Thorium(IV) adsorption onto multilayered Ti₃C₂T_x MXene: a batch, X-ray diffraction and EXAFS combined study

The interlayer regulation of layered environmental adsorption materials such as two-dimensional early transition metal carbides and carbonitrides (MXenes) plays an important role in their purification performance for specific pollutants. Here the enhanced uptake of Th^IV by multilayered titanium carbides (Ti₃C₂T_x) through a hydrated intercalation strategy is reported. Th^IV adsorption behaviors of three Ti₃C₂T_x samples with different c lattice parameters were studied as a function of contact time, pH, initial concentration, temperature and ion strength in batch experiments. The results indicated that the Th^IV uptake was pH and ionic strength dependent, and the adsorption process followed the pseudo-second-order kinetics and the heterogeneous isotherm (Freundlich) model. Thermodynamic data suggested that the adsorption process of all MXene samples was a spontaneous endothermic reaction. The dimethyl sulfoxide intercalated hydrated Ti₃C₂T_x featured the largest interlayer space and exhibited the highest Th^IV adsorption capacity (162 mg g⁻¹ at pH 3.4 or 112 mg g⁻¹ at pH 3.0), reflecting the significant increase in available adsorption sites from Ti₃C₂T_x interlayers. The adsorption mechanism has been clarified based on adsorption experiments and spectroscopic characterizations. An ion exchange process was proposed for the interaction between hydrated MXenes and Th^IV, where H⁺ from surface [Ti−O]⁻H⁺ groups were the primary active sites on Ti₃C₂T_x. Extended X-ray absorption fine structure (EXAFS) fitting results, in combination with X-ray diffraction (XRD) and Fourier transform infrared spectroscopy (FTIR) analyses, clearly indicated that Th^IV mainly formed the outer-sphere complexes on Ti₃C₂T_x surface through electrostatic interaction under strong acid conditions, while at pH > 3.0 the adsorption mechanism was determined by inner-sphere coordination and electrostatic interaction together.

1. Introduction

As a long-lived natural radioactive element, thorium is a potential nuclear fuel for the next-generation reactors (e.g. liquid fluoride thorium reactor) with advantages of large abundance, low cost, low risk of nuclear proliferation and almost no waste (Wang, Brown et al., 2018; Wang, Liu et al., 2018). In addition to the nuclear industry, thorium and its compounds are widely used in the applications of efficient catalysis, refractory materials, high-quality lenses and aviation alloy manufacturing (Rao et al., 2006). In terms of environmental concerns, a significant amount of thorium-containing waste has been discharged by ore processing in the rare earth industry, coal burning in thermal power plants and the use of chemical fertilizers. The radioactive nature and heavy metal toxicity of thorium may lead to great hazards for public health and ecosystems (Zhou et al., 2021). In aqueous solutions, thorium usually exists in the form of stable Th^IV, which can be deposited in bone, kidney and liver (Liu et al., 2019; Rao et al., 2006), thus causing permanent damage to the human body. In addition, Th^IV is often used as a chemical analog of other tetravalent actinides to simulate their migration and transformation behavior in the environment (Chen & Wang, 2007). The development of novel materials featuring high uptake capacity, good stability and excellent selectivity for Th^IV capture is urgently desired with regard to the effective utilization of thorium resources and radioactive pollution remediation.

Because of the large pre-concentration coefficient and facile operation, adsorption is one of the most attractive approaches for thorium extraction from aqueous solution. In the past few decades, a large number of solid-phase adsorbents such as metal oxide particles (Reiller et al., 2005; Rojo et al., 2009; Tan et al., 2007; Yusan et al., 2016), inorganic porous materials (Alahabadi et al., 2020; Sheng et al., 2008; Misaelides et al., 1995), functionalized mesoporous silica (Lebed et al., 2011; Yuan et al., 2014; Zhang, Ma et al., 2020), polymers (Yuan et al., 2019), organic–inorganic hybrids (Abbasizadeh et al., 2013; Kaygun & Akyil, 2007) and metal–organic framework (MOF) materials (Moghaddam et al., 2018; Zhang et al., 2017) have been widely examined to evaluate Th^IV separation from artificial and/or environmental wastewater. Apart from the above materials, the layered adsorbents provide numerous surface atoms and have unique structures with intralayer covalent bonding and interlayer van der Waals bonding (Wang et al., 2020), which allow the facile regulation of surface and interlayer properties to prepare multifunctional adsorbents for Th^IV elimination. Compared with the natural layered minerals (e.g. phlogopite, bentonite, illite and sepiolite) (Erden & Donat, 2017; Fralova et al., 2021; Wu et al., 2018; Zhang et al., 2016; Zhao, 2008), the emerging two-dimensional materials such as graphene oxide (Bai et al., 2014; Li, Yang et al., 2018) and layered sulfides (Li, Li et al., 2018; Xu et al., 2021) exhibit much higher thorium removal capacity and better selectivity under acidic conditions, which reflects the great potential of novel lamellar materials in radionuclide remediation.

Two-dimensional transition metal carbides and carbo­nitrides (MXenes) are a family of layered materials with hexagonal structure and the general chemical formula of M_n+1X_nT_x discovered in the past decade, where M represents an early transition metal, X represents C and/or N, T_x denotes surface termination groups, and n = 1–3 (Naguib et al., 2011, 2012). So far, MXene has been widely used in the application fields of catalysis, energy storage, sensing and environment (VahidMohammadi et al., 2021; Li et al., 2019), and the specific performance of this material largely depends on its chemical bonding and surface termination properties (Magnuson & Näslund, 2020; Naslund et al., 2021). Benefiting from the variable element composition, abundant oxygen-containing surface groups and good hydrophilicity, MXenes have proved to be promising environmental remediation materials in the field of water purification (Rasool et al., 2019; Ihsanullah, 2020; Chen et al., 2020). In terms of radionuclide elimination, various MXenes (e.g. Ti₃C₂T_x, V₂CT_x, T₂CT_x) and their composites and derivatives have been synthesized to investigate the corresponding adsorption behavior for Cs^I (Jun, Jang et al., 2020; Khan et al., 2019; Shahzad et al., 2020), Sr^II (Jun, Park et al., 2020), Ba^II (Mu et al., 2018), Pd^II (Mu et al., 2019), Eu^III (Zhang, Wang et al., 2020), U^VI (Wang, Song et al., 2018; Wang et al., 2016, 2021), Re^VII/Tc^VII (Wang et al., 2019) and I⁻ (Huang et al., 2020). For Th^IV, only one preliminary study has reported its adsorption on Ti₂CT_x MXene (Li et al., 2019), and results indicated that the adsorption had a high distribution coefficient and was independent of ionic strength. However, the easily degradable nature of Ti₂CT_x in environmental atmosphere and water greatly limits its environmental application. Moreover, the rational utilization of the spontaneous intercalation behavior of MXenes is of great importance, because the uptake performance may be significantly improved by weakening the interlayer interaction of multilayer structures. Herein, we report the adsorption of Th^IV on more stable Ti₃C₂T_x MXene as extended research. The effect of interlayer regulation of Ti₃C₂T_x on the adsorption behavior has been examined in detail by batch experiments. Furthermore, the underlying adsorption mechanism is also deciphered by EXAFS (extended X-ray absorption fine structure), XRD (X-ray diffraction) and FTIR (Fourier transform infrared) spectroscopy analyses.

2. Experimental

2.1. Reagents

All the chemicals used as-received in this study were analytical grade and obtained from Sinopharm Chemical Reagent Co. Ltd. A 10 g l⁻¹ thorium stock solution was prepared by dissolving an appropriate amount of thorium nitrate pentahydrate [Th(NO₃)₄·5H₂O] in deionized water acidified with nitric acid.

2.2. Preparation of Ti₃C₂T_x MXenes

The pristine and intercalated Ti₃C₂T_x samples were synthesized according to the literature (Wang et al., 2017). Briefly, a pristine Ti₃C₂T_x water suspension (Ti₃C₂T_x-H) was prepared by etching Ti₃AlC₂ with 15 wt% HF and washing with deionized water. Dry Ti₃C₂T_x sample (labeled as Ti₃C₂T_x-D) was obtained by filtering and vacuum drying Ti₃C₂T_x-H MXene at room temperature. The filtered Ti₃C₂T_x-H was dispersed in dimethyl sulfoxide (DMSO) for 2 days, and then washed with deionized water to prepare a water suspension of hydrated DMSO-intercalated Ti₃C₂T_x (labeled as Ti₃C₂T_x-DMSO-H). To avoid the slow oxidation, Ti₃C₂T_x-H and Ti₃C₂T_x-DMSO-H were stored in polypropylene bottles filled with Ar at 277 K (4°C) for later use. The mass concentrations of hydrated samples were measured by calculating the weight of MXene after filtering and drying the suspension with a fixed volume.

2.3. Batch adsorption experiments

A series of adsorption experiments for Th^IV uptake in aqueous solution by various Ti₃C₂T_x samples were performed as a function of contact time, pH, initial Th^IV concentration (C₀), temperature, ionic strength and other competing metal cations. The Th^IV stock solution was diluted to initial concentrations ranging from 5 to 150 mg l⁻¹ to carry out adsorption experiments by a batch method. Typically, 4 mg MXene was mixed with an appropriate amount of deionized water to keep the total volume of suspension as 5 ml in a beaker. Then 5 ml of solution containing Th^IV or multi-metal ions was added such that the concentration of sorbent was 0.4 g l⁻¹. The pH value of the solution was adjusted with small amounts of 0.1 M NaOH and 0.1 M HNO₃, and then the solution was stirred at room temperature for a specified time. The solid adsorbent was removed after Th^IV adsorption to obtain the supernatant sample by using a polyethersulfone syringe filter (0.22 µm, ANPEL Scientific Instrument Co. Ltd, Shanghai). The supernatant was diluted with 5 wt% HNO₃ before the final concentration determination. For the ionic strength experiment, NaClO₄ (0.1–500 mmol l⁻¹) was selected as representative electrolytes. In the selective adsorption test, all initial concentrations of Th^IV and other competing metal ions, including Co²⁺, Ni²⁺, Zn²⁺, Sr²⁺, La³⁺, Nd³⁺, Sm³⁺, Gd³⁺ and Yb³⁺, were 0.5 mmol l⁻¹. The residual concentrations of thorium and other metal elements were determined by an inductively coupled plasma optical emission spectrograph (ICP-OES, Horiba JY2000-2, France).

For the above batch adsorption experiments, the uptake capacity Q_e (mg g⁻¹) and distribution coefficient K_d (ml g⁻¹) were calculated using the following equations:

where C₀ and C_e are the initial and final equilibrium concentration of cations, respectively, and V and m are the volume of solution and the mass amount of solid sorbent in the batch adsorption tests, respectively. All the adsorption experiments have been carried out at least twice, and the uncertainty of measurement is less than 5%.

2.4. Material characterization

The morphologies and chemical compositions of the MXene samples were analyzed by a Hitachi S-4800 field-emission scanning electron microscope equipped with Horiba 7593-H energy-dispersive X-ray spectroscopy (EDS). A Bruker D8 Advance diffractometer (Cu Kα radiation, λ = 1.5406 Å) was used to collect powder XRD patterns. The step size for XRD pattern scanning was 0.02°. A Bruker Tensor 27 spectrometer was used to measure the FTIR spectra of samples by a potassium bromide pellet method.

The L₃ edge (16300 eV) EXAFS spectra of Th were collected at beamline 1W1B of the Beijing Synchrotron Radiation Facility (BSRF) using a silicon (111) double-crystal monochromator. Th(NO₃)₄·5H₂O (reference) was measured in transmission mode while other samples were measured in fluorescence mode using a Lytle-type ion chamber detector. EXAFS oscillation data were extracted, analyzed and fitted using computer programs Athena and Artemis, which were parts of the IFEFFIT program package (Ravel & Newville, 2005). A k range of ∼1.8−11.5 Å⁻¹ and a background frequency cutoff parameter (Rbkg) of 1.2 were used for acquisition of the Fourier transform of k³-weighted EXAFS data. To fit parameters such as coordination number (CN), atomic distance (R) and Debye–Waller factor (σ²), the theoretical phase and amplitude functions for the scattering pathways of Th–O and Th–N were calculated based on the crystal structures of Th(NO₃)₄·5H₂O. The fitting procedures were carried out in R space from 1.2 to 3.5 Å for Th(NO₃)₄·5H₂O and 1.2 to 3.0 Å for other samples with a fixed amplitude reduction factor (S₀²) of 0.97.

3. Results and discussion

3.1. Characterization

MXene samples of Ti₃C₂T_x-D, Ti₃C₂T_x-H and Ti₃C₂T_x-DMSO-H were prepared by drying, wetting and intercalating multilayered Ti₃C₂T_x. Their powder XRD patterns are shown in Fig. 1. The (002) peak of Ti₃C₂T_x-D is centered at 2θ of 8.8°, corresponding to a c lattice parameter (c-LP) of 20.0 Å. Because of the hydration effect and co-intercalation of H₂O/DMSO molecules inside the MXene interlayer, the (002) peak position shifts to 6.8° and 4.6° for Ti₃C₂T_x-H and Ti₃C₂T_x-DMSO-H, respectively, which results in much larger c-LPs (25.9 and 38.1 Å). Furthermore, the good reproducibility of intercalated MXene preparation has been proved using Ti₃C₂T_x-DMSO-H as a demonstration (Fig. S1 in the supporting information). The SEM (scanning electron microscopy) images in Fig. 2 indicate that Ti₃C₂T_x-D, Ti₃C₂T_x-H and Ti₃C₂T_x-DMSO-H have similar morphology, i.e. they feature micrometre-sized bulk particles composed of nano-lamellar structures. The above results are in good agreement with our previous study (Wang et al., 2017).

Figure 1 XRD patterns of the pristine dry, hydrated and DMSO-intercalated Ti3C2Tx.

Figure 2 SEM images and EDS results of different MXene samples before (a)–(c) and after (d)–(i) Th^IV adsorption. (a), (d), (g) Ti3C2Tx-D; (b), (e), (h) Ti3C2Tx-H; (c), (f), (i) Ti3C2Tx-DMSO-H.

3.2. Adsorption kinetics

The adsorption kinetics experiments of radioactive Th^IV onto the three MXene samples were carried out at a concentration of 100 mg l⁻¹ and pH 3.4, with the results shown in Fig. 3(a). The adsorption of Th^IV on Ti₃C₂T_x-D reached a plateau in 30 min with an uptake capacity of 37 mg g⁻¹. The hydrated and intercalated MXenes showed much higher Th^IV removal performance but slower kinetics, and the equilibration times for Ti₃C₂T_x-H and Ti₃C₂T_x-DMSO-H were 90 and 300 min, respectively. Previous literature (Wang et al., 2017) and XRD measurements in this study (see below, Section 3.7) have demonstrated that the adsorption of cations by dry MXene sample mainly occurs on the external surface because its interlayer space is too narrow, while the hydrated and intercalated MXene could provide a large amount of interlayer adsorption sites besides the exterior surface adsorption sites; therefore Ti₃C₂T_x-H and Ti₃C₂T_x-DMSO-H exhibited much larger Th^IV uptake capacities than Ti₃C₂T_x-D in this work. Our results suggest that the adsorption of Th^IV inside MXene interlayers corresponds to a slower internal diffusion process; this is because the `deep' adsorption sites in multilayer MXene particles may require sufficient diffusion of Th^IV in the confinement space.

Figure 3 Th^IV adsorption kinetics of MXene samples and linear fitting data with the pseudo-second-order kinetic model. (C0 = 100 mg l^−1, m/V = 0.4 g l^−1, pH = 3.4 ± 0.1.)

In order to clarify the adsorption process, the pseudo-second-order kinetic model was used to analyze the experimentally observed kinetic data and the linear fitting results are shown in Fig. 3(b). The linearized form of the model is given as follows:

where Q_e (mg g⁻¹) and Q_t (mg g⁻¹) are the quantities of the sorbed Th^IV at equilibrium time and time t, respectively, and k (g mg⁻¹ min⁻¹) is the pseudo-second-order adsorption rate constant. The model parameters and correlation coefficients obtained by the model are listed in Table 1.

Table 1
Kinetics model parameters and correlation coefficients for Th^IV adsorption on Ti₃C₂T_x-D, Ti₃C₂T_x-H and Ti₃C₂T_x-DMSO-H

	Pseudo-second-order model			Equilibrium capacity (mg g^−1)
	Qe (mg g^−1)	k (g mg^−1 min^−1)	R^2
Ti3C2Tx-D	37	0.0368	0.998	37
Ti3C2Tx-H	103	0.0074	0.999	102
Ti3C2Tx-DMSO-H	138	0.0020	0.999	138

It can be clearly seen that t/Q_t has a good linear relationship with t for the adsorption kinetics of all three MXene samples, and the correlation coefficients (R²) are close to 1. Additionally, the calculated removal capacities of Ti₃C₂T_x-D, Ti₃C₂T_x-H and Ti₃C₂T_x-DMSO-H based on the model are similar to the experimental equilibrium capacities (Table 1). The pseudo-second-order kinetic model describes the removal of Th^IV on Ti₃C₂T_x MXenes quite well, revealing that the kinetics are determined by two components which may be related to the diffusion of Th^IV and the occupation of adsorption sites on MXene surface.

3.3. Effect of pH

Fig. 4 represents the removal of Th^IV by Ti₃C₂T_x-D, Ti₃C₂T_x-H and Ti₃C₂T_x-DMSO-H as a function of pH. The uptake of Th^IV for the three MXene samples shows an overall increasing trend with the increase in solution pH, which is ascribed to the changes in the surface charge of MXene and the species of Th^IV in the solutions. Since pristine Ti₃C₂T_x and its intercalated products had similar zero charge points around pH 2.0–2.5 (Ying et al., 2015), their surfaces were positively charged under very acidic conditions. In this case, the adsorption of cationic Th^IV on Ti₃C₂T_x MXenes was greatly inhibited due to the electrostatic repulsion and the competition of the high concentration of H⁺. Nonetheless, Ti₃C₂T_x-DMSO-H still exhibited a considerable adsorption capacity (∼45 mg g⁻¹) at pH 1.5, implying that its largest interlayer space promotes the ion exchange of protons on the MXene surface (in the form of [Ti–O]⁻H⁺) with Th^IV. As the pH increased above 2.5, the surface of Ti₃C₂T_x became negatively charged, and the electrostatic interaction between adsorbent and Th^IV resulted in an increase in adsorption capacity. The species distribution of Th^IV as a function of pH was calculated using the Medusa/Hydra software. It can be seen from Fig. S2 that although there is no Th(OH)₄ precipitation under the experimental conditions, the partial hydrolysis species of Th^IV including Th(OH)³⁺ and Th(OH)₂²⁺ formed with the increase in pH. The fractions of Th(OH)³⁺ and Th(OH)₂²⁺ increase sharply at pH > 2.0 and pH > 3.0, respectively, which also contributes to the increase of Th^IV adsorption capacity at high pH conditions.

Figure 4 Th^IV adsorption by Ti3C2Tx-D, Ti3C2Tx-H and Ti3C2Tx-DMSO-H as a function of solution pH. (C0 = 100 mg l^−1, m/V = 0.4 g l^−1.)

3.4. Adsorption isotherms

The adsorption isotherm experiments were conducted to assess the effect of Th^IV equilibrium concentration on the adsorption behavior of the MXenes. As shown in Fig. 5, over the initial concentration range of 5–150 mg l⁻¹, the obtained maximum experimental adsorption capacities (Q_max) of Th^IV for Ti₃C₂T_x-D, Ti₃C₂T_x-H and Ti₃C₂T_x-DMSO-H at pH 3.4 were 46, 126 and 162 mg g⁻¹, while the corresponding Q_max at pH 3.0 were 17, 86 and 112 mg g⁻¹, respectively. Ti₃C₂T_x-DMSO-H had the best Th^IV enrichment performance among the three MXene samples, which is attributed to the most available active adsorption sites stemming from its largest interlayer space.

Figure 5 Adsorption isotherms of Th^IV uptake onto Ti3C2Tx-D, Ti3C2Tx-H and Ti3C2Tx-DMSO-H. (a) pH 3.4 ± 0.1; (b) pH 3.0 ± 0.1; solid line: fitted by Freundlich model; dashed line: fitted by Langmuir model.

To further evaluate the adsorption mechanism, Langmuir and Freundlich isotherm models were used to fit the experimental data. The former mainly describes monolayer adsorption and the latter is suitable for the interpretation of multilayer adsorption. The equations of the two models are expressed as follows in order:

For the Langmuir model [equation (4)], Q_m is the maximum adsorption capacity (mg g⁻¹) corresponding to a complete monolayer coverage and k_L is a constant indirectly related to the sorption capacity and energy of adsorption (l mg⁻¹), which characterizes the affinity of the adsorbate to the adsorbent. As for the Freundlich model [equation (5)], Q_e (mg g⁻¹) is the equilibrium adsorption amount, k_F and n are the Freundlich constants related to the adsorption capacity and the adsorption intensity, respectively. Based on Fig. 5 and the fitting parameters in Table 2, it is apparent that the Freundlich model could better describe the adsorption process of Th^IV on the three MXene samples. Because the fitting of pH 3.0 and pH 3.4 led to similar results, we concluded that the heterogeneous adsorption of Th^IV by Ti₃C₂T_x was a pH-independent intrinsic characteristic. The heterogeneous adsorption mechanism could be explained by the various termination groups (–OH, –O and –F) and unique multilayer structures of Ti₃C₂T_x. It was also noted that Ti₃C₂T_x-H and Ti₃C₂T_x-DMSO-H had larger values of parameter n than Ti₃C₂T_x-D, implying a more favorable adsorption of Th^IV by the hydrated and intercalated MXene.

Table 2
Isotherm parameters fitted by the Langmuir and Freundlich models for Th^IV adsorption onto Ti₃C₂T_x-D, Ti₃C₂T_x-H and Ti₃C₂T_x-DMSO-H

	Langmuir model			Freundlich model
	qm (mg g^−1)	kL (ml mg^−1)	R^2	kF (mg g^−1)	n	R^2
Ti3C2Tx-D pH 3.4	40.8	0.0367	0.714	3.20	1.95	0.868
Ti3C2Tx-H pH 3.4	126.4	0.173	0.985	29.5	3.01	0.980
Ti3C2Tx-DMSO-H pH 3.4	157.7	0.178	0.971	43.3	3.5	0.988
Ti3C2Tx-D pH 3.0	22.3	0.0192	0.720	1.89	2.35	0.870
Ti3C2Tx-H pH 3.0	85.7	0.181	0.989	31.2	4.79	0.997
Ti3C2Tx-DMSO-H pH 3.0	110.7	0.307	0.990	46.7	5.31	0.995

3.5. Effect of temperature and thermodynamic study

The adsorption of Th^IV onto the MXene samples was carried out at 280, 290, 300 and 310 K to investigate the effect of temperature. The uptake capacities of Ti₃C₂T_x-D, Ti₃C₂T_x-H and Ti₃C₂T_x-DMSO-H increased with increasing temperature [Fig. 6(a)], suggesting that a higher temperature is favorable for the adsorption of Th^IV by Ti₃C₂T_x. Thermodynamic evaluation of an adsorption process is necessary to conclude whether the process is spontaneous or not. The Gibb's free energy change (ΔG⁰, kJ mol⁻¹), enthalpy change (ΔH⁰, kJ mol⁻¹) and entropy change (ΔS⁰, J mol⁻¹ K⁻¹) can be calculated by equations (6) and (7):

where R is the universal gas constant (8.314 J mol⁻¹ K⁻¹), T is absolute temperature (K) and K_d represents the distribution coefficient (ml g⁻¹). Combining equations (6) and (7), ΔH⁰ and ΔS⁰ can be determined from the linear relationship of lnK_d versus 1/T, described as follows:

Figure 6 (a) Effect of temperature on the adsorption of Th^IV by Ti3C2Tx-D, Ti3C2Tx-H and Ti3C2Tx-DMSO-H. (C0 = 100 mg l^−1, pH = 3.0±0.1, m/V = 0.4 g l^−1); (b) Van't Hoff equation plots for the adsorption of Th^IV onto Ti3C2Tx-D, Ti3C2Tx-H and Ti3C2Tx-DMSO-H.

Fig. 6(b) shows the above linear plots and Table 3 lists the thermodynamic parameters calculated from the fitting lines. ΔG⁰ is negative at all the conditions applied, confirming that the adsorption of Th^IV by Ti₃C₂T_x is spontaneous. Ti₃C₂T_x-DMSO-H has the most negative ΔG⁰ among the three MXene samples, which reflects the thermodynamically favorable nature of the hydrated intercalated MXene for Th^IV adsorption. ΔH⁰ > 0 indicates that the adsorption is an endothermic process. Ti₃C₂T_x-D has a larger ΔH⁰ than the other two samples, which might be related to the hydration/dehydration states of MXene. The positive value of ΔS⁰ suggests the increased randomness at the adsorbent/adsorbate interface during the adsorption. Additionally, compared with the thermodynamic data of previously studied hydrated Ti₂CT_x (Li et al., 2019), Ti₃C₂T_x-H renders a lower ΔH⁰ and ΔS⁰, which indicates that the affinity of Ti₃C₂T_x to Th^IV is not as strong as that of Ti₂CT_x, although the two MXenes have similar surface termination and layer structure.

Table 3
Fitting results of thermodynamic parameters for Th^IV adsorption onto Ti₃C₂T_x-D, Ti₃C₂T_x-H and Ti₃C₂T_x-DMSO-H

	ΔH (kJ mol^−1)	ΔS (J mol^−1 K^−1)	ΔG (kJ mol^−1)
			280 K	290 K	300 K	310 K
Ti3C2Tx-D	21.39	111.6	−9.86	−10.97	−12.09	−13.21
Ti3C2Tx-H	9.58	90.3	−15.7	−16.61	−17.51	−18.41
Ti3C2Tx-DMSO-H	11.30	99.6	−16.59	−17.58	−18.58	−19.58

3.6. Effect of ionic strength and selectivity test

Particular attention was also paid in this study to Th^IV adsorption under various ionic strengths and competing cations to evaluate the potential of Ti₃C₂T_x in practical applications of radionuclide separation and thorium purification.

As can be seen in Fig. 7, the removal efficiency of Th^IV by the hydrated MXenes decreased with the increase in ionic strength, which reflected that the adsorption process might be controlled by the ion exchange mechanism. Nevertheless, Ti₃C₂T_x-DMSO-H still exhibited good Th^IV uptake performance at the medium ionic strength conditions. The K_d extracted from Fig. 7 was found to be larger than 5000 ml g⁻¹ when the ionic strength was ≤0.1 mol l⁻¹, confirming Ti₃C₂T_x-DMSO-H is an excellent adsorbent for Th^IV removal from common wastewater. It is interesting that the removal efficiency of Ti₃C₂T_x-D increased with the increase in ionic strength, which is contrary to the behavior of the hydrated Ti₃C₂T_x MXenes. A reasonable explanation is that the presence of a high concentration of Na⁺ in aqueous solution may activate the dried Ti₃C₂T_x, leading to the intercalation of hydrated ions at the edge of nano-lamellar MXene, thereby increasing the adsorption sites for Th^IV.

Figure 7 Effect of ionic strength towards Th^IV adsorption onto Ti3C2Tx-D, Ti3C2Tx-H and Ti3C2Tx-DMSO-H. (C0 = 40 mg l^−1, pH = 3.4±0.1, m/V = 0.4 g l^−1.)

To evaluate the ion selectivity of the MXene adsorbents, a competing adsorption experiment was also performed in a mixed solution containing Th^IV and nine other metal cations. The initial concentration of each metal ion was 0.5 mmol l⁻¹. As shown in Fig. 8(a), Ti₃C₂T_x-D had almost no selectivity because the adsorption performance of Th^IV (mmol g⁻¹) was comparable with other metal ions. The Th^IV uptake capacities at pH 3.0 for Ti₃C₂T_x-H and Ti₃C₂T_x-DMSO-H were 0.33 and 0.40 mmol g⁻¹, and at pH 3.4 they were 0.39 and 0.45 mmol g⁻¹, respectively, whereas the uptake capacities for other competing ions were as low as less than 0.07 mmol g⁻¹. The selectivity coefficients, which we defined as the ratio of K_d of Th^IV to that of the competing ions, were calculated to be larger than 8.3 and 9.4 for Ti₃C₂T_x-H and Ti₃C₂T_x-DMSO-H at pH 3.0, respectively, suggesting a desirable selectivity of the hydrated and intercalated Ti₃C₂T_x for Th^IV.

Figure 8 Effect of competing metal cations on the removal of Th^IV by Ti3C2Tx-D (a), Ti3C2Tx-H (b) and Ti3C2Tx-DMSO-H (c) at pH 3.0 and 3.4. (C0 = 0.5 mmol l^−1 for all cations, m/V = 0.4 g l^−1.)

3.7. Adsorption mechanism

To clarify the underlying adsorption mechanism, the MXene samples loaded with Th^IV were vacuum dried and then subjected to SEM, EDS and XRD characterization. SEM images in Figs. 2(d)–2(f) indicated that the multilayer structures of Ti₃C₂T_x MXene were well retained after adsorption. EDS results in Figs. 2(g)–2(i) clearly confirmed that the content of Th in the MXene samples increased rapidly with the expansion of Ti₃C₂T_x interlayer space, which is consistent with the results of batch adsorption experiments. The comparison of XRD patterns can reflect the interaction behavior between layered materials and guest ions. According to Fig. 9(a) and Fig. S3, the (002) peak position of Ti₃C₂T_x-D remained unchanged after the adsorption, suggesting that no hydrated Th^IV ions entered the interlayers, and the uptake of Th^IV only occurred on the exterior surface of MXene. For hydrated MXenes including Ti₃C₂T_x-DMSO-H and Ti₃C₂T_x-H, the (002) peak shifted to 6.6° after the adsorption, revealing the successful intercalation of Th^IV in these samples. Since the solution pH had a slight decrease after the adsorption, we attribute the intercalation to the replacement of hydrated H⁺ by Th^IV in the interlayers of MXene, namely, there may be an ion exchange reaction. An estimation of the size of intercalated ions was carried out by calculating the enlarged interlayer space (1/2 Δc-LP) between Th-loaded Ti₃C₂T_x-DMSO-H/Ti₃C₂T_x-H and Ti₃C₂T_x-D samples. The dimension of intercalated Th^IV was calculated to be about 3.4 Å, which is smaller than the previously reported size of intercalated uranyl ions (3.9 Å) (Wang et al., 2017). This difference in size is reasonable because the hydrated thorium ion has a spherical configuration while the hydrated uranyl ion prefers equatorial plane coordination. Fig. 9(b) shows the effect of pH on the c-LP of Th-loaded Ti₃C₂T_x-DMSO-H. As the reaction pH decreased, the c-LP of MXene decreased from 26.7 to 25.1 Å, and the intensity of the (002) peak was also significantly weakened. This result shows that the interlayer adsorption amount and the chemical species of Th^IV under different pH together determine its intercalation behavior inside MXene interlayers.

Figure 9 (a) XRD patterns of the Th-loaded MXene samples at pH 3.4; (b) XRD patterns of the Th-loaded Ti3C2Tx-DMSO-H at different solution pH.

The EXAFS technique can probe the local environment around Th atoms (Shi et al., 2014), and therefore provide important information for understanding the microscopic mechanism of Th^IV adsorption on the Ti₃C₂T_x MXene. The Th L₃ edge k³-weighted EXAFS spectra of the Th-loaded hydrated MXenes and the references are shown on the left side of Fig. 10. The EXAFS data for the Ti₃C₂T_x-D sample are not available due to the poor quality of the spectrum caused by low Th^IV adsorption capacity. The oscillation periods of Ti₃C₂T_x-DMSO-H at pH 2.0 are consistent with that of Th^IV aqueous solution, but are significantly different from Th(NO₃)₄·5H₂O, implying the presence of highly hydrated Th^IV ions in the adsorbent and there were no coordinated nitrates around Th atoms. The amplitudes at k > 8.5 Å become lower under higher pH (e.g. pH 3.0 and 3.4), which reflects a lack of long-range order of Th^IV in the MXene interlayers. Fig. 10 (right side) represents the associated Fourier transform (FT) spectra in R space, and the fitting results are shown in Table 4. The FT spectra exhibit a strong peak at around 2 Å (without phase-shift correction), which can be ascribed to the first oxygen coordination shell surrounding the Th^IV atom. The average Th—O distance was fitted as 2.46 Å with a CN of 10.7 for Th^IV aqueous solution, which is in good agreement with the literature (Dähn et al., 2002; Rothe et al., 2002; Zhang et al., 2017). The sample of Ti₃C₂T_x-DMSO-H at pH 2.0 had 10.1 oxygen atoms at 2.46 Å. Although the CN is slightly reduced due to the confinement of the MXene interlayer, this result is similar to that of Th^IV aqueous solution, which indicates that electrostatic interaction is dominant during the adsorption process under low pH conditions. When the solution pH increased to 3.0 and 3.4, the Th—O distance shifted to 2.43 Å, while concomitantly the CN decreased to 9.5 and 8.9, respectively. It has been reported that the compact Th—O bond at 2.40–2.42 Å and the CN of 8.0–9.0 usually correspond to the formation of stable inner-sphere complexes for Th^IV (Seco et al., 2009; Zhang et al., 2017). Therefore, our fitting results indicate that the contribution of surface complexation to Th^IV adsorption on Ti₃C₂T_x cannot be ignored at such pH conditions. Additionally, the fitting parameters of Ti₃C₂T_x-H are very close to that of Ti₃C₂T_x-DMSO-H at pH 3.4, reflecting that the different intercalated guest in this study (water or a mixture of water and DMSO molecules) will not significantly affect the underlying adsorption mechanism. The FTIR spectra in Fig. 11 provide further information to support the above conclusion. After Th^IV uptake, the intensity of the absorption peak at 649 cm⁻¹ decreased and the absorption peak at 561 cm⁻¹ clearly shifted to 568 cm⁻¹ for both hydrated MXenes, whereas the absorption peaks for Ti₃C₂T_x-D were almost unchanged. The absorption bands in the range of 560–650 cm⁻¹ can be ascribed to the Ti–O interaction of Ti₃C₂T_x MXene (Peng et al., 2014). The changes of FTIR spectra apparently demonstrate the presence of strong affinities between Ti–O and Th^IV ions in Ti₃C₂T_x-H and Ti₃C₂T_x-DMSO-H. Combined with the fact that the ionic strength could influence the uptake capacity of MXene, we conclude that under higher pH conditions (pH > 3.0), the adsorption of Th^IV on hydrated Ti₃C₂T_x is jointly determined by inner-sphere complexation and electrostatic interaction.

Table 4
Fitting parameters extracted from least-squares fitting analysis of EXAFS spectra

CN, coordination number; R, radial distance; ΔE, energy shift relative to the calculated Fermi level; σ², Debye–Waller factor; R factor, goodness-of-fit indicator.

Sample	Path	CN	R (Å)	ΔE (eV)	σ^2 (Å^2)	R factor
Th(NO3)4·5H2O	Th–O	11.0†	2.52±0.02	5.32±1.39	0.009	0.016
	Th–N	4.0†	3.08±0.04		0.007
Th^4+ (aq)	Th–O	10.7±0.7	2.46±0.01	7.67±0.58	0.009	0.009
Ti3C2Tx-DMSO-H pH 2.0	Th–O	10.1±0.8	2.46±0.01	6.75±0.65	0.010	0.011
Ti3C2Tx-DMSO-H pH 3.0	Th–O	9.5±0.8	2.43±0.01	5.51±0.69	0.013	0.010
Ti3C2Tx-DMSO-H pH 3.4	Th–O	8.9±0.8	2.43±0.01	5.09±0.77	0.013	0.011
Ti3C2Tx-H pH 3.4	Th–O	9.0±0.7	2.44±0.01	5.39±0.64	0.012	0.017

†Fixed during fitting.

Figure 10 Left: Th L3 edge k3-weighted EXAFS spectra (dotted lines) and the best fits (solid lines) for the MXene samples under different solution pH. Right: corresponding non-phase-shift-corrected Fourier transforms. (a) Th(NO3)4·5H2O reference; (b) aqueous Th^IV solution; (c) Ti3C2Tx-DMSO-H pH 2.0; (d) Ti3C2Tx-DMSO-H pH 3.0; (e) Ti3C2Tx-DMSO-H pH 3.4; (f) Ti3C2Tx-H pH 3.4.

Figure 11 FTIR spectra of the MXene samples before (dotted lines) and after (solid lines) Th^IV adsorption.

4. Conclusion

In summary, this work systemically studied the adsorption behavior of Th^IV by dry, hydrated and intercalated Ti₃C₂T_x MXene samples. Because of the full utilization of the active sites inside the interlayers, the intercalation adsorption of Th^IV was dominant for the hydrated MXenes, while only external surface adsorption was available for the dry MXene, so Ti₃C₂T_x-DMSO-H and Ti₃C₂T_x-H exhibited higher uptake capacities and slower kinetics. Our work demonstrated that the elimination of Th^IV by Ti₃C₂T_x was a spontaneous endothermic reaction which followed a heterogeneous adsorption model. Ti₃C₂T_x showed a lower affinity for binding Th^IV compared with its titanium-based analog Ti₂CT_x, and the adsorption process could be influenced by the ionic strength. Nevertheless, Ti₃C₂T_x-DMSO-H still exhibited considerable removal capacity and desirable selectivity for Th^IV at moderate concentration of ionic strength and with competing cations, reflecting its application potential in the purification of thorium-containing radioactive wastewater. Additionally, by analyzing the coordination environment and adsorption sites, we confirmed that the adsorption of Th^IV on the hydrated Ti₃C₂T_x MXenes was determined by a combination of electrostatic interaction and surface complexation. These findings can greatly enrich our understanding of the microscopic interactions between MXenes and actinide ions.

Footnotes

‡These authors contributed equally to this work.