A negatively-charged supramolecular trap for precisely catching strontium ion

Abstract

Due to the analogous physicochemical properties and weak coordination ability of alkali and alkaline earth metals, accurate separation of radioactive ⁹⁰Sr from groundwater or seawater still presents a big challenge in environmental radioactivity remediation. Here we mimic the complexation behavior of molecular crown-ether carboxylic acids to construct an elegant negatively charged supramolecular trap in an anionic crown ether-based metal-organic framework (ZJU-X99) for precisely catching Sr²⁺. Owing to the synergistic effects of electrostatic interactions arising from the In(COO)₄^- nodes and supramolecular host-guest recognition from the 18-crown-6 rings, ZJU-X99 exhibits rapid adsorption kinetics (1 min), high adsorption capacity (263 mg/g), and exceptional selectivity for Sr²⁺ even when 1000-fold of Na⁺, K⁺ and Cs⁺ coexist. Relative to alkali metals, Sr²⁺ ions are intricately ensconced within the supramolecular trap, resulting in lowest binding energy and minimal structural alterations. Dynamic column experiments and radioactive ⁹⁰Sr decontamination trials further validate its practical application prospects. Our findings offer valuable insights into the design of supramolecular frameworks featuring tailored binding sites for targeted ions.

Introduction

The separation of alkali and alkaline earth metals is of utmost importance in various fields such as energy storage¹, water purification², spent fuel reprocessing³, and medical isotope preparation⁴. For instance, ⁹⁰Sr generated from the nuclear waste serves as a β-radiation emitter releasing high-energy β particles. The release of ⁹⁰Sr into the environment from the nuclear accidents has caused certain contamination of groundwater, rivers, and seawater, attracting widespread concern⁵. Upon entering the human body, it readily associates with bones, potential causing bone cancer⁶. Conversely, ⁹⁰Y, primarily derived from the β decay of ⁹⁰Sr, holds vast prospects in nuclear medicine⁷. Given the adjacency of alkali and alkaline earth metals in the periodic table and their closely resembling physicochemical properties, coupled with their weak coordination ability and high hydration energy, the attainment of selective separation poses a significant challenge⁸.

Over the past few decades, various techniques have been emerged for the efficient separation of Sr²⁺ ions, including ion exchange⁹, liquid-liquid extraction¹⁰, precipitation¹¹, and membrane separation¹², but these methods suffer from several drawbacks such as secondary waste generation, high energy consumption, complex operations, and poor selectivity¹³. In contrast, adsorption stands out due to its energy efficiency, low cost, simplicity, and recyclability, being regarded as a green technology¹⁴. A multitude of adsorbent materials have been explored for the separation of alkali and alkaline earth metals, encompassing metal sulfides¹⁵, zeolites¹⁶, zirconium phosphate¹⁷, metal-organic frameworks (MOFs)¹⁸, crystalline silicotitanates (CST)¹⁹, and hexacyanoferrates²⁰. Nevertheless, most of them lacked specific binding sites tailored for Sr²⁺, resulting in diminished adsorption performance particularly in the presence of high concentrations of competing ions. Addressing these challenges necessitates the development of selective recognition sites and the enhancement of strong binding affinity for Sr²⁺.

The strategic construction of selective recognition sites via host-guest chemistry holds significant promise for enhancing ion separation. Crown ethers, as macrocyclic molecules with selective complexation properties, have found widespread application in ion recognition²¹, providing opportunities for nuclear waste treatment and medical isotopes separation²². To augment ion selectivity, the integration of crown ethers into solid-state materials has emerged as a promising strategy. Traditional methods of incorporating these crown ethers, such as direct physical adsorption or chemical grafting, often resulted in molecular aggregation and pore blockage within the adsorbents²³. This, in turn, could impede adsorption kinetics and reduce overall capacity. An approach involving embedding crown ethers within the structured frameworks of MOFs provides an effective solution to this issue²⁴. Different crown ether units could be modularly introduced as host-guest functional groups into MOFs and single-crystal analysis has revealed that the cavity of 18-crown-6 and Sr²⁺ was well matched²⁵. Despite this, the competitive ion environment and the inherently weak binding affinity of crown ethers necessitate further optimization to improve both selectivity and capacity. Recent studies suggested that macrocyclic chelating agents with negatively charged carboxylic groups could significantly enhance the binding efficiency for heavier alkaline earth metals²⁶. Given that the atomic radius of Sr²⁺ closely resembles those of Na⁺ and K⁺, but with a notably higher charge density, endowing supramolecular crown ether functional groups together with a negatively charged skeleton within a solid matrix is hypothesized to strengthen the interactions between Sr²⁺ ions and the adsorbent’s framework (Fig. 1a).

Fig. 1: Design strategy and crystal structure of ZJU-X99.

a Structure design of ideal Sr²⁺ adsorbent. b Anionic In(COO)₄⁻ node and ligand used in ZJU-X99 (TBADB-18Cr6 = 4,4’,5,5’-terabenzoic acid dibenzo-18-crown-6). c View of the three-dimensional structures of ZJU-X99. d Negatively-charged supramolecular trap of ZJU-X99. e Supramolecular trap presented by electrostatic potential surface region which less than −0.17 eV (blue). f The size of the pore 1 and pore 2 in ZJU-X99. g The solvent-accessible void volume of ZJU-X99. Atom colors: In, orange; O, red; C, light gray; H, white.

Inspired by this, we proposed a synthesis strategy to construct a negatively charged supramolecular trap in an anionic indium-based MOF containing crown ether units (ZJU-X99), which was assembled by negatively charged In(COO)₄⁻ nodes and 4,4’,5,5’-terabenzoic acid dibenzo-18-crown-6 (18Cr6) linkers. The unique combination of suitably sized crown ether units and the anionic framework rendered ZJU-X99 exceptionally proficient at selectively adsorbing Sr²⁺ ions, achieving remarkably high adsorption capacities. This study represents the synergistic effects of electrostatic interactions and host-guest recognition mechanisms within crown ether-based anionic frameworks targeting alkali and alkaline earth metals.

Results

Synthesis and crystal structure

Colorless transparent hexagonal flake-shaped (Fig. 2a) crystals of ZJU-X99 ([NH₂Me₂][In(TBADB-18Cr6)]·H₂O, TBADB-18Cr6 = 4,4’,5,5’-terabenzoic acid dibenzo-18-crown-6) were obtained by heating a mixture of 0.071 mmol (21.4 mg) In(NO₃)₃·xH₂O and 0.071 mmol (60 mg) TBADB-18Cr6 in a 6 mL solution of N,N-dimethylformamide (DMF), acetonitrile, and acetic acid (volume ratio = 2 mL/2 mL/2 mL) at 120 °C for 72 h (85.4% yield based on In). Single-crystal X-ray diffraction analysis revealed that ZJU-X99 crystallized in the monoclinic space group C2/c, exhibiting an open 2-fold interpenetrated 3D anionic framework structure. Its asymmetric unit consists of one In³⁺ ion, one TBADB-18Cr6 molecule, one water molecule, and one [NH₂Me₂]⁺ (Supplementary Fig. 1). Each In³⁺ ion is coordinated by eight oxygen atoms from four different TBADB-18Cr6 ligands in a bidentate manner, forming In(COO)₄⁻ polyhedral clusters (Fig. 1b). Every TBADB-18Cr6 ligands serve as 4-connected nodes and coordinate with four In³⁺ to construct a single-layer network with a pore size of 7.8 Å × 11.3 Å (Supplementary Fig. 2). Two single-layer network duplexes 2-fold interpenetrated architecture, resulting in a 3D anionic framework (Fig. 1c). [NH₂Me₂]⁺ cations are located within the channel as charge balancing species (Supplementary Fig. 3). Remarkably, the close proximity (3.0 Å and 3.3 Å) between the O2 and O14 atoms on the In(COO)₄^- clusters and the crown ether plane suggests that the negatively charged supramolecular trap formed by the In(COO)₄^- clusters and 18Cr6 ring can be utilized to selectively capture cations with specific sizes (Supplementary Fig. 4 and Fig. 1d). Electrostatic potential surface (ESP) mappings further highlighted the presence of a negatively charged supramolecular trap wrapped between In(COO)₄⁻ and TBADB-18Cr6, exhibiting the most negative electrostatic potential distribution (< -0.17 eV) (Fig. 1e). Regarding to previous supramolecular crown-ether MOFs, the neutral clusters and crown ether units are too distant to cooperatively construct such unique supramolecular trap. In addition, there are two types of channels along the b-axis, with dimensions of 6.4 Å × 8.3 Å and 4.6 Å × 7.3 Å, facilitating rapid ion exchange (Fig. 1f). Topological analysis indicated that In atom and TBADB-18Cr6 unit act as 4-connected nodes respectively, resulting in a binodal 4,4-connected framework with a topology of pts (Supplementary Fig. 5). The solvent accessible volume is 16.8% calculated by PLATON (Fig. 1g).

Fig. 2: Sr²⁺ adsorption performance of ZJU-X99.

a Optical image of ZJU-X99. b Powder X-ray diffraction patterns of ZJU-X99 after soaked in the pH range from 0 to 11 and β radiation. c CO₂ adsorption/desorption isotherms and pore-size distribution profiles (Inset) for ZJU-X99. d Adsorption kinetics curves of ZJU-X99 toward Sr²⁺ (C₀ = 10 mg/L). e Adsorption isotherm curves of Sr²⁺ by ZJU-X99 compared with ZJU-X100. f Effect of Na⁺, K⁺, and Cs⁺ on the adsorption of Sr²⁺ by ZJU-X99 (C₀ = 10 mg/L). g A comparison of Sr²⁺ uptake capacity with reported adsorbents. h Dynamic adsorption and elution curves of ZJU-X99 (C₀ = 8–12 mg/L).

PXRD patterns of the synthesized material matched the simulated pattern well (Fig. 2b), confirming the phase purity of the sample. The permanent porosity of ZJU-X99 was confirmed through CO₂ adsorption/desorption isotherms, with a surface area of 293 m²/g (Fig. 2c). TG and variable-temperature PXRD analysis revealed that ZJU-X99 remained stable below 300 °C, indicating its high thermal stability (Supplementary Fig. 7a, b). To assess the stability of the material under practical application conditions, ZJU-X99 was soaked in solutions with pH values ranging from 0 to 11 for 12 h or exposed to 200 kGy β and γ radiation. The PXRD patterns, In³⁺ leaching rate, weight retention percentage, FTIR spectra, and CO₂ adsorption/desorption isotherms of sample were tested. The PXRD patterns and FT-IR spectra of ZJU-X99 remained nearly unchanged under pH conditions of 1 to 11 and radiation exposure (Fig. 2b and Supplementary Fig. 8). Under these conditions, the low In³⁺ leaching rate, high weight retention percentage and the minimal decline in specific surface area (Supplementary Figs. 9–12) confirm the excellent stability of ZJU-X99 across a broad pH range and under irradiation conditions.

Efficient capture of Sr²⁺ ions

Adsorption kinetics experiments demonstrated that ZJU-X99 reached adsorption equilibrium within 1 min, exhibiting a remarkable removal efficiency of 99% for Sr²⁺ (Fig. 2d). As illustrated in Supplementary Table 3, the adsorption rate of ZJU-X99 was significantly faster compared to the majority of other adsorbents. Such rapid adsorption kinetics and high removal efficiency demonstrated the excellent suitability of ZJU-X99 for nuclear emergency accident remediation. This is attributed to the electrostatic interaction and the host-guest recognition of the negatively charged supramolecular trap constructed by In(COO)₄^- cluster and the 18Cr6 unit, which synergistically enhanced the adsorption of Sr²⁺. Simultaneously, the 18Cr6 unit served as a one-dimensional channel to facilitate the entry of Sr²⁺ into the framework. The Langmuir model fitted the adsorption isotherm data of ZJU-X99 for Sr²⁺ well, yielding a maximum adsorption capacity of 263 mg/g (Fig. 2e and Supplementary Fig. 13), surpassing those of most previously reported crown ether composite materials (DtBuCH18C6/Sipolymer²⁷, DCH18C6 functionalized resin²⁸, Fe₃O₄@SiO₂@DtBuCH18C6²⁹, TNTs@DCH18C6³⁰), anionic MOFs (SZ-7³¹, SZ-4⁹, SZ-6³², FJSM-InMOF¹⁸), metal sulfides (ZnSnS-1³³, NaTS³⁴, KNbS³⁵, KZrTS³⁶), and zeolites³⁷ (Fig. 2g). It is worth noting that the maximum adsorption capacity of ZJU-X99 for Sr²⁺ exceeded those of crown ether-based neutral MOFs (SNU-200³⁸, MOF-18Cr6²⁴, and ZJU-X100²⁵), attributable to the substitution of neutral metal nodes with anionic In(COO)₄⁻, thereby increasing negative charge density of the framework. ZJU-X99 exhibited excellent Sr²⁺ removal performance within the pH range of 3–11, with a removal efficiency exceeding 99% (Supplementary Fig. 14). After irradiation with β and γ radiation at 200 kGy, the removal efficiency of Sr²⁺ could still maintain at above 96% (Supplementary Fig. 15). Even after 5 cycles of adsorption-desorption, the material could still remove over 97% of Sr²⁺ (Supplementary Fig. 16).

The interference resistance is a crucial consideration in screening adsorbent materials. The selectivity capability of ZJU-X99 in removing Sr²⁺ was tested in the presence of Na⁺, K⁺, Rb⁺, Cs⁺, and Mg²⁺. In the presence of a 10-fold concentration of monovalent cations, the removal efficiency of Sr²⁺ remained above 99%. When Mg²⁺ was present at a 10-fold concentration, the removal efficiency of Sr²⁺ was maintained at over 93% (Supplementary Fig. 17). Considering the excess competing ions such as Na⁺ and K⁺ present in real environments, the removal efficiency of Sr²⁺ was investigated as a function of the concentration of competing ions. Surprisingly, when the concentrations of Na⁺, K⁺, and Cs⁺ were increased to 1000-fold, ZJU-X99 was still able to remove 99% of Sr²⁺ (Fig. 2f). Even in the presence of complex competing ions including 1000-fold Na⁺, 1000-fold K⁺, 1000-fold Cs⁺, 1-fold Mg²⁺ and 1-fold Ca²⁺, more than 99% Sr²⁺ was captured by the adsorbent (Supplementary Fig. 18). Compared to the reported layered metal sulfides³⁹, anionic metal-organic frameworks^9,18,40, and anionic inorganic adsorbents^41,42, ZJU-X99 maintains excellent adsorption percentage for Sr²⁺ in the presence of higher concentrations of competing ions (Supplementary Table 6). It is reported that macrocyclic chelating agents with large negative charges are advantageous for binding heavy alkaline earth metals²⁶. Consistent with this finding, the synergism of the electrostatic interactions from the negatively charged In(COO)₄^- nodes and supramolecular host-guest interaction from 18Cr6 groups effectively enhanced the high selectivity of Sr²⁺.

To assess the material’s practical application potential, ZJU-X99 was utilized as the stationary phase and packed into an adsorption column with an inner diameter of 5.7 mm for dynamic adsorption experiments (Supplementary Fig. 20). Sr²⁺ solution was passed through the adsorption column via a peristaltic pump at a flow rate of 0.3 mL/min. Effluent was collected by an automatic collection for analysis of residual Sr²⁺ concentration. After the adsorbent was fully saturated, Sr²⁺ was desorbed using a 0.1 M HCl solution, followed by rinsing the column with deionized water for the next adsorption process. As shown in Fig. 2h, upon reaching the breakthrough point (C_t/C_o = 5%), the treatment volume was approximately 285 mL (2850 bed volume), indicating that only 0.1 mL (0.1 g) of ZJU-X99 could process 285 mL of Sr²⁺ solution. This substantial reduction in waste volume underscores its high removal efficiency, primarily attributed to the rapid adsorption kinetics and high adsorption capacity of ZJU-X99 towards Sr²⁺. The treatment volume of ZJU-X99 remained above 2490 bed volumes throughout the three cycles, demonstrating the material’s stability and effectiveness in removing Sr²⁺. In the dynamic breakthrough experiment with 10-fold excess Na⁺, K⁺, and Cs⁺, the treatment volume of ZJU-X99 for Sr²⁺ remained at 2580 mL (Supplementary Fig. 21), indicating the material’s selectivity for Sr²⁺. In addition, various solid-to-liquid ratios of ZJU-X99 were employed to examine the application potential for the removal of real radioactive ⁹⁰Sr, under the condition of a solid-to-liquid ratio of 10:1, ZJU-X99 demonstrated the capability to remove over 95% of ⁹⁰Sr (Supplementary Fig. 22), indicating its promising practical application prospects for radioactive waste management.

Revealing the adsorption process

To investigate the adsorption mechanism of Sr²⁺ in ZJU-X99, scanning electron microscopy coupled with energy-dispersive X-ray spectroscopy (SEM-EDS), powder X-ray diffraction (PXRD), and X-ray photoelectron spectroscopy (XPS) were used to analyze ZJU-X99 before and after adsorption of Sr²⁺. As shown in Supplementary Fig. 23, the SEM-EDS results confirmed the uniform distribution of Sr within the crystal. The PXRD peak of ZJU-99 retained after the adsorption of Sr²⁺ (Supplementary Fig. 24), indicating excellent chemical stability of ZJU-X99 attributed to the high chemical bond energy of In-O and the interpenetrated 3D structure. Thermogravimetric (TG) test results showed that Sr²⁺⊂ZJU-X99 maintained a high decomposition temperature (300 °C), which meant that the adsorption of Sr²⁺ did not affect the thermal stability of the material (Supplementary Fig. 25). In the XPS spectrum of ZJU-X99 after Sr²⁺ adsorption, characteristic peaks of Sr 3p and Sr 3d (135.5 eV and 133.8 eV correspond to Sr 3d_3/2 and Sr 3d_5/2 peaks, respectively, Supplementary Fig. 26 and Fig. 3a) newly appeared, demonstrating the successful incorporation of Sr²⁺ into the ZJU-X99 structure²⁴. The significant reduction of the N 1s peak at 399.2 eV (Supplementary Fig. 26) in Sr²⁺⊂ZJU-X99 confirmed the ion exchange reaction between [NH₂Me₂]⁺ and Sr²⁺ ions⁴⁰. In the high-resolution O 1s XPS spectrum after Sr²⁺ adsorption, a 0.3 eV shift of the C-O peak was observed and a new O-Sr peak appeared at 532.1 eV for ZJU-X99 (Fig. 3b). This suggests the existence of interactions between Sr²⁺ and the 18Cr6 moiety²⁴. In comparison to ZJU-X99, the high-resolution C 1s XPS spectrum of Sr²⁺⊂ZJU-X99 showed a 0.1 eV shift of the C-O peak (Supplementary Fig. 27), further confirming the binding of Sr²⁺ with the 18Cr6 unit.

Fig. 3: Characterization of ZJU-X99 before and after Sr²⁺ adsorption.

a High-resolution Sr 3d spectrum of Sr²⁺⊂ZJU-X99. b High-resolution O 1s spectra of ZJU-X99 before and after Sr²⁺ adsorption. c Solid-state ¹H NMR spectra of ZJU-X99 before and after Sr²⁺ adsorption. d Solid-state ¹³C NMR spectra of ZJU-X99 before and after Sr²⁺ adsorption. e Sr K-edge XANES spectra for Sr²⁺⊂ZJU-X99 and SrO. f EXAFS analysis of the k-space spectrum of Sr²⁺⊂ZJU-X99. g Fourier transforms of K-edge EXAFS spectra of Sr²⁺⊂ZJU-X99. h WT contour plots for Sr²⁺⊂ZJU-X99. i WT contour plots for SrO.

Solid-state ¹H and ¹³C NMR were employed to investigate the adsorption of Sr²⁺ onto ZJU-X99 before and after adsorption. In the solid-state ¹H NMR spectra (Fig. 3c), the peak at 2.7 ppm was attributed to the hydrogen atoms on the crown ether⁴³. Upon Sr²⁺ adsorption, the ¹H NMR peak of 18Cr6 moiety shifted and broadened, indicative of a coordination interaction between Sr²⁺ and the 18Cr6 unit, which restricted the motion of the crown ether unit^43,44. The chemical shift at 7.0 ppm corresponded to the hydrogen atoms on the phenyl group⁴³. After Sr²⁺ adsorption, the signal of hydrogen atoms on the phenyl ring shifted, possibly due to the exchange out of [NH₂Me₂]⁺ that originally interacted with the benzene groups in the pore. In the solid-state ¹³C NMR spectra (Fig. 3d), the signal at 65–70 ppm was attributed to the carbon atoms of the crown ether unit⁴⁵. Distinct signals in the range of 150–110 ppm were assigned to the carbon atoms on the phenyl group⁴⁶. The signal at 35.8 ppm was attributed to the [NH₂Me₂]⁺⁴⁷, which decreased after Sr²⁺ adsorption, indicating the successful exchange of [NH₂Me₂]⁺ by Sr²⁺. Moreover, compared to ZJU-X99, the carbonyl peak at 177.0 ppm in Sr²⁺⊂ZJU-X99 changed from a single peak to a double peak, indicating interactions between carbonyl oxygen and Sr²⁺, further confirming that Sr²⁺ entered the cavity of 18Cr6 groups and coordinated with the oxygen atoms of In(COO)₄⁻ after exchanging with [NH₂Me₂]⁺.

XAFS spectroscopy measurements were conducted on Sr²⁺⊂ZJU-X99 to investigate the coordination numbers and distances of Sr with surrounding atoms. The XANES spectrum showed an absorption edge energy of 16119 eV for Sr²⁺⊂ZJU-X99 (Fig. 3e), indicating no change in the oxidation state of Sr⁴⁸. Figure 3f, g showed that the Fourier-transformed extended X-ray absorption fine structure (FT-EXAFS) spectra of Sr²⁺⊂ZJU-X99 in k-space and R-space. The fitted parameters are presented in the Supplementary Table 7. Analysis of the spectra indicated the presence of eight O atoms surrounding the Sr atom. Specifically, six Sr-O bonds exhibited an average distance of 2.8 Å, while the remaining two O exhibited an average distance of 3.1 Å from the central Sr atom. Additionally, there were twelve C atoms around the Sr atom. Among them, eight C atoms had an average distance of 3.6 Å from the central Sr atom, while the remaining four C atoms had an average distance of 3.9 Å from the central Sr atom. These findings were consistent with the data on Sr-O and Sr-C bond lengths obtained from the following single-crystal structure analysis. The Wavelet Transform (WT)-EXAFS of Sr K-edge provided molecular structural information in k-space and R-space, yielding relevant information about the coordination environment of Sr²⁺. As depicted in the Fig. 3h, i, the maximum signal in the WT spectrum of Sr²⁺⊂ZJU-X99 was similar to the Sr-O bond in SrO, suggesting the formation of Sr-O bonds in ZJU-X99 upon Sr²⁺ adsorption.

Single-crystal-to-single-crystal transformation

Some crown ether units exhibit flexibility and render the resulting MOFs unsuitable for single-crystal XRD characterization, leading to a lack of in-depth exploration of atomic-level adsorption mechanisms. ZJU-X99 exhibited outstanding chemical stability and a strong affinity for Sr²⁺ ions. The single-crystal structure of ZJU-X99 after Sr²⁺ adsorption could be readily obtained through immersion in saturated Sr²⁺ solution. Single-crystal-to-single-crystal (SC-SC) transformation occurred in ZJU-X99 after Sr²⁺ adsorption (Fig. 4a), which revealed the exchange of [NH₂Me₂]⁺ in the channels with Sr²⁺ (Supplementary Figs. 3 and 4a). The asymmetric unit consists of Sr²⁺⊂ZJU-X99 comprises one In atom, one TBADB-18Cr6 molecule, and one Sr²⁺ (Supplementary Fig. 28a). It belongs to C2/c space group and after adsorption the crystal lattice parameters were changed from 21.784 Å and 10.078 Å to 21.9165 Å and 9.7073 Å along the a and b-axis, respectively. Sr²⁺ forms a SrO₈ polyhedral structure coordinated with six oxygen atoms from the 18Cr6 ring and two oxygen atoms from the In(COO)₄⁻ node, with Sr-O distances ranging from 2.8 Å to 3.1 Å (Fig. 4e, i), which is exactly consistent with the EXAFS analysis. Sr²⁺ is located 1.2 Å above the plane of oxygen atoms in the 18Cr6 (Supplementary Fig. 28b). The crown ether maintained the shallow boat conformation, with bending angles slightly changing from 24.7° to 28.4° (Supplementary Fig. 29), indicating a highly tailored fit of this negatively charged supramolecular trap in ZJU-X99 for Sr²⁺.

Fig. 4: Crystal structures of ZJU-X99 after adsorption of metal ions.

View of the three-dimensional structures of Sr²⁺⊂ZJU-X99 (a), Na⁺⊂ZJU-X99 (b), K⁺⊂ZJU-X99 (c), Cs⁺⊂ZJU-X99 (d). Representation of Sr²⁺ (e), Na⁺ (f), K⁺ (g), Cs⁺ (h) in the negatively charged supramolecular trap. Conformation of 18Cr6 unit in ZJU-X99 after complexation with Sr²⁺ (i), Na⁺ (j), K⁺ (k), Cs⁺ (l). Atom colors: In, orange; O, red; C, light gray; H, white; Sr, navy blue; Na, sky blue; K, light blue; Cs, powder blue.

To elucidate the high selectivity of ZJU-X99 towards Sr²⁺, we also obtained single-crystal structures after adsorption of Na⁺, K⁺, and Cs⁺ ions (Fig. 4b–d). Minimal variations in unit cell parameters of ZJU-X99 upon adsorption of metal ions with different sizes (Supplementary Table 8) suggest negligible changes in the overall framework of the crystal. The lengths of M-O bonds (M = Na, K, Cs), as presented in the Fig. 4j, k, l, exhibit close numerical values across different ions. Na⁺ within ZJU-X99 exhibits a positional disorder and was split into two positions during structure optimization, with the occupancy rates of 0.85 and 0.15 (Fig. 4j and Supplementary Fig. 31). The 18Cr6 unit undergoes distortion for the portion with an occupancy rate of 0.15, transitioning from a shallow boat conformation to a planar structure. A certain distortion of the crown ether rings is also observed for K⁺⊂ZJU-X99, with one O-C-O bond distorted by 90° (Fig. 4k and Supplementary Fig. 32). Cs⁺ within ZJU-X99 exhibits a positional disorder, with occupancy rates of 0.25 for two sites (Fig. 4l and Supplementary Fig. 33). The 18Cr6 unit undergoes some distortions, transitioning to a planar structure. The lower occupancy rate and distortion suggest that the specific ion recognition sites of ZJU-X99 are not well-suited for the larger Cs⁺ ion (167 pm)⁴⁹. Compared to Sr²⁺⊂ZJU-X99, all other three ions entering the trap need to overcome corresponding energy barriers for structural distortions and conformational transitions, indicating that the negatively charged supramolecular trap in ZJU-X99 is better suited for Sr²⁺. This is consistent with the previous results that macrocyclic chelating agents with large negative charges favored binding to heavy alkaline earth metals²⁶.

Density functional theory calculations analysis

The exceptional selectivity of ZJU-X99 for Sr²⁺ was further identified by periodic density functional theory (DFT) calculations compared to other alkaline earth metal ions (Mg²⁺, Ca²⁺), alkali metal ions (Na⁺, K⁺, Cs⁺) and water molecule. The optimized negatively charged supramolecular complex closely resembles their crystal structures, with small root mean square deviations (RMSDs) and bond lengths varying within 0.06 angstroms. The reaction energy ΔE for the hydrate metal ions M(H₂O)₆^+/2+ adsorbed by ZJU-X99 and generate M^+/2+⊂ZJU-X99 and water molecules were calculated since the solvent seriously affects the cation recognition of crown ether^9,50. As shown in Fig. 5a, the ΔE adheres to a recognition selectivity of M²⁺ > M⁺ > H₂O. Moreover, Sr²⁺ cation shows stronger ΔE than Mg²⁺ and Ca²⁺ with different ionic radii suggesting its higher selectivity to ZJU-X99. The reaction Gibbs free energy ΔG were calculated to further evaluate the thermodynamic feasibility (Fig. 5b). Sr²⁺ cation gives the lowest ΔG value of −155.98 kcal/mol which indicates the remarkable thermodynamic stability of Sr²⁺⊂ZJU-X99 complex and the exceptional selectivity of ZJU-X99 for Sr²⁺ compared to Mg²⁺ and Ca²⁺. For the alkali metal ions, the ΔG adheres to a recognition selectivity of K⁺ > Cs⁺ > Na⁺ which is consistent with the results in literature⁵⁰ confirming the reliability of theoretical calculations. The calculated charge density difference of M^+/2+⊂ZJU-X99 (Supplementary Figs. 34–38) show the charge accumulation from ZJU-X99 to Sr²⁺ (Fig. 5c) markedly stronger than that of Cs⁺⊂ZJU-X99 (Fig. 5d). The M²⁺⊂ZJU-X99 show similar charge density difference (Fig. 5e) while the corresponding ΔG of ZJU-X99 to Sr²⁺, Ca²⁺ and Mg²⁺ are −155.98, −151.55 and −148.75 eV which indicates the higher adsorption ability of ZJU-X99 to Sr²⁺ than Ca²⁺ and Mg²⁺. In the meanwhile, the ZJU-X99 show relatively weaker adsorption to M⁺ ions (Supplementary Fig. 39) and water molecule according to the less charge transfer between them. Besides, the average Mayer bond orders of Sr-O bonds shorter than 0.20 (Supplementary Table 13) and the overlap between Sr 4d and O 2p orbitals are relatively weak (Supplementary Fig. 40), which is in accordance with the highly ionized characteristics of Sr²⁺, and further demonstrates the dominance of charge-driven ionic bonds.

Fig. 5: Density functional theory analysis.

a, b The reaction energy ΔE and Gibbs free energy ΔG of ZJU-X99 to metal ions and water molecule. The calculated charge density difference of Sr²⁺⊂ZJU-X99 (c) and Cs⁺⊂ZJU-X99 (d). e The extracted charge density difference of M^+/2+⊂ZJU-X99 and H₂O⊂ZJU-X99 and the corresponding ΔG (isovalue = 0.02, 0.01 and 0.0005 a.u. for M²⁺, M⁺ and H₂O, respectively).

Discussion

In summary, we discovered a highly efficient and selective adsorbent material, ZJU-X99, with a unique negatively charged supramolecular trap for precisely catching Sr²⁺ from complex mixtures. ZJU-X99 exhibited exceptional stability even under high doses of irradiation and across a wide pH range. Moreover, it offered several notable advantages in the realm of Sr²⁺ ions separation, including a maximum adsorption capacity of 263 mg/g, rapid attainment of adsorption equilibrium within 1 min and 99% removal efficiency of Sr²⁺ from solutions containing 1000 times the concentration of Na⁺, K⁺, or Cs⁺. These advantages were stemmed from the synergism of electrostatic interactions arising from the negatively charged In(COO)₄^- nodes and supramolecular host-guest recognition from 18Cr6 groups. In addition, ZJU-X99 exhibited satisfactory performance in treating radioactive ⁹⁰Sr solutions and could significantly reduce the volume of contaminated water when used as a column packing material. In comparison to monovalent alkali metal ions, the entrapment of Sr²⁺ within ZJU-X99 induced minimal structural distortion and conformational alterations as evidenced by single-crystal structure analysis. Further theoretical calculations corroborated these observations, elucidating the exceptional selectivity of ZJU-X99 towards Sr²⁺ ions over other alkaline and alkaline earth ions originated from thermodynamically favorable complexation reaction and strong charge-driven electrostatic attractions. This work established a design strategy for supramolecular macrocycle-based MOFs featuring anionic frameworks, providing a comprehensive understanding the factors governing ion affinity and selectivity. These insights offer valuable guidance for the design and optimization of advanced materials tailored for efficient ion separation applications.

Methods

Materials and characterizations

All starting materials were purchased from commercial companies and used without further purification. Powder X-ray diffraction (PXRD) patterns were recorded on a Rigaku Ultima IV diffractometer equipped with Cu Kα radiation from 3° to 50° with a step of 0.02°. Various-temperature PXRD was obtained on a Rigaku SmartLab X-ray diffraction instrument at 30 – 300 °C from 3° to 50° with a step of 0.02°. The porous properties of the samples were measured by CO₂ physisorption at 273 K on an automated specific surface area and micropore analyzer (TriStar II Plus). TG analysis was performed in a nitrogen atmosphere with a rate of 10 °C min⁻¹ using a TA Instruments SDT Q600. X-ray photoelectron spectroscopy (XPS) spectra were acquired on a Thermo Scientific ESCALAB 250Xi. The β irradiation was carried out using electron beams generated from an electron accelerator. The γ irradiation was carried out using ⁶⁰Co γ-ray source apparatus. EDS images were acquired on a Scanning Electron Microscopy with an Energy Dispersive X-ray Spectrometer (SEM–EDS, Hitachi SU8010). The concentrations of Sr²⁺ in solutions were analyzed by an ICP–MS (Agilent Technologies 7800). The solid-state ¹H and ¹³C NMR spectra were collected with a Bruker AVANCE NEO 600WB spectrometer. EXAFS analyses were performed at the Shanghai Synchrotron Radiation Facility. The single-crystal X-ray diffraction data collection of ZJU-X99, Sr²⁺⊂ZJU-X99, Na⁺⊂ZJU-X99, K⁺⊂ZJU-X99 and Cs⁺⊂ZJU-X99 were recorded on a Bruker D8-Venture diffractometer with a Turbo X-ray Source (Cu–Kα radiation). The data frames were collected using the program APEX 3 and processed using the program SAINT routine in APEX 3. The crystal structures were refined by the full-matrix least-squares on F² using the SHELXTL-2018 program⁵¹.

Synthesis of ZJU-X99

The ligand of TBADB-18Cr6 (60 mg, 0.071 mmol) and In(NO₃)₃·xH₂O (21.4 mg, 0.071 mmol) were added in 6 mL of DMF/acetonitrile/acetic acid (v/v/v, 2 mL/2 mL/2 mL) solution. The mixture was placed into a Teflon-lined stainless-steel autoclave and heated at 120 °C for 72 h. After cooling to room temperature, the crystals were isolated by filtration, then washed with DMF, EtOH and dried at 50 °C for 12 h (Yield: 85.4% based on In atoms). The crystal data of ZJU-X99 is presented in Supplementary Table 1.

Adsorption experiments

Adsorption Kinetics Experiments. The adsorption kinetics experiments were carried by dispersing ZJU-X99 into Sr²⁺ solution with a solid-liquid ratio of 1:1, respectively. The mixture was stirred at room temperature. The supernatant solution was collected by syringe in different time intervals from 1 to 120 min, then filtered by 0.22 µm nylon membrane filters. The filtrates of residual Sr²⁺ were diluted with 5% HNO₃ and the concentrations of Sr²⁺ were determined by ICP–MS. The removal percentages were calculated by the equation as follows:

$${R}_{{{{\rm{e}}}}}=\frac{({c}_{0}-{c}_{{{{\rm{t}}}}})}{{C}_{0}}\times 100\%$$

(1)

where C₀ (mg/L) is the initial concentration, C_t (mg/L) is the concentration of Sr²⁺ after stirring at different times.

Adsorption Isotherm Experiments. The sorption isotherm experiments were measured by exposing ZJU-X99 into different concentrations of Sr²⁺ solution with a solid-liquid ratio of 1:1. The samples were stirred for 12 h at room temperature and filtered by 0.22 µm nylon membrane filters. The concentrations of residual Sr²⁺ in aqueous solution were determined by ICP–MS. The sorption capacity of ZJU-X99 toward Sr²⁺ was calculated by the following equation:

$${q}_{{{{\rm{e}}}}}=\frac{\left({c}_{0}-{c}_{e}\right)}{m}{{{\rm{V}}}}$$

(2)

Where C₀ (mg/L) is the initial concentration, C_e (mg/L) is the equilibrium concentration, m (g) is the mass of sorbent, and V (L) is the volume of the solution.

Langmuir and Freundlich sorption models were used to fit the sorption isotherms. The Langmuir isotherm model is expressed as given below:

$$\frac{{C}_{{{{\rm{e}}}}}}{{q}_{{{{\rm{e}}}}}}=\frac{1}{{q}_{{{{\rm{m}}}}}{K}_{{{{\rm{L}}}}}}+\frac{{C}_{{\mbox{e}}}}{{q}_{{\mbox{m}}}}$$

(3)

where q_e (mg/g) is the sorption capacity at equilibrium, q_m (mg/g) is the maximum sorption capacity and K_L is the Langmuir constant, C_e (mg/L) is the equilibrium concentration of Sr²⁺. The Freundlich isotherm model is shown as follows:

$${{{\mathrm{ln}}}}{q}_{{{\rm{e}}}}={{{\mathrm{ln}}}}{K}_{{{{\rm{F}}}}}+\frac{1}{n}{{\mathrm{ln}}}{C}_{{{{\rm{e}}}}}$$

(4)

where K_F is a constant about sorption capacity and n is related to the intensity of sorption.

Adsorption Selectivity Experiments. The adsorption selectivity experiments of ZJU-X99 toward Sr²⁺ were performed by adding ZJU-X99 into 10 ppm of Sr²⁺ solution containing corresponding proportions of competing anions with a solid-liquid ratio of 1:1, respectively. The solution was stirred for 12 h at room temperature. The samples were filtered and the concentrations of residual Sr²⁺ were analyzed by ICP–MS, respectively.

Reusability test

ZJU-X99 was firstly immersed in Sr²⁺ (10 ppm) solution with a solid-liquid ratio of 1:1 for 3 h at room temperature. Then the samples were filtered and the solid was desorbed by 0.1 M HCl solution for the regeneration of ZJU-X99. The regenerated adsorbents were used in subsequent adsorption experiments.

Column experiments

0.1 g of ZJU-X99 was placed in a polyethylene column with an inner diameter of 5.7 mm. The loading height is approximately 4 mm, resulting in a bed volume of 0.1 mL. To prevent solid sample loss, two sieve plate with a pore size of 20 μm was placed at the top and bottom of the column. A solution containing 8–12 mg/L Sr²⁺ was then introduced into the column at a flow rate of 0.3 mL/min. Solution flow rate and sample collection were regulated by a peristaltic pump and automatic collector. After the material reached full saturation, Sr²⁺ was desorbed using a 0.1 M HCl solution, followed by rinsing the column with deionized water to prepare it for the subsequent adsorption cycle.

⁹⁰Sr decontamination test

Caution! ⁹⁰Sr is a high-energy β radiation emitter that poses significant health risks and all experiments involving ⁹⁰Sr are conducted exclusively in authorized laboratories designed for radiological research. Strict adherence to established safety precautions and protocols for handling radioactive substances is mandatory. The ⁹⁰Sr decontamination test was conducted by adding ZJU-X99 to ⁹⁰Sr solutions (3704 cpm) with different solid-liquid ratios of 1:1, 10:1, 25:1, 50:1 and 100:1. After shaking for 10 min, the supernatant was filtered and the activity of ⁹⁰Sr was measured by a liquid scintillation counting (LSC, PerkinElmer Tri-carb 4910TR) technique.

Density functional theory calculations

The periodic density functional theory (DFT) calculations were performed with PBE⁵² exchange-correlation functional within the generalized gradient approximation (GGA) as implemented in the CP2K 9.1^53,54 software package. The norm-conserving⁵⁵ GTH pseudopotentials were performed for core electrons while the polarizable double-zeta quality basis sets were performed for valence electrons with 4, 5, 6, 9, 10, 13 valence electrons for C, N, O, alkali metal ions, alkaline earth metal and In, respectively. Considered the existence of Na⁺, the long-range electrostatics terms were calculated with a plane-wave basis set with a 1000 Ry cutoff and the Grimme’s DFT-D3 dispersion corrections were used with Becke-Johnson damping function^56,57. Because of the large unit cell size, only the Gamma point was considered. The global geometry optimization was performed while for the frequency calculations, the vibration of metal ions and [In(COO)₄-TBADB-18Cr6] were calculated and other atoms were constrained. For the cluster model, the structures of ZJU-X99 containing TBADB-18Cr6, In(COO)₄^- polyhedral clusters, and coordinated metal were extracted from their respective crystal structures. The geometry optimization was performed at the B3LYP level^58,59,60 using the Gaussian16 package⁶¹. The heavy atoms of the aromatic ring were positionally constrained. The C, H, and O atoms were described using the 6–31 g(d, p) basis set⁶², while metal ions were described using the Stuttgart small-core effective core potential and their corresponding SEG basis set^63,64.