Bright luminescent lithium and magnesium carbene complexes

We report on the convenient synthesis of a CNC pincer ligand composed of carbazole and two mesoionic carbenes, as well as the corresponding lithium- and magnesium complexes. Mono-deprotonation affords a rare “naked” amide anion. In contrast to the proligand and its mono-deprotonated form, tri-deprotonated s-block complexes show bright luminescence, and their photophysical properties were therefore investigated by absorption- and luminescence spectroscopy. They reveal a quantum yield of 16% in solution at ambient temperature. Detailed quantum-chemical calculations assist in rationalizing the emissive properties based on an Intra-Ligand-Charge-Transfer (ILCT) between the carbazolido- and mesoionic carbene ligands. (Earth-)alkali metals prevent the distortion of the ligand following excitation and, thus, by avoiding non-radiative deactivation support bright luminescence.

The chemical shifts δ are calculated in ppm; the solvent residual signals of incompletely deuterated solvent molecules were used as an internal reference for the 1 H NMR spectra and the carbon solvent signals for 13 C NMR spectral data or external standards ( 7 Li NMR was referenced to 1M LiCl in D2O and 15 N NMR spectra to ammonia). Solvents were purified using a two-column solid-state purification system (Glass Contour System, Irvine, CA), transferred to the glovebox without exposure to air and stored over a mirror of potassium or activated molecular sieves, respectively. NMR solvents were obtained dry and packaged under argon and as well stored over a mirror of potassium or activated molecular sieves.

3:
The synthetic protocol for the internal cyclisation was adapted from a literature known protocol to design 1,2,3-triazolium derived ionic liquids 4 : A mixture of 2 (975 mg, 1.64 mmol, 1 eq.) and potassium iodide (5.46 g, 32.9 mmol, 20 eq.) was heated in acetonitrile (50 mL) at 90 °C for 44 h. All volatiles were then removed under reduced pressure. The remaining solid was extracted with dichloromethane (3 x 50 mL) and filtered. The filtrate was concentrated to 20 mL and added dropwise to a stirred solution of diethylether (300 mL). The precipitate was filtered off and washed with diethylether (100 mL) and npentane (30 mL). The beige-colored solid was dried in vacuo to yield the desired product in 96% (1.57 mmol, 1. 22

Synthesis of MgBr 5 with MeMgBr
Ligand 3 (30 mg, 0.04 mmol, 1.0 eq.) was suspended in 1 mL of diethylether and 1M MeMgBr in diethylether (13.5 µL, 0.12 mmol, 3.0 eq.) was added. The suspension turned immediately yellow. The solvent was removed in vacuo to give a yellow residue. The 1 H NMR spectroscopic analysis in DMSO-d 6 corroborated the clean and quantitative conversion to MgBr 5. Note, no signals for a methyl complex (between 0 and −50 ppm) was found, whereas the in-situ reaction in deuterated benzene ( Fig S19) revealed a singlet at 0.16 ppm, which can be assigned to free methane.

Photospectroscopic Investigations
UV/vis absorption spectra between 300 and 1000 nm were recorded on a Lambda2 (PerkinElmer) dual beam absorption spectrometer with a scan rate of 480 nm/min and a resolution of 1 nm. Luminescence spectra were recorded on a Fluoromax-3 spectrometer (Horiba Yobin Yvon) with 1 nm spectral bandwidth and 0.1 s integration time, if not stated otherwise. Quantum yields (Φ em ) were obtained using Coumarin-102 (Φ em = 76% in EtOH 5 ) and Coumarin-151 (Φ em = 49% in EtOH 6 ). TCSPC experiments were measured with a FS5-TCSPC spectrofluorometer from Edinburgh Instruments equipped with a photomultiplier R928P emission detector. Samples were excited by a VISUV versatile picosecond laser module from Picoquant. The repetition rate was 8 MHz. All samples for photo spectroscopic investigations of the air-and moisture sensitive compounds Li 5, MgBr 5, and 4 were prepared in an argon-filled glovebox and hermetically sealed prior to the measurement. All glass containers (including vials and quartz glass cuvettes) were silylated prior to the transfer into the glovebox using standard procedures with dimethyldichlorosilane (DMDCS). Thereby, all glassware was soaked in a 10% DMDCS solution in toluene for 30 minutes and then rinsed twice with toluene. Subsequently, it was soaked in methanol for 15 minutes, rinsed twice with methanol and then finally dried in an oven at 160 °C for at least two hours before transferring them into the glovebox. Benzene for photo spectroscopic investigation was distilled over a NaK alloy and stored over a potassium mirror.

Fig. S20
3D steady-state fluorescence spectroscopy of Li 5 in benzene. The fluorescence intensity is depicted as color code. The white bar masks signals stemming from the scattered excitation light.

Fig. S21
Comparison between the absorption spectrum of Li 5 (grey) and the excitation spectrum that leads to photoluminescence at 510 nm (black). The last absorption that leads to photoluminescence is located at the 460 nm shoulder and was used to determine the Stokes-shift of Li 5. The spectra were measured in benzene.

Fig. S22
Comparison between the absorption spectrum of MgBr 5 (grey) and the excitation spectrum that leads to photoluminescence at 510 nm (black). The spectra were measured in benzene.

Fig. S23
Normalized absorption spectra of Li 5 at different stages of decomposition compared to an absorption spectrum of 4 (light red). As dissolved (black), slightly decomposed (dark blue and blue), and fully decomposed (light blue) after exposure to air.

Fig. S24
Comparison of absorption spectra of 4 in benzene in a sealed container (light red) and after 2 hours of exposure to air (green, dashed). Clearly, no spectral changes are discernible.

Fig. S25
Top: Luminescence decay (bright blue dots) and fit (blue line) of Li 5 in benzene (λem = 500 nm) obtained by time-correlated single-photon counting upon excitation at 355 nm, including the instrument response function (IRF, black dots). A mono-exponential fit was sufficient to fit the dataset (lifetime given in legend). Bottom: Corresponding residual to the TCSPC fit.

Fig. S26
Top: Luminescence decay (blue dots) and fit (blue line) of MgBr 5 in benzene (λem = 500 nm) obtained by time-correlated single-photon counting upon excitation at 355 nm, including the instrument response function (IRF, black dots). A bi-exponential fit was sufficient to fit the dataset (lifetimes and contribution given in legend). The second luminescent species is due to slight decomposition of the highly reactive compound during the time of the experiment and corresponds to the decomposition product. Bottom: Corresponding residual to the TCSPC fit.

Diffusion NMR (DOSY)
The diffusion NMR experiments (Bruker pulse sequence "ledbpgp2s") were performed on a Bruker Avance III 300 instrument at a probe temperature of 22 °C. Deuterated benzene was used as solvent and an array of 16 runs (32 scans each; 1 s relaxation delay; 50 ms diffusion time) with a gradient pulse of 1.8 ms and the magnetic field gradient ranging from 10 mT m -1 -508 mT m -1 was recorded. Each experiment was reproduced once and the values reported in Table S1 are averaged over these two runs.
Sample preparation: -Complex Li 5 was suspended in benzene and filtered into a J-Young NMR tube to obtain a saturated solution. Note, Li 5 is only very moderately soluble in weakly-coordinating benzene; thus, only the singlet relating to the two tert-butyl substituents gives reliable integration.
-Complex MgBr 5 was also suspended in benzene and filtered in a J-Young NMR tube. As for Li 5, the low solubility allowed for reliable integration of the tert-butyl signal only.
-As a reference for the monomer, we used the mono-deprotonated ligand 4, which required the addition of a stoichiometric amount of NaBAr F to achieve sufficient solubility in benzene.

S18
Further notes: -Measuring MgBr 5 and Li 5 in one sample for direct comparison is not feasible due to anion scrambling.
-Measuring MgBr 5 together with mono-deprotonated ligand 4 in the presence of NaBAr F also leads to shifts for MgBr 5, which indicates exchange reactions; however both molecules showed the same drift properties which nevertheless suggests that both molecules are monomers.
The volume of a sphere with the corresponding hydrodynamic radius is calculated according to Eq.
S2.  The volumes of MgBr 5 and 4 are similar and the ratio between the volumes suggests that both are monomers (4 seems to be a bit larger which is agreement with the presence of the bulky anion − BAr F ). The relation of the volumes of MgBr 5 or and Li 5 confirms that MgBr 5 is a monomer and Li 5 a dimer in solution.

S21
The lattice benzene molecules were found to be strongly disordered and needed to be modeled using DFIX restraints fixing the bond distances of the carbon atoms.

Computational Details -General
The calculations were performed with ORCA 4.2.1. 11,12 All calculated structures were verified as true minima by the absence of negative eigenvalues in the harmonic vibrational frequency analysis. Potential conformers were assessed for all structures, but only the most stable ones are given below. Tighter than default convergence criteria (tightopt), grid values (grid5, finalgrid6) were chosen for both the optimization of the structural parameters and the scf (tightscf). The geometry optimizations were performed at the B3LYP-D3(BJ)/def2-SVP level of theory. [13][14][15][16][17][18] The RIJCOSX approximation (gridx5) and the related auxiliary basis set def2/J were used to speed up the calculations. [19][20][21] The absorption spectra were