Luminescent Anion Sensing by Transition‐Metal Dipyridylbenzene Complexes Incorporated into Acyclic, Macrocyclic and Interlocked Hosts

Abstract A series of novel acyclic, macrocyclic and mechanically interlocked luminescent anion sensors have been prepared by incorporation of the isophthalamide motif into dipyridylbenzene to obtain cyclometallated complexes of platinum(II) and ruthenium(II). Both the acyclic and macrocyclic derivatives 7⋅Pt, 7⋅Ru⋅PF6, 10⋅Pt and 10⋅Ru⋅PF6 are effective sensors for a range of halides and oxoanions. The near‐infra red emitting ruthenium congeners exhibited an increased binding strength compared to platinum due to the cationic charge and thus additional electrostatic interactions. Intramolecular hydrogen‐bonding between the dipyridylbenzene ligand and the amide carbonyls increases the preorganisation of both acyclic and macrocyclic metal derivatives resulting in no discernible macrocyclic effect. Interlocked analogues were also prepared, and preliminary luminescent chloride anion spectrometric titrations with 12⋅Ru⋅(PF6)2 demonstrate a marked increase in halide binding affinity due to the complementary chloride binding pocket of the [2]rotaxane. 1H NMR binding titrations indicate the interlocked dicationic receptor is capable of chloride recognition even in competitive 30 % aqueous mixtures.


C NMR and HRMS spectra General Considerations
Commercial grade chemicals and solvents were used without further purification. Where anhydrous solvents were used, they were degassed with N 2 and passed through an MBraun MPSP-800 column Where degassed solvents were used, they were degassed via bubbling of N 2 gas through the solution unless stated otherwise.
Triethylamine was distilled from and stored over potassium hydroxide. De-ionised water dispensed from a Millipore Milli-Q purification system was used in all cases. Tetrabutylammonium (TBA) salts were stored under vacuum in a desiccator. Microwave reactions were carried out using a Biotage Initiator 2.0 microwave. All synthetic procedures have been reliably repeated multiple times. Routine 300 MHz NMR spectra were recorded on a Varian Mercury 300 spectrometer, 1 H NMR operating at 300 MHz, 13 C{ 1 H} at 76 MHz, 19

F at 283
MHz and 31 P at 121 MHz . Where the solubility of the compounds were too low, or not enough compound existed, a Bruker AVII500 with 13 C Cryoprobe spectrometer was used for obtaining 13 C{ 1 H} at 126 MHz, however in some cases a compete 13 C{ 1 H} spectrum could not be obtained. All 500 MHz 1 H Spectra and all 1 H NMR titrations were recorded on a Varian Unity Plus 500 spectrometer. All chemical shift (δ) values are given in parts per million and are referenced to the solvent. In cases where solvent mixtures are used, the main solvent is used as the reference. Where an apparent multiplet (e.g. app. t.) is quoted, J app is given. Low resolution ESI mass spectra were recorded on a Micromass LCT Premier XE spectrometer. Accurate masses were determined to four decimal places using Bruker μTOF and Micromass GCT spectrometers. UV/visible experiments were carried out on a PG instruments T60U spectrometer at 293 K. Steady-state fluorescence spectra were recorded on JobinYvon-Horiba Fluorolog-3 spectrophotometer or a Varian Cary-Eclipse spectrometer.

Luminescent Anion binding titrations
Luminescence experiments were carried out on a Varian Cary-Eclipse spectrometer for the platinum (II) receptors using an excitation wavelength of 320 nm at 293 K. To a 2.5 mL, 1 × 10 -5 M solution of each receptor was added aliquots of the tetrabutylammonium salts dissolved in a stock solution made up with the receptor, such that the same concentration of the host was maintained throughout the titration experiments.
Luminescence experiments were carried out on a Horiba Fluorolog spectrometer for the ruthenium (II) receptors using an excitation wavelength of 530 nm at 293 K. To a 1 mL, 1 × 10 -4 M solution of each receptor was added aliquots of the tetrabutylammonium salts dissolved in a stock solution made up with the receptor, such that the same concentration of the host was maintained throughout the titration experiments. In both cases the titration data was analysed and association constants determined using the SPECFIT program. 10,11

1 H NMR anion titration data
Initial NMR sample volumes and concentrations were 500 μL and 2.0 mM respectively. Solutions (100 mM) of anion were added as their tetrabutylammonium salts. Spectra were recorded at 0, 0.2, 0.4, 0.6, 0.8, 1.0, 1.2, 1.4, 1.6, 1.8, 2.0, 2.5, 3.0, 4.0, 5.0, 7.0 and 10 equivalents. In all cases where association constants were calculated, bound and unbound species were found to be in fast exchange on the NMR timescale. Association constants were obtained by analysis of the resulting data using the WinEQNMR2 computer program. Binding stoichiometry was investigated by visual analysis of the titration data, and using approximations of Job plots.
Estimates for the association constant and the limiting chemical shifts were added to the program's input file.
The parameters were refined by non-linear least-squares analysis using WINEQNMR2 12 to achieve the best fit between observed and calculated chemical shifts. The input parameters for the final chemical shift and association constant were adjusted based on the program output until convergence was reached. Comparison of the calculated and experimental binding isotherms demonstrated that an appropriate model with an appropriate stoichiometry were used.

Single crystal X-ray crystallography
Data was collected at 150 (2)  When using the Nonius machine, a series of ω-scans were typically performed in such a way as to collect every independent reflection to a maximum resolution of 0.77 Å, aiming for 99.5% completeness. Cell parameters and intensity data (including inter-frame scaling) were processed using the DENZO-SMN package. When using synchrotron radiation, ω-scans were performed such that a half-sphere of data was collected to a maximum resolution of 0.77 Å. Cell refinement, data reduction and scaling were performed using the CrystalClear package. 13 The structures were solved by direct methods using the SIR92 software 14 or by charge flipping using Superflip. 15 Structures were refined using full-matrix least-squares on F 2 within the CRYSTALS suite. 16 All non-hydrogen atoms were refined with anisotropic displacement parameters unless specified otherwise. Disordered portions were modelled using refined partial occupancies. Geometric and vibrational restraints were applied where appropriate to ensure physically reasonable models.
In some cases, the molecular structure within solvent voids could not be resolved in the difference map and PLATON SQUEEZE 17 was therefore used to account for the residual electron density in the refinement.
After the construction of a stable, physically reasonable and complete model, the weights were optimised, 18 analogous reflections were omitted and absent high-angle data (in the case of poorly diffracting samples) were pruned using the Wilson plot. This generally led to convergence of the refinement, giving the final structure.
For the more challenging structures, in which the refinement did not converge immediately, initially half-shifts, then restraints and finally rigid body refinement were used to overcome the problem. IUCr CheckCIF/PLATON 19 was used to validate the structures and warnings were dealt with as appropriate or justified using validation reply forms.
Single crystals of were grown from vapour diffusion of methanol into a chloroform solution of the compound over several days.