Phosphonate-Decorated Covalent Organic Frameworks for Actinide Extraction: A Breakthrough Under Highly Acidic Conditions

Open AccessCCS ChemistryRESEARCH ARTICLE1 Aug 2019Phosphonate-Decorated Covalent Organic Frameworks for Actinide Extraction: A Breakthrough Under Highly Acidic Conditions Jipan Yu, Liyong Yuan, Shuai Wang, Jianhui Lan, Lirong Zheng, Chao Xu, Jing Chen, Lin Wang, Zhiwei Huang, Wuqing Tao, Zhirong Liu, Zhifang Chai, John K. Gibson and Weiqun Shi Jipan Yu Laboratory of Nuclear Energy Chemistry, Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 100049 (China) , Liyong Yuan Laboratory of Nuclear Energy Chemistry, Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 100049 (China) , Shuai Wang Laboratory of Nuclear Energy Chemistry, Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 100049 (China) College of Nuclear Science and Engineering, East China University of Science and Technology, Nanchang 330013 (China) , Jianhui Lan Laboratory of Nuclear Energy Chemistry, Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 100049 (China) , Lirong Zheng Beijing Synchrotron Radiation Facility, Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 100049 (China) , Chao Xu Collaborative Innovation Center of Advanced Nuclear Energy Technology, Institute of Nuclear and New Energy Technology (INET), Tsinghua University, Beijing 100084 (China) , Jing Chen Collaborative Innovation Center of Advanced Nuclear Energy Technology, Institute of Nuclear and New Energy Technology (INET), Tsinghua University, Beijing 100084 (China) , Lin Wang Laboratory of Nuclear Energy Chemistry, Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 100049 (China) , Zhiwei Huang Laboratory of Nuclear Energy Chemistry, Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 100049 (China) , Wuqing Tao Laboratory of Nuclear Energy Chemistry, Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 100049 (China) , Zhirong Liu College of Nuclear Science and Engineering, East China University of Science and Technology, Nanchang 330013 (China) , Zhifang Chai Laboratory of Nuclear Energy Chemistry, Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 100049 (China) , John K. Gibson Chemical Sciences Division, Lawrence Berkeley National Laboratory (LBNL), Berkeley, CA 94720 (USA) and Weiqun Shi *Corresponding author: E-mail Address: [email protected] Laboratory of Nuclear Energy Chemistry, Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 100049 (China) https://doi.org/10.31635/ccschem.019.20190005 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesDownload CitationsTrack Citations ShareFacebookTwitterLinked InEmail Although solid-phase extraction is a useful approach for metal ion separation from aqueous solutions, existing sorbents suffer from low extraction efficiencies and/or instability when in contact with strong acidic media. We report here the first study on rational design and fabrication of phosphonate-decorated covalent organic frameworks, COF-IHEP1 and COF-IHEP2, for efficient and selective extraction of of uranium (VI) [U(VI)] and plutonium(IV) [Pu(IV)] from highly acidic solutions. We found that the negatively charged frameworks with excellent stability under harsh conditions and strong chelating ability from incorporated phosphonate moieties make the COFs superb U(VI) and Pu(IV) sorbents. In 1 M HNO3 solution, COF-IHEP1 achieved recorded U(VI) uptake of 112 mg·g−1, with extremely high selectivity toward Pu(IV) even in the presence of large excess of competing metal cations. Our mechanistic study confirms a sandwich-type microstructure of hydrated U(VI) and Pu(IV) cations bound by the oxygen sites of phosphonate groups. These newly fabricated tailored hybrid porous materials could afford new opportunities to achieve highly efficient actinide extraction from acidic nuclear fuel effluents. Download figure Download PowerPoint Introduction The recent rapid expansion of the nuclear industry presents significant challenges, and thus, opportunities for developing novel materials to improve the nuclear fuel cycle. In particular, advances in the fabrication of new materials and techniques are critically needed to extract uranium [U(VI)] and plutonium [Pu(IV)] from acidic wastewater and mine drainage.1 Although conventional liquid–liquid extractions (e.g., plutonium–uranium recovery by extraction [PUREX]), have been well documented and widely used for processes relative to U(VI) and Pu(IV) separation, practical issues such as the use of toxic or flammable solvents, formation of emulsions, and generation of large volumes of secondary hazardous organic wastes are becoming increasingly tricky. Nonetheless, solid-phase extraction, based on the utilization of solid sorbents with high sorption capacities, has proven advantages of simplicity, reliability, and low solvent consumption, as well as potentially high enrichment factors.2 Traditional ion-exchange resins, for example, are commercial products that exhibit efficient uptake of U(VI) and Pu(IV),3 but their radiation resistances and chemical stabilities under harsh conditions require attention.4 Furthermore, relatively slow sorption kinetics and low ion capacity greatly limit their practical applications.2 Many neutral sorbents based on mesoporous silicas,5 carbons,6 polymers,7 and others8 have also been evaluated for uranium extraction in the last decades, and a current review of this topic is available.9 Under highly acidic conditions, such as what is observed with nuclear fuel cycle-related wastewater treatment, none of the reported materials used were effective, which is most likely due to sorbent disassembly and/or surface protonation.10,11 Thus far, there are no solid materials that fulfill all the requirements for real actinide separations under highly acidic conditions. In recent years, a new family of organic porous and crystalline materials that incorporate strong covalent linkages known as covalent organic frameworks (COFs) has emerged.12,13 Owing to the diversity of skeletons and porous structures, COFs have been exploited in applications such as gas storage and separation,14,15 energy conversion and storage,16 catalysis,17 semiconduction,18 photoemissions,19 proton conduction,20,21 optoelectronics,22,23 and so on.24,25 COFs also offer a platform for designing versatile materials for addressing environmental issues.26–33 Meanwhile, there have been insufficient promising results on this important topic. COFs can be readily prepared and modified by designing suitable precursors and reactions. The strong covalent bonds in COFs endow them with desirable stability under various harsh conditions and are made entirely from light elements (H, B, C, N, and O), thus, have a low density, which offers a clear benefit for improving adsorption capacities. All of these merits make COFs excellent candidates for actinide sorbents under highly acidic conditions, although no clear data on this topic have been disclosed. Herein, we rationally designed and constructed two new COFs, COF-IHEP1 and COF-IHEP2, in which structural units are linked through hydrazone bonds to form extended two-dimensional (2D) nanosheets. Hydrazone linkage was chosen because of readily available phosphonate precursors and mild reaction conditions, as well as a suitable hydrogen-bond system to ensure desirable stability under highly acidic conditions. More importantly, the weaker alkalinity of hydrazone linkage versus amide and amine hinders proton combination and ensures a negatively charged or neutral surface at a low pH, which, undoubtedly, facilitates metal cation trapping from an acidic solution. Phosphonate groups were decorated into the skeletons of COFs to provide active chelating sites, given that phosphoryl groups show desirable selectivity toward U(VI) and Pu(IV) as evidenced in the PUREX process. The prepared COFs exhibit exceptional stability under various harsh conditions, including gamma irradiation, and negatively charged surfaces over a wide pH range. Moreover, the side arm of COF phosphonate units exhibit favorable and selective coordinating properties for U(VI) and Pu(IV), achieving a record of high U(VI) uptake and unprecedented selectivity toward Pu(IV) in 1 M HNO3 solution. To the best of our knowledge, this is the first report on substantial actinide uptake by solid-phase extraction under highly acidic conditions, including the first Pu(IV) trapping in metal–organic frameworks/COFs and also, the first evaluation of radiation stability of COFs. Results and Discussion Material design, synthesis, and characterization The key intermediate, 2,5-bis[2-(diethoxy-phosphoryl)ethoxy]-terephthalohydrazide ( 1), was prepared following four steps with 63% yield (see for details). COF-IHEP1 and COF-IHEP2, which appeared as yellow powders in 84% and 88% yields, respectively, were synthesized via condensation reactions between ( 1) and 1,3,5-Triformylbeneze ( 2) or 1,3,5-Tris(4-formylphenyl)benzene ( 3) in flame-sealed tubes at 120 °C for 3 days; 3∶1 mesitylene/dioxane was used as the solvent, and 6 M acetic acid was the catalyst (Figure 1 and details in ). The resultant two COF-bearing phosphonate units in an ordered porous structure were expected to be promising candidates for U(VI) and Pu(IV) chelation. Scanning electron microscopy images of the prepared COFs were recorded (), and only one morphologically unique crystallite could be found for each COF, suggesting the desirable phase purity. Fourier-transform infrared spectra of the two COFs were compared with those of the precursors ( and ). The characteristic vibrational peaks for C=N bonds at 1618 cm−1 34 for COF-IHEP1 and 1668 cm−1 for COF-IHEP2 were indications of condensation reactions. The blueshift for the C=N bond in COF-IHEP2 was related to the extended π-conjugation in the framework. The thermal stabilities of COF-IHEP1 and COF-IHEP2 were evaluated by thermogravimetric analysis under air flow, which demonstrated that both COFs are thermally stable for up to 240 °C when partial degradation of the phosphate side chain occurred ( and ). Crystalline structures of COF-IHEP1 and COF-IHEP2 were analyzed using powder X-ray diffraction (PXRD) with Cu Kα radiation in conjunction with multiscale computational simulations and Pawley refinement (Figure 2a,b and ). COF-IHEP1 exhibited a strong PXRD peak at 3.55°, weaker peaks at 6.20° and 6.97°, and low-intensity broad features around 12.23°, 15.38°, and 26.56°, corresponding to the (100), (110), (200), (220), (310), and (001) facets, respectively (Figure 2a). Peak broadening was observable at > 5°, which could be ascribed to the disturbing effect of the phosphate side chain, given that the counterpart of COF-IHEP1 without phosphate sidechains (COF-IHEP3) was readily prepared with high crystallinity (). To elucidate the detailed structures of these two COFs, we constructed a series of structural models with AA binarized neural network (bnn) topology and AB grade (gra) topology stacking modes based on molecular mechanics simulations. Our results revealed that for each structural topology, there were two conformations, with cis and trans orientations of the phosphonate side chains (details in ). On the whole, four possible structures of COF-IHEP1 were derived from the simulations: bnn with cis, bnn with trans, gra with cis, and gra with trans. Periodic density functional theory (DFT) calculations were then performed to obtain theoretically accurate lattice parameters and atomic coordinates. Pawley refinements with the PXRD data were carried out to optimize the crystal structures with the AA and AB packing modes. The results suggest that the crystal structure in the AA stacking mode provided the best fit to the PXRD pattern of COF-IHEP1 (Figure 2a, red plot). A hexagonal unit cell (P3) with refined parameters a = b = 28.574 Å, c = 3.598 Å, α = β = 90°, and γ = 120° was deduced for COF-IHEP1 in eclipsed bnn packing mode with a cis orientation of the phosphonate side chains. Changing the orientation of the phosphonate side chains only minimally affected the framework structure and PXRD pattern. The (001) facet corresponded to a π–π stacking distance of about 3.6 Å along the stacking direction perpendicular to the 2D layers ( and and and ). In contrast, the AB stacking mode resulted in small pores covered by neighboring layers and a predicted PXRD pattern that poorly matched the experimental pattern ( and ). At the present resolution, either topology bnn or gra could be assigned. Figure 1 | Synthetic procedures for COF-IHEP1 and COF-IHEP2. Construction of COF-IHEP1 was accomplished by condensation of 2,5-bis[2-(diethoxyphosphoryl)ethoxy]terephthalohydrazide (1) and 1,3,5-Triformylbeneze (2). Formation of COF-IHEP2 was accomplished by treating (1) with 1,3,5-Tris(4-formylphenyl)benzene (3) under solvothermal conditions. Download figure Download PowerPoint Figure 2 | Characterizations of the COFs. Experimental (black) and predicted (red) PXRD patterns of COF-IHEP1 (a) and COF-IHEP2 (b) in bnn packing mode. Inset: AA stacking mode of the COFs. C, gray; N, blue; P, purple; O, red. H atoms are omitted for clarity. N2 adsorption and desorption isotherms of COF-IHEP1 (c) and COF-IHEP2 (d) measured at 77 K. (e) PXRD patterns of COF-IHEP1 before and after the treatment in various solvents. (f) pH-Dependent zeta potentials of COF-IHEP1. Download figure Download PowerPoint COF-IHEP2 exhibits a strong PXRD peak at 2.50°, a weaker peak at 4.83°, and broad features around 15.17° and 25.99°, which were assigned to the (100), (200), (600), and (001) facets of the P3 space group, respectively (Figure 2b). The corresponding π–π stacking distances between neighboring layers were deduced as 3.6 Å. Following the same procedure above, we obtained three possible structures of COF-IHEP2: bnn with trans, gra with cis, and gra with trans. The experimental pattern matches reasonably well with the simulated pattern (Figure 2b, red curve) for the AA stacking structure ( and ); Despite the absence of the predicted peak around 7°, the match is much better than for alternative structures. The refined lattice parameters for COF-IHEP2 in eclipsed bnn packing mode are a = b = 42.912 Å, c = 3.629 Å, α = β = 90°, and γ = 120°. The porosity of both COFs was determined by measuring N2 adsorption–desorption isotherms at 77 K (Figure 2c,d). The Brunauer–Emmett–Teller (BET) surface areas of COF-IHEP1 and COF-IHEP2 were estimated as 110 and 330 m2·g−1, respectively. These values are much smaller than those for COF-IHEP3 (without phosphate side chain, ), which likely originated from the existing phosphate side chain in the pore disturbing transport of the gas molecules. A nonlocal density functional theory model was fitted to the isotherms of COF-IHEP1 and COF-IHEP2 to estimate their pore size distributions (). The average pore diameters of 1.0 nm for COF-IHEP1 and 2.3 nm for COF-IHEP2 were in good agreement with the expected pore sizes from the crystal simulations based on the bnn topology. Chemical stability of the COFs was assessed by soaking COF-IHEP1 in a variety of organic solvents, as well as in 3 M aqueous HNO3 and 1 M NaOH. Comparison of PXRD patterns suggested that COF-IHEP1 is stable in common organic solvents, such as dimethylformamide, tetrahydrofuran, acetone, and hexane, as well as in boiling water. In addition, COF-IHEP1 retains its crystalline structure without discernible changes in the dominant (100) PXRD peak position and intensity even after 24 h treatment with 3 M HNO3 or 1 M NaOH, revealing excellent chemical stability (Figure 2e). In contrast, COF-IHEP2 exhibits relatively poor chemical stability as evidenced by dramatic changes in PXRD features after equal treatment (), likely resulting from its larger pore size. Zeta potentials of the COF surfaces as a function of solution pH were estimated to elucidate the COF surface properties, especially under low pH and highly acidic conditions. Figure 2f shows the zeta potentials of COF-IHEP1 at various pH levels. The value of the zeta potential was found to be zero at pH 1.2 and revealed a negatively charged surface of COF-IHEP1 over a pH range of 1.2–7, undoubtedly facilitating metal cations trapping. U(VI) and Pu(IV) sorption studies After confirming the porosity, structural integrity, and high density of chelating sites of COF-IHEP1 and COF-IHEP2, we examined their abilities to capture U(VI) and Pu(IV) from aqueous solutions. The sorption kinetics data at pH 1.0 ( and ) suggested that U(VI) sorption into both COFs is fast during the initial 20 min, followed by a slower process before reaching equilibrium at ∼ 3 h, and the kinetics was independent of the initial U(VI) concentration (50–188 ppm). The pseudo-second-order model fitted well into the experimental data with a correlation coefficient > 0.999 and rate constants (k) of ca. 1.0 × 10−3 (), revealing fast chemical sorption processes.35 Sorption isotherms for U(VI) into the two COFs were also determined at pH 1.0 by varying the initial U(VI) concentrations from 5 to 200 mg·L−1 (). Both COF-IHEP1 and COF-IHEP2 exhibited very steep sorption profiles for U(VI) with an increase of U(VI) concentration at equilibrium (Ce, positively correlated to initial U(VI) concentrations). At an initial U(VI) concentration over 120 mg/L corresponding to a Ce of > 60 mg/L, U(VI) sorption into the two COFs reached equilibrium from which the derived saturated sorption capacities are 160 and 140 mg·g−1 for COF-IHEP1 and COF-IHEP2, respectively. The new value of 160 mg·g−1 represents the highest U(VI) capture by solid extraction at such a low pH. Both sorption isotherms are well fitted by the Langmuir model with a correlation coefficient of more than 0.99 (), indicating a monolayer uniform sorption mode.36 To further assess the effect of solution acidity on U(VI) removal by the two COFs, U(VI) sorption into COF-IHEP1 from highly acidic HNO3 media and aqueous solutions at various pH was studied. The results (Figure 3a) revealed a strong influence of solution acidity on sorption. In particular, at pH 5, the U(VI) uptake into COF-IHEP1 reached 257 mg·g−1, corresponding to almost complete U(VI) uptake for a 10 mL of 100 mg·L−1 U(VI) solution using only 4 mg of the sorbent. As the solution acidity increased from pH 3.0 to pH 1.0, the U(VI) uptake into COF-IHEP1 decreased to 239 and 155 mg·g−1, respectively. These values, however, were much higher under the same conditions than for most benchmark materials, such as UiO-68-P(O)(OEt)2,1 bio-inspired nanotraps,37 SZ-3,38 NP10,39 KIT-6-80-P,2 MIPAF-11 c40, and MXene.41 More significantly, when highly acidic media were used as is typical for nuclear fuel reprocessing and wastewater treatment, high U(VI) uptake of 112 mg·g−1 in 1 M HNO3 and 70 mg·g−1 in 2 M HNO3 were obtained. These values set a new record for U(VI) uptake under highly acidic conditions. This remarkable performance of U(VI) could be rationalized based on accessible phosphonate active sites well dispersed in the COF channels and zeta potentials of the COF surfaces. Although the surface charge of the COF sorbents changed from negative to positive under acidic conditions (pH < 1.0), the zeta potential was not high enough to repel all the positively charged U cations. In this case, the strong coordination between active phosphonate groups and U cations played a key role, thus leading to a considerable U(VI) uptake. To test adsorbability of the COFs toward other actinides, 239Pu(IV) and 241Am(III) sorption into COF-IHEP1 and COF-IHEP2 from highly acidic HNO3 media and aqueous solutions at various pH was studied. Both COFs achieve Pu(IV) removal as high as ∼ 90% regardless of the solution acidity (Figure 3b), whereas almost no Am(III) sorption occurs in all the solutions. This result suggested that, in addition to U(VI), COF-IHEP1 and COF-IHEP2 could serve as superb sorbents for Pu(IV) in strongly acidic solutions, which is reasonable, given that these COFs bear phosphonate groups that are known to efficiently bind both U(VI) and Pu(IV) in the PUREX process. Figure 3 | U(VI) and Pu(IV) sorption studies with multiple cycling, selectivity, and radiation stability assessments. (a) U(VI) sorption into COF-IHEP1 for a wide range of acidity and comparison with that of other benchmark materials. ★ denotes the value in this work. (b) Removal percentage of Pu(IV) with COF-IHEP1 and COF-IHEP2 for a wide range of acidity. (c) Recycle use of COF-IHEP1 and COF-IHEP2 for U(VI) uptake at pH 1.0. (d) Selective sorption of U(VI) and Pu(IV) with COF-IHEP1 and COF-IHEP2 from a solution containing competing metal ions at pH 1.0. (e) PXRD patterns of COF-IHEP1 before and after gamma irradiation at different doses. (f) U(VI) sorption by COF-IHEP1 before and after irradiation. "Dry" denotes dry samples and "In water" denotes water-saturated samples. Download figure Download PowerPoint Multiple cycling, selectivity, and radiation stability So far, our results suggest the feasibility of utilization of the two COFs, especially COF-IHEP1, for U(VI) and Pu(IV) capture from acid media. To fully realize such applications requires consideration of additional attributes. First, the sorbent should be stable during the sorption process, and the sorbed metal ions should be easily desorbed, with the reclaimed sorbent reusable. BET surface areas and FTIR of COF-IHEP1 after U(VI) uptake from 1 M HNO3 solution were measured, and the results were compared with those for freshly synthesized COFs ( and ). Except for the appearance of the absorption band of UO22+ at 921 cm−1, there were no discernable changes in the FTIR of COF-IHEP1 after harsh conditions, confirming its excellent chemical stability. The surface area of COF-IHEP1 decreased from 110 to 65 m2·g−1, which could be attributed to the incorporation of U(VI) ions into the pores of the COFs. Then, the U(VI)-loaded COF-IHEP1 was regenerated readily on rinsing with saturated sodium carbonate solution, followed by a simple filtration, which resulted in a high demetallation rate of 93%. The reclaimed COF-IHEP1 was dried in vacuum overnight and subjected to the next round of U(VI) sorption. Notably, the sorption capacity of COF-IHEP1 toward U(VI) diminished only very gradually after the second sorption–desorption cycle, with 92% of the sorption capacity retained after four cycles (Figure 3c), thus demonstrating reusability. After five sorption–desorption cycles, COF-IHEP1 was characterized by FTIR, XRD, and BET surface area measurement (, , and ). FTIR spectra showed no discernable changes, whereas the weakened and broadened peaks in the PXRD pattern, suggested a decreased crystallinity of the COFs. However, the porous structure of the COFs was maintained, with a surface area of 47 m2·g−1. The sorbent should also show substantial selectivity toward target metal ions over competing metal ions to enable efficient utilization. Herein, the selectivity test was performed at pH 1.0 using a multimetal ions solution. The results in Figure 3d and show that the two COFs, especially COF-IHEP1, effectively and selectively removed U(VI) and Pu(IV) ions in the presence of competing metal ions with an equal concentration of 0.5 mM. The distribution coefficient (Kd) for Pu(IV) is as high as 2.0 × 104 mL·g−1, which is two orders of magnitude higher than that for competitive lanthanides, indicating an unprecedented selectivity of this material toward Pu(IV) in strongly acidic solutions. A third key consideration is that the sorbents should be robust to high radiation doses such as may be encountered in nuclear fuel reprocessing. To assess the radiation stability of the two COFs, dry samples and samples soaked in water were subjected to 60Co gamma irradiation using a dose rate of 60–80 Gy·min−1, up to a maximum dose of 200 kGy. There were no discernable changes in PXRD patterns (Figure 3e), indicating maintenance of the COF framework during irradiation, whereas FTIR and thermogravimetric analysis (TGA) results suggested that there was no significant radiolysis of the phosphonate groups ( and , respectively). More importantly, no significant decrease in U(VI) sorption (Figure 3f) was observed for any of the test samples even at 200 kGy irradiation, which confirmed excellent radiation stability. This outcome is understandable in that the COFs exhibit extended π-conjugation frameworks and close π–π stacking structures. Specifically, the extended π-conjugation framework allows radiant energy to be well dispersed throughout the COF layers, whereas the close π–π stacking enables radiant energy transfer between layers. Both of these effects provided resistance to the breaking of bonds in the COFs, thus achieving good resilience toward gamma irradiation. The above findings suggested that abundant opportunities for phosphonate-modified COFs as efficient and practical U(VI) and Pu(IV) sorbents. In particular, the stable nature and outstanding performance of COF-IHEP1 in HNO3 substantiated the applicability of this material for U(VI) and Pu(IV) separation in used nuclear fuel reprocessing. Sorption mechanism To understand the sorption mode, 31P CP/MAS NMR was used to assess the interaction between U(VI) and COF-IHEP1. Figure 4a shows the comparison of 31P CP/MAS NMR spectra of COF-IHEP1 before and after U(VI) loading. After U(VI) sorption, the signal at 36 ppm, due to P atoms of phosphonate42 (purple curve), split downfield with the second peak and shifted to 41 ppm (blue curve), indicating a strong interaction between phosphonate groups and U(VI) cations. We prepared the phosphonate-free COF-IHEP3 as a reference to assess the contribution of the skeleton structure to U(VI) uptake. COF-IHEP3 was obtained as a yellow powder in 89% yield and characterized using FTIR, TGA, PXRD, and NMR (details in ). The saturated absorption capacity of this reference sample was only 5 mg·g−1 at pH 1.0, which revealed a minimal contribution of the unmodified skeleton structure of the COFs to U(VI) uptake and further demonstrates the critical role of phosphonate groups. Figure 4 | Sorption mechanism studies. (a) 31P CP/MAS NMR spectra of COF-IHEP1 (purple curve) and U(VI)-loaded COF-IHEP1 (blue curve). The favored binding modes of UO22+ (b) and Pu4+ (c) in COF-IHEP1 as predicted by DFT calculations. C, cyan; P, purple; O, red; H, white; U, yellow. (d) Left: raw U LIII-edge k3-weighted EXAFS spectra of U(VI)-loaded COF-IHEP1 (A/orange fit) including the best theoretical fits, and autunite (B/green fit) as a reference. Right: corresponding Fourier transforms (not corrected for phase shift). Download figure Download PowerPoint Considering our overall results, it is apparent that U(VI) sorption into COF-IHEP1 occurs mainly through binding to phosphonate groups anchored in the COF channels. To further elucidate the detailed coordination structure at the interface, probable binding modes of UO22+ to COF-IHEP1 were predicted from DFT calculations (details in ). The electron exchange and correlation energy were calculated using the Generalized Gradient Approximation-Perdew Burke Ernzerh (GGA-PBE) method. Also, the plane-wave cutoff of 400 eV and the Gaussian electron smearing method with σ = 0.05 eV were used. During all simulations, the U 6s26p65f36d17s2 valence electrons were treated explicitly. To elucidate the influence of the periodic boundary, and model the pore channel, a large p(1 × 1 × 2) supercell with 477 atoms was adopted in all calculations. Figure 4b shows the optimized binding modes of UO22+ in COF-IHEP1. It is apparent that the uranium atom adopted four-coordination to two P=O oxygen atoms from neighboring layers and two oxygen atoms from water molecules. The computed average U–OP=O bond distance is ∼ 2.5 Å. The binding mode of Pu4+ in COF-IHEP1 was also predicted using the same approach. The results in Figure 4c suggest a similar coordination structure of Pu4+ in COF-IHEP1 except that the plutonium atom tended to achieve seven coordination to two P=O oxygen atoms from neighboring layers and five oxygen atoms from water molecules, and exhibited slightly shorter Pu–OP=O bond distances of ∼ 2.4 Å. Extended X-ray absorption fine structure (EXAFS) spectra of U(VI) sorbed into COF-IHEP1 at pH 1 were recorded to verify the simulated binding mode of UO22+ with COF-IHEP1 (