Highly Efficient Uptake of Cs+ by Robust Layered Metal-Organic Frameworks with a Distinctive Ion Exchange Mechanism.

137Cs with strong radioactivity and a long half-life is highly hazardous to human health and the environment. The efficient removal of 137Cs from complex solutions is still challenging because of its high solubility and easy mobility and the influence of interfering ions. It is highly desirable to develop effective scavengers for radiocesium remediation. Here, the highly efficient uptake of Cs+ has been realized by two robust layered metal-organic frameworks (MOFs), namely [(CH3)2NH2]In(L)2·DMF·H2O (DMF = N,N'-dimethylformamide, H2L= H2aip (5-aminoisophthalic acid) for 1 and H2hip (5-hydroxyisophthalic acid) for 2). Remarkably, 1 and 2 hold excellent acid and alkali resistance and radiation stabilities. They exhibit fast kinetics, high capacities (q m Cs = 270.86 and 297.67 mg/g for 1 and 2, respectively), excellent selectivity for Cs+ uptake, and facile elution for the regeneration of materials. Particularly, 1 and 2 can achieve efficient Cs+/Sr2+ separation in a wide range of Sr/Cs molar ratios. For example, the separation factor (SF Cs/Sr) is up to ∼320 for 1. Moreover, the Cs+ uptake and elution mechanisms have been directly elucidated at the molecular level by an unprecedented single-crystal to single-crystal (SC-SC) structural transformation, which is attributed to the strong interactions between COO- functional groups and Cs+ ions, easily exchangeable [(CH3)2NH2]+, and flexible and robust anionic layer frameworks with open windows as "pockets". This work highlights layered MOFs for the highly efficient uptake of Cs+ ions in the field of radionuclide remediation.

Introduction

The rapid development of the nuclear industry meets the ever-increasing energy demand; however, it brings the challenge of effectively treating radioactive wastes. 137Cs is one of the common radioactive elements in nuclear waste produced by fission of 235U. It is a high-level thermal fission nuclide with a half-life of about 30.17 years, emitting strong γ-radiations.[1,2]137Cs+ possesses a high solubility and a strong migration ability and biological toxicity. Once it enters the human body, 137Cs+ mainly accumulates in the soft tissues and forms internal radiation, resulting in the damage of the liver, kidneys, and central nervous system.[3] In addition, 137Cs is widely used in medical, industrial, and agricultural fields .[4−6] Therefore, it is of great importance to effectively remove and recover radiocesium from radioactive wastewater for human health, environmental protection, and resource recycling. However, the efficient capture 137Cs+ from complex solutions still remains a serious challenge due to its easy mobility and the influence of interfering ions.

A number of methods have been studied to remove Cs+ ions, mainly including coprecipitation,[7] solvent extraction,[8] reverse osmosis,[9] adsorption,[10] and ion exchange.[11] The ion exchange method has received widespread attention due to its convenient operation, low cost, and high efficiency.[12] To date, various ion exchange materials have been investigated for the removal of Cs+ ions such as zeolites,[13] titanates,[14] clays,[15] layered double hydroxides,[16] and metal ferrocyanides and ferricyanides.[17] However, these materials show some disadvantages, such as a slow adsorption rate, a low adsorption capacity, or a poor selectivity. Therefore, it is urgent to design and construct new highly efficient ion-exchange materials with fast kinetics and high selectivities for the uptake of Cs+ ions.

Over the past a few decades, metal–organic frameworks (MOFs) have attracted great attention due to their simple preparation, high porosity, large specific surface area, and strong structural designability.[18,19] MOFs are also considered promising adsorbents for radioactive ions because of their highly ordered porous structure and various functional groups.[20−24] A few MOFs have been explored for the removal of Cs+ ions from solutions, such as the three-dimensional (3D) actinide-based MOF [(CH3)2NH2][UO2(L2)]·0.5DMF·15H2O (H3L = 3,5-di(4′-carboxylphenyl) benzoic acid),[25] the 2D MOF (NC4H12)(NC2H8)2[In3(pydc)6]·13.1H2O (SZ-6; pydc = 2,5-pyridinedicarboxylic acid),[26] and classical 3D MOFs like MIL-101-SO3H[27] and UiO-66-NH–SO3H-3.[28] Our group reported the 3D FJSM-In MOF with an excellent radiation stability, a high removal rate, and a high selectivity for Cs+.[29]

Noteworthy is that the reported MOFs for adsorbing Cs+ mostly focus on 3D structures, and it is still rare for 2D layered MOF to be used for the uptake of Cs+, except SZ-6.[26] Compared with 3D MOFs, layered MOFs have more exposed active sites and a larger surface area, which benefit them by shortening the diffusion distance of target ions and thus the response time.[30] Moreover, layered MOFs have a variety of stacking modes and adjustable interlayer spacing, which make them structures with high flexibility. In some cases, the flexibility of layered MOFs may lead to a structural transformation,[31] which is greatly helpful in clarifying the capturing mechanism of radionuclides. All these characteristics of layered MOFs are beneficial for the effective adsorption of Cs+ ions.

Numerous studies have shown that the angle of the ligand effectively affects the structural dimensionality of MOFs.[32−36] For example, Banerjee et al. reported two layered MOFs constructed from isophthalic acid,[34] and Wang et al. used a flexible 5-(4-pyridyl)-methoxylisophthalic acid ligand to prepare three layered MOFs.[33] On the other hand, our previous research developed an effective strategy that ion exchange materials were constructed by inducing the anionic skeletons with small protonated amines such as [(CH3)2NH2]+ and [CH3NH3]+ as charge balancing cations, which easily escape from anionic networks.[29,37−42]

Herein, we prepared two MOFs with anionic 2D layer structures by introducing the derivatives of isophthalic acid, namely [(CH3)2NH2]In(L)2·DMF·H2O (DMF = N,N′-dimethylformamide, H2L= H2aip (5-aminoisophthalic acid) for 1 and H2hip (5-hydroxyisophthalic acid) for 2). They have excellent acid and alkali resistances and radiation stability. 1 and 2 possess outstanding adsorption performances for the Cs+ ion with fast kinetics, high uptake capacities, excellent selectivity, and facile elution. The Cs+ uptake capacities of 270.86 mg/g for 1 and 297.67 mg/g for 2 outcompete those of most reported solid sorbent materials. Particularly, 1 can achieve Cs+ and Sr2+ separation in a wide range of Sr/Cs molar ratio with a separation factor (SFCs/Sr) up to ∼320. Additionally, it can selectively remove Cs+ in complex environments with the interference of Na+, K+, Ca2+, and Mg2+ ions. It is fortunate that single-crystal X-ray crystallography helped verify that their Cs+ adsorbed products were Cs(H2O)3In(aip)2·2H2O (denoted as 1-Cs) and Cs0.72(H3O)0.28(H2O)4In(hip)2·H2O (denoted as 2-Cs). Impressively, we also obtained single crystals of the K+ elution product for 1-Cs, i.e., K(H2O)3In(aip)2·2H2O (denoted as 1-K). Therefore, the Cs+ adsorption–elution mechanisms by MOF have been directly “visualized” for the first time by studying the unprecedented single-crystal to single-crystal (SC to SC) structural transformation.

Experimental Section

Chemical Materials

InCl3·4H2O (99% Shanghai Xianding Biological Technology Co., Ltd.), 5-aminoisophthalic acid (H2aip, 98%, Anaiji Chemical Technology (Shanghai) Co., Ltd.), 5-hydroxyisophthalic acid (H2hip, 99%, J&K Scientific Ltd.), HNO3 (65–68%, Sinopharm Chemical Reagent Co., Ltd.), NaOH (AR, Tianjin Guangfu Fine Chemical Industry Research Institute), CsCl (4N, Shanghai Longjin Metallic Material Co., Ltd.), NaCl (AR, Sinopharm Chemical Reagent Co., Ltd.), KCl (AR, Tianjin Fuchen Chemical Reagent Co., Ltd.), CaCl2·2H2O (74%, Shanghai Silian Chemical Plant), MgCl2 (AR, Adamas Reagent Co., Ltd.), SrCl2 (AR, Tianjin Guangfu Fine Chemical Industry Research Institute) and N,N′-dimethylformamide (DMF, AR, Sinopharm Chemical Reagent Beijing Co., Ltd.) were purchased from commercial sources and used without further purification.

Preparation of [(CH3)2NH2]In(aip)2·DMF·H2O (1)

A mixture of InCl3·4H2O (0.050 g, 0.17 mmol), 5-aminoisophthalic acid (0.060 g, 0.33 mmol), and 6 mL of DMF was transferred to a 20 mL Teflon-lined steel autoclave, which was heated at 200 °C for 12 h and then slowly cooled to room temperature (RT). The product was washed with DMF several times and then dried naturally. Brown plate-like crystals of 1 were obtained (35 mg yield, 33.74% based on In). Anal. Calc. for 1 (C21H27N4InO10): C, 41.33%; H, 4.46%; N, 9.18%. Found: C, 40.62%; H, 4.36%; N, 8.92%.

Preparation of Cs(H2O)3In(aip)2·2H2O (1-Cs)

Crystals of 1 (40 mg) were immersed in 40 mL of a 3000 ppm Cs+ aqueous solutions, and the mixture was shaken for 12 h at RT. Then, the brown plate-like crystals of 1-Cs were isolated after being washed with water and ethanol and dried naturally. Anal. Calc. for compound 1-Cs (C16H20N2CsInO13): C, 28.34%; H, 2.68%; N, 4.13%. Found: C, 28.40%; H, 2.73%; N, 4.11%.

Preparation of K(H2O)3In(aip)2·2H2O (1-K)

1-Cs crystals (5 mg) were mixed with 5 mL of a 0.3 mol/L KCl solution, and the mixture was shaken for 12 h at RT. Then, the brown plate-like crystals of 1-K were separated after being washed with water and ethanol and dried naturally. Anal. Calc. for 1-K (C16H20N2KInO13): C, 31.91%; H, 3.35%; N, 4.65%. Found: C, 31.32%; H, 2.98%; N, 4.68%.

Preparation of [(CH3)2NH2]In(hip)2·DMF·H2O (2)

A mixture of InCl3·4H2O (0.20 g, 0.68 mmol), 5-hydroxyisophthalic acid (0.24 g, 1.32 mmol), and 12 mL of DMF was transferred to a 28 mL Teflon-lined steel autoclave, heated at 200 °C for 12 h, and slowly cooled to RT. The product was washed with DMF several times and then dried naturally. Then, the colorless block-like crystals were obtained (135 mg yield, 32.43% based on In). Anal. Calc. for compound 2 (C21H25N2InO12): C, 41.20%; H, 4.12%; N, 4.58%. Found: C, 40.40%; H, 4.03%; N, 4.28%.

Preparation of Cs0.72(H3O)0.28(H2O)4In(hip)2·H2O (2-Cs)

Crystals of 2 (10 mg) were added to 10 mL of a 1000 ppm Cs+ aqueous solution, and the mixture was shaken for 5 h at RT. Then, the colorless flake-like crystals and microcrystals of 2-Cs were isolated after being washed with water and ethanol and dried naturally.

Characterization Techniques

Powder X-ray diffraction (PXRD) patterns of samples were obtained on a Rigaku Miniflex-II diffractometer by using Cu Kα radiation (λ = 1.54178 Å) in a 2θ range of 5–60°. The results of elemental analyses (EA) were recorded using a German Elementary Vario EL III instrument. Thermogravimetric curves were recorded in a N2 atmosphere on a thermal gravimetric analysis (TGA) instrument with a heating rate of 10 K/min. The optical adsorption spectra were measured with a Shimadzu 2600 UV/vis spectrometer. Energy-dispersive spectroscopy (EDS) and elemental distribution mapping were performed with a JEOL JSM-6700F scanning electron microscope. The morphologies of prepared samples were examined with a field emission scanning electron microscope (SEM, HITACHI FE-SEM SU8010). X-ray photoelectron spectroscopy (XPS) was conducted on an ESCALAB 250Xi type X-ray photoelectron spectrometer. Inductively coupled plasma-optical emission spectroscopy (ICP-OES) was performed on an ICAP-7400 system to measure the sodium, potassium, calcium, and magnesium concentrations. The concentrations of cesium and strontium were measured by inductively coupled plasma-mass spectrometry (ICP-MS) on a XSerise II instrument. 1 and 2 were irradiated with γ-rays at total doses of 100 (1.1 kGy/h for 95 h) and 200 kGy (1.1 kGy/h for 181.8 h) using a 60Co irradiation source (two million curies) provided by the Detection Center of Suzhou CNNC Huadong Radiation Co., Ltd., China.

Single-crystal X-ray diffraction (SCXRD) data were recorded on a SuperNova CCD diffractometer with graphite-monochromated Cu Kα radiation (λ = 1.54178 Å) at 100 K for 1 and 1-Cs and Mo Kα radiation (λ = 0.71073 Å) at 100 K for 2. SCXRD data were collected on a Rigaku Hypix system with graphite-monochromated Ga Kα radiation (λ = 1.3405 Å) at 300 and 100 K for 1-K and 2-Cs, respectively. The structures were solved by direct methods and refined by full-matrix least-squares on F2 using the SHELX-2018 program package. All non-hydrogen atoms were refined anisotropically. Detailed crystallographic data and structure-refinement parameters are summarized in Table S1. CCDC nos. 2124172 for 1, 2124173 for 1-Cs, 2124174 for 1-K, 2124175 for 2, and 2124176 for 2-Cs contain the supplementary crystallographic data.

Ion Exchange Experiments

To aqueous solutions of CsCl with the different concentrations of Cs+ (10 mL) was added the ground polycrystalline powder of 1 or 2 (10 mg). The mixture was kept under magnetic stirring at RT for 10 h, filtered by the 0.22 μm Millipore filter on the syringe, then diluted with dilute nitric acid to meet the concentration range of the test instrument. The concentrations of ions were determined by ICP-OES or ICP-MS. More details about the ion exchange experiments are provided in the Supporting Information.

Results and Discussion

Crystal Structures

SCXRD shows that 1 crystallizes in the monoclinic space group P21/c. Its asymmetric unit contains one In3+ ion, two aip2– ligands, one DMF moleucle, one [(CH3)2NH2]+ ion, and one lattice water (Figure S1a). Each In3+ ion is seven-coordinated by six oxygen atoms from three chelating carboxyl groups and one oxygen atom from one monodentate carboxyl group, giving rise to a distorted [InO7] polyhedron (Figure S2a). The lengths of the In–O bonds range from 2.122(3) to 2.531(3) Å (Table S2). A long In–O distance (In(1)–O(6) = 2.731(4) Å) is not considered. Each [InO7] polyhedron connects to four adjacent [InO7] polyhedrons through four aip2– ligands to form an anionic layer of [In(aip)2] with the square-like window of 13.30 Å × 13.96 Å (Figure a). The In–L–In (where L is the center of benzene ring on the aip2– ligand linking two In atoms) angle is 112.67°, resulting in a waved anionic layer of [In(aip)2] (Figures b, S3a, and S4a). TOPOS analysis indicates the anionic layer is a sql topological type when the In3+ ions are considered as four-connected nodes and the aip2– ligands are considered as bridging linkers (Figure S3b).

Figure 1

(a) A square-like window with one [(CH3)2NH2]+ cation, one DMF molecule, and one lattice water in 1. (b) The packing arrangement of the layers in 1 viewed along the a-axis. DMF, lattice water molecules, and hydrogen atoms are omitted for clarity. (c) The packing arrangement of the layers in 1 viewed along the a-axis, where the [In(hip)2] layers are simplified to topological structures. Lattice water molecules are omitted for clarity. (d) A square-like window with one Cs+ ion in 1-Cs. (e) The packing arrangement of the layers in 1-Cs viewed along the a-axis. Lattice water molecules and hydrogen atoms are omitted for clarity. (f) The packing arrangement of the layers in 1-Cs viewed along the a-axis, where the [In(hip)2] layers are simplified to topological structures. Lattice water molecules are omitted for clarity. (g) A square-like window with one K+ ion in 1-K. (h) The packing arrangement of the layers in 1-K viewed along the a-axis. Lattice water molecules and hydrogen atoms are omitted for clarity. (i) The packing arrangement of the layers in 1-K viewed along the a-axis, where the [In(hip)2] layers are simplified to topological structures. Lattice water molecules are omitted for clarity. Color codes are as follows: In, turquoise; O, red; C, gray; N, blue; Cs, yellow; K, purple.

The layers are stacked along the c-axis in an AB fashion (Figure b and c). The adjacent layers are embedded in each other, and the amino groups of the aip2– ligands point to the interlayer space. The shortest In–In distance of the adjacent anionic layers in 1 is 7.74 Å. The protonated dimethylamine cations from the in situ decomposition of DMF are located between the layers, compensating the negative charge of the anionic network.[25,37] Furthermore, [(CH3)2NH2]+ cations interact with DMF and lattice H2O molecules through hydrogen bonds (Figure a and Table S3), which further interact with anionic frameworks via hydrogen bonds. Thus, layers are further linked by hydrogen bonds to form a 3D supramolecular structure. To a certain degree, abundant hydrogen bonding enhances the skeleton stability of 1. Although it has been reported that a structurally similar compound [(CH3)2NH2][In(aip)2]}·xG (G = guest) (space group P1̅) could adsorb methylene blue (MB) by the ion exchange method,[43] radionuclide remediation by a thus-layered framework has not been documented.

2 belongs to the monoclinic space group C2/c. Its asymmetric unit contains half a formula unit (Figure S1d). Compared with the seven-coordination of In3+ in 1, the In3+ ion in 2 is coordinated by eight oxygen atoms from four hip2– ligands in a bidentate chelating fashion to form a [InO8] polyhedron. The adjacent [InO8] polyhedra are linked by hip2– ligands to form a square-like window with a size of 13.25 Å × 14. 67 Å (Figure S5a), which is further extended by edge-sharing to form an anionic layer structure of [In(hip)2] (Figure S5b). The In–L–In (where L is the center of benzene ring on the hip2– ligand linking two In atoms) angle is 127.661(1)° in 2 as opposed to that of 112.671(4)° in 1, resulting in a more flatter anionic layer in 2 (Figure S4a and d). The hydroxyl groups of hip2– ligands point to the interlayer space similar to that of amino groups in 1 (Figure S5c). The anionic layers in 2 are also stacked in the AB fashion. The shortest In–In distance of the adjacent anionic layers in 2 is 7.53 Å, which is closer than 7.74 Å in 1. The negative charge of the [In(hip)2] network is balanced with [(CH3)2NH2]+ cations, which interact with the anionic layers via N–H···O hydrogen bonds (Table S5 and Figure S5a). DMF and lattice water molecules are also distributed between the layers (Figure S5d). 2 has been reported before by Yong et al.,[44] where it showed outstanding robustness in a wide range of pH levels and excellent adsorption for positive charge dyes, but it has not been applied to radionuclide remediation. 1 and 2 have the anionic layers, exchangeable [(CH3)2NH2]+ cations, and carboxyl functional groups. These structural features will provide convenience for ion exchange between charge-balancing cations and Cs+ ions.

The phase-pure crystals of 1 were obtained without manual selection as verified by PXRD, and the acid and alkali resistances of 1 were evaluated. 1 retained its layered structure after being immersed in aqueous solutions with various pH values (in the range from 2.98 to 11.05) for 10 h (Figure S7a). Furthermore, the leaching percentages of In for 1 were lower than 0.2% in the pH range of 2.98–12.06 (Figure S7b and Table S6). These results demonstrate that 1 exhibits a good stability in a wide pH range.

Structural Transformation

Fortunately, we obtained single crystals for the Cs+ adsorbed products of 1 and 2, namely 1-Cs and 2-Cs, respectively. SCXRD analyses clearly reveal that the asymmetric unit of 1-Cs contains one formula unit (Figure S1b). Although both space groups of the pristine 1 and 1-Cs are in the space group P21/c, the lengths of crystallographic a- and c-axis are reduced from 10.0110(15) and 14.956(3) Å in the pristine 1 to 9.8620(3) and 14.3580(5) Å in 1-Cs, respectively (Table S1). The coordination mode of the In3+ ion varies from seven-coordination in 1 to eight-coordination in 1-Cs (Figure S2d), originating from the carboxyl groups of one aip2– ligand changing from the monodentate chelating mode in 1 to the bidentate chelating mode in 1-Cs (Figure S2f). The square-like window in the anionic layer shrunk from 13.30 Å × 13.96 Å in 1 to 13.49 Å × 13.53 Å in 1-Cs, which was induced by the substitution of [(CH3)2NH2]+ and DMF with Cs+ and H2O molecules, respectively; additionally, Cs+ ions were trapped between the layers in 1-Cs (Figures a, d, and e). The size of the In–L–In angle increased from 112.67° in the pristine 1 to 113.21° in 1-Cs (Figures S4a and b). The Cs+ ion is surrounded by three water molecules and six oxygen atoms of the six COO– groups from six aip2– ligands to give a [CsO9] polyhedron in 1-Cs. The Cs–O distances range from 3.001(4) to 3.311(4) Å (Table S2). Thus, Cs+ incorporation into the original interlayer of 1 results in a significant contraction of the [In(aip)2] interlayer space and is accompanied by the departure of [(CH3)2NH2]+ ions and DMF molecules. The shortening of the shortest In–In distance from 7.74 Å in 1 to 7.38 Å in 1-Cs (Figures c and f), which is mainly attributed to the strong interaction of Cs···COO– groups, reflects this and indicates that 1 possess structural flexibility. Importantly, these structural transformation analyses confirm the successful entrance of Cs+ ions by the ion exchange mode and clearly reveal the Cs+ uptake mechanism of 1.

2 undergoes a similar ion exchange process for Cs+ uptake compared to that for 1. SCXRD shows that the space group of 2-Cs still remains C2/c. Its asymmetric unit contains half a In3+ ion, one hip2– ligand, half a cation (Cs0.36(H3O)0.14, Cs0.36 = Cs10.34 + Cs1B0.02), and two and a half water molecules (Figure S1e). The In3+ ion is still eight-coordinated, forming a [InO8] polyhedron. It is evident that the lengths of the crystallographic a- and b-axes change from 14.6706(14) and 13.2520(13) Å in pristine 2 to 13.0956(4) and 14.77(4) in 2-Cs, respectively (Table S1). This change is accompanied by the change of the window sizes from 13.25 Å × 14.67 Å in pristine 2 to 13.10 Å × 14.77 Å in 2-Cs (Figures S5a and e). The In–L–In angle is basically unchanged (Figures S4d and e). Cs+ ions are trapped between the [In(hip)2] layers (Figure S5f), which are surrounded by four oxygen atoms from four COO– groups and four water molecules to form [CsO8] polyhedra. The Cs–O distances are in the range of 2.78(3)–3.62(2) Å (Table S4). The strong interactions between Cs+ and oxygen atoms of COO– groups from [In(hip)2] layers lead to a shortening of the In(hip)2] interlayer space, with the closest In–In distance of adjacent layers changing from 7.53 Å in 2 to 7.18 Å in 2-Cs (Figure S5g). These changes fully confirm the structural flexibility of 2 and clarify the Cs+ uptake mechanism of 2.

Strikingly, the absorbed Cs+ ions of 1-Cs can be easily eluted by a KCl solution. The SCXRD, EA, and TGA analyses of eluted product imply that its formula is K(H2O)3In(aip)2·2H2O (1-K). 1-K also crystallizes in the monoclinic space group P21/c (Table S1) with one formula in the asymmetric unit (Figure S1c). The connection types of the In3+ ion and the aip2– ligand in 1-K are the same as those in compound 1-Cs. The K+ ion is surrounded by six oxygen atoms from COO– functional groups and three water molecules (Figure S2g). It is distinct that the size of the square-like window changes from 13.49 Å × 13.53 Å in 1-Cs to 13.40 Å × 13.63 Å in 1-K, and K+ ions are also trapped between the layers in 1-K (Figures d, g, and h). The interlayered In–In distance is further shortened to 7.13 Å (Figure i) due to the interaction between K+ ions and [In(aip)2] layers via O atoms of carboxyl groups from aip2– ligands. The results show that 1-Cs can be easily eluted in the presence of high concentrations of K+ ions, further demonstrating the robustness and flexibility of the [In(aip)2] layer framework in 1. Note that the clear Cs+ elution mechanism was first revealed to be the replacement of Cs+ with K+ from the view of structural chemistry for MOFs materials.

Characterizations for Ion-Exchanged Products

The ion exchange of [(CH3)2NH2]+ and Cs+ in 1 and 2 was confirmed by various characterization methods, including ICP-OES, ICP-MS, EDS, SEM, and XPS. Photographs of crystals for 1, 1-Cs, 1-K, 2 and 2-Cs are depicted in Figure a and e. The color of the samples before and after the Cs+ ion exchange of 1 as well as the eluted product 1-K remain brown, which is consistent with the results of UV–vis spectra; that is, the optical adsorption edges of 1 (3.06 eV), 1-Cs (3.11 eV), and 1-K (3.07 eV) are close (Figure S9a). Photographs of crystals of 2 and 2-Cs obviously show that the crystal shape changes from the block-like shape of pristine 2 to the flake-like shape or microcrystals of 2-Cs (Figure e). The optical adsorption edge of 2 is 3.55 eV and that of 2-Cs shows a blue-shift to 3.64 eV (Figure S9b). SEM images show that the crystal of 1 has smooth surfaces. 1-Cs retains the pristine shapes but its surface is lightly eroded, while the surface of 1-K becomes rougher than that of 1-Cs (Figures b, c, and d, respectively). Compared with the pristine 2, the surfaces of 2-Cs have obvious cracks (Figures f and g, respectively).

Figure 2

(a) Photographs of crystals of 1, 1-Cs, and 1-K. SEM images of crystals of (b) 1, (c) 1-Cs, and (d) 1-K. (e) Photographs of crystals of 2 and 2-Cs. SEM images of crystals of (f) 2 and (g) 2-Cs.

The results of EDS and elemental distribution maps of ion-exchanged products verify that Cs+ ions entered the samples of 1 and 2 (Figure S10) and were homogeneously distributed in the samples (Figures a and b, respectively). The Cs+ exchanged products exhibit evident characteristic peaks of 3d5/2 and 3d3/2 for Cs+ at 724.5 and 738.6 eV (Figures c and d) for 1, respectively, and at 724.7 and 738.6 eV (Figures e and f) for 2 in the XPS survey spectra.[45] The peaks at 399.3 and 402.1 eV can be assigned to N 1s of the amino groups[46] and dimethylamine cations in 1,[29] respectively (Figure S11a). Contrasted with the pristine 1 and 2, the characteristic peaks for N 1s of the dimethylamine cation for 1-Cs and 2-Cs at 402.1 and 402.3 eV, respectively, almost disappear (Figure S11). This evidence suggests that Cs+ ions have successfully entered the structures of 1 and 2 by the ion exchange mode, accompanied by the departure of [(CH3)2NH2]+ cations.

Figure 3

SEM images of (a) 1 and (b) 2 and their corresponding elemental distribution maps of In and Cs. (c) X-ray photoelectron survey spectra for 1 (black line) and 1-Cs (pink line). (d) X-ray photoelectron spectrum of cesium for 1-Cs. (e) X-ray photoelectron survey spectra for 2 (black line) and 2-Cs (red line). (f) X-ray photoelectron spectrum of cesium for 2-Cs.

Kinetic Studies

The effects of the contact time for Cs+ ion exchange with 1 and 2 were investigated. The kinetic curves show that the concentration of Cs+ steeply decreases with time (Figures a and b). The removal rate of Cs+ (RCs, equation S1) by 1 rapidly reached 92.92% within 1 min and then gradually reached equilibrium within 5 min (Table S7). RCs of 2 can reach 89.23% within 1 min. As the time increased, the adsorption process of Cs+ ions by 2 gradually reached equilibrium within 10 min (Table S8). The Cs+ adsorption kinetics of 1 and 2 are significantly faster than those of other adsorption materials, such as [(CH3)2NH2][UO2(L2)]·0.5DMF·15H2O (20 min),[25] [(CH3)2NH2]4[(UO2)4(TBAPy)3]·18DMF·17H2O (H4TBAPy = 1,3,6,8-tetrakis(p-benzoic acid)pyrene, 30 min),[47] FJSM-InMOF (3 h),[29] zeolite A (90–120 min),[48] KMS-2 (10–15h),[2] H–CST (1 h),[49] and K-SGU-45 (15 h).[50] The rapid kinetics of 1 and 2 toward Cs+ ions are mainly ascribed to the strong interactions between COO– groups and Cs+ ion and , the flexibility of layered structures with open windows. The kinetic data were fitted with pseudo-first-order (equation S2) and pseudo-second-order kinetics models (equation S3). Both kinetics data were found be fit better by the pseudo-second-order kinetic model, with high correlation coefficients R2 of 0.99997 for 1 and 0.99994 for 2 (Figure S12).

Figure 4

Kinetic curves of (a) 1 and (b) 2 for Cs+ ion exchange plotted as the Cs+ concentration (mg/L, red line) and RCs (blue line) vs time t (min), respectively. Equilibrium data for the Cs+ ion exchange of (c) 1 and (d) 2 were fitted with the Langmuir and Langmuir–Freundlich isotherm models. Isotherm curves are derived from the concentrations of Cs+ at equilibrium plotted against the capacity (milligrams of ions removed per gram of an ion-exchanger).

Adsorption Isotherm Studies

To evaluate the Cs+ uptake capacities of 1 and 2, we performed the Cs+ uptake isotherm study with various cesium concentrations at RT (Tables S9 and S10). q (mg/g) is the ion exchange capacity and is described by equation S4. The Cs+ equilibrium curves were fit by the Langmuir (equation S5) and Langmuir–Freundlich (equation S6) models (Figures c and d). The higher correlation coefficients R2 of the Langmuir models for 1 (0.9229) and 2 (0.98714) indicate that the adsorption behaviors of 1 and 2 for Cs+ ions can be better fitted with the Langmuir model (Table S11). The maximum cesium exchange capacity (qmCs) of 1 and 2 are 270.86 and 297.67 mg/g, respectively. Notably, the maximum Cs+ uptake capacities of 1 and 2 are more than 3.3× higher that of the commercial AMP-PAN (81 mg/g).[51]1 and 2 also outperform the commercial scavengers TAM-5 (197.8 mg/g);[52] a series of MOFs materials including [(CH3)2NH2][UO2(L2)]·0.5DMF·15H2O (about 145 mg/g; H3L = 3,5-di(4′-carboxylphenyl)benzoicacid),[25] SZ-6 (129 mg/g),[26] UiO-66-NH–SO3H-3 (233.4 mg/g),[28] and FJSM-InMOF (198.63 mg/g);[29] and MOF-based composite materials HKUST-1/KNiFC (153 mg/g)[53] and GO/Co-MOF (192.41 mg/g).[54] Additionally, the qmCs values of 1 and 2 are remarkably higher than those of metal sulfides (KMS-1, 226 mg/g;[55] FJSM-SbS, 143.74 mg/g;[42] and FJSM-SnS-2, 266.5 mg/g[56]) and oxide materials (MgAl2.32Si5.2O14.88·18.23H2O, 102.34 mg/g;[57] SnSiMo, 16 mg/g;[58] and NaFeTiO, 52.8 mg/g[59]). These results indicate that both compounds show a high adsorption capacity for Cs+ ions.

Competitive Studies

Selectivity is one of the essential indicators for efficient ion exchange materials. There are often excessive interfering ions, such as K+, Na+, Ca2+, and Mg2+ in wastewater, which generally has a huge effect on the selective capture of Cs+. Hence, the selectivity for Cs+ was comprehensively examined for 1 and 2. RCs values of 1 and 2 in the presence of individual excessive alkali and alkaline-earth ions were first investigated. As shown in Figure a and b and Tables S12 and S13, when the Na/Cs molar ratios are ∼93 and 73, RCs of 1 and 2 can reach 84.29% and 75.40%, respectively. However, RNa is lower than 1% at the same time. The distribution coefficient (Kd) was calculated using equation S7 and was used to evaluate the affinity and selectivity of the material for the target ions. KdCs values for 1 and 2 are over 103 mL/g under the Na/Cs molar ratios of ∼93 and 73, while KdNa values are only 6.26 and 8.09, respectively. These results indicate that 1 and 2 can retain their good selectivity for removing Cs+ ions even in the presence of excess Na+ ions. The two compounds also have a good selectivity for Cs+ ion in the presence of K+ ions. When the K/Cs molar ratios are ∼20 for 1 and ∼47 for 2, KdCs and RCs values can reach 8.84 × 103 mL/g and 89.84% for 1, and 5.0 × 103 mL/g and 83.33% for 2, respectively, while KdK values are only 522.29 for 1 and 177.38 for 2, respectively. Furthermore, alkaline-earth metal Mg2+ and Ca2+ ions were selected as the representatives of divalent interfering metal ions. In the presence of individual excess Ca2+, KdCs values were 8.55 × 103 for 1 and 4.14 × 103 for 2, respectively, both of which are far higher than that of Ca2+. The KdCs can still remain 4.92 × 103 mL/g for 1 and 4.69 × 103 for 2 when the Mg/Cs molar ratio is 62.22 (RCs = 83.10%) for 1 and 36.75 (RCs = 82.44%) for 2. The results imply that both compounds still have strong interactions for Cs+ in the presence of individual excess amounts of Ca2+ and Mg2+.

Figure 5

KdCs and RCs of (a) 1 and (b) 2 in the presence of individual excess amounts of Na+, K+, Ca2+, and Mg2+ ions. KdCs, KdCs, and SFCs/Sr of (c) 1 and (d) 2 under different Sr/Cs molar ratios. (e) Kd of various metal ions in 1 and 2 with the coexistence of Cs+, K+, Na+, Ca2+, Mg2+, and Sr2+. (f) RCs of compounds 1 and 2 in tap water (Fuzhou, Fujian), Minjiang water (Fuzhou, Fujian), river water (Longyan, Fujian), and simulated groundwater.

Meanwhile, the effective separation of 137Cs/90Sr is important for the recovery and application of radionuclidesy.[60,61] Therefore, the separation of Cs+ and Sr2+ by 1 and 2 under the different Sr/Cs molar ratio was investigated (Figure c and d and Tables S14 and S15). When the Sr/Cs molar ratio is 0.11, RCs values reach 95.63% for 1 and 87.36% for 2. Accordingly KdCs values are 2.19 × 104 and 6.91 × 103 mL/g for 1 and 2, respectively, while RSr values are only 13.56 for 1 and 11.99 for 2. Even at the Sr/Cs molar ratio of 75.55, the RCs value of 1 can remain 83.53%, which is 53× higher than RSr (1.57%). Although the RCs value of 2 is only 66.67% at the Sr/Cs molar ratio of 75.55, it is still higher than RSr (4.72%). The separation factor (SF) can be calculated from equation S8 and can be utilized to judge whether the absorbent can separate two ions from each other. The highest value of SFCs/Sr for 2 is close to 80 in the Sr/Cs molar ratio range of 0.03–75.55. In particular, SFCs/Sr value of 1 are always above 100, and the highest SFCs/Sr is about 317, indicating that 1 has the excellent selectivity for Cs+ over Sr2+ and can achieve the effective separation of Cs+/Sr2+.

Additionally, the ion exchange performance with the coexistence of various alkali and alkaline-earth ions were further explored (Figure e and Table S16). Under the mixed K+, Na+, Mg2+, Ca2+, and Sr2+ ions, the RCs value of 1 is still as high as 85.87%, while the RCs value of 2 is 67.73%. KdCs values for 1 and 2 are 6.08 × 103 and 2.10 × 103 mL/g, respectively, which are higher than those of K+, Na+ Mg2+, Ca2+, and Sr2+. Moreover, the two compounds were evaluated for the removal efficiency of Cs+ in the simulated Cs+ contaminated water environments (Figure f and Tables S17 and S18). RCs values of 1 in tap water, Minjiang water, and river water were 87.84%, 89.27%, and 90.11%, respectively, and KdCs exceeded 7.00 × 103 mL/g. RCs values of 2 in tap water, Minjiang water, and river water were 87.84%, 87.27% and 80.19%, respectively, and KdCs exceeded 4.00 × 103 mL/g. In the simulated groundwater with K+, Mg2+, Ca2+, and high-concentration Na+, KdCs of 1 and 2 reach the level of 103 mL/g, and both RCs values are higher than 70%. The above results suggest that 1 and 2 can still maintain their excellent selectivity for Cs+ ions in the tap water, Minjiang water, river water, and simulated groundwater. However, RCs values of 1 and 2 in seawater are only 1.65% and 0.83%, respectively, which indicate that the absorption behaviors of 1 and 2 for Cs+ are greatly affected by the higher concentration of salts in seawater.

Irradiation Resistance

Considering that materials will be exposed to a strong radiation environment to handle the radioactive waste, the radiation stabilities of 1 and 2 were investigated. PXRD patterns of samples before and after radiation show that the original structures of 1 and 2 were maintained after 100 and 200 kGy γ irradiation (Figure S13). This demonstrates that they possess excellent radiation-resistant properties. We also tested the Cs+ ion exchange properties of two compounds before and after irradiation (Figure a and Table S19). RCs of 1 before and after irradiation can remain in the range of 88–92% under different initial concentrations of Cs+, and KdCs are more than 103 mL/g. Similarly, RCs and KdCs values of 2 before and after irradiation remain almost unchanged. These results demonstrate that irradiation did not affect the Cs+ adsorption performance of 1 and 2, further indicating their potential applications in radioactive waste treatment.

Figure 6

(a) RCs of 1 and 2 before and after 100 and 200 kGy γ at the following Cs+ concentrations: 21.4 mg/L Cs+ for (A) 1 and (B) 2 and 41.675 mg/L Cs+ for (C) 1 and (D) 2. (b) RCs of 1-K for three adsorption–elution cycles at different Cs+ concentrations.

Elution and Reuse

It is worth noting that the Cs+ exchanged products of 1 and 2 can be easily eluted with a 0.3 M KCl solution. EDS analyses of both eluted products confirm that Cs+ ions could be entirely replaced by K+ ions (Figure S14). Elemental distribution mapping reveals that K+ ions show the homogeneous distribution in the eluted productions (Figure S15a and b). We took 1-Cs as a representative to study its elution in detail. The XPS spectrum of 1-K exhibits characteristic 2p peaks for K+ at 293.3 and 296.0 eV (Figures S15c and d). The characteristic peaks of Cs+ disappear after elution, indicating the successful elution of Cs+. The RCs value of 1-K can remained in the range of 84.89–87.41% under different initial concentrations of Cs+ after two adsorption–elution process recycles. Even in the third cycle, the RCs value can still reach 80.48% (Figure b and Table S20). Additionally, PXRD shows 1-K could still maintain the crystallinity and stability of its layered framework after three adsorption–elution cycles (Figure S16a-d). EDS analyses of eluted products confirm that Cs+ ions could be entirely eluted by K+ ions (Figure S16e and 16f). Therefore, easy-to-operate elution and regeneration highlight the potential of the current layered compounds for the enrichment and recycle of cesium.

Mechanism on Cs+ Uptake

Although MOFs materials have been reported for the uptake of Cs+ ions,[25−29] it is still a challenge to deeply understand the ion exchange mechanism at an atomic level due to the worse crystallinity of MOF materials after adsorbing ions. In particular, it is still rare for a layered MOF to be used for the uptake of Cs+. 1 and 2 were prepared using the derivatives of isophthalic acid and exhibited excellent Cs+ uptake performance. The derivatives of isophthalic acid are powerful assistants for constructing layered structures and bring structural advantages for ion exchange, such as functional groups, adjustable interlayer spacing, and flexible frameworks. The charge balancing agent, [(CH3)2NH2]+ cations with an easy exchangeable ability, are useful for directing anionic frameworks. Impressively, 1 and 2 are robust enough for single crystals to be obtained after the exchange of Cs+ ions. SC-SC structural transformations clearly reveal that Cs+ ions are embedded in the interlayer through collaborative interactions with the anionic layers. It is crucial for the ion exchange process that there are the strong interactions between Cs+ ions and COO– groups from anionic layers. Simultaneously, XPS analyses confirm the disappearance or absence of N 1s of the dimethylamine cations in both Cs+ exchanged products, indicating the easy escape of [Me2NH2]+ cations. Moreover, compared with the pristine compounds, the shortest In–In distances between adjacent layers and the square-like window sizes within the layers of 1-Cs, 1-K and 2-Cs changed, confirming structural flexibility and accommodation ability of the current layered compounds. In other words, the excellent Cs+ ion exchange performances of 1 and 2 are attributed to the strong interactions between functional COO– groups and Cs+ ions, easily exchangeable [(CH3)2NH2]+, and flexible and robust anionic layers with open windows. In the case of MOF materials, the Cs+ adsorption–elution mechanisms are clearly unveiled for the first time at the molecular level.

Conclusion

In conclusion, two layered MOFs materials with wide pH durabilities and excellent radiation stabilities were prepared using the derivatives of isophthalic acid. 1 and 2 exhibit not only high adsorption capacities for Cs+ ions but also ultrafast removal abilities for Cs+ ions. Particularly, both compounds can selectively remove Cs+ in the presence of interfering ions and achieve Cs+/Sr2+ separation in a wide range of Sr/Cs molar ratios. Additionally, the Cs+ exchange products of 1 can be regenerated and recycled by a cost-affordable method. More significantly, this work systematically reveals the Cs+ uptake and elution mechanisms of layered MOF materials from the view of structural chemistry, with the fascinating structural transformation from layered structures to three-dimensional frameworks. This work not only develops layered MOF materials as radionuclide scavengers but also deepens the understanding of the relationship between the structure of materials and the property for the removal of a radionuclide. Eventually, this work provides insight into the design of new ion exchange materials for application in the treatment of radioactive contamination.