Naphthalene-Pillared Benzene Triimide Cage: An Efficient Receptor for Polyhedral Anions and a General Tool for Probing Theoretically-Existing Anion-π Binding Motifs

Abstract

Open AccessCCS ChemistryRESEARCH ARTICLE5 Aug 2022Naphthalene-Pillared Benzene Triimide Cage: An Efficient Receptor for Polyhedral Anions and a General Tool for Probing Theoretically-Existing Anion-π Binding Motifs De-Hui Tuo, Yu-Fei Ao, Qi-Qiang Wang and De-Xian Wang De-Hui Tuo Beijing National Laboratory for Molecular Sciences, CAS Key Laboratory of Molecular Recognition and Function, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190 University of Chinese Academy of Sciences, Beijing 100049 , Yu-Fei Ao Beijing National Laboratory for Molecular Sciences, CAS Key Laboratory of Molecular Recognition and Function, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190 University of Chinese Academy of Sciences, Beijing 100049 , Qi-Qiang Wang *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] Beijing National Laboratory for Molecular Sciences, CAS Key Laboratory of Molecular Recognition and Function, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190 University of Chinese Academy of Sciences, Beijing 100049 and De-Xian Wang *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] Beijing National Laboratory for Molecular Sciences, CAS Key Laboratory of Molecular Recognition and Function, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190 University of Chinese Academy of Sciences, Beijing 100049 https://doi.org/10.31635/ccschem.021.202101366 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesDownload CitationsTrack Citations ShareFacebookTwitterLinked InEmail A triangular prism cage 2 with benzene triimide (BTI) as the base and 2,7-naphthalene dimethylene as the supporting pillars was designed and synthesized efficiently. The two parallel BTI planes constitute a cavity of 7.58 Å which allows the inclusion of anions of various sizes. So far, cage 2 has shown the strongest binding affinity to a series of polyhedral (tetrahedral and octahedral) anions including PF 6 − (24651 M−1), CIO 4 − (28234 M−1), CH 3 SO 3 − (30571 M−1), BF 4 − (31613 M−1), and HSO 4 − (84623 M−1). The obtained 11 crystal structures of 2⊃X− complexes with different anions demonstrated that multiple and cooperative anion-π interactions contribute synergistically to strong complexation. The series of complex structures systematically suggested that cage 2 is a good, general tool for probing anion-π binding motifs that have only been predicted in theoretical calculations. For anions of triangular ( NO 3 − ) and polyhedral ( BF 4 − , ClO 4 − , PF 6 − ) shapes, the as-predicted most stable anion-π binding motifs with charge-neutral π systems were experimentally observed for the first time. Download figure Download PowerPoint Introduction Noncovalent interactions involving aromatic rings are ubiquitous in chemical and living systems.1,2 In particular, as a young noncovalent driving force, diverse applications of anion-π interactions have been recently demonstrated, including anion recognition, anion-directed self-assembly, stimulus-responsive aggregation and catalysis.3–14 Though still largely underexplored compared to cation-π interactions, the importance of anion-π interactions in biocatalysis, biomolecular recognition and in the structure of macromolecules is increasingly recognized.15–25 One distinguished property of anions is their wide shape variety, from spherical, linear to triangular and polyhedral. However, the strength of anion-π interactions involving polyhedral, tetrahedral, or octahedral anions is generally weak owing to the large size and low electron density of these anions.26–28 On the other hand, the polyhedral shape of anions would, in principle, lead to diverse binding motifs when the anion approaches above an aromatic π surface. Indeed, theoretical calculations suggest that different binding motifs (motifs a– c in Figure 1) exist, with different binding energy, binding geometry, and anion-π interacting distance.29–31 Among them, motif c is suggested to be most energetically favorable. However, while the less stable motifs a and b have been frequently observed in crystal structures of complexes26–28,32–36 and in biological anion-π pairs,21 the most stable motif c has barely been found. In the meantime, the subtle difference in the binding motifs could cause a large influence on relevant functions, for example, catalysis37,38 or macromolecular structuring.21 Therefore, highly efficient anion-π binding of polyhedral anions can be realized and their structural nature are important problems to be addressed. To approach the answer, establishment of an ideal anion-π host and understanding of the comprehensive structural relevance between anions and associated aromatic rings is highly desirable. Figure 1 | (a–c) Predicted different anion-π binding motifs for polyatomic anions. Download figure Download PowerPoint Here, we aim to provide a wide-scope molecular tool to quantify and picture the different binding behavior in anion-π interacting systems. The current design is based on the use of benzene triimide (BTI)34 and the theoretical guidance for ion-π interactions,39 following three considerations: (1) BTI bearing a high quadruple moment (QZZ = 14.5 B) enables enhanced anion-π interactions; (2) the extending surface of BTI provides multiple binding sites (both the six- and five-membered rings are available to accept an anion), which should be advantageous for the accommodation of differently-shaped and polydentate anions; (3) appropriate separation between the two BTI planes is crucial: a small distance is good for achieving high specificity toward given anions like N 3 − as previously demonstrated with small cage 1,34 while an excessively large distance would weaken the additivity or cooperativity. Thus, a triangular prism cage 2 with BTI as the base and 2,7-naphthalene dimethylene as the spacer is designed (Figure 2). This cage is estimated to form a cavity of ∼7.6 Å as determined by the rigid 2,7-naphthalene methylene spacer. Thus, it allows inclusion of anions with a wide range of radius and provides a wide-scope molecular probe. Figure 2 | Design of BTI-based molecular cage 2. Download figure Download PowerPoint Experimental Methods Detailed synthetic procedures are listed in the Supporting Information (see Supporting Information Schemes S1 and S2). The compounds were characterized by 1H NMR, 13C NMR, high-resolution electrospray ionization (ESI) mass spectroscopy, elemental analyses, and single-crystal X-ray diffraction (see Supporting Information Figures S1 and S45–S50). UV–vis titrations of cage 2 and anions were carried out through continuous addition of anion solution to the cuvette containing a solution of cage 2 (1.0 × 10−3 M, 2 mL in corresponding solvent) (see Supporting Information Figures S14–S19). 1H NMR titrations of cage 2 and anions were carried out through continuous addition of anion solution to the NMR tube containing a solution of cage 2 (1.0 × 10−3 M, 500 μL in N,N-dimethylacetamide (DMA)-d9) (see Supporting Information Figures S21–S32). The crystals of [ 2⊃X]− complexes were cultivated through slow diffusion of tetrahydrofuran (THF), acetone, or 1,1,2,2-tetrachloroethane (TCE) into host–guest solution in DMA or acetone. The crystal data are summarized in the Supporting Information (see Supporting Information Figures S2–S12 and Table S2). Results and Discussion Synthesis, structure, and cyclic voltammetry The synthesis of cage 2 is quite straightforward (Scheme 1). The reaction between parent BTI 3 and 2,7-bis(bromomethyl)naphthalene 4 in N,N-dimethylformamide (DMF) with diisopropylethylamine (DIPEA) as the base through either a one-pot (Route I) or stepwise pathway via intermediate 5 (Route II) afforded the target cage product. For comparison, a tri(naphthalen-2-ylmethyl) substituted BTI fragment compound 6 was also synthesized through substitution of 3 with 2-(bromomethyl)naphthalene (see Supporting Information Scheme S2). Scheme 1 | Synthesis of cage 2. Download figure Download PowerPoint The cage 2 shows good crystallization property. Placing a solution of 2 in dimethylsulfoxide (DMSO) (>1 mM) at room temperature afforded single crystals of good quality. X-ray analysis revealed that the top and bottom BTI planes are parallel, facing each other in a fully eclipsed fashion and separated by 7.58 Å. The three naphthalene planes are perpendicular to BTI planes, with their lower rims directed inward toward the cavity (Figures 3a and 3b). The electron-deficient BTI and electron-rich naphthalene moieties form intermolecular π–π stacking interaction, resulting in ordered self-assembly in solid state (see Supporting Information Figure S1). It is worth noting that the larger cavity of 2 compared to 1 permits inclusion of several solvent molecules such as DMSO, showing that solvent molecules can be strongly competitive species in anion complexation in solution. Figure 3 | Crystal structure of cage 2. (a) Top view, the cavity is occupied by DMSO molecules and (b) side view. Hydrogen atoms were omitted for clarity. Download figure Download PowerPoint As electron deficiency is crucial for anion-π binding, the cyclic voltammetry (CV) of cage 2 was measured (see Supporting Information Figure S13). In DMF or DMSO containing n-Bu4NPF6 (0.1 M) as the electrolyte, two irreversible redox couples were recorded. The first reduction potential (E1/2 = −0.96 V) is almost identical to the single BTI fragment 6, indicating the unaltered electron-deficiency of BTI after assembled into cage. The existence of two irreversible redox couples looks strange at first glance and is quite different from that of the three and six reversible redox couples observed for BTI fragment 6 and cage 1, respectively (see Supporting Information Figure S13).34 This irregular redox behavior could be caused by inclusion of electrolyte anion PF 6 − inside the large cavity of 2 (see below binding study). Anion binding in solution To evaluate the authentic recognition ability of cage 2 in solution, the first question to be addressed is whether the vast solvent molecules compete with the guest anions to occupy the cavity. The inclusion of DMSO molecules in the cavity was indeed observed in the crystal (vide supra). Hence, different solvents of varied molecular size and shape, including dichloromethane (DCM), DMSO, DMF, TCE, and DMA, were investigated (see Supporting Information Figures S14–S19). N3−, I−, and SCN− (as tetrabutylammonium salts) were taken as representative guest anions. Anion binding was monitored by UV–vis titrations. As summarized in Supporting Information Table S3, in DCM and DMSO, the binding constants (Ka) of these anions are generally small (Ka < 10 M−1). Increased but moderate binding (10∼100 M−1) was obtained in medium-sized solvents, like DMF and TCE. Pleasantly, dramatically enhanced binding, 14878 M−1 for SCN−, 2220 M−1 for I−, and 599 M−1 for N3−, was obtained in DMA at a larger size. This suggested solvent has a pronounced effect and DMA is the best choice to provide information about independent complexation. In addition, the use of different anion salts (tetrabutylammonium and sodium salts) was compared to evaluate the countercation effect (see Supporting Information Figure S15 and Table S4). Similar binding was observed for the two salts, suggesting a negligible solvation and ion pairing effect. Using DMA as solvent, the binding of cage 2 with a series of anions (as tetrabutylammonium salts), including spherical Cl−, Br−, I−, linear N 3 − , SCN−, triangular NO 3 − , AcO−, tetrahedral BF 4 − , ClO 4 − , HSO 4 − , CH 3 SO 3 − , and octahedral PF 6 − , was investigated. For all anions tested, a UV–vis spectral response was observed, indicating general host–guest interaction (Figures 4a and 4b and Supporting Information Figure S14). For anions such as I−, N 3 − , SCN−, AcO−, a new charge transfer (CT) band at 400–500 nm, along with a visible color change of the solution was observed. For Br−, only a weak charge transfer band and no visible color change was observed. For the remaining anions including Cl−, NO 3 − , BF 4 − , ClO 4 − , HSO 4 − , CH 3 SO 3 − and PF 6 − , the above charge transfer band was not observed, but instead with a sole hypochromic effect at 360–450 nm. The different spectral responses probably imply different locations of anions over the BTI surface. The noncovalent anion-π binding motif where the anion stays right above the BTI centroid should result in little orbital mixing and thus negligible charge transfer, while σ-type interactions where the anion resides over the periphery of the aromatic ring is likely to facilitate charge transfer.40 Figure 4 | UV–vis titration spectra of cage 2 (1.0 × 10−3 M in DMA) upon addition of (a) n-Bu4NSCN, (b) n-Bu4NPF6 (Insets are job's plot with total concentration of 1.0 × 10−3 M) and 1H NMR titration spectra of cage 2 (1.0 × 10−3 M in DMA-d9) upon addition of (c) n-Bu4NCl, (d) n-Bu4NPF6. Download figure Download PowerPoint The binding constants for various anions were determined by fitting UV–vis spectrometric titration data with 1∶1 stoichiometry (determined by Job's plot and ESI-MS, Figures 4a and 4b, Supporting Information Figures S33–S44) (Table 1). Gratifyingly, cage 2 shows a particularly high binding affinity toward polyhedral anions (tetrahedral and octahedral), despite their large size and relatively low electron density. The binding constants are in the range of PF 6 − (24651 M−1), ClO 4 − (28234 M−1), CH 3 SO 3 − (30571 M−1), BF 4 − (31613 M−1) to HSO 4 − (84623 M−1), representing the strongest anion-π complexation for a single polyhedral anion. In contrast to the generally high affinity for polyhedral anions, the binding strength with linear anions is strongly dependent. While thiocyanate (SCN−) is efficiently bound (14878 M−1), N 3 − gives only moderate binding (599 M−1). For spherical halides, binding is clearly affected by the size. Iodide is strongly complexed (2220 M−1), bromide is only moderate (58 M−1), and no credible binding was observed for chloride. The complexation with triangular anions such as AcO− and NO 3 − is weak to moderate (73 and 604 M−1, respectively). Table 1 | Association Constants (Ka, M−1) of Cage 2 with Various Anionsa Cl− Br− I− N 3 − SCN− AcO− N.D. 58 ± 1 2220 ± 10 599 ± 1 14878 ± 91 73 ± 1 NO 3 − BF 4 − ClO 4 − HSO 4 − CH 3 SO 3 − PF 6 − 604 ± 3 31613 ± 259 28234 ± 1177 84623 ± 2970 30571 ± 1596 24651 ± 858 aDetermined by fitting UV–vis titration data with Hyperquad program. Tetrabutylammonium salts were used. The solvent is DMA. Considering the large cavity of 2 and the extending BTI surface, the size match between the cavity and the anion is most likely the dominant factor determining binding strength (see Supporting Information Figure S20). When the anions are too small, lower than 1.9 Å ( NO 3 − , Cl−, Br−), the binding is weak as it could not make use of the cooperativity or additivity of the two BTIs (the single BTI is a weak anion receptor as previously demonstrated34). For large and polyhedral anions such as HSO 4 − (2.21 Å), BF 4 − (2.30 Å), ClO 4 − (2.40 Å), and PF 6 − (2.95 Å), the strong binding can be ascribed to a matched size and multiple, cooperative anion-π interactions. When the anion is too large, like PF 6 − , it would start to suffer from hindrance disadvantage when cramped within the cavity, and thus the binding is lowered. The size effect is also obvious in halide binding which follows the order I− (2.20 Å) > Br− (1.88 Å) > Cl− (1.81 Å). Similarly, the dramatically decreased binding of N 3 − (1.95 Å) in the large cavity of 2 (7.6 Å, 599 M−1) than in the small cage 1 (5.0∼6.2 Å, 11098 M−1),34 and the greater binding of 2 with SCN− than with N 3 − (14878 M−1 vs 599 M−1) are also size effects. Besides, for anions with similar size, the polyatomic nature probably leads to more efficient interaction with the extended BTI surface. For example, binding of NO 3 − (1.79 Å, 604 M−1) is stronger than Cl− (1.81 Å, N.D.) and SCN− (2.13 Å, 14878 M−1) is stronger than I− (2.20 Å, 2220 M−1). The interaction between cage 2 and anions was further investigated by 1H NMR (Figure 4c and 4d and Supporting Information Figures S21–S32). Upon the addition of PF 6 − , two sets of distinct signals corresponding to the free host and the host-anion complex arose, reflecting a slow-exchange process (Figure 4d). This suggested that the PF 6 − anion is probably tightly trapped within the cage cavity and its entry/release is slow under NMR timescale. The upfield shift of the inward proton signal (Ha) and slight downfield shift of the outward protons (Hb, Hc) on naphthalene spacers could be caused by the shielding/deshielding effect of the included PF 6 − .41 A similar slow-exchange process was also observed for I−, SCN−, BF 4 − , ClO 4 − , and CH 3 SO 3 − (see Supporting Information). On the other hand, anions such as HSO 4 − , Cl−, Br−, N3−, and NO 3 − only caused gradual downfield shifts of Ha and upfield shifts of Hb, Hc, and Hd, exhibiting a fast-exchange process. In addition to the size-dependent exchange behavior under NMR timescale, it is interesting to note that different shift trends of signal Ha were observed. Larger anions such as BF 4 − , ClO 4 − , PF 6 − , and CH 3 SO 3 − caused a shielding effect on proton Ha due to their dispersive electron density, while the smaller anions with compacted electron density tend to form weak hydrogen bonds with the proton. Accordingly, upfield and downfield shifts of the signal were resulted, respectively. Anion binding in solid state The systematic and insightful structural information on anion-π interactions was obtained by X-ray crystal diffraction. High-quality single crystals of complexes of 2 with 11 different anions were cultivated through slow diffusion of THF or acetone into host–guest solutions in DMA. The crystal parameters and refinement are summarized in Supporting Information Tables S1 and S2. Crystal structures with spherical halides Cl−, Br−, I− The structures of 2⊃X− (X = Cl, Br, I) are depicted in Figures 5a–5c. The three complexes show a similar sandwich structure, and the anion separates the two BTI planes almost equally. To fit the binding, the two BTIs change from a fully eclipsed to slightly staggered the intermolecular interaction to be by π–π stacking of than (see Supporting Information Figures the similar binding subtle in interaction for different halide anions. for example, resides above of the benzene to the BTI planes at about Å and Å). stays the periphery of benzene ring and approaches of the an distance of Å to the two Accordingly, Cl− is to of the naphthalene resulting in a weak hydrogen in a between that of and and in with its size. The σ-type binding the of charge transfer, which could the observed for and The of the band for is probably due to its small size and electron density. Figure 5 | Crystal structures of (a) (b) and (c) Hydrogen atoms and were omitted for clarity. Download figure Download PowerPoint Crystal structures with linear SCN− and N 3 − The complex structures of 2 with linear anions SCN− and N 3 − are depicted in Figures and the that anions between the two BTI planes, the structural are quite in the perpendicular to the of the cage and with It a = than a linear as observed in a anion-π The over the benzene while and atoms on the Hydrogen between the SCN− and sites of naphthalene is In contrast to the observed in N 3 − a low anions are from the and only with One resides over the and the other two the periphery of benzene Besides, and form hydrogen bonds with the sites of A of the two complex structures that cooperative anion-π interactions from the two BTIs most likely the stronger binding of SCN− over N 3 − The smaller N 3 − to with only of the BTI planes, while the larger SCN− the cavity and cooperative anion-π binding with BTI planes, an enhanced binding This was also by stronger binding of N 3 − by cage 1 (Ka = 11098 than by cage 2 (599 M−1), where the smaller cavity of 1 permits cooperative anion-π binding for small N 3 − by Figure 6 | Crystal structures of (a) and (b) N 3 − Hydrogen atoms and were omitted for clarity. Download figure Download PowerPoint Crystal structures with triangular NO 3 − The NO 3 − structure is depicted in Figures and In contrast to the observed anion-π motifs a and b (Figure complex a into the cage cavity with its parallel to the of the cage. bottom atoms on BTI and the to the other BTI anion-π interactions. In such a binding anion-π in the range of Å are the to NO 3 − in which motif is it is that the right cavity of 2 the of the binding motif and the stronger binding than Figure | and Crystal structure of NO 3 − Hydrogen atoms and were omitted for clarity. Download figure Download PowerPoint Crystal structures with tetrahedral and octahedral BF 4 − , ClO 4 − , PF 6 − The complex structures of 2 with polyhedral anions BF 4 − , ClO 4 − , PF 6 − are depicted in Figures In the three anions are included the of the cavity. noncovalent anion-π interactions and c binding motif generally BF 4 − shows a high with of the anion almost to that of the cage. The between the of the two side benzene atoms at side over BTI with to at Å, and the approaches to the centroid of the other BTI at Å. Besides, between and naphthalene protons suggested the of hydrogen Similar interactions in ClO 4 − , for a slight of the of the anion from that of the cage. In the of PF 6 − , the PF 6 − almost right on the of the cage. anion-π interactions corresponding to c binding motifs are In addition, multiple weak hydrogen bonds are also It is worth that BF 4 − , ClO 4 − , and PF 6 − are as anions and as independent in recognition Here, the multiple, cooperative anion-π interactions provide a to the strong binding for these anions. Besides, the noncovalent anion-π interactions in these complexes imply less orbital The host–guest is also can the of the band and the observed hypochromic effect in solution. Figure | Crystal structures of (a) BF 4 − (b) ClO 4 − and (c) PF 6 − Hydrogen atoms and were omitted for clarity. Download figure Download PowerPoint Crystal structures with HSO 4 − and CH 3 SO 3 − The anion-π binding in HSO 4 − is different to other tetrahedral anions, as shown in Figures and b binding motifs are suggested by two atoms interacting with each BTI The to range from to Å, representing quadruple interactions. atoms form weak hydrogen bonds with the In CH 3 SO 3 − , the SO 3 − into the cage cavity, the One resides over the and the other two is on the periphery of the benzene The to at Å interactions. Besides, the three atoms form hydrogen bonds with the sites of Figure | Crystal structures of (a) CH 3 SO 3 − and (b) HSO 4 − Hydrogen atoms and were omitted for clarity. Download figure Download PowerPoint have demonstrated that the 2,7-naphthalene BTI molecular cage 2 is an efficient host for polyhedral anions. The binding constants in the range of to M−1 the strongest anion-π binding of charge-neutral π toward an polyhedral anion. and crystal structures clearly suggest that multiple and cooperative anion-π interactions synergistically contribute to the strong complexation. cage 2 provides a general tool to anion-π binding motifs for anions with Though the binding associated with different anion-π motifs for polyatomic anions has been in theoretical the influence of different anion-π motifs on binding has not been Here, we provide for the first the of the most stable anion-π motifs involving charge-neutral and that the stable anion-π motif be important factor to the strong binding ability of 2 toward polyhedral anions. that can applications in the highly efficient recognition of polyhedral, anions that is for also provides new to whether and different anion-π motifs can functions, for example, or state in the or structural of when anion-π are Supporting Information Supporting Information is available and 1H NMR and UV–vis electrospray ionization mass of NMR spectra of all the data for data of N 3 − , NO 3 − , HSO 4 − , CH 3 SO 3 − , BF 4 − , ClO 4 − and PF 6 − of is no of to Information This was by National of and and Chinese Academy of in and and 2. π for Wang Wang Anion-π and in of Anion-π The of in and Design for Anion-π in Crystal with Anion-π and of a 1, of Anion-π in Using in Anion-π into for between Anions and Binding of Anion-π to Anion-π in The of in 6, in a in of in the of by Anion-π in of Anion-π to the of of and in the for Anion-π of in Anion-π the Tuo Wang Wang Ao Wang Wang of Anion-π for Anions with Wang Wang Binding and Wang of for Anion and Anion-π and for of Anion-π of with and as Anion-π Receptor for and and of and of Using with Wang Wang and of and with Anions in the Tuo Ao Wang Wang Triimide Cage as a of the Binding of a Cage by Anion-π as in of with Anion-π to on Ao Wang Wang Anion-π for Efficient and with Molecular into the and Design for for the Design of Anion The of with Wang of and with of Anion-π Information Chinese cage

Topics & Concepts