Sulfate Recognition by Persistent Crystalline Capsules with Rigidified Hydrogen‐Bonding Cavities

Abstract

Locked up: The sulfate anion is selectively encapsulated into rigid hydrogen-bonded capsules upon crystallization of a simple and flexible urea-functionalized ligand in the presence of Mg2+ ions and water (see structure; sulfate: space-filling). The resulting framework persists in the presence of different-shaped anions, allowing the shape selectivity of hydrogen-bonding hosts to be evaluated, and providing a simple way to separate the environmentally relevant sulfate from highly competitive mixtures of aqueous anions. Selectivity is a fundamental property of pervasive importance in chemistry and biology as reflected in phenomena as diverse as membrane transport, catalysis, sensing, adsorption, complexation, and crystallization. Although the key principles of complementarity and preorganization governing the binding interactions underlying such phenomena were delineated long ago,1 truly profound designed selectivity has proven elusive, in part because synthetic molecular architectures are neither maximally complementary for binding target species nor sufficiently rigid. Even if a host molecule possesses a high degree of complementarity for a guest species, with almost no exceptions it can distort its structure or even rearrange its conformation altogether to accommodate competing guests. One approach taken to overcome this challenge has been to devise three-dimensional rigid molecules that bind species within complementary cavities. Although rare examples have been reported to demonstrate the principle,1b such cases are not generally of practical utility, because of synthetic inaccessibility and often poor kinetics. Alternatively, more readily synthesized flexible architectures are generally employed,2 but then the challenge becomes one of locking their structures in place. This locking has been somewhat easier to accomplish in self-assembled capsules,3, 4 though other challenges emerge, such as finding means to internally functionalize the cavities of the capsules.5 Taking a cue from natural binding agents6 that derive their rigidity from a network of molecular interactions, especially hydrogen bonding, we present herein an example of a crystalline capsule that recognizes sulfate by a highly complementary array of rigidified hydrogen bonds. We have been employing crystalline host environments functionalized with anion-coordinating groups as a means to obtain maximal three-dimensional complementarity and rigidity.7 In the present study, we focused on the problem of sulfate recognition and separation,8, 9 motivated in part by its high relevance to environmental remediation and nuclear-waste cleanup10 but also by the challenge of overcoming the Hofmeister bias that disfavors transfer of densely charged, highly hydrated anions.7d The crystalline capsules described herein display hydrogen-bonding cavities that are highly complementary to tetrahedral divalent oxoanions, and remarkably persistent in the presence of other oxoanions of various shapes and basicities, despite being made from rather simple and flexible building blocks. This system thus presents a rare opportunity to examine anion selectivity under the ideal condition of a constant hydrogen-bonding host environment, and, at the same time offering a potentially practical solution for sulfate separation from highly competitive aqueous salt solutions. The tripodal tris-urea ligand L1 1 was selected for this study based on analogous urea derivatives of tris(2-aminoethyl)-amine that exhibited recognition of sulfate in solution and the crystalline state.9a, 11 Notably, the m-CN-Ph substituted tris-urea ligand encapsulated sulfate in a 2:1 complex when incorporated into a silver coordination polymer, with the resulting formation of 12 hydrogen bonds representing the highest coordination number reported for SO42− ions in a synthetic receptor.9a However, selective and efficient separation of sulfate from aqueous solutions with this or other analogous crystalline frameworks proved elusive so far owing to the low stability and relatively high solubility of the frameworks in water. We reasoned that functionalization of this tripodal scaffold with stronger coordinating groups, such as pyridines, would facilitate the incorporation of this anion receptor into more robust crystalline frameworks with superior sulfate-separation abilities. Crystallization of L1 with various MSO4 salts (M=Zn, Cd, Co, Mg) from 1:1 H2O/MeOH solutions afforded single crystals with the composition MSO4(L1)2(H2O)6, (1) as determined by X-ray diffraction and elemental analysis. Crystal-structure determination showed that, regardeless of M, all the solids are isostructural with the same space group (P21/n) and very similar unit cell and geometrical parameters (Supporting Information). The following study was therefore solely focused on the Mg2+ system (1 a), which appeared most practical for sulfate separation. The structure of 1 a12 consists of centrosymmetric SO42−(L1)2 anionic units with the sulfate encapsulated by 12 hydrogen bonds from six urea groups (Figure 1 a), which may be considered the optimal coordination number for sulfate.9a The SO42− ion coordination geometry deviates however from the ideal C3 symmetry predicted by molecular modeling (one chelating urea for each of the six O-S-O edges),13 with the most notable difference that one of the urea groups binds an O vertex rather than an O-S-O edge of the sulfate. This situation is most likely the result of the centrosymmetric geometry imposed by the crystal symmetry, which also causes the sulfate to be rotationally disordered over two positions to compensate for its lack of an inversion center. The SO42−(L1)2 complex is incorporated in a hydrogen-bonded capsule by 12 additional hydrogen bonds from six water molecules, each bridging, through its two protons, a pyridine N and a urea CO acceptor (Figure 1 b). The resulting quasioctahedral arrangement of the water molecules around the hydrogen-bonded capsules matches the coordinating geometry of the metal cations, present as Mg2+(H2O)6 hydrates, which leads to the formation of a three-dimensional framework with distorted NaCl topology (Figure 1 c,d). This structure thus represents a rare example of a hydrogen-bonded network with octahedral connectivity.14 We note that the sulfate encapsulation is dependent on the formation of this crystalline framework, as only a 1:1 SO42−/L1 complex could be detected in solution (CH3OD/D2O, 1:1) by an NMR Job's plot analysis (Supporting Information), in agreement with other solution binding studies of similar receptors.11 This result reinforces the strength and uniqueness of the crystallization approach to anion recognition and separation, which allows exclusive binding modes that are otherwise difficult to achieve in solution. Crystal structure of 1 a. a) Sulfate encapsulation by 12 hydrogen bonds (dashed lines) from six urea groups; yellow S, red O, blue N. b) Hydrogen-bonded capsule assembled from two L1 (stick model), one SO42− ion (space filling model), and six bridging water molecules (stick model). c) A slice of the hydrogen-bonded framework obtained by self-assembly of the anionic capsules with Mg2+ ions (green). d) 3D representation of the octahedral framework with NaCl topology; magenta SO42−, green Mg2+(H2O)6, gray L1. Attempts to crystallize similar frameworks under the same conditions from monocations such as Na+ or K+ in the presence of different monoanions of various shapes and basicities, such as ClO4−, NO3−, HCO3−, HSO4−, and H2PO4−, resulted in no crystals. Likewise, monoanionic MgX2 salts, where X=F−, Cl−, Br−, I−, ClO4−, and NO3−, failed to produce any crystals in the presence of L1 under similar conditions. It thus appears that the presence of a dication/dianion pair, which provides a stronger electrostatic stabilization, is critical for the formation of 1 from water. This feature allowed us to separate sulfate from excess nitrate by selective crystallization of 1 a. Indeed, competitive crystallization experiments indicated that 1 a could be isolated in high yield (88 %) from a 100-fold excess of NaNO3,15 as determined by FTIR and elemental analysis (Supporting Information). This promising result is particularly relevant to the problem of nuclear-waste cleanup, where it is desirable that sulfate be removed from nitrate-rich nuclear wastes prior to their vitrification, owing to the low solubility in the glass and the corrosive nature of SO42−.10 It is also a striking demonstration of the power of crystallization techniques to effect essentially infinite selectivity for the crystal-forming species among competing species. To assess the sulfate selectivity against more competitive anions of similar charge, we investigated the ability of SO32−, CO32−, and SeO42− to form similar hydrogen-bonded frameworks in the presence of Mg2+ and L1. While the first two anions have different shapes than SO42− (trigonal-pyramidal and trigonal-planar, respectively), SeO42− has similar tetrahedral shape but is slightly larger than sulfate (rXO=1.49 and 1.65 Å for X=S and Se, respectively).16 This series is thus ideal for probing the shape and size selectivity of 1 a (taking into account the differences in anion basicity and hydration). All three anions were found to form hydrogen-bonded frameworks that are isostructural to 1 a, with the same general formula of MgX(L1)2(H2O)6 (X=SO32− (1 b), CO32− (1 c), SeO42− (1 d)), same space group, and slightly different unit-cell parameters.17 To assess the sulfate selectivity against these anions, we performed pairwise competitive crystallization experiments with 1:1 molar mixtures of MgSO4 and Na2X salts (X=SO32−, CO32−, SeO42−) in the presence of two equivalents of L1, and analyzed the anionic compositions of the precipitated solids by FTIR spectroscopy and elemental analysis (Supporting Information). Qualitative analysis of the infrared spectra (Figure 2) clearly indicates that sulfate is selectively separated in each case. Table 1 summarizes the selectivities observed as determined quantitatively by elemental analysis. FTIR analysis of the competitive crystallization experiments. a) SO42−/SO32− competition: 1 a is shown in green, 1 b in blue, and the solid crystallized from MgSO4/Na2SO3 in red. b) SO42−/CO32− competition: 1 a is shown in green, 1 c in red, and the solid crystallized from MgSO4/Na2CO3 in blue. c) SO42−/SeO42− competition: 1 a is shown in green, 1 d in blue, and the solid crystallized from MgSO4/Na2SeO4 in red. Anion (X)/crystal Sep. factor (α)[a] Δ(ΔElat)[b] PA[c] ΔG0 h(X)[d] SO42−/1 a - 0 −1889 −1080 SO32−/1 b 372 75.3 −1971 −1295 CO32−/1 c 53 58.7 −2042 −1315 SeO42−/1 d 5 33.8 −1863 −900 As can be seen from Table 1, the competition experiments established the following anion selectivity order: SO42−>SeO42−>CO32−>SO32−. This trend differs from the orders of hydrogen-bond acceptor abilities of the anions (estimated from the gas-phase proton affinities): CO32−>SO32−>SO42−>SeO42−, or of the Hofmeister series (based on ΔGh): SeO42−>SO42−>SO32−>CO32−. Rather, the observed anion selectivity order agrees with the relative crystal stabilities of 1 a>1 d>1 c>1 b, as found from lattice-energy calculations using the experimental crystal structures (Supporting Information). For comparison, the selectivity expected from crystallization of the corresponding Mg salts is: MgCO3>MgSO3>MgSO4,19 which follows the increasing anion radii order. The combined experimental and computational data thus indicate peak selectivity for sulfate against both stronger bases, such as carbonate and sulfite, and less hydrophilic, larger anions, such as selenate. The remarkable selectivity for SO42− ions appears to be determined both by the nearly ideal complementarity of the binding site for sulfate and the rigidity of its hydrogen-bonding environment. First, the complementarity provides for shape and size recognition through the urea functional groups in 1 a, which, as shown in Figure 3 a, completely encapsulate the sulfate with an array of favorably positioned hydrogen bonds.20 In direct contrast, the CO32− and SO32− ions, with their different shape, do not fit very well within these hydrogen-bonding cavities, as they require a different geometric arrangement of the urea groups for optimal binding.13 In fact, these anions display repulsive interactions (shown in red in Figure 3 a) between one NH proton and the positively charged C and S atoms, with contact distances of 2.66 and 2.75 Å that are significantly shorter than the corresponding sums of the van der Waals radii of 2.90 and 3.00 Å, respectively. Such contacts are absent in the case of the tetrahedral SO42− and SeO42− ions because of the effective shielding by the fourth O atom in these anions. Second, L1 shows remarkable rigidity when constrained within these water-bridged crystalline frameworks (Figure 3 b,c), despite its otherwise flexible structure. Such rigidity prevents the receptor from adjusting its geometry to optimize the binding of the competing anions, which contributes to the observed discrimination between sulfate and selenate despite their nearly identical shape and size. This system is thus reminiscent of protein receptors, which despite their inherent flexibility can achieve remarkable selectivities by forming well-defined and rigid binding sites, whose structural integrity is dependent on the presence of multiple noncovalent interactions and the inclusion of water. It is also important to note that the sulfate recognition in 1 a is distinct from anion templating21 or induced-fit phenomena, since virtually identical structures are obtained with the other anions despite the wide variation in their size and shape. This structural persistence is apparently a result of the hydrogen-bonded water bridges that rigidify the framework in much the same way as found in calixarene capsules sustained by incorporated water molecules.3c a) Anion binding sites in 1 a (left), 1 b (middle), and 1 c (right). Hydrogen bonds are shown as black dotted lines and repulsive interactions in 1 b,c as red dotted lines. b) Overlay of 1 a (magenta) and 1 d (green). Mg(H2O)6 cations were removed for clarity. c) Overlay of the L1 structures in 1 a–d. We have illustrated herein a striking example of anion recognition and separation based on crystallization of rigid hydrogen-bonded capsule frameworks made from a simple, flexible ligand functionalized with complementary urea groups, which allowed us to selectively separate the environmentally relevant sulfate from highly competitive mixtures of aqueous anions. The unusual persistence of a common framework structure in the presence of different anions of various sizes, shapes, and basicities provided a unique opportunity for assessing anion selectivities in hosts with a virtually constant hydrogen-bonding environment. While we specifically targeted sulfate separation in this particular study, this approach could be applicable to separation of other guest molecules or ions through appropriate crystal design. Supporting information for this article is available on the WWW under http://www.wiley-vch.de/contents/jc_2002/2008/z704937_s.pdf or from the author. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.