Structural and Luminescent Properties of Homoleptic Silver(I), Gold(I), and Palladium(II) Complexes with nNHC-tzNHC Heteroditopic Carbene Ligands.

Novel silver(I), gold(I), and palladium(II) complexes were synthesized with bidentate heteroditopic carbene ligands that combine an imidazol-2-ylidene (nNHC) with a 1,2,3-triazol-5-ylidene (tzNHC) connected by a propylene bridge. The silver(I) and gold(I) complexes were dinuclear species, [M2(nNHC-tzNHC)2](PF6)2 (M = Ag or Au), with the two bidentate ligands bridging the metal centers, whereas in the palladium(II) complex [Pd(nNHC-tzNHC)2](PF6)2, the two ligands were chelated on the same metal center. Because of the presence of two different carbene units, isomers were observed for the gold(I) and palladium(II) complexes. The molecular structures of the head-to-tail isomer for gold(I) complexes, with a twisted or folded-syn conformation of the bridge between the carbene units, were determined by X-ray diffraction analysis. The study was completed with a systematic structural investigation through density functional theory (DFT) calculations. For palladium(II) species, the head-to-head form was structurally characterized. The dinuclear gold(I) complexes were emissive in the solid state in the blue region (PLQY up to 8%); time-dependent density functional theory (abbreviated as TD-DFT) calculations disclosed that the absorption bands have metal-to-ligand-charge-transfer character and evidenced that the emission occurs from the T1 level (phosphorescence).

Introduction

Since the isolation of the first imidazol-2-ylidene in the early 90s,[1] the interest of the scientific community toward N-heterocyclic carbenes (NHCs) has continuously increased. Nowadays, NHCs are employed as efficient organocatalysts,[2,3] although their main use remains as ligands for transition-metal centers,[4−9] and the resulting NHC complexes have found application in several research areas, that is, catalysis, bioinorganic chemistry, and material science.[4,8,10−13] Also, bidentate di-NHCs have received quite a bit of attention,[14−16] especially considering that, in this latter case, it is possible to tune the stereoelectronic properties of the ligand by changing the wingtip and backbone substituents of the heterocyclic ring, but also by modifying the linker between the carbene units. Furthermore, the nature of the bridge between the heterocyclic moieties can be a key element in influencing the bridging rather than the chelating coordination mode of the ligand to the metal centers: with short or constrained linkers, a bridging coordination of the ligand between different metals gives polynuclear structures, whereas with more flexible bridges, the chelating coordination occurs preferentially.[17,18] The majority of the examples reported in the literature regards di-NHC based on two imidazol-2-ylidene rings (nNHC) connected by an aliphatic or aromatic bridge; however, several heterocyclic scaffolds are available to prepare heterocyclic carbenes, spanning from the classical imidazole-, imidazoline-, or benzimidazole-ylidene to other five-member rings, like pyrazole-, triazole-, or tetrazole-ylidene, to six-, seven-, or four-member rings, or to more exotic ones.[19,20] Among them, it is interesting to mention the 1,3,4-trisubstituted-1,2,3-triazol-5-ylidene compounds (tzNHCs), whose triazolium salt precursors can be easily isolated via a click reaction, that is, the copper-catalyzed alkyne azide cycloaddition (CuAAC), followed by the alkylation of the nitrogen in the N-3 position.[21] The synthesis of heteroditopic carbene ligands, that is, ligands combining two types of N-heterocyclic carbenes, is appealing because, in principle, it should allow one to isolate original heterobimetallic species, taking advantage of the different acidity of the heterocyclic rings.[22−25] Moreover, the properties of a complex having a bidentate ligand with mixed donors might not be intermediate with respect to those of the corresponding homoditopic compounds. For example, a palladium(II) complex with a dicarbene ligand having two different NHC donors was found to display enhanced catalytic properties with respect to the corresponding complexes with the homodicarbene ligands; this was explained in terms of “electronic asymmetry” in the complex.[26,27]

In this work, the influence of a mixed nNHC-tzNHC dicarbene ligand was evaluated on the photoemission properties of the corresponding dinuclear silver(I) and gold(I) complexes. The employed ligands were designed with a propylene linker between the nNHC and the tzNHC donors because we demonstrated that this bridge has the right length and flexibility to promote intramolecular metallophilic interactions, thus influencing the luminescence properties[28] and the reactivity of the gold(I) complexes.[29] The coordination of these mixed dicarbenes was also extended to a palladium(II) complex, which, compared to the gold(I) complexes, should present a different coordination geometry (square planar vs linear) and a chelate coordination of the ligand rather than a bridging one. The synthesis of the gold(I) and palladium(II) complexes was readily achieved through transmetalation of the dicarbene ligand from the corresponding silver(I) complexes. Interestingly, in the case of homobimetallic gold(I) complexes, as well as for the mononuclear palladium(II) complex, two isomers were observed in solution. For gold(I) complexes, the luminescence properties as well as the nature of the different possible isomers were investigated through relativistic DFT and TD-DFT calculations.

Results and Discussion

Synthesis of the Imidazolium–Triazolium Proligands

Two different proligands with imidazolium and 1,2,3-triazolium rings, connected by a propylene bridge, were successfully synthesized following the procedure reported by Liebscher and described in Scheme .[30,31]

Scheme 1

Synthesis of the Proligands 1-PF and 2-PF

The procedure involves three steps from 1-(pent-4-ynyl)-1H-imidazole: (i) formation of the triazole ring via the CuAAC reaction; (ii) methylation of both the imidazole N3 and triazole N1 nitrogen atoms using excess of methyl iodide, providing 1-I and 2-I; and (iii) anion exchange between I– and PF6– to give 1-PF and 2-PF. The last step favors the solubility of the proligands in polar solvents (DMSO and mostly CH3CN). The formation of the desired imidazolium/triazolium salt was easily assessed by the 1H NMR spectra, which present two different signals located at δ > 8.0 ppm attributable to the protons in position 2 of the imidazolium moiety and in position 5 of the triazolium ring.

Synthesis of Silver(I) Complexes

The silver(I) di(N-heterocyclic carbene) complexes 3 and 4 were synthesized by reaction of the proper diazolium bis(hexafluorophosphate) salts, 1-PF and 2-PF, respectively, with excess of silver(I) oxide in acetonitrile at 85 °C for 48 h (Scheme ).

Scheme 2

Synthesis of Silver(I) Complexes 3 and 4

Complex 4 was obtained as a stable off-white solid, whereas complex 3, although spectroscopically pure, is an oil that slowly decomposes. The disappearance of the signals at δ above 8.0 ppm in the 1H NMR spectra is consistent with the deprotonation at the positions C2-H of the imidazole moiety and C5-H of the triazole ring. In the 13C NMR spectra, two different signals are present at δ ca. 165 and 180 ppm, attributable to the carbene carbons of the tzNHC-Ag and nNHC-Ag moieties, respectively, in the typical range reported in the literature.[32] Finally, the presence of the peak corresponding to the fragment [Ag2L2PF6]+ in the electrospray ionisation mass spectrometry (ESI-MS) spectra confirms the dinuclear dicationic structure of the complexes, with the experimental isotopic distribution in agreement with the simulated one. All these data suggest a dinuclear structure for the silver complexes, with the two dicarbene ligands bridging two different silver(I) centers; the structure with an nNHC facing a tzNHC appears the most probable by analogy with related complexes reported in the literature,[33] although a dynamic behavior in solution involving more than one isomer cannot be excluded.

Synthesis of Gold(I) Complexes

Gold(I) complexes were synthesized via two methods: (i) transmetalation of the dicarbene ligand from the corresponding silver(I) complexes 3 and 4 using a Au/Ag 1:1 ratio, or (ii) deprotonation of the diazolium salt with sodium acetate in the presence of the gold(I) precursor AuCl(SMe2). The transmetalation procedure was employed starting from both silver(I) complexes (Scheme ), although with the ligand having the benzyl substituent at the triazole ring, the reaction was performed in situ, that is, without the preisolation of silver(I) complex 3, considering its oily nature and instability. For these reasons, the gold(I) species, analogous to silver(I) complex 3, was also synthesized via deprotonation of the diazolium salt (Scheme ).

Scheme 3

Synthesis of Gold(I) Complexes 5/5′ and 6/6′

Scheme 4

Synthesis of Gold(I) Complexes 5/5′

Interestingly, the NMR experiments of the isolated solids via transmetalation show two sets of signals in a close to 1:1 ratio, both in the 1H and in 13C NMR spectra, likely due to the presence of the two isomers associable to the different coordination modes of the heteroditopic ligands. An analogous result was also obtained from the deprotonation reaction starting from the diazolium salt although the observed ratio was 5:1 (complex 5 as the major isomer). This difference in the molar ratio registered in the two synthetic procedures can be ascribed to the different reaction temperatures,[34] which should favor the formation of the most stable isomer.

Due to the intrinsic ditopic nature of the ligands and the possible orientations of the propylene linker in the ligand, several isomers could be figured out. The main difference regards the coordination sphere around the metal centers (constitutional isomerism, Chart a): (i) in the head-to-tail species, each gold(I) center is coordinated by one nNHC and one tzNHC; (ii) in the head-to-head species, one gold(I) coordinates two nNHC and the other one binds the two tzNHC. We labeled the crude mixture of complexes 5/5′ and 6/6′, where the apostrophe indicates the head-to-head isomer. Furthermore, considering the possible arrangements of the bridge of the bidentate ligand, these two coordination isomers may be present at least in four conformations (conformational isomerism): the twisted (t), the folded-syn (fs), the folded-anti (fa), and the stretched-out ones (s) (Chart b, represented only for the head-to-tail coordination).

Chart 1

Schematic Representation of the Possible Isomers for the Gold(I) Dinuclear Complexesa

(a) Head-to-head and head-to-tail constitutional isomers and (b) the conformational isomers (twisted, folded-syn, folded-anti, and stretched-out) represented only for the head-to-tail coordination.

From this point of view, the two sets of signals might be associated with two species differing in the arrangement of the bridge in the bidentate ligands[34,35] (conformational isomerism) or in the head-to-head or head-to-tail coordination (constitutional isomerism). In our opinion, the latter appears more convincing considering the following factors. (i) The 1H NMR spectrum registered after heating the solution of the mixture 6/6′ at 70 °C in an NMR tube remains unchanged and also the relative ratio of the two sets of signals remains constant; if the two species differ on the conformation of the bridge, upon increasing the temperature a conversion or an exchange between the two species is expected, considering the flexibility of the propylene linker. The unchanged spectra support indirectly the presence of two constitutional isomers, which cannot convert upon heating. (ii) The different conformations of the bridges might not be retained in solution.[36] (iii) A similar di(nNHC) gold(I) complex, that is, [Au2(MeIm(CH2)3ImMe)2]2+ (Im = imidazole-2-ylidene), presents the same NMR spectrum profile in the folded-syn or twisted conformation (Chart b).[28,37] (iv) Starting from the mixture 6/6′, we were able to crystallize complex 6 (head-to-tail coordination of the ligand with folded-syn conformation of the bridges; see further in the text) and characterized it in solution. Complex 6 is stable in solution, and no conversion to the second species was observed; furthermore, the 1H NMR analysis of the remaining solution, where the crystals were grown, confirmed an increase of the 6′/6 ratio, thus supporting the presence of two different species like two constitutional isomers are. (v) During a slightly modified synthesis, the head-to-head isomeric form was isolated by crystallization (see below and the structure of complex 7).

The formation of the gold(I) complexes could be confirmed by the presence of signals at ca. δ 170 and 185 ppm related to the carbene carbons in positions 2 and 5 of the nNHC and tzNHC, respectively, slightly downfield-shifted from those observed for the related silver(I) complexes and in the range of ppm reported in the literature for nNHC and tzNHC gold(I) complexes.[32,38] The dinuclear dicationic structure of complexes 5/5′ and 6/6′ was confirmed by ESI-MS spectra, which present a signal attributed to the species [Au2L2PF6]+ with the L = nNHC-tzNHC dicarbene ligand.

By slow diffusion of diethyl ether into a solution of the mixtures 5/5′ and 6/6′ in acetonitrile, few crystals were obtained, and the molecular structure of the crystallized compound was determined by X-ray diffraction. In both cases, that is, starting from mixture 5/5′ or 6/6′, only crystals of the head-to-tail isomer were isolated. Complexes 5 and 6 have a dinuclear dicationic nature with two dicarbene ligands bridging the two gold atoms, and the two gold(I) centers are coordinated to one tzNHC and one nNHC.

The ORTEP views of the cationic complexes are shown in Figures and 2 together with the atomic numbering scheme. Relevant bond distances and angles are reported in the captions. The flexible propylene linkers allow a backfolded conformation of the bridges in both structures;[39] however, in complex 5, they adopt the folded-syn conformation (fs, Chart b), whereas in compound 6 the observed conformation is the twisted one (t, Chart ). From this point onward, also the arrangement of the bridge will be specified in the notation of the complex. In the crystals of the cationic compound 5(f), PF6– anions are present, whereas in those of 6(t), acetonitrile solvent molecules are also found. Both cationic complexes present a dinuclear structure, with the two gold(I) centers linearly dicoordinated, a geometry usually observed for metal centers having a d10 electronic configuration. The bond angles Ccarb-Au-Ccarb are close to the linearity [176.1(2)° in 5(f) and 176.7(3) and 175.5(3)° in 6(t)], and the bond distances are comparable to those reported for analogous di(NHC) gold(I) complexes.[28,40] The torsion angle Ccarb-Au···Au-Ccarb in 5(f) is 20.2(1)°, and in 6(t) it is close orthogonal [79.0(4) and 83.1(4)°]. Moreover, the Ccarb-Au···Au angle is close to orthogonal in 6(t) [89.3(2), 94.5(2), 94.6(2), 82.8(2)°] and bigger than 90° in 5(f) [111.7(1)°]. This affects the Au···Au separation, which is significantly different in the two complexes: in 5(f), the gold atoms are separated by 3.3479(4) Å, whereas in 6(t), the 3.0680(5) Å gold-to-gold distance, being much below the sum of the van der Waals radii for gold metal, is one of the shortest ever observed in analogous di(NHC) gold(I) complexes so far.[40] This clearly suggests the presence of aurophilic interaction between the two metals. In both structures, the mean planes of the two imidazol rings coordinated to the same Au atom are slightly twisted, the dihedral angles being 30.8(1)° in 5(f) and 18.6(1) and 17.7(1)° in 6(t). In the molecular and crystal structures of 5(f) and 6(t), no stacking is observed between the planar ring moieties, except between the phenyl and triazole rings in 6(t) labeled by atoms C11, C12, C13, C14, C15, C16 and N8, N9, N10, C24, C25 (evidenced in Figure by the ball and stick drawing). In fact, these groups lie on almost parallel planes [dihedral angle of 8.1(4)°] at a distance between the centroids of the rings of 3.764(6) Å. The bridging coordination mode of the dicarbene ligands imposes coordination chirality to complex 6(t). Nevertheless, the complex crystallizes in the C2/c space group and so both enantiomers are present in the crystals.

Figure 1

ORTEP-style view of the cationic compound 5(f), with atomic numbering scheme; ellipsoids are drawn at their 30% probability level. Hydrogen atoms and PF6– anions have been omitted for clarity. Selected bond distances (Å) and angles (deg): C1–N2 1.349(9), C1–N1 1.371(7), C1–Au 2.005(5), C8–C10 1.397(6), C10–N3 1.371(6), C10′–Au 2.026(5), Au···Au′ 3.3479(4), N3–N4 1.316(7), N4–N5 1.325(6); N2–C1–N1 103.0(5), N2–C1–Au 127.9(4), N1–C1–Au 129.0(5), N3–C10–C8 102.0(4), N3–C10–Au′ 124.1(3), C8–C10–Au′ 133.3(4), C1–Au–C10′ 176.1(2), C1–Au–Au′ 71.47(15), and C10′–Au–Au′ 111.71(14). Code for atoms: ′ = −x, y, 0.5 – z.

Figure 2

ORTEP-style view of the cationic compound 6(t), with atomic numbering scheme; ellipsoids are drawn at their 30% probability level. Hydrogen atoms and PF6– anions have been omitted for clarity. Selected bond distances (Å) and angles (deg): Au1–C1 2.020(10), Au1–C25 2.040(10), Au1–Au2 3.0680(5), Au2–C9 2.020(9), Au2–C17 2.012(9), N1–C1 1.365(11), N2–C1 1.331(12), N3–N4 1.351(10), N4–N5 1.330(11), N6–C17 1.365(12), N7–C17 1.357(11), N8–N9 1.331(10), N9–N10 1.305(11); C1–Au1–C25 176.7(3), C1–Au1–Au2 82.8(2), C25–Au1–Au2 94.5(2), C9–Au2–C17 175.5(3), C9–Au2–Au1 89.3(2), and C17–Au2–Au1 94.6(2).

Figure 3

View of the cationic part of complex 6(t), emphasizing the parallel planes of the phenyl ring and the triazole one.

As stated above, the gold complexes 5/5′ were synthesized without isolation of the silver complex, and, in particular, the gold precursor AuCl(SMe2) was simply added to the acetonitrile solution of the silver complex 3 once filtered from the silver oxide in excess. In one of the first attempts to synthesize the gold complexes, we performed the same reaction but adding the gold(I) precursor directly to the reaction mixture, still containing the silver oxide. By slow diffusion of diethyl ether in the final acetonitrile solution, very few single crystals of 7 adequate for X-ray diffraction analysis were formed. In the crystals of 7, two slightly different molecules of trinuclear bimetallic compounds (molecule A and molecule B) of formula [Au2(nNHC-tzNHC)2Ag(CH3CN)2]3+, PF6– anions, and diethyl ether molecules are present. The crystal structures of the two cationic compounds are shown in Figure , together with a selected atomic labeling scheme. Relevant bond distances and angles of the two molecules are also reported in the caption. The trinuclear compounds are composed of two gold and one silver metal atoms; interestingly, the two gold atoms are linked by two head-to-head ligands in a twisted conformation, and the fragment Ag(CH3CN)2 interacts with one of the two Au centers. The Ag(CH3CN)2 fragment in the two molecules presents different geometrical parameters: in fact, in molecule A, the two acetonitrile molecules are disposed in a nearly linear mode (N–Ag–N bond angle 167.3(6)°), whereas in molecule B, the same fragment presents a nearly trigonal planar arrangement (N–Ag–N bond angle 129.6(8)°). This affects also the Ag···Au distance, which is longer in molecule A (2.8860(16) Å) than in molecule B (2.7165(18) Å). The Au···Au distances (3.1195(8) Å and 3.0516(8) Å for molecules A and B, respectively) are very similar in both complexes and in agreement with the values measured in analogous bimetallic dicarbene gold(I) complexes.[28,37,39−41] The two molecules are adjacent in the crystal structure with terminal Au2 atoms at a distance of 3.4295(7) Å. Moreover, the imidazole rings of the ligands are stacked in almost parallel planes (dihedral angle of 9.6(5)°), and the distance between the centroids is 3.834(6) Å. The Au and Ag atoms of both molecules are aligned on the same twofold crystallographic axis. Thus, the complexes are symmetric, with one-half of both structures being correlated with the other half by the twofold axis. A disordered diethyl ether molecule separates the two stacked molecules. In both molecules, the benzyl substituents of the bridging ligands point toward the Ag(CH3CN)2 fragment to minimize the steric hindrance. The imidazole rings coordinated to the same Au atom (Au2A and Au2B) are slightly twisted with dihedral angles of 34.3(6) and 29.7(16)°, whereas the triazole rings coordinated to the same Au atom (Au1A and Au1B) are almost coplanar [8.7(1) and 8.8(6)°].

Figure 4

ORTEP-style view of cationic compound 7 [molecules A (bottom) and B (top)], with selected atomic numbering scheme; ellipsoids are drawn at their 30% probability level. Hydrogen atoms have been omitted for clarity. Selected bond distances (Å) and angles (deg): C2A–Au2A 2.020(9), C1A–Au1A 2.002(11), C2B–Au2B 2.025(11), C1B–Au1B 2.028(9), N1SB–Ag1B 2.157(14), N1SA–Ag1A 2.085(13), Au1B–Ag1B 2.7165(18), Au1B–Au2B 3.0516(8), Au2A–Au1A 3.1195(8), Au1A–Ag1A 2.8860(16); C1B–Au1B–C1′B 179.5(5), C1B–Au1B–Ag1B 89.7(3), C1B–Au1B–Au2B 90.3(3), C2B–Au2B–C2′B 178.1(5), C2B–Au2B–Au1B 89.0(3), C2A–Au2A–C2′A 173.7(5), C2A–Au2A–Au1A 86.8(2), C1A–Au1A–C1′A 178.8(5), C1A–Au1A–Ag1A 89.4(3), C1A–Au1A–Au2A 90.6(3), N1SA–Ag1A–N1′SA 167.3(6), and N1SB–Ag1B–N1′SB 129.6(8). Code for atoms: ′ = 1 – x, y, 1.5 – z.

Luminescence Properties

The photoluminescence properties of silver(I) complex 4 and gold(I) complexes 5(fs) and 6(t), the isomers isolated by crystallization, were analyzed at room temperature in the solid state as films obtained by drop casting from acetonitrile solutions on a quartz substrate. The silver complex 4 was found to be non-emissive, and a darkening of the sample was observed after the analysis, consistent with the known sensitivity of silver compounds to light exposure. Both gold complexes displayed instead intense blue luminescence (Figure ). The emission maximum of 5(fs), centered at 430 nm, is blue-shifted of 50 nm with respect to that of 6(t). We observed comparable absolute emission quantum yields with maximum value of 8% for 6(t) that has also the shorter Au···Au distance (3.068 Å, Table ). This suggests a stronger aurophilic interaction than for 5(fs).

Figure 5

Solid-state excitation (PLE, line and symbol) and emission (PL, red line) spectra of complexes 5(f) (left) and 6(t) (right).

Table 1

Photophysical Data of the Gold(I) Complexes 5(f) and 6(t) in Solid State at Room Temperature

	excitation λmax (nm)	emission λmax (nm)a	Φ (%)a	τ (μs)b	Au···Au distance (Å)
5(fs)	330	430	5	0.23	3.348
6(t)	360	480	8	1.21	3.068

Excited at wavelength corresponding to the maximum of PLE spectra.

The emission follows a three-component exponential decay; the reported τ is the average of the three values.

The photoluminescence excitation spectra (PLE, Figure ) are centered in the UV region with a low energy tail extending in the visible up to ca. 420 nm for 6(t) and are attributed to metal-to-ligand transitions, as confirmed also by TD-DFT calculations of the lowest electronic absorption (see below). In the case of complex 6(t), PL and PLE bands are red-shifted relative to those of the corresponding diimidazolium salt 2-PF (Figure S1). This is mainly ascribed to the perturbation provided by coordination of the ligand to the metal centers. Instead, 1-PF did not show appreciable luminescence. The emission decay curves of both 5(f) and 6(t) were interpolated with a three-exponential function, resulting in average lifetimes of 0.23 and 1.21 μs, respectively.

In the solid state, the emission wavelength can be correlated to the aurophilic interaction, which in turn might depend on the Au···Au distance found in the single-crystal X-ray diffraction studies. In fact, a stronger aurophilic contact is expected to lead to a lower emission energy. Accordingly, we observed the emission band of 6(t) red-shifted compared to that of 5(f).

The emission properties of gold(I) complexes depend both on the extent of aurophilic interactions and on molecular stacking in the lattice that can be tailored by the ligands thanks, for example, to the presence of supramolecular interactions. In this case, we did not observe the formation of particular supramolecular arrangements in the solid state: in the dinuclear complex 6(t), the Au···Au distance (3.068 Å) is indeed shorter than that observed in highly emissive samples (PLQY = 96%, 3.272 Å) having however a peculiar staggered arrangement in the unit cell.[28] The absence of similar spatial arrangement of the complexes in the crystals of 5(f) and 6(t) could account for the moderate quantum yields registered for these complexes. These results further support the difficulty to foresee the extent of the luminescence of this type of complexes because several factors (aurophilicity, crystal packing, π–π interactions, etc.) play a role preventing any a priori prediction.

DFT Analysis

As described in the previous sections and in particular in the description of the single-crystal structures, the dinuclear gold(I) complexes can differ not only for the coordination around gold (head-to-tail or head-to-head coordination of the heteroditopic ligand) but also for the relative orientation of the propylene bridges: twisted (t), folded-syn (fs), folded-anti (fa), and stretched-out (s) (Chart ). Consequently, at least eight isomers might be obtained for each gold(I)–gold(I) complex. First, we optimized the structures of the four possible conformers (t, fs, fa, s) of the head-to-tail complex 6 at the ZORA-BLYP/TZ2P level of theory (Figure S2a). For complex 6(t), the crystallographic coordinates were used as initial guess, whereas the remaining three conformers were built manually. The twisted and both the folded (syn and anti) conformers (6(t), 6(f), and 6(f)) are close in energy, differing by ca. 3 kcal mol–1, whereas the energy of the stretched-out 6(s) is 11.3 kcal mol–1 higher than the most stable one, 6(t). Importantly, this last complex is also the one isolated by crystallization. Considering these results and the obtained crystal structures, reported in the previous section, the complexes 5, 5′, and 6′ were also optimized at the ZORA-BLYP/TZ2P level of theory but considering only the twisted and folded-syn conformation of the bridges (Figure S2b). For 5(f), the crystallographic coordinates were used as initial guess, whereas the other structures were built manually. The comparison between the computed and crystallographic structures reveals a nice agreement; the Au–Au distance is only slightly overestimated. In all the complexes, the Au–Au distance is in the range 3.01–3.29 Å. The position of the propylene bridges affects the overall geometry since it imposes a more spherical shape (twisted conformer) or a hemispherical shape (folded-syn species), leading to an antiparallel and an almost parallel orientation of the C-Au-C axes, respectively. Notably, the shortest Au–Au distances (3.01–3.17 Å) are calculated for the twisted structures [5(t), 5′(t), 6(t), and 6′(t)] and the longest ones (3.24–3.31 Å) in the remaining complexes. In general, the antiparallel orientation of the C-Au-C axes (twisted conformer) increases the stability of the complex; only 5(f) lies 2.0 kcal mol–1 below 5(t), as reported in Table S1. In fact, besides aurophilicity, the secondary interactions of the bulky ligands must be taken into account (see Figure S3 and comments in the Supporting Information for details). The unexpected stability of 5(f) compared to 5(t) is likely due to the pronounced shift of the Au biscarbene moieties and the onset of two Au−π interactions with both phenyl rings (the distance between the metal and the centroids is 3.7 Å), which are arranged so that the overall molecular symmetry is almost perfectly C2 (binary axis orthogonal to the Au-Au axis).

Focusing on complexes 5(f) and 6(t), the presence of symmetry in the former reflects in a greater stability in terms of the HOMO–LUMO gap (Table ). The frontier Kohn–Sham MOs in 6(t) have strong metal (HOMO) and strong ligand (LUMO) character, respectively; conversely, in 5(f), HOMO lobes are significantly expanded on the ligands (51.01% Au nuclei and 28.28% ligands), whereas LUMO is strongly ligand centered (Figure ).

Table 2

HOMO–LUMO Gaps (eV), HOMO and LUMO Compositions (Percentage Contributions of Au and Ligands) of Complexes 5(f) and 6(t); Level of Theory: ZORA-BLYP-D3(BJ)/TZ2Pa

		HOMOb		LUMOb
	HOMO–LUMO	Au	ligands	Au	ligands
5(fs)	3.032	51.01	28.28	5.74	78.6
6(t)	2.891	72.14			87.46

A complete analysis for all the conformers is reported in Table S4.

Only the relevant contributions are summed, explaining why the total sum is not 100%.

Figure 6

Kohn–Sham HOMOs (left) and LUMOs (right) of complexes 5(f) and 6(t). Level of theory: ZORA-BLYP-D3(BJ)/TZ2P; orbital isosurfaces: 0.03.

To understand the luminescence properties of 6(t) and 5(f), we first calculated their absorption spectra in the gas phase with the aim of identifying the lowest and strongest excitations from the ground state. The SAOP model combined with all-electron TZ2P basis sets was employed, and scalar relativistic effects were taken into account.

For 6(t), the lowest singlet–singlet absorption (S1 ← S0) falls in the near-visible region at 390.6 nm (f = 0.09) and corresponds to an almost purely monoelectronic HOMO–LUMO transition (98.3%) with net metal-to-ligand-charge-transfer (MLCT) character. A bunch of absorptions at close wavelengths is then found in the range 375–346 nm. In these transitions, the MLCT character is less pronounced because the filled MOs of the couples involved in the dominant excitations have a significant ligand percentage contribution. Another group of close peaks falls in the range 346–328 nm with similar or even pure ligand-to-ligand character. The strong absorption at 310 nm is due to ligand-based transitions, but the lack of dominant contributions precludes a precise assignment.

The lowest absorption in the UV–vis spectrum of 5(f) is blue-shifted if compared to that of 6(t); it is computed at 378 nm (f = 0.002). It is ascribed to an almost pure HOMO–LUMO monoelectronic transition (99.4%) with net MLCT character as observed for 6(t). Another absorption of similar intensity (f = 0.004) is computed at 365 nm and has monoelectronic composition HOMO – 1 – LUMO + 1 (99.4%) with ligand-to-ligand character. In the transitions corresponding to the intense peaks computed at 337, 322, and 315 nm, the MLCT character decreases since the contributions of the Au lobes on the involved filled MOs are rather small. Finally, the very intense absorptions at 289 and 283 nm (f = 0.07 in both cases) involve couples of delocalized MOs with significant metal and ligand contributions.

The composition of the lowest singlet–triplet absorption (T1 ← S0) is almost identical to the singlet–singlet one, that is, 99% HOMO–LUMO, and falls at 384 nm (5(f)) and at 397.3 nm (6(t)). Both 5(f) and 6(t) were fully optimized in the lowest triplet state, and the phosphorescence wavelengths were estimated subtracting the energy of the ground state (level of theory: ZORA-BLYP/TZ2P). Emission as phosphorescence is calculated at 384 nm (5(f)) and 431 nm (6(t)). These wavelengths are blue-shifted with respect to the experimental ones, but their separation is in very good agreement (experimental 47 nm vs calculated 50 nm), considering also that the SAOP potential, which is optimized for excited states, was not used because it not available for the triplet geometry optimization.

The Au···Au interaction is greatly enhanced in the T1 excited state, with much shorter Au···Au distances compared to those measured in the ground state (see Table S3). This was already observed in other dinuclear gold(I) complexes having bridging dicarbenes or diphosphines coordinated to the metal centers.[28,37,42]

Synthesis of Palladium(II) Complex

The study on the coordinating properties of this type of nNHC-tzNHC ligands was also extended to palladium(II), a metal center characterized by a coordination geometry different from linearity, with the aim to explore the possibility of obtaining isomers even when this type of ligands is in a chelate fashion. The transmetalation of the dicarbene ligand was successfully used for the syntheses of palladium(II) complexes 8/8′ starting from the silver(I) complex 4. [PdCl2(COD)] (COD = 1,5-cyclooctadiene) was used as the palladium precursor due to the labile nature of the diene ligand, and the reaction was performed in acetonitrile at room temperature for 3 h with an Ag/Pd 2:1 molar ratio (Scheme ).

Scheme 5

Synthesis of the Complexes 8/8′

As already observed for gold complexes, two sets of signals can be detected in both the 1H and 13C NMR spectra, associable to the two isomers 8 and 8′ (Scheme ) in a 3:2 ratio (referred to the CH3 signals). This might be explained with a different stability of the two species, related to the different steric hindrance around the metal center. In fact, in complex 8′, the phenyl rings are in close proximity, and therefore this conformer is expected to be less stable.

The chelating coordination of the ligands on the metal center is supported by the presence of several multiplets associated with the CH2 protons of the bridge as a consequence of the formation of a rigid 8-membered metallacycle. The 13C NMR spectrum further supports the coordination of the dicarbene ligand to the palladium(II) center: the nNHC carbene carbon signal is observed at δ ca. 171 ppm, coherent with the literature.[43−45] The ESI-MS spectra accredit the coordination of two ligands on the same metal center due to the presence of the fragment [PdL2PF6]+ (with L = nNHC-tzNHC dicarbene ligand) at m/z 813.19.

By slow diffusion of diethyl ether into an acetonitrile solution of complexes 8/8′, single crystals were obtained; the crystal structure was solved by X-ray diffraction studies. In Figure , the molecular structure of the palladium complex 8 is shown.

Figure 7

ORTEP-style view of cationic compound 8, with atomic numbering scheme; ellipsoids are drawn at their 30% probability level. Hydrogen atoms have been omitted for clarity. Selected bond distances (Å) and angles (deg): Pd1–C9 2.058(5), Pd1–C1 2.062(6), N1–C1 1.354(7), N2–C1 1.364(7), N3–N4 1.337(6), N4–N5 1.311(7); C9′–Pd1–C1 92.4(2), and C9–Pd1–C1 87.6(2). Code for atoms: ′ = −x, −y, −z.

It confirms the dicationic nature of complex 8 with two dicarbene ligands chelating on the palladium center and forming two 8-membered rings. The cationic compound crystallizes with the PF6– anion and acetonitrile solvent molecules. The structure of the cationic complex is centrosymmetric, with the metal atom located on the inversion center. The palladium(II) presents, as expected, a square planar configuration: the C1-Pd1-C9 and C9-Pd1-C1 deviated slightly from the ideal 90° value (92.4(2) and 87.6(2)°). The distances between the palladium(II) center and the carbene carbon atoms are 2.062(6) Å (nNHC) and 2.058(5) Å (tzNHC), in the expected range of distances for this type of compounds.[26,46−48] The dihedral angle between the phenyl substituent and the triazole ring is twisted by 71.84°, thus minimizing the steric hindrance.

Conclusions

Gold(I) and palladium(II) complexes based on bidentate heteroditopic nNHC-tzNHC dicarbene ligands were successfully synthesized via transmetalation of the dicarbene ligand from the corresponding silver(I) complex. The gold(I) complexes are dinuclear species with the two bidentate ligands bridging the two metal centers, whereas the palladium(II) complex is mononuclear with the two dicarbene ligands chelating the metal. In the case of gold(I) and palladium(II) complexes, two isomers, that is, head-to-tail or head-to-head, were obtained due to the two different carbene moieties on the ligand. The stability of the other plausible gold(I) isomers was analyzed through DFT calculations, which confirm that the species that crystallize are the most stable ones. The presence of the two isomers was not evident for silver(I) complexes probably because of the well-known dynamic behavior of the silver carbene species.[34] Interestingly, the gold(I) complexes were found to emit (PLQY 8%) in the blue region; the comparison between the measured luminescence and the results of the DFT calculations for the S0 ← T1 transition is in good agreement.

Considering the presence of two different carbene functions on the ligand, we are currently working on the possibility of obtaining heterobimetallic complexes via step-by-step deprotonation of the azolium rings of the dicarbene proligand.

Experimental Section

Materials and Methods

Reagents and solvents were high-purity commercially available products and generally used as received. All manipulations were carried out using standard Schlenk techniques under an atmosphere of argon or dinitrogen. NMR spectra were recorded on a Bruker Avance 300 MHz (300.1 MHz for 1H, 121.5 for 31P, and 75.5 for 13C) or on a Bruker DMX 600 (599.98 MHz for 1H NMR and 150.07 MHz for 13C NMR) or on a Bruker DRX 400 (400.13 MHz for 1H and 100.62 MHz for 13C); chemical shifts (δ) are reported in units of ppm relative to the residual solvent signals. High-resolution mass spectroscopy (HR-MS) analyses were performed in the ESI mode using a microTOF instrument (Bruker Daltonics). ESI-MS analyses were performed using an LCQ-Duo (Thermo-Finnigan) operating in the positive ion mode. Compounds were dissolved in acetonitrile, and the sample solutions were directly infused into the ESI source by a syringe pump. Elemental analyses were carried out with a Thermo Scientific FLASH 2000 apparatus. 1-(Pent-4-ynyl)-1H-imidazole, proligand 1-I was prepared according to the literature procedure.[31] In a few cases, the elemental analysis results exceed the range accepted for establishing the analytical purity of a compound, but they represent the best obtained values. The purity of the compounds was also established on the basis of NMR spectra (the products are completely soluble in the NMR solvent, and only the signals of the analyzed species are recorded) and HR-ESI measurements.

Synthesis of Imidazolium–Triazolium Salts

Synthesis of 4-[3-(1H-Imidazol-1-yl)propyl]-1-phenyl-1H-1,2,3-triazole

1-(Pent-4-ynyl)-1H-imidazole (0.30 g, 2.25 mmol), sodium-l-ascorbate (0.24 g, 1.18 mmol), and copper(II) sulfate pentahydrate (0.29 g, 1.16 mmol) were placed in a two-neck round-bottom flask. Methanol (10 mL) and a solution of phenyl azide (0.5 M in t-butyl-methylether, 4.5 mL, 2.25 mmol) were then added, and the reaction mixture was left under stirring for 72 h at 25 °C. Afterward, methanol was removed under vacuum; the residue was dissolved in CH2Cl2 (50 mL) and washed with water (3 × 50 mL). The organic phase was dried over Na2SO4, and the solvent was then removed, affording a yellow oil, which was further washed with diethyl ether (3 × 5 mL) (yield 33%). 1H NMR (CDCl3, 300 MHz) δ = 2.30 (s br, 2H, CH2), 2.65 (s br, 2H, CH2), 4.10 (s br, 2H, CH2), 7.26–7.42 (m, 9H, CH + CH Ar + CH tz + CH im).

Synthesis of 3-(Phenyl)-1-methyl-5-[3-(3-methyl-1H-imidazol-3-ium-1-yl)propyl]-3H-1,2,3-triazol-1-ium diiodide 2-I

CH3I (0.30 mL, 4.80 mmol) was added to a solution of 4-[3-(1H-imidazol-1-yl)propyl]-1-(phenyl)-1H-1,2,3-triazole (0.11 g, 0.45 mmol) in CH3CN (5 mL). The reaction mixture was refluxed for 18 h, obtaining an orange solution. The solvent was removed under vacuum, and the orange oil was washed with diethyl ether, giving an orange solid (yield 82%). 1H NMR (CD3CN, 300 MHz) δ = 2.50 (t, 3JHH = 7.5 Hz, 2H, CH2), 3.13 (t, 3JHH = 7.5 Hz, 2H, CH2), 3.89 (s, 3H, CH3), 4.44 (s, 3H, CH3), 4.48 (t, 3JHH = 7.5 Hz, 2H, CH2), 7.45 (s, 1H, CH), 7.69 (m, 3H, CH Ar), 7.76 (s, 1H, CH), 7.99 (m, 2H, CH Ar), 9.24 (CH tz), 9.44 (CH im). 13C{1H} NMR (CD3CN, 75 MHz): δ = 14.9 (CH2), 20.4 (CH2), 36.4 (CH3), 39.0 (CH3), 48.3 (CH2), 122.3 (CH), 123.9 (CH), 128.2 (CH tz), 129.2 (CH Ar), 129.3 (CH Ar), 129.7 (CH Ar), 132.1 (C Ar), 136.3 (CH im), 144.0 (C).

General Synthesis of 3-(Substituted)-1-methyl-5-[3-(3-methyl-1H-imidazol-3-ium-1-yl)propyl]-3H-1,2,3-triazol-1-ium bis(hexafluorophosphate) 1-PF and 2-PF

A solution of KPF6 (1.96 mmol) in H2O (3 mL) was added to a solution of 1-I or 2-I (0.37 mmol) in MeOH (3 mL). The mixture was stirred for 24 h at room temperature, affording a light-brown solid that was filtered and dried under vacuum. 1-PF. Yield 99%. 1H NMR (CD3CN, 300 MHz) δ = 2.14 (t, 3JHH = 7.5 Hz, 2H, CH2), 2.78 (t, 3JHH = 7.5 Hz, 2H, CH2), 3.83 (s, 3H, CH3), 4.09 (s, 3H, CH3), 4.20 (t, 3JHH = 7.5 Hz, 2H, CH2), 5.68 (s, 2H, CH2), 7.36 (s, 1H, CH), 7.39 (s, 1H, CH), 7.46 (s, 5H, CH Ar), 8.19 (s, 1H, CH tz), 8.45 (s, 1H, CH im). 13C{1H} NMR (CD3CN, 75 MHz): δ = 20.8 (CH2), 27.9 (CH2), 37.0 (CH3), 38.7 (CH3), 49.1 (CH2), 57.9 (CH2), 123.3 (CH), 124.8 (CH), 129.2 (CH tz), 130.1 (CH Ar), 130.2 (CH Ar), 130.6 (CH Ar), 133.1 (C Ar), 137.3 (CH im), 144.3 (C). 31P{1H} NMR (CD3CN, 121 MHz): δ = −144.6 (heptet, PF6). 2-PF. Off-white solid, yield 81%. 1H NMR (CD3CN, 300 MHz) δ = 2.32 (t, 3JHH = 7.5 Hz, 2H, CH2), 2.93 (t, 3JHH = 7.5 Hz, 2H, CH2), 3.86 (s, 3H, CH3), 4.24 (s, 3H, CH3), 4.29 (t, 3JHH = 7.5 Hz, 2H, CH2), 7.39 (s br, 1H, CH), 7.46 (s br, 1H, CH), 7.72 (m, 3H, CH Ar), 7.86 (m, 2H, CH Ar), 8.55 (s, 1H, CH tz), 8.74 (s, 1H, CH im). 13C{1H} NMR (CD3CN, 75 MHz): δ = 20.8 (CH2), 27.8 (CH2), 37.0 (CH3), 39.0 (CH3), 49.1 (CH2), 122.5 (CH), 123.3 (CH), 125.0 (CH tz), 127.7 (CH Ar), 131.5 (CH Ar), 133.0 (CH Ar), 136.0 (C Ar), 137.2 (CH im), 144.8 (C). 31P{1H} NMR (CD3CN, 121 MHz): δ = −144.7 (heptet, PF6).

Synthesis of the Silver(I) Complexes 3 and 4

Synthesis of 3

The proligand 1-PF (0.10 g, 0.17 mmol) and Ag2O (0.20 g, 0.87 mmol) were placed in a two-neck round-bottom flask. Subsequently, CH3CN (15 mL) was added, and the reaction mixture was left under stirring for 48 h at 85 °C. Afterward, it was filtered on Celite, and the solvent was removed under vacuum, giving 3 as a light-brown oil. 1H NMR (CD3CN, 300 MHz) δ = 2.39 (t, 3JHH = 6.6 Hz, 2H, CH2), 2.77 (t, 3JHH = 6.6 Hz, 2H, CH2), 3.51 (s, 3H, CH3), 4.07 (s, 3H, CH3), 4.11 (m, 2H, CH2), 5.40 (s, 2H, CH2), 7.25 (m, 6H, CH + CH Ar), 7.45 (s, 1H, CH). 13C{1H} NMR (CD3CN, 75 MHz): δ = 22.1 (CH2), 29.2 (CH2), 37.1 (CH3), 38.8 (CH3), 50.5 (CH2), 59.6 (CH2), 121.8 (CH), 124.2 (CH), 128.9 (CH Ar), 129.6 (CH Ar), 129.8 (CH Ar), 136.0 (C Ar), 148.5 (C), 164.2 (CAg tzNHC), 181.1 (CAg nNHC). Traces of acetic acid are present (as indicated by the signals at ca. 20 and 173 ppm) probably due to the hydrolysis of the acetonitrile solvent. HR-MS (positive ions, CD3CN): m/z calcd for C17H21AgN5 (25%, [AgL]+) 402.0842, found 402.0958, m/z calcd for C34H42Ag2F6N10P (100%, [Ag2L2PF6]+) 951.1331, found 951.1356; with L = nNHC-tzNHC dicarbene ligand.

Synthesis of 4

Proligand 2-PF (0.051 g, 0.09 mmol) and Ag2O (0.12 g, 0.50 mmol) were placed in a two-neck round-bottom flask. Acetonitrile (7 mL) was added, and the reaction mixture was left under stirring for 48 h at 85 °C. Afterward, it was filtered on Celite, and the volatiles were removed at low pressure. An off-white solid formed upon addition of diethyl ether (20 mL), and it was filtered and dried under vacuum (yield 78%). 1H NMR (CD3CN, 300 MHz) δ = 2.36 (m, 2H, CH2), 2.81 (m, 2H, CH2), 3.50 (s, 3H, CH3), 4.07 (t, 3JHH = 6.6 Hz, 2H, CH2), 4.15 (s, 3H, CH3), 7.09 (s, 1H, CH), 7.20 (s, 1H, CH), 7.48 (m, 3H, CH Ar), 7.73 (m, 2H, CH Ar). 13C{1H} NMR (CD3CN, 75 MHz): δ = 21.8 (CH2), 28.3 (CH2), 37.5 (CH3), 38.7 (CH3), 50.0 (CH2), 121.3 (CH), 123.8 (CH), 124.6 (CH Ar), 130.6 (CH Ar), 131.3 (CH Ar), 140.7 (C Ar), 148.2 (C), 164.5 (CAg tzNHC), 181.5 (CAg nNHC). HR-MS (positive ions, CD3CN): m/z calcd for C32H38Ag2F6N10P (100%, [Ag2L2PF6]+) 923.1021, found 923.1026, with L = nNHC-tzNHC dicarbene ligand.

Synthesis of the Gold(I) Complexes 5/5′ and 6/6′

Synthesis of 5/5′

Method A

Proligand 1-PF (0.049 g, 0.086 mmol) and Ag2O (0.10 g, 0.44 mmol) were placed in a two-neck round-bottom flask. Acetonitrile (10 mL) was added, and the reaction mixture was left under stirring for 48 h at 85 °C. Afterward, the mixture was filtered on Celite, and a solution of AuCl(SMe2) (0.026 g, 0.088 mmol) in acetonitrile (5 mL) was added to the filtrate. The reaction mixture was left under stirring for 3 h at room temperature and then filtered on Celite; finally, the solvent was removed under vacuum in order to obtain an oil. Recrystallization by CH3CN/Et2O affords the product as an off-white solid (global yield 78%). Anal. Calcd for C34H42N10Au2P2F12: C, 32.02; H, 3.32; N, 10.99%. Found: C, 31.18; H, 3.25; N, 10.72%. ESI-MS (positive ions, CH3CN): m/z 1129.27 [100%, Au2L2PF6]+; with L = nNHC-tzNHC dicarbene ligand. The 1H NMR spectra show two set of signals, attributable to the complexes 5 and 5′, in a 1:1 ratio. 5. 1H NMR (CD3CN, 600 MHz,) δ = 2.53 (m, 2H, CH2), 2.86 (m, 2H, CH2), 3.42 (s, 3H, CH3), 4.10 (s, 3H, CH3), 4.23 (m, 2H, CH2), 5.44 (s, 2H, CH2), 7.13 (d, 1H, CH), 7.25 (d, 1H, CH), 7.33 (m, 5H, CH Ar). 13C{1H} NMR (CD3CN, 150 MHz): δ = 22.0 (CH2), 28.0 (CH2), 37.7 (CH3), 38.0 (CH3), 49.9 (CH2), 59.1 (CH2), 121.2 (CH), 124.4 (CH), 128.6 (CH Ar), 129.0 (CH Ar), 129.7 (CH Ar), 135.5 (C Ar), 147.8 (C), 170.1 (CAu tzNHC), 185.9 (CAu nNHC). 5′. 1H NMR (CD3CN, 600 MHz,) δ = 2.52 (m, 2H, CH2), 2.87 (m, 2H, CH2), 3.62 (s, 3H, CH3), 4.06 (s, 3H, CH3), 4.26 (m, 2H, CH2), 5.33 (s, 2H, CH2), 7.17 (d, 1H, CH), 7.20 (d, 1H, CH), 7.28–7.38 (m, 5H, CH Ar). 13C{1H} NMR (CD3CN, 150 MHz): δ = 22.2 (CH2), 28.7 (CH2), 37.8 (CH3), 38.3 (CH3), 50.1 (CH2), 58.8 (CH2), 121.8 (CH), 124.6 (CH), 128.6 (CH Ar), 129.8 (CH Ar), 129.9 (CH Ar), 135.6 (C Ar), 148.0 (C), 171.7 (CAu tzNHC), 184.7 (CAu nNHC).

Method B

A mixture of sodium acetate (0.015 g, 0.183 mmol), the diimidazolium salt 1-PF (0.048 g, 0.082 mmol), and AuCl(SMe2) (0.024 g, 0.082 mmol) in dimethylformamide (10 mL) was heated and maintained at 120 °C for 3 h. The mixture was then filtered on Celite, and n-hexane (70 mL), dichloromethane (10 mL), and diethyl ether (50 mL) were added to the filtrate, affording a white solid. The precipitate was filtered and dried under vacuum (yield 77%). The 1H NMR spectra show two sets of signals, attributable to the complexes 5 and 5′, in a ca. 4:1 ratio. Crystals of complex 5 were obtained by slow diffusion of diethyl ether into an acetonitrile solution of the crude mixture.

Synthesis of 6/6′

A solution of AuCl(SMe2) (0.033 g, 0.113 mmol) in acetonitrile (5 mL) was added to a solution of silver(I) complex 4 (0.060 g, 0.056 mmol) in the same solvent (5 mL). The reaction mixture was left under stirring for 3 h at room temperature; afterward, it was filtered on Celite and the solvent was removed under vacuum in order to obtain a white solid (yield 75%). Anal. Calcd for C32H38N10Au2P2F12: C, 30.83; H, 3.07; N, 11.24%. Found: C, 30.40; H, 3.15; N, 10.86%. ESI-MS (positive ions, CH3CN): m/z 1101.23 [100%, Au2L2PF6]+, L = nNHC-tzNHC dicarbene ligand. The 1H NMR spectra show two sets of signals, attributable to the complexes 6 and 6′, in a 1:1 ratio. 6. 1H NMR (CD3CN, 400 MHz) δ = 2.35 (m, 2H, CH2), 2.83 (m, 2H, CH2), 3.46 (s, 3H, CH3), 4.16 (m, 2H, CH2), 4.18 (s, 3H, CH3), 7.00 (d, 3JHH = 2.0 Hz, 1H, CH), 7.18 (d, 3JHH = 2.0 Hz, 1H, CH), 7.45 (m, 2H, CH Ar), 7.57 (m, 1H, CH Ar), 7.70 (m, 2H, CH Ar). 13C{1H} NMR (CD3CN, 100 MHz): δ = 21.6 (CH2), 27.5 (CH2), 38.0 (CH3), 38.2 (CH3), 49.2 (CH2), 121.0 (CH) 124.2 (CH Ar), 124.5 (CH), 130.5 (CH Ar), 131.5 (CH Ar), 139.9 (C Ar), 148.1 (C), 169.7 (CAu tzNHC), 183.8 (CAu nNHC). Crystals of complex 6 were obtained by slow diffusion of diethyl ether into a solution of the complexes 6 and 6′ in acetonitrile. 6′. 1H NMR (CD3CN, 400 MHz) δ = 2.65 (m, 2H, CH2), 2.99 (m, 2H, CH2), 3.34 (s, 3H, CH3), 4.16 (m, 2H, CH2), 4.19 (s, 3H, CH3), 7.13 (d, 3JHH = 2.0 Hz, 1H, CH), 7.27 (d, 3JHH = 2.0 Hz, 1H, CH), 7.47 (m, 2H, CH Ar), 7.54 (m, 1H, CH Ar), 7.70 (m, 2H, CH Ar). 13C{1H} NMR (CD3CN, 100 MHz): δ = 22.4 (CH2), 27.9 (CH2), 38.0 (CH3), 38.2 (CH3), 49.9 (CH2), 121.2 (CH), 124.3 (CH Ar), 124.7 (CH), 130.6 (CH Ar), 131.4 (CH Ar), 139.9 (C Ar), 147.3 (C), 169.3 (CAu tzNHC), 184.5 (CAu nNHC).

Synthesis of the Palladium(II) Complexes 8/8′

A solution of [PdCl2(COD)] (0.014 g, 0.049 mmol) in acetonitrile (5 mL) was added to a solution of silver(I) complex 4 (0.052 g, 0.049 mmol) in the same solvent (5 mL). The reaction mixture was left under stirring for 3 h at room temperature; afterward, it was filtered on Celite, and the solvent was removed under vacuum, giving a white solid (yield 74%). HR-MS (positive ions, CD3CN): m/z calcd for C32H38N10PdPF6 (100%, [PdL2PF6]+) 813.1957, found 813.2056, L = nNHC-tzNHC dicarbene ligand. The 1H NMR spectrum shows two sets of signals, attributable to the complexes 8 and 8′, in a 3:2 ratio. 8. 1H NMR (CD3CN, 400 MHz) δ = 2.35 (m, 2H, CH2), 2.89 (m, 2H, CH2), 3.71 (s, 3H, CH3), 3.92 (s, 3H, CH3), 4.23 (m, 2H, CH2), 7.01 (d, 3JHH = 1.35 Hz, 1H, CH), 7.04 (d, 3JHH = 1.35 Hz, 1H, CH), 7.17 (m, 2H, CH Ar), 7.70 (m, 3H, CH Ar). 13C{1H} NMR (CD3CN, 400 MHz): δ = 25.8 (CH2), 27.4 (CH2), 37.2 (CH3), 39.3 (CH3), 52.0 (CH2), 124.0 (CH), 124.2 (CH), 126.9 (CH Ar), 131.0 (CH Ar), 132.0 (CH Ar), 140.73 (C Ar), 145.4 (C), 157.0 (CPd tzNHC), 170.6 (CPd nNHC). 8′. 1H NMR (CD3CN, 400 MHz) δ = 2.89 (m, 2H, CH2), 3.40 (m, 2H, CH2), 3.71 (s, 3H, CH3), 3.92 (s, 3H, CH3), 4.13 (m, 2H, CH2), 6.90 (d, 3JHH = 1.35 Hz, 1H, CH), 7.01 (d, 3JHH = 1.35 Hz, 1H, CH), 7.18 (m, 2H, CH Ar), 7.70 (m, 3H, CH Ar). 13C{1H} NMR (CD3CN, 400 MHz): δ = 23.8 (CH2), 26.1 (CH2), 37.1 (CH3), 39.3 (CH3), 53.1 (CH2), 123.8 (CH), 125.3 (CH), 127.1 (CH Ar), 130.7 (CH Ar), 132.1 (CH Ar), 140.3 (C Ar), 146.1 (C), 170.1 (CPd nNHC), (CPd tzNHC not detected). Crystals of complex 8 were obtained by slow diffusion of diethyl ether into a solution of the complexes 8 and 8′ in acetonitrile. The assignment of the signals to 8 or 8′ has been made on the abundance in the 1H NMR spectra, attributing the more intense signals to complex 8, the one that crystallizes.

Crystal Structure Determination of Compounds 5, 6, 7, and 8

The crystallographic data for compounds 5, 6, 7, and 8 were collected on a Bruker Apex II single-crystal diffractometer working with monochromatic Mo Kα radiation and equipped with an area detector.[49] The structures were solved by direct methods and refined against F2 with SHELXL-2014/7 with anisotropic thermal parameters for all non-hydrogen atoms,[50] except the carbon and oxygen atoms of the diethyl ether solvent molecule in compound 7 that have been refined isotropically. In some cases, the propyl chains of the ligands have been found disordered in two positions and refined with the appropriate occupancy factor. Idealized geometries were assigned to the hydrogen atoms. For the molecular graphics, Program ORTEP-3 for Windows has been used.[51] Crystal data for the four compounds are reported in Table . Crystallographic data have been deposited with the Cambridge Crystallographic Data Centre as supplementary publication. Copies of the data can be obtained free of charge on application to the CCDC, 12 Union Road, Cambridge CB2 1EZ, U.K. (fax, (+44) 1223 336033; e-mail, deposit@ccdc.cam.ac.uk).

Table 3

Crystal Data for Compounds 5, 6, 7, and 8a

compound	5	6	7	8
formula	C34H42Au2F12N10P2	C34H41Au2F12N11P2	C40H50AgAu2F18N12O0.5P3	C36H44F12N12P2Pd
molecular weight	1274.66	1287.65	1643.63	1041.17
crystal system	monoclinic	monoclinic	monoclinic	monoclinic
space group	C2/c	C2/c	C2/c	P21/c
a [Å]	8.3231(4)	24.845(2)	22.366(2)	11.917(3)
b [Å]	18.3439(9)	14.7302(12)	23.882(2)	12.117(3)
c [Å]	26.7035(13)	24.302(2)	21.841(2)	15.264(4)
β [deg]	91.709(2)	105.8310(10)	106.733(2)	97.907(5)
V [Å3]	4075.2(3)	8556.8(12)	11 172.3(18)	2183.0(10)
temperature [K]	200	173	293	173
Z	4	8	8	2
Dcalc [g·cm–3]	2.078	1.999	1.954	1.584
μ [cm–1]	7.367	7.019	5.776	0.593
F(000)	2448	4944	6320	1056
reflections collected	35 864	55 629	66 312	23 253
independent reflections	6192	10 024	11 118	3843
reflections in refinement	5667	8133	6478	2915
R(int)	0.0539	0.0571	0.0752	0.0618
refined parameters	274	564	686	289
R1 [I > 2σ(I)]	R1 = 0.0578	R1 = 0.0542	R1 = 0.0530	R1 = 0.0596
	wR2 = 0.1760	wR2 = 0.1079	wR2 = 0.1399	wR2 = 0.1610
wR2 [all data]	R1 = 0.0605	R1 = 0.0706	R1 = 0.1003	R1 = 0.0785
	wR2 = 0.1796	wR2 = 0.1133	wR2 = 0.1671	wR2 = 0.1814
GOF	1.061	1.184	1.025	1.049
CCDC	1873190	1873191	1873192	1873193

R1 = ∑|Fo – Fc|/∑(Fo); wR2 = [∑[w(Fo2 – Fc2)2]/∑[w(Fo2)2]]1/2.

Luminescence Studies

Luminescence spectra were recorded at room temperature on films obtained by drop casting a sample solution in acetonitrile on a quartz substrate using an optical fiber bundle coupled to the spectrofluorimeter (Fluorolog-3, Horiba JobinYvon) equipped with a double-grating monochromator on both the excitation and emission sides. A 450 W Xe arc lamp and an R928P Hamamatsu photomultiplier were employed as the excitation source and the detector, respectively. The emission spectra were corrected for detection and optical spectral response of the spectrofluorimeter supplied by the manufacturer. The excitation spectra were corrected for the spectral distribution of the lamp intensity using a photodiode reference detector. Absolute photoluminescence quantum yields (PLQYs) were measured by means of a Spectralon-coated integrating sphere accessory (4″, F-3018, Horiba Jobin-Yvon) fitted in the fluorimeter sample chamber. Three independent measurements were carried out on each sample, and the error on PLQY was 20%.

Photoluminescence decay curves were obtained through single-photon experiments using a 295 nm pulsed LED as excitation sources (Horiba NanoLED). The collected data were analyzed with Horiba DAS6 Decay Analysis Software.

Computational Details

Density functional theory (DFT) calculations were carried out with the Amsterdam density functional (ADF) program.[52−54] The BLYP[55−58] density functional was used, and the Grimme dispersion was added (BLYP-D3(BJ)).[59] The zeroth-order regular approximation (ZORA)[60] was chosen, including the scalar relativistic effects, as recommended in the presence of heavy nuclei.[61−63] The TZ2P basis set, which is a large uncontracted set of Slater-type orbitals (STOs), of triple-ζ quality and augmented with two sets of polarization functions on each atom (2p and 3d in the case of H, 3d and 4f in the case of C and N, 6g and 7h in the case of Au), was employed. The frozen-core approximation was used for the core electrons: up to 1s for C and N and up to 4d in the case of Au. For the numerical integration, the Becke grid was used.[64,65] This level of theory gave reliable results in recent studies on gold complexes.[66−68] TD-DFT calculations were performed on the optimized geometries of 5(f) and 6(t) in the gas phase using all-electron TZ2P basis sets for all of the atoms. SAOP, which is an appropriate exchange-correlation potential with the statistical averaging of (model) orbitals, was chosen to calculate the excitation energies[69,70] because it reliably describes the excited states of organometallic compounds.[71−73] The lowest 30 singlet–singlet as well as singlet–triplet excitations were computed. The lowest triplet states of 5(f) and 6(t) were fully optimized in order to estimate the phosphorescence wavelengths of 5(f) and 6(t); the BLYP-D3(DJ) functional combined with the TZ2P small core basis set for all the atoms was used in these unrestricted calculations; scalar relativistic effects were included with the ZORA formalism.