution XPS spectra of the adsorption of U(VI) and Th(IV) on halloysite nanotubes under the condition of absorbing U(VI) and Th(IV) respectively and simultaneously. As can be seen from Fig. 13 and Table 6, the changes of peak position of U4f or (and) Th4f were observed. Especially, simultaneous adsorption of U(VI) and Th(IV), U4f5/2 and U4f7/2 of XPS peak position shifted lower compared with only adsorbing U(VI), while Th4f5/2 and Th4f7/2 of XPS peak position shifted higher compared with only adsorbing U(VI), which illustrated that Th(IV) has a stronger affinity for halloysite nanotubes than U(VI). It is well known, the adsorption capacity of metal ions on mineral surface is related to the binding energy of the inner electrons of the adsorbed ions. The higher the binding energy of the inner electrons of metal ions, the higher the affinity of ions to minerals, and the stronger the adsorption capacity of minerals to the metal ions. Therefore, the higher the binding energy of 4f electrons of Th(IV) ions adsorbed on the surface of minerals, the higher the affinity of Th(IV) ions with this mineral, and the stronger the adsorption capacity of minerals to Th(IV) ions.

Fig. 13
figure 13

High-resolution XPS spectra of the adsorption of U(VI) and Th(IV) on the halloysite nanotubes. a High-resolution scans of U4f after U(VI) adsorption, b high-resolution scans of Th4f after Th(IV) adsorption, c, d high-resolution scans of U4f and Th(IV) after U(VI) and Th(IV) adsorption, C0 = 1.5 × 10−4 mol/L, I = 0.1 mol/L, m/V = 5 g/L, pH = 4.0 ± 0.1, T = 298 ± 1 K

Full size image

In a word, Th(IV) has a considerable complexation capacity for halloysite nanotubes compared with U(VI). When uranium and thorium are simultaneously adsorbed by halloysite nanotubes, the adsorption mechanism of thorium on halloysite nanotubes was most mainly based on the complex by the interlayer Al–OH sites of halloysite and edge sharing with [Al(O,OH)6] octahedra and/or [Si(O,OH)4] tetrahedral sites, while that of uranium was only based on the complex by edge sharing with [Al(O,OH)6] octahedra and/or [Si(O,OH)4] tetrahedral sites.

Conclusions

In summary, the adsorption of U(VI) and Th(IV) on halloysite nanotubes was conducted under various physicochemical conditions. The rate of adsorption of U(VI) and Th(IV) to the halloysite nanotubes conformed well to the pseudo-second order kinetics model. The adsorption process of U(VI) and Th(IV) ions from aqueous solution by the halloysite nanotubes was an endothermic and spontaneous behavior and in good agreement with Freundlich and D–R adsorption isotherm. The influence of ionic strength on the adsorption of thorium on halloysite was much greater than that on uranium. By XRD, FT-IR and XPS analysis, the adsorption mechanism of thorium on halloysite nanotubes was most probably based on the formation of outer-sphere complex by the interlayer Al–OH sites of halloysite, while that of uranium was based on the formation of inner-sphere complex by edge sharing with [Al(O,OH)6] octahedra and/or [Si(O,OH)4] tetrahedral sites. The higher adsorption efficiency of the natural halloysite for Th(IV) could be utilized for selective separation of Th(IV) from U(VI) from aqueous.