Self‐Assembled Anion‐Binding Cryptand for the Selective Liquid–Liquid Extraction of Phosphate Anions

Abstract The ligands L1 and L2 form trinuclear self‐assembled complexes with Cu2+ (i.e. [(L1)2Cu3]6+ or [(L2)2Cu3]6+) both of which act as a host to a variety of anions. Inclusion of long aliphatic chains on these ligands allows the assemblies to extract anions from aqueous media into organic solvents. Phosphate can be removed from water efficiently and highly selectively, even in the presence of other anions.

Synthesis of ligand L 2 .
(2) -Ph), 7.56 (t, 3H, J = 7.4, -Ph), 7.42 (t, 6H, J = 7.7, -Ph), 4.35 (m, 3H, cy-CH), 2.44 (mcoincident with solvent, 3H, broad, -CH), 1.59 (q, 3H, J = 11.2 Hz, cy-CH). 13   Synthesis of 2. To a stirred suspension of the tribenzoylated trithiourea (1) (1.88 g, 3.04 mmol) in water (40 mL) sodium hydroxide (1.22 g, 30.4 mmol) was added. The reaction was then heated at 60°C for 12 h, during which a colourless precipitate formed. After this time the mixture was allowed to cool to ambient temperature and the precipitate isolated by vacuum filtration, which was washed with water and diethyl ether, affording the tri-thiourea derivative (2) as a white powder (635 mg, 68% The 1 H NMR contains a series of broad peaks which is common with compounds that contain multiple thiourea units and arises for intra-and inter-molecular interactions. ref Regardless the 1 H NMR contains no starting material and only a minor contamination of benzoic acid and was used without purification in the next step.    and pyridine (0.38 mL, 4.77 mmol) in anhydrous dichloromethane (20 mL) was added dropwise hexanoyl chloride (0.67 mL, 4.77 mmol) under an atmosphere of dinitrogen. This was allowed to stir at room temperature for 2 h, until analysis via TLC (SiO2, 1% methanol in dichloromethane) showed no starting material remained. The mixture was further diluted with dichloromethane (20 mL) and washed with saturated sodium bicarbonate solution (15 mL). The combined organic layer was dried over anhydrous magnesium sulfate and solvent removed under reduced pressure, giving the crude product as a slightly pink oil. This was purified by column chromatography (SiO2, 1% methanol in dichloromethane), affording the product (4) as a colourless oil (820 mg, 83%    Synthesis of L 1a . To a solution of 2-(α-bromoacetyl)pyridine derivative (5)

Ion Chromatography Experiments.
Calibration standards of the concentrations 0.2, 0.4, 0.6, 0.8 and 1.0 mM were prepared by the standard dilution of a 20 mM stock of each mixed salt solution used. These are detailed below: Calibration for Experiment 1. Competitive extraction of common anions (NaCl, NaNO3, NaHSO4 and NaH2PO4).
Calibration for Experiment 2. Competitive extraction of common anions (NaCl, NaNO3, Na2SO4 and Na2HPO4) and total phosphate concentration experiments.
In a typical experiment; To a solution of L 1a (10 mg, 0.010 mmol) and Cu(trif)2 (5.37 mg, 0.015 mmol) in 3% MeOH in DCM (3 mL) and ultrapure H2O (2 mL, 18.2 MΩ-cm) was added 1 mL of a mixed salts solution consisting of NaHSO4·H2O (68.3 mg, 0.495 mmol), NaH2PO4 (59.4 mg, 0.495 mmol), NaCl (29.0 mg, 0.495 mmol) and NaNO3 (42.1 mg, 0.495 mmol) in ultrapure H2O (100 mL, 18.2 MΩ-cm) and this was set to stir at RT for 18 hours. After this time, 2 mL of the aqueous layer was taken and adjusted volumetrically to 5 mL with ultrapure H2O for analysis by IC (theoretical maximum concentration of each anion 0.66mM). The remaining experiments were carried out in an identical manner but using different mixed salts solutions and the details of these are below.
Further experiments (experiments 4 -8) were carried out in an identical manner but using differing amounts of L 1a and L 2a and the resulting change in the stoichiometric amount of Cu(trif)2 to form the complex and the details of these are tabulated below. Experiment no.

Crystallography.
Single crystal X-ray diffraction data was collected at 150(2) K on a Bruker D8 Venture diffractometer equipped with a graphite monochromated Mo(K) radiation source and a cold stream of N2 gas. Solutions were generated by conventional heavy atom Patterson or direct methods and refined by full-matrix least squares on all F 2 data, using SHELXS-97 and SHELXL software respectively. C1 Absorption corrections were applied based on multiple and symmetry-equivalent measurements using SADABS.

Computational Details
Geometry optimisations were carried out using Gaussian09 (Revision D.01) and the default values for the calculation parameters and convergence criteria. M1 Cu centres were describes using the Stuttgart effective core potentials and basis set, M2 whilst the Pople 6-311G basis set was employed for all other atoms, and the anionic moieties. M3 Optimisations were carried out using the B3LYP density functional, M4-8 with dispersion interactions computed through Grimme's D3 parameter set with Becke-Johnson damping. M9 Solvent effects were computed by employing the integral equation formalism polarised continuum model for DCM. M10 All stationary points were fully characterised through analytical frequency calculations as minima i.e. with positive eigenvalues. All energies were recomputed with a larger basis set featuring LANL2DZ for the Cu centers M11 and DEF2TZVP for all other atoms. M12 The optimized geometries are presented below. The formation energy of the complex (comprised of the cage and the anionic moiety) was calculated from the individual empty cage and the ionic moiety as reference; negative values represent a thermodynamically favourable formation.