Homo- and heterobimetallic group 11 and 12 n-heterocyclic carbene complexes: a combined synthetic, spectroscopic and theoretical investigation

Submission note: A thesis submitted in total fulfilment of the requirement for the degree of Doctor of Philosophy In Chemistry to the Department of Chemistry and Physics, School of Molecular Sciences, College of Science, Health and Engineering, La Trobe University, Bundoora.

A series of six homobimetallic Au(I) complexes with bidentate N-heterocyclic carbene (NHC) bridging ligands have been prepared. Bis-imidazolium pro-ligands were prepared by alkylation of bis-imidazole precursors in which the imidazole moieties were linked by different arene groups (benzene, pyridine and pyrazine). The cationic homobimetallic Au(I) complexes were prepared by treating the bis-imidazolium salt with (THT)AuCl in the presence of the weak base NaOAc. Five of the six complexes prepared were sufficiently soluble to be purified and characterised by 1H and 13C NMR spectroscopy and mass spectrometry. Single crystal X-ray crystal structures were obtained for these five complexes (PF6− salts) and one of the complexes as a bromide salt. Two different conformations were observed for the Au(I) complexes in the solid state; complexes of the benzene-linked ligand, adopt only an open (pseudo-planar) conformation with a long Au Au distance (ca. 7 angstrom) while the pyrazine-linked complex was observed to adopt only a twisted (helical-type) conformation with a short Au Au distance (ca. 3.4 angstrom - classified as an aurophilic interaction). Conversely, the pyridine-linked complexes were observed to adopt both open and twisted conformations in the solid state. All complexes which adopt the twisted conformation (pyridine or pyrazine-linked) were found to be emissive in the solid state at RT in the visible region of the spectrum (514, 578 and 579 nm), and one of the complexes which adopts the open conformation was found to be emissive in the solid state at 77 K but at higher energy (398 nm). All of the complexes were non-emissive in solution at RT, however they became emissive upon cooling a solution to form a glassy matrix at 77 K. The benzene-linked complex was emissive at only at high energy (400 nm), the pyrazine-linked complex was emissive at lower energy (475 nm), whereas the pyridine-linked complexes were all found to exhibit dual emission bands (ca. 400 nm and ca. 530 nm). These results show a clear relationship where high energy emission is observed for complexes in the open conformation and low energy emission is seen for complexes in the twisted conformation. In addition, it is hypothesised that each of the high and low energy emission bands observed for the pyridine-linked complexes are associated with the open and twisted conformations, respectively, which have been trapped in the glassy matrix. It is also thought that the benzene-linked complexes are not able to twist into the helical conformation due to an unfavourable steric interaction between the proton which is ortho to the NHC donor groups, and the Au(I) centres. It is likely that this accounts for only the open conformation being observed in the solid state and only high-energy emission observed in glassy matrix. A new synthetic methodology has been developed in order to prepare Group 11 and 12 heterobimetallic complexes. This method was established so that heterobimetallic complexes could be synthesised with the same symmetrical pyridine-linked ligand as that used to prepare the homobimetallic Au(I) complexes. Although in this case the methyl substituents were replaced with ethyl substituents to improve solubility. AgAu and AuHg heterobimetallic complexes were synthesised using this method, along with analogous homobimetallic Cu(I), Ag(I), Au(I) and Hg(II) complexes. Single crystals for all of the complexes (apart from the homobimetallic Ag(I) complex) were grown and analysed by X-ray diffraction. In all cases, the complexes exhibited the same twisted conformation observed for the homobimetallic Au(I) complexes, with short M···M distances of 2.955(1) angstrom (Cu Cu), 3.5412(1) angstrom (Au Au), 3.6942(5) angstrom (Hg Hg), 3.29(4) angstrom (Ag Au), and 3.21(4) angstrom (Au Hg). The Cu Cu and Hg Hg distances are longer than the sum of the van der Waals radii for each of these metals, but the Au Au, Ag Au and Au Hg distances are shorter, and can therefore be classified as metallophilic interactions. Each of the complexes that exhibits a metallophilic interaction is emissive in the solid state at room temperature and the homobimetallic Ag(I) complex which has not been crystallographically characterised, was found to be non-emissive at RT but emissive upon cooling to 77 K. Density functional theory (DFT) and ab initio calculations were carried out to help confirm the earlier hypothesis. Initially, a benchmarking process was undertaken to investigate which DFT or ab initio method was most suitable for calculating different parameters (e.g. geometry optimisations and relative energies). M06-L/SVP was found to be the optimal level of theory for geometry optimisations, and RI-SCS-MP2/TZVP (at the M06-L/SVP optimised geometries) was used for calculating the relative energies of each conformations for all complexes. Through examination of molecular orbital (MO) contour plots, MO energies and time-dependent density functional theory (TD-DFT) calculations, CAM-B3LYP was found to be the most suitable density functional for calculation of the electronic absorption spectra. The relative energy calculations (with the inclusion of acetonitrile solvent) for the homobimetallic Au(I) complexes show that the stability of the twisted conformation increases with respect to the arene linker in the order benzene [less than] pyridine [less than] pyrazine, with the benzene-linked complexes being more stable in the open conformation, matching experimental observations. MO and TD-DFT calculations show that the twisted pyridine and pyrazine-linked Au(I) homobimetallic complexes give rise to smaller HOMO-LUMO gaps and lower energy MLCT and ML-LCT transitions, compared to open conformers which are associated with transitions from occupied orbitals with greater ligand contribution (LC and ML-LCT). Homobimetallic complexes of the other Group 11 metals are found to be very similar in terms of the TD-DFT results, but the Group 12 Hg(II) homobimetallic complex displays different characteristics. The unsymmetrical nature of the heterobimetallic complexes are reflected in the calculated electronic absorption transitions. The AgAu and AuHg complexes give rise to transitions which originate from orbitals with greater metal character in the twisted conformation than in the open conformation, which is similar to homobimetallic counterparts. The copper-catalysed azide alkyne cycloaddition reaction or ‘click’ chemistry was used in an attempt to prepare heterobimetallic coinage metal complexes by linking together two monometallic NHC complexes, functionalised with an alkyne and an azide group. An Au(I)-NHC complex was prepared from the corresponding azido imidazolium salt, and was structurally characterised by X-ray crystallography. However, despite several attempts analogous metal complexes (Ag(I) or Au(I)) of the alkyne functionalised imidazlium salt could not be isolated. The two imidazolium salts were successfully linked using the ‘click’ reaction to make a triazole linked bis-imidazolium salt, however the same reaction between the azido-functionalised Au(I)-NHC complex and the alkyne functionalised imidazolium salt resulted in decomposition of the complex. Attempts were made to prepare a new series of monodentate and bidentate acyclic (alkyl)(amino)carbene ligands, to allow the preparation of coinage metal complexes with different chemical and photophysical properties. The mono-aldiminium salt precursors were prepared by alkylation of mono-imines, as this was found to be the most effective route. However, the same method, was not successful in preparation of the bis-aldiminium salt. Deprotonation of the mono-aldiminium salts were also unsuccessful, and therefore metal complexes of these pro-ligands could not be synthesised. DFT calculations have been used to investigate the reasons behind the failure of these deprotonation reactions.