Photoluminescence control by atomically precise surface metallization of C-centered hexagold(i) clusters using N-heterocyclic carbenes

The properties of metal clusters are highly dependent on their molecular surface structure. The aim of this study is to precisely metallize and rationally control the photoluminescence properties of a carbon(C)-centered hexagold(i) cluster (CAuI6) using N-heterocyclic carbene (NHC) ligands with one pyridyl, or one or two picolyl pendants and a specific number of silver(i) ions at the cluster surface. The results suggest that the photoluminescence of the clusters depends highly on both the rigidity and coverage of the surface structure. In other words, the loss of structural rigidity significantly reduces the quantum yield (QY). The QY in CH2Cl2 is 0.04 for [(C)(AuI-BIPc)6AgI3(CH3CN)3](BF4)5 (BIPc = N-isopropyl-N′-2-picolylbenzimidazolylidene), a significant decrease from 0.86 for [(C)(AuI-BIPy)6AgI2](BF4)4 (BIPy = N-isopropyl-N′-2-pyridylbenzimidazolylidene). This is due to the lower structural rigidity of the ligand BIPc because it contains a methylene linker. Increasing the number of capping AgI ions, i.e., the coverage of the surface structure, increases the phosphorescence efficiency. The QY for [(C)(AuI-BIPc2)6AgI4(CH3CN)2](BF4)6 (BIPc2 = N,N′-di(2-pyridyl)benzimidazolylidene) recovers to 0.40, 10-times that of the cluster with BIPc. Further theoretical calculations confirm the roles of AgI and NHC in the electronic structures. This study reveals the atomic-level surface structure–property relationships of heterometallic clusters.


Physical measurements and instrumentation
NMR data were recorded on a Bruker Avance III spectrometer (500 MHz). When CDCl3 was used, tetramethylsilane was used as the internal standard (0 ppm) for 1 H NMR. In the cases when CD2Cl2 was used, a mono-protonated solvent signal of CD2Cl2 (CDHCl2, 5.32 ppm) and a solvent signal of CD2Cl2 ( 13 CD2Cl2, 53.84 ppm) were used as the internal standards for 1 H and 13 C NMR measuremants, respectively. Abbreviations: s, singlet; d, doublet; sept, septet; dd, double doublet; br, broad. ESI-TOF-MS spectrum were recorded on a Micromass LCT Premier XE. Experimental conditions: ion mode, positive; desolvation temperature, 150 °C; source temperature, 80 °C. UV-vis spectra were recorded on a Jasco V-770 spectrophotometer. Luminescence was measured on a Jasco FP-8300 spectrofluorometer. Luminescence quantum yields were determined on a Hamamatsu C9920-02G spectrometer. Luminescence lifetimes were determined on a Hamamatsu C11367-02 spectrometer. Elemental analysis was conducted in the Microanalytical Laboratory, Department of Chemistry, Graduated School of Science, the University of Tokyo. Under an atmosphere of nitrogen, a mixture of benzimidazole (1.18 g, 10 mmol) and potassium carbonate (1.52 g, 11 mmol) was suspended in CH3CN (30 mL) in a Schlenk flask and stirred at ambient temperature for 1 h. Then, isopropyl bromide (2.2 mL, 20 mmol) was added into the suspension and the reaction mixture was heated at reflux for 36 h. After the solvent was removed under vacuum the residue extracted with H2O/CH2Cl2 (20:15 mL). Removal of the solvent from the organic layer gave 1-isopropylbenzimidazole as an oil. The resulting oil, 2-(chloromethyl)pyridine hydrochloride (1.64 g, 10 mmol) and potassium carbonate (2.76 g, 20 mmol) were then suspended in CH3CN (50 mL) in a Schlenk flask under nitrogen atmosphere. This mixture was heated at reflux for 48 h. The solvent was then removed in vacuo, and the residue was dissolved in CH2Cl2 (50 mL). Filtration and removal of the solvent from the filtrate yielded a viscous residue. Colorless block-like crystals of BIPc·HCl were obtained by layering Et2O on a filtered CH2Cl2/CH3OH (9:1, v:v) solution of product in tubes. Yield: 1.30 g (45%, based on benzimidazole).  Complexes [(C)(Au I -BIPc)6](BF4)2 (5) and [(C)(Au I -BIPc 2 )6](BF4)2 (6) were synthesized according to the literature procedures with modifications. 1,5,6 Ag 2 O (35.0 mg, 0.15 mmol) was added to a solution of benzimidazolium chloride (BIPc·HCl (80.4 mg, 0.30 mmol) for 5; BIPc 2 ·HCl (98.7 mg, 0.30 mmol) for 6) in CH2Cl2/CH3OH (7.5 mL/2.5 mL). The suspension was stirred for 2.5 h in the dark and then filtered through Celite. After tht-AuCl (96.0 mg, 0.30 mmol) was added, the solution was stirred overnight (~12 h) in the dark. The suspension was again filtered through Celite, and the solvents were then removed using a rotary evaporator. After adding NaBF4 (165 mg, 1.5 mmol) and CH3OH (5 mL), the suspension was stirred for 5 min. CH2Cl2 (15 mL), a solution of KOH (28.0 mg, 0.50 mmol) in CH3OH (3 mL), a solution of AgBF4 (58.5 mg, 0.30 mmol) in CH3OH (1 mL), and H2O (50 μL) were then sequentially dropwise added to the mixture under stirring, which leads to a brown suspension. After another stirring for 5 min, the suspension was again filtered through Celite and evaporated to dryness. The solid was then transferred to a Schlenk flask with nitrogen atmosphere, and dry CH2Cl2 (5 mL), Et3N (30.0 μL, 0.20 mmol) and a 2.0 M solution of Me3SiCHN2 in n-hexane (48.0 μL, 0.10 mmol) were added. The resulting mixture was stirred for another 1 h. After filtration into a tube, a layer of dry Et2O was added on the CH2Cl2 solution, which gave the products as colorless block-like crystals. Yields: 66.1 mg (46%, based on tht-AuCl) for 5; 68.1 mg (43%, based on tht-AuCl) for 6.

X-ray crystallography
Intensity data of compounds BIPc·HCl, (3), and [(C)(Au I -BIPc 2 )6Ag4(CH3CN)2](BF4)6 (4) were collected on a Rigaku XtaLAB Synergy-DW system (CuKa) at 93 K. The structures were solved by direct methods, and non-hydrogen atoms except for the disordered BF4 − in 6 and the central carbon atom in 4 were refined anisotropically by the least-squares on F 2 using the SHELXTL program. As a result, 4 has one level B alert (Isotropic non-H Atoms in Main Residue). The hydrogen atoms of organic ligands were generated geometrically. Squeeze tool of PLATON and absorption correction using WinGX were applied to 3, due to the large solvent voids and heavy absorption.

Computation details
DFT and TD-DFT calculations were performed for all clusters using the B3LYP functional. 10 Relativistic effective core potential LANL2DZ 11 was used for Au and Ag atoms and the basis sets of other atoms were 6-31G*. 12 Optimizations were carried out based on the crystal structures and vibrational frequency analyses were conducted to verify the stationary points to be local minima on the potential energy surface. 4* was constructed by removing the two acetonitrile groups of 4, and a highly symmetric structure was obtained after optimization.
For simulating absorption spectra, 100 excited states were solved to cover the spectrum in the energy range up to about 200 nm and the rotatory strength was calculated in the velocity form. TD-DFT calculation was conducted including the solvent effects of CH2Cl2 with the polarizable continuum model (PCM) and the nonequilibrium linear response scheme. 13 To calculate phosphorescence energies, we obtained the optimal geometry of the lowest triplet excited state (T1) and, at this geometry, calculated the emission energy with the Δ selfconsistent-field (ΔSCF) approach. All calculations were conducted using the Gaussian 16 suite of programs. 14 The natural population analysis (NPA) charges were calculated at the level of B3LYP/6-31G* using NBO 3.1 as implemented in Gaussian 16. 14 Bader charges and orbital compositions were obtained by the Multiwfn program. 15 The radiative rate constants of 3 and 4* by ZORA method including spin-orbit interaction in the perturbative method 16 implemented in ADF program package. 17 The B3LYP functional combined with DZ basis set was utilized.
Note that some metal-metal bonds of the triplet structure of 3 are extraordinarily elongated (dissociated) during optimization in the gas phase, which provides the problem for DFT calculations with a single determinant. Therefore, the lowest triplet excited state was obtained by optimization including the solvent effects. Meanwhile, it is practically impossible to locate MECP of S0/T1, which makes the discussion of non-radiative decay difficult theoretically.

Additional results and discussion on the homometallic CAu I 6 clusters 5 and 6
Single crystal X-ray diffraction (ScXRD) analysis was successfully performed on the hexagold(I) clusters [(C)(Au I -BIPc)6](BF4)2 (5) and [(C)(Au I -BIPc 2 )6](BF4)2 (6). As shown in Fig. S2, both of them have an octahedral structure with a carbon tetraanion in the center. Each NHC ligand BIPc or BIPc 2 in the complexes coordinates to one gold atom. The Au-Au distances in 6 were found in a range from 2.8585(3) to 3.1911(3) Å, which shows slight larger deviation than other known hexagold(I) clusters (Table S1). 1,5,6,17 As a result, the arrangement of ligands BIPc 2 in 6, especially their wingtip groups, are less ordered. Other key structure parameters such as Au-C (center) and Au-C (ligand) distances in 5 and 6 are comparable with known NHC-protected CAu I 6 clusters. Global characterizations were performed on 5 and 6, including NMR, UV-vis, ESI MS, luminescence etc ( Figures  S4-17). First, there is only one set of signals in the NMR spectra of both compounds. That is to say, the six Au + -BIPc or Au + -BIPc 2 moieties are equivalent when the clusters are dissolved in solution. The slightly altered octahedral structure of 6, as indicated by ScXRD, is probably due to the bulkiness and large number of picolyl wingtip groups, which affected the packing of molecules when forming crystals. Second, both 5 and 6 have triplet absorption peaks at around 350 nm in CH2Cl2, which are very similar to reported hexagold(I) clusters protected by NHC ligands. 1,5,6 Metal-to-ligand charge transfer (MLCT) from the metal kernel to NHC ligands is a highly possible origin. Third, sharp peaks correspond to [(C)(Au I -BIPc)6] 2+ and [(C)(Au I -BIPc 2 )6] 2+ etc. were observed in the mass spectra. The experimental and simulated pattern fit well with each other. Last but not the least, both 5 and 6 show green luminescence in the solid state, which is similar to [(C)(Au I -BIiPr)6](BF4)2 but very different from [(C)(Au I -IPy)6](BF4)2 and [(C)(Au I -BIPy)6](BF4)2 ( Figure S1). 1,6 Previously, we found that the phenyl part of the BIiPr in [(C)(Au I -BIiPr)6](BF4)2 is significantly involved in the lowest unoccupied molecular orbital (LUMO). However, the installation of aromatic wingtip group (pyridyl) will dramatically change the electronic structure of cluster. The pyridyl groups, instead of the phenyl parts in [(C)(Au I -BIPy)6](BF4)2, primarily participate in the LUMOs of cluster. As a result, the luminescence shows obvious blue-shift. In the cases of 5 and 6, the emissions were restored to green. This suggests that the introduction of methylene linker between benzimidazolyl and pyridyl groups may efficiently decouple the electron resonance.
To try to explain the UV-vis absorption and luminescence of 5 and 6, time-dependent density functional theory (TD-DFT) calculations were carried out. The absorption spectra of 5 and 6 were theoretically simulated as shown in Figures S19 and S20, respectively, which well reproduced the experimental ones in CH2Cl2. It is confirmed that the lowest peaks around 350 nm are mainly attributed to the MLCT transition from the Au kernel to ligands, while the strong peaks around 260 nm mainly come from the ππ* transitions of the ligands. (Tables S2 and S3) In addition, the involved molecular orbitals of 5 and 6 are illustrated in Figures S21 and S22. Indeed, the HOMOs (highest occupied molecular orbitals) are mainly located in the CAu I 6 cores of clusters, and more importantly, the HOMOs are located in the benzimidazolylidene moieties of NHC ligands, with almost no contribution from picolyl groups. These results indicate that electronic structures of the whole cluster can be tuned by employing alkyl or aromatic wingtip groups.

Additional results and discussion on the complexation experiments of 5 and 6 with AgBF4 in CH2Cl2/CH3OH (9:1, v:v)
In our first trials, the complexations of 5 and 6 with AgBF4 were conducted in CH2Cl2/CH3OH (9:1, v:v), and the complexation processes were monitored by using ESI MS. As shown in Figure S31, after adding 1 equiv. of AgBF4 into 5 in a mixed solution of CH2Cl2/CH3OH (v:v = 9:1), the peaks correspond to 5 almost immediately disappeared. 2+ emerged, as well as a peak around m/z = 1094.5. This peak became dominant when the amount of AgBF4 was further increased. According to the isotope pattern, it can be concluded that this mono cationic peak contains no Ag + or BF4 -. We assign this peak to [HAu2(BIPc)(picolylbenzimidazolylidene)2(CH3OH)] + , which may contain protic NHCs and is rather abnormal. Although ionization conditions of ESI MS may lead to dissociation of cluster, such fragmentation of NHC ligands indicates potential decomposition of heterometallic cluster with the presence of both CH3OH and silver ions.
The complexation of 6 and AgBF4 was also conducted in CH2Cl2/CH3OH ( Figure S32). Similar to that of 5, bright luminescence was observed when the secondary ions were added. In addition, a strong peak containing a decomposed ligand was also detected in the ESI MS spectra. Note that the difference between this monocationic peak (m/z 1143.5) and the signal observed in the complexation of 5 and AgBF4 (m/z 1094.5) is ~49, which is exactly the difference between ligands BIPc and BIPc 2 . Considering the monocationic nature and large mass-to-charge ratio of these two peaks, more than one organic ligand or motif should be included. This further backed up our presumption that protic ligands were formed during the ionization process with the presence of CH3OH and silver ions. Similarly, no heterometallic product was isolated under this condition.

Additional results and discussion on the theoretical calculation of phosphorescence lifetimes and radiative rate constants of 3 and [(C)(Au I -BIPc 2 )6Ag I 4](BF4)6 (4*)
Except for the absorption and phosphorescence energy, we have also evaluated the phosphorescence lifetimes and the radiative rate constants of 3 and 4* using the ZORA method with spin-orbit interaction in a perturbative way implemented in ADF program package. The results were compared with that of 2. The calculated kr is obtained as kr = 1/τ.
As collected in Table S9, the three lowest-lying spin-orbit states of 2, 3 and 4* are degenerate with rather close energies but different oscillator strength values (f) and phosphorescence lifetimes (τ), which mainly contributes to the phosphorescence. The lifetime of the spin-orbit state with the largest f value is 27.97 μs for 3, which is smaller than the corresponding values of 2 (35.25 μs) and 4* (63.10 μs). Thus, the trend of lifetime is partly consistent with that observed in experiments. The calculated kr values show a weak correlation with the experimental ones. The combination of several states may contribute to the observed phosphorescence, resulting in the disagreement in the experiment and theory.
The wavefunction and spin-orbit coupling of the low-lying spin-orbit states were also analyzed (Table S10). These states are mainly composed of T1 or S1, with small contributions from other excited states. It can be seen that the Ag atoms coordinated to the Au kernel significantly changed the main component of each state and the coupling of the spin-orbit states, leading to the different kr values of these compounds.

Supplementary schemes, figures, and tables
Scheme S1. Aromatic wingtip groups of (benz)imidazolylidene may significantly contribute to the photophysical properties of the corresponding metal clusters Scheme S2. Coordination modes of ligands, structures of metal kernels, and QYs (in CH2Cl2) of CAu I 6-based clusters protected by NHC and phosphine ligands