Dinuclear Cu(I) Halides with Terphenyl Phosphines: Synthesis, Photophysical Studies, and Catalytic Applications in CuAAC Reactions

Dinuclear Cu(I) Halides with Terphenyl Phosphines: Synthesis, Photophysical Studies and Catalytic Applications in Cu- AAC Reactions Álvaro Beltrán,† Inmaculada Gata,† Celia Maya,‡ João Avó,§ João Carlos Lima,# César A. T. Laia, # Riccardo Peloso,*,‡ Mani Outis*,# and M. Carmen Nicasio*,†. †Departamento de Química Inorgánica, Universidad de Sevilla, Aptdo 1203, 41071 Sevilla, Spain ‡Instituto de Investigaciones Químicas (IIQ), Departamento de Química Inorgánica and Centro de Innovación en Química Avanzada (ORFEO-CINQA), Consejo Superior de Investigaciones Científicas (CSIC) and Universidad de Sevilla, Avenida Américo Vespucio 49, 41092 Sevilla, Spain § IBB-Institute for Bioengineering and Biosciences, Instituto Superior Técnico, Universidade de Lisboa, 1049-001 Lisboa, Portugal. #LAQV-REQUIMTE, Departamento de Química, Faculdade Ciências e Tecnologia, Universidade NOVA de Lisboa, 2829-516 Monte de Caparica, Portugal. KEYWORDS. Cu(I) complexes, phosphines, CuAAC, iodoalkynes, Phosphorescent Cu(I) Complexes, Non-adiabatic Coupling ABSTRACT: Several dinuclear terphenyl phosphine copper(I) halide complexes of composition [CuX(PR2Ar’)]2 (X = Cl, Br, I; R = hydrocarbyl, Ar’ = 2,6-diarylterphenyl radical), 1-5, have been isolated from the reaction of CuX with 1 equiv of the phosphine ligand. Most of them have been characterized by X-ray diffraction studies in the solid state, thus allowing comparative discussions of different structural parameters, namely CuCu and CuAryl separations, conformations adopted by coordinated phosphines, planarity of the Cu2X2 cores. Centrosymmetric complexes [CuI(PMe2ArXyl2)]2, 1c and [CuI(PEt2ArMes2)]2, 3c, despite their similar structures, show very distinct photoluminescence (PL) in powder form at room temperature. The photophysical behavior of these compounds in liquid solution, solid-solid Zeonex solution and powder samples at room temperature and 77 K have been investigated and supported by DFT calculation. Identification of vibronic coupling modes, done by group theory calculations and the technique of projection operators, shows that the manifestation of these modes is conditioned by crystal packing. Complexes [CuI(PMe2ArXyl2)]2, 1c and [CuI(PEt2ArMes2)]2, 3c display remarkable activity in copper-catalyzed azide-alkyne cycloaddition reactions involving preformed and in situ-made azides. Reactions are performed in H2O, under aerobic conditions, with low catalyst loadings and tolerate the use of iodoalkynes as substrates. INTRODUCTION Cu(I) halides with phosphines display a variety of coordination numbers and geometries primarily dictated by the size of the phosphine and, to a lesser degree, by the nature of the halide.1 Thus for a 1:1 stoichiometry, tetranuclear [CuX(PR3)]4 with cubane-like structure, are frequently encountered for less bulky phosphines such as PEt3,2 PMe2Ph3 and PPh3.4 Furthermore, for the larger halides (Br and I)5 less common stepped cubane structures have also been found. As the size of phosphine increases, dinuclear [CuX(PR3)]2 species with three-coordinate Cu(I) centers are stabilized, as noted for PCy3,6 P(o-Tol)3,7 PBz3,8 (Bz = benzyl) among others.9 Mononuclear two-coordinate [CuX(PR3)] are obtained for the highly sterically demanding phosphines PMes310 (Mes = mesityl) and P(tmp)3 (tmp = 2,4,6-trimethoxyphenyl).11 Due to their flexible coordination properties, Cu(I) halides aggregates have been incorporated, as building blocks, in coordination polymers,1,12 which exhibit a range of photophysical properties.13 In this context, discrete {CunIn} clusters display interesting properties such as polymorphism, thermochromism and thermally activated delayed fluorescence (TADF).14-17 Concerning the Relativistic Nephelauxetic Effect through covalent character of Cu-I, bonds and large spin-orbit coupling (SOC) constant of I18, a remarkable advantage of these compounds can be thought as conferring heaviness to a light metal based core. First row (3d) transition metals are not heavy enough for an efficient SOC but, a high SOC is required for rapid Inter-System-Crossing (ISC) in order to achieve phosphorescent metal complexes. Third row (5d) transition metal based complexes like those of Ir(III) and Pt(II) exhibit the desirable fast ISC as a high SOC is induced by the heavy metal centre itself. These phosphorescent complexes have been the focus of many investigations especially for organic light emitting diodes (OLEDs)19,20 .However these are very expensive metals and their sources are increasingly scarce. Thus, luminescent Cu(I) halides aggregates represents an attractive sustainable alternative to those based on heavy precious metals. Additionally, Cu(I) complexes enjoy a rich catalytic chemistry, including cross-coupling reactions,21 oxidations,22 borylation reactions,23 C-H functionalization24 and 1,3-dipolar cycloadditons, in particular, the Cu-catalyzed Azide-Alkyne Cycloaddition (Cu- AAC).25 The latter is the most prominent example of click chemistry.26 Nitrogen-containing ligands and N-heterocyclic carbenes, albeit to a lesser extent, have commonly been used as supporting ligands in CuAAC reactions.25,27 However, Cu(I) complexes stabilized by phosphine ligands have scarcely been applied.28 Recently, Díez-González and co-workers have employed mononuclear and tetranuclear Cu(I) halides bearing phosphorus(III) ligands as catalysts in efficient CuAAC reactions performed in water under mild conditions.29 We have developed a family of heteroleptic phosphine ligands bearing a bulky terphenyl substituent, PR2Ar’ (Ar’ = terphenyl radical) (Chart 1) and we have evaluated their electronic and steric parameters.30,31 Herein, we report on the synthesis and structures of a series of dinuclear [CuX(PR2Ar’)]2 complexes. Two new phosphorescent complexes containing the Cu2I2 core have been identified and their excited state dynamics where low-lying electronic transitions are mainly (M+X)LCT are discussed. In addition, the catalytic activity of these complexes in CuAAC reactions is also examined. Chart 1. Terphenyl phosphines used in this work. RESULT AND DISCUSSION Synthesis and structural studies of Cu(I) Complexes. Dinuclear complexes 1-5, of the general formula CuX(PR2Ar’) (PR2Ar’ = L1, X = Cl, Br, I: 1a, 1b, 1c; PR2Ar’ = L2, X = Cl, Br, I: 2a, 2b, 2c; PR2Ar’ = L3; X = Cl, Br, I: 3a, 3b, 3c; PR2Ar’ = L4, X = I: 4c; PR2Ar’ = L5 X = Cl, Br, I: 5a, 5b,5c) were synthesized by mixing equimolar amounts of the corresponding copper(I) halide, CuX, and the dialkyl terphenyl phosphine, PR2Ar’, in dichloromethane at room temperature. The transformation of the metal precursor into the product occurred with concomitant disappearance of the colorless copper halide, which was almost insoluble in the reaction medium. Reaction times of about 12 h were required to achieve high yields. Compounds 1-5 were isolated as air-stable colorless solids. They are highly soluble in chlorinated and aromatic solvents and poorly soluble in saturated hydrocarbons such as pentane, heptane, or petroleum ether. 31P{1H} NMR spectra of complexes 1-5 revealed that the broad resonances of the P nuclei are slightly or moderately shifted to higher frequencies (ca. 0.5 to15 ppm) upon coordination to the metal center, analogously to what observed for related Ni,31 Pt, Rh, Ir,30 and Au32 compounds. The 1H NMR and 13C{1H} NMR spectra of CuX(PR2Ar) species in CDCl3 do not show any relevant differences compared to the free ligands, which is consistent with a monodentate coordination mode of the P ligand in conjunction with fast rotations around the P—Cu and/or P—Caryl bonds. Nevertheless, the previously observed 1-P,1-C coordination of the terphenyl phosphine30-32 cannot be ruled out in solution, given that a fast interchange of the two flanking aryl rings of the terphenyl group in the coordination environment of the metal would also account for the observed NMR properties. If only mononuclear and dinuclear complexes are considered, four possible structures in solution can be envisaged (Fig 1). Figure 1. Possible coordination modes of PR2Ar’ ligands and geometries around the Cu centre for mononuclear and dinuclear CuX(PR2Ar’) species in solution based on NMR data. Arrows indicate fast rotations on the NMR time scale. (a) 1-P monodentate coordination, dinuclear, trigonal planar geometry at Cu; (b) 1-P,1-C bidentate coordination, dinuclear, tetrahedral geometry at Cu; (c) 2-P,C bidentate coordination, mononuclear, trigonal planar geometry at Cu; (d) 1-P monodentate coordination, mononuclear, linear geometry at Cu. DOSY 1H NMR experiments of CDCl3 solutions of complexes 1c, 2a, 3c and 5b were performed in order to get more information about the nuclearity of the CuX(PR2Ar’) species in solution (see Figure 1 in SI). In all cases only one species was detected with calculated diffusion coefficients of ca. 6-7∙10-10 m2∙s-1, in accord with previously reported data for complexes of similar size.33 Moreover, the diffusion coefficients were used to calculate the corresponding hydrodynamic radii of the molecules with the Stokes-Einstein equation giving values of ca 6-7 Å in good agreement with the equivalent radii estimated from the X-Ray molecular structures of the dimers. X-ray diffraction analyses were conducted on single crystals of complexes 1, 2c, 3b, 3c, 4c, 5c, revealing that their molecular structures in the solid consist in bimetallic dihalo-bridged neutral complexes, [Cu(-X)(PR2Ar’)]2, with the metal centers lying in a distorted trigonal planar geometry. This type of complex has numerous precedents in the literature.6-9 ORTEP diagrams Table 1. Selected structural parameters for [Cu(-X)(PR2Ar’)]2 complexes 1-5 in the solid state. Complex X R Ar’ Cu--Cu (Å) Cu-X-Cu (º) [Cu2X2] planarity Cu--Ar (shortest Cu--C) Sum of bond angles at Cu Point group* Conformation of PR2Ar’ 1a Cl Me ArXyl2 3.04 82a no 2.76 356 C1 A and C 1b Br Me ArXyl2 2.88 73a no 2.86a 352 Cs (C1) C 1c I Me ArXyl2 2.62 60.7 yes 2.96 352 C2h(Ci) C 2c I Me ArDipp2 2.69 63a no 3.20a 354a Cs (C1) B 3b Br Et ArMes2 3.15 80.4 yes 2.83 353 C2h(Ci) C 3c I Et ArMes2 3.35 80.5 yes 2.90 352 C2h(Ci) C 4c I iPr ArXyl2 3.12 73.7 yes 3.82 354 C2h(Ci) C 5c I Cyp ArXyl2 3.90 90.0 yes 2.69 352 C2h(Ci) C aMean. *Symmetry of the [CuX(PR2Ar’)]2 complexes reached after slight rotation of the aryl groups around the P-Cipso and C-C axis and/or translations of the X ligands (point group found in the X-ray structure). See discussion. Figure 2. ORTEP views for the molecular structures of complexes 1c (top) and 3c (bottom). Selected bond distances (Å) and angles (°) for 1c: Cu1-P1 2.2171(9), Cu1-I1 2.5853(5), Cu1-I1´ 2.5886(5), Cu1- Cu´2.6175(6), Cu1-I1-Cu1´ 60.78(1), P1-Cu1-I1 115.03(3), P1-Cu1- I1´ 117.55(3), I1-Cu1-I1´ 119.22(2).; for 3c: Cu1-P1 2.217(1), Cu1- I1 2.6125(5), Cu1-I1´ 2.5795(6), Cu1-I1-Cu1´ 80.45(2), P1-Cu1-I1 126.60(3), P1-Cu1-I1´ 126.97(3), I1-Cu1-I1´ 99.55(2). for the molecular structures of complexes 1c and 3c are depicted in Fig. 2 as representative examples. Interestingly, for all structures, the slightly pyramidalized copper atoms lie at ca. 0.3-0.4 Å above the plane defined by the two halides and the P atom and points towards one of the lateral rings of the terphenyl group. This observation may suggest the existence of week bonding interactions, reasonably of electrostatic nature, between the copper centre and one or two carbon atoms of the aromatic rings, especially for complexes 1a and 5c, which exhibit the shortest Cu…Caryl separations (ca. 2.7 Å). Based on the comparison of the Cu…Caryl distances reported in Table 1 with the sum of the covalent radii of C(sp2) and Cu (ca. 2.0 Å),34 the existence of “classical” copper-carbon covalent bonding is ruled out clearly. For the sake of comparison, the extent of copper-car- bon interactions in complexes 1-5 appears much smaller than in related unsaturated platinum complexes [PtMe(OH2)(PMe2ArDipp2)]+,30b [PtMe2(PMe2ArDipp2)]+,30b [PtCl2(PMe2ArDipp2)]+,30a and [Pt(tBuCH=CH2)(PMe2ArDipp2)],30c in which shorter Pt…Caryl metal-carbon separations of ca. 2.2-2.3 Å are observed. Another point of interest of the structural properties of complexes 1-5 is the existence, in some cases, of typical d10- d10 interactions35 resulting in Cu—Cu separations of approximately 2.6-2.7 Å. For the iodide complexes 1c and 2c, which contain the dimethyl phosphines PMe2ArXyl2 L1 and PMe2ArDipp2 L2, the Cu—Cu distance matches almost perfectly the double of the covalent radius of copper (1.32 Å),34 thus indicating a clear Cu(I)- Cu(I) bond. The correlation between the metal-metal separation and the CuXCu bond angle for complexes 1-5 is plotted in Figure S8 (see Supporting Information), and discloses a fair linear dependence, particularly for acute angles. In general, R substituents in PR2Ar’ bulkier than CH3 seem to impede the formation of a metal-metal bond, reasonably for steric reasons, with copper-cop- per separations larger than 3 Å. The solid state structures of complexes 1-5 deserve some additional comments on their molecular symmetry and the conformation of the dicopper [Cu2X2] cores and the phosphine ligands. Although most of the [Cu(-X)(PR2Ar’)]2 species studied in this work show Ci symmetry (C1 in three cases), by slight rotations of the aryl rings around the C—C and P—C bonds and slight translations of the X ligands for non-planar [Cu2X2] cores, a C2h symmetry would be easily reached. Figure 3 provides a schematic representation of the C2h-like molecular structure and the symmetry elements showed by the majority of [Cu(-X)(PR2Ar’)]2 dimers in an idealized geometry. Figure 3. C2h-like molecular structure and the symmetry elements in most of the [Cu(-X)(PR2Ar’)]2 dimers described in this work. The conformations adopted by the terphenyl phosphines used in this work have recently been analyzed by some of us.31 Three possible conformations have been identified (Figure 4), being the type C the most common by far in the dicopper complexes of this work. This observation could reasonably be explained by the existence of some Cu…C interaction. Anyway, taking into account that the NMR spectra of all complexes are consistent with fast rotations around the P—Cu and/or P—C bonds, the energy separations between the different conformations in solution are likely small. The neutral [Cu2X2] cores are planar or quasi-planar (deviations from the middle plane less than 0.13 Å) with no apparent relation with other metric parameters. Figure 4. Conformations observed in the solid state by the Cu-coor- dinated terphenyl phosphines PR2Ar’ used in this work Photophysical Studies. The nature of transitions within a sole isolated molecule can be assigned through UV-Vis spectra of complexes 1c and 3c as well as the spectra of their corresponding free ligands shown in Fig. 5. The complexes display an additional electronic transition at longer wavelengths (above 300 nm) assigned to Metal and Halide to Ligand Charge Transfer band ((M+X)LCT), which has been confirmed with TD-DFT calculations (see below). Intra-ligand electronic transitions are observed at lower wavelengths in the complexes, matching the observed transition in the UV-Vis spectra of the respective free ligands. The photoluminescence (PL) of 1c and 3c is negligible in solution at room temperature, even in degassed conditions. Thus, to allow the characterization of their optical properties, the PL of these complexes must be studied in solid state (powder, dispersion in polymer matrix and in frozen solution at 77K). Both complexes are emissive in powder solid phase, however they display different features at different temperatures. At 77 K the excitation and emission spectra of both complexes are very similar (Fig. 6). At room temperature, the emission band of 3c becomes slightly broader without significant change of its emission maximum with Φ PL of 19%, while 1c has only very residual emission. The diffuse reflectance of powder samples with Kubelka- Munk transformation function f(r∞) are shown in Fig. 6, which exhibits barely resolved absorption bands at 285 nm and 335 nm coincident with the excitation and UV-Vis bands in solution. This same position confirms the localization of the excitation and thus emission from isolated molecules.36 Figure 5. UV-Vis spectra of complexes 1c and 3c (bold line) and their corresponding free ligands (dashed line) in cyclohexane. Figure 6. Diffuse Reflectance spectra of the complexes under Kubelka-Munk (KM) transformation function (f(r∞)bold lines) overlapped with the excitation spectra at 77 K collected at 515 nm for 1c and at 525 nm for 3c (pointed lines). Emission spectra of powder samples excited at 350 nm at 77 K and 300 K . As isolated molecules can indeed be emissive in solid state, we did the characterization also in Zeonex solid-solid solution at room temperature. In this rigid matrix, at room temperature, 1c and 3c display almost identical spectral behaviour (Fig. 7) as those found for powder samples at 77 K. This behaviour discards the possibility of aggregation induced emission. Room temperature decay time measurement performed on films placed in vacuum exhibit a mono-exponential decay with lifetime of 56 s for 1c and 82 s for 3c (Fig. S9 in SI). In air saturated environment one can clearly notice the quenching effect of O2. The long life time and the sensitivity to oxygen strongly suggest that luminescence arises from excited triplet states. Figure 7. Excitation and Emission spectra of complexes 1c and 3c in solid film dispersion in vacuum and in the presence of air. The narrowing of the emission band at 77 K without a noticeable shift of observed emission (Fig 6 top) and the large Stokes shift (ca. 1.42 eV) is compatible with the assignment of the observed emission to phosphorescence from lower triplet state. For a better understanding of the emission properties a deeper analysis with theoretical calculations using TD-DFT were carried out. TD-DFT calculations. Calculated excitation energies for optimized geometries in cyclohexane are in good agreement with the experimental UV-Vis spectra (Fig S10 in SI). TD-DFT calculations reveal an essentially (M+X)LCT character of low-lying electronic transitions with some contribution of Intra-Ligand Charge transfer (ILCT) through participation of P atoms. In both 1c and 3c which are centrosymmetric complexes, the S1 state would be achieved by a mainly HOMO→LUMO transition, forbidden by Laporte symmetry selection rule (u→u). However the S2 state is reached by a mainly HOMO→LUMO+1, which is a symmetry allowed transition (u→g). DFT calculations show also that LUMO and LUMO+1 are degenerate molecular orbitals (Fig. 8). Figure 8. Frontier Orbitals, their respective parity and energies. LUMO and LUMO+1 are degenerated Molecular Orbitals. These theoretical results confirms that once the molecule achieves the excited state any Internal-Conversion (IC) from S2 to S1 would lead to a non-radiative decay path as S1 to S0 is a forbidden transition. So the population of the triplet manifold should be done upon ISC from S2. In both cases the closest triplet state in energy is T8 which is of mainly (HOMO,LUMO+1) configuration as like as S2 (Table 2). The same configuration of S2 and T8 has important consequences: (i) a very small energy gap between the two charge transfer states due to decreased exchange integral between the unpaired electrons; (ii) same electronic configuration does not involve different type of orbitals thus ISC between them would be weak via vibronic Spin-Orbit-Coupling.37 However, it is expectable that ISC from S2 to the nearly isoenergetic T8 state could be performed due to the Internal Heavy Atom Effect of Iodine acting on the unpaired electron of HOMO centred on Cu2I2 moiety so that ⟨𝑆2|𝐻 𝑆𝑂|𝑇8⟩ ≠ 0 .38 Table 2. Electronic configuration of first, second Singlet excited state and the closest Triplet state in energy as well as the main involved Molecular Orbitals (f and E are the calculated oscillator strength and excitation energy respectively). Complex State f E (nm) MO (%) 1c S1 0.0008 339 HOMO→LUMO (80%) S2 0.0078 338 HOMO→LUMO+1(79%) T8 -* 340 HOMO→LUMO+1(20%) 3c S1 0.0000 335 HOMO→LUMO (92%) S2 0.0080 334 HOMO→LUMO+1(92%) T8 -* 339 HOMO→LUMO+1(80%) * N/A as calculations were done without SOC inclusion. The density of the triplet manifold in the sense of spatial character can be discussed as well. In the case of complex 3c, 80% of the closest triplet state to S2 is (HOMO,LUMO+1) in configuration interaction. In the case of the complex 1c, however, the (HOMO,LUMO+1) contribution is just 20% but there are other Tn (n<8) very close in energy one to each other which also have some contribution of (HOMO,LUMO+1) in their configuration interaction set (Table S3 in SI). Quantum Mechanical properties of these systems like the participation of atomic p orbitals of I in the lowest charge transfer state and small EST explain the ISC and the experimentally observed phosphorescence. Computational results show that ISC from the initially Franck-Condon (FC) region after excitation is possible without considerable structural change of the excited molecule. Given the spatial character of nearly degenerate charge transfer states the ISC rate constant is expected not to be large enough to be dominant over other possible competitive pathways,38a namely 𝑆2 𝐼𝑆𝐶 → 𝑇8 versus 𝑆2 𝐼𝐶 → 𝑆1. Although it is quite logical to accept that ISC from S2 to triplet manifold is favoured in rigid media as in this kind of environments vibronic couplings responsible for IC from S2 to S1 would be more hindered, this type of analysis of the configuration of the excited states does not explain why powder samples of 1c are not emissive at room temperature. Our DFT calculations provide further useful information regarding Molecular Orbital energy levels. Accordingly, LUMO and LUMO+1 are quite close in energy, the energetic difference between them being 0.008 eV for 1c and 0.001 eV for 3c, close enough to be considered degenerate in both cases. Despite of degeneracy we have treated LUMO and LUMO+1 individually and not as a pair because by symmetry operations of the C2h point group they cannot be transformed into each other or any linear combination of them. Moreover, LUMO and LUMO+1 have different parities (Fig. 8) so, although they are automatically orthogonal Molecular Orbitals of the same kind, they do not have the same symmetry (⟨𝜋𝑢 ∗|𝜋𝑔 ∗⟩ = 0) and cannot transform as an E type irreducible representation (irrep).39 Almost simultaneous transition to two potential energy surfaces S1 and S2 (Table 2) means vertical transition corresponds to a molecular geometry point where these states are nearly degenerate. So a conical intersection (CI) between S1 and S2 within the FC region can be expected. S1 and S2 have 1Bg and 1Au symmetries respectively so the degeneracy we are looking for would be between states which belong to different irreps of the same Abelian point group. In this case the CI is not required by symmetry, as in the Jahn–Teller case, but occur rather due to the nature of the two electronic states involved and shall be considered an accidental intersection case.40 Unlike homoleptic [Cu(N^N)2]+ complexes which exhibit flattening Jahn-Teller distortion41 in these three-coordinate systems pseudo Jahn-Teller effect cannot be expected. The excited state of the heteroleptic dinuclear complexes 1c and 3c is formed by promotion of an electron from HOMO which mainly corresponds to the Cu2I2 moiety (Fig. 8). Due to the highly covalent character of Cu-I bonds on the one hand and great participation of iodine atoms in HOMO on the other, it is not plausible to assume a pure d9 configuration for any Cu centres of the [Cu2I2].+ core. However, in three coordinated Cu(I) complexes the possibility of other types of distortion in the excited state cannot be ruled out.42 The existence of a CI between the optically allowed S2(1Au) and the dark S1(1Bg) in the vicinity of the FC zone helps us to postulate a more accurate mechanism of luminescence. We already saw that the ground state geometry of both 1c and 3c belong to the C2h abelian point group and does not contain degenerate irreps, but upon vertical excitation a degenerate electronic state is imposed. As mixing between electronic states along nuclear displacements become significant at near degeneracy critical geometries, the excited molecules at FC zone (from which ISC would occur) can undergo non-adiabatic transition from S2 to S1 through vibronic coupling.43a-b More importantly since the geometry corresponding to the CI is not expected to be quite different from that at FC point, electronic coupling around the CI point can be approximated by low-order expansion around the FC geometry.43b-c In this way nuclear dynamics occurring within the S1- S2 can be approached as a pairwise interacting states by the wellestablished Linear Vibronic Coupling (LVC) model. In LVC model the symmetry of the vibrational coordinates that couple electronic states of an Abelian point group is determined by the direct product of the irreps of the electronic states.40a,43a In our case the interstate vibronic coupling between S1 and S2 will not vanish when the coupling modes are of bu symmetry as 𝐵𝑔⨂𝐴𝑢= 𝐵𝑢. Taking FC point as reference geometry is a very useful approximation making us possible to figure out how these coupling modes would look like. To do so let us focus on dimeric coordination sphere with 6 atoms [Cu2I2(P)2] for the sake of simplicity. By group theory calculation it is known that among twelve (3N- 6) internal coordinates in C2h point group, four of them will be bu non-totally symmetric odd normal modes. Using the technique of projection operators and being aware of the presence of the Cu2I2 four-membered ring44 one can determine how these symmetry adapted vibrational normal modes look like (Fig. 9). Figure 9. Four vibrational modes of bu symmetry : asymmetric P-Cu stretching and P^(Cu-Cu) bending (top) Cu2I2 in plane ring deformation and out of plane ring puckering (bottom),. These Group Theory results can explain why unlike other rigid media, in room temperature powder solid phase the PL behavior of 1c and 3c is so different. Complexes 1c and 3c interact distinctly with adjacent molecules through short contacts. In 3c there are short contacts stablished through Iodine atoms so the Cu2I2 core is somehow stuck within the crystalline network while short contacts through Iodine feature is not present in 1c (Fig. S11 in SI) and thus individual molecules at 300 K have enough space and flexibility to undergo Cu2I2 ring deformation breaking down the geometry at FC zone which the molecule needs to undergo ISC and emit light. The packing force is strong enough to block this deactivating vibrational modes significantly as in the Raman spectra one can notice a characteristic resolved band at 126 cm-1 observable in 1c but not clear in the spectrum of 3c. The 126 cm-1 wavenumber is consistent with the Cu-I stretching in Raman spectroscopy45 and by the other hand this band is not observable in the free ligands so we can assign it to some Cu-I stretching in 1c. Although it can correspond only to an even stretching mode (ag or bg), experimental evidence of the flexibility of Cu2I2 core conditioned by solid state packing is presented by Raman Spectroscopy. (Fig. 10). Figure 10. Raman spectra of 1c and 3c compared to their corresponding free ligands L1 and L3 in order to confirm the assignment of the resolved band at 126 cm-1 to Cu-I stretching mode. Based on spectroscopic observations supported by theoretical calculations we suggest a mechanism of luminescence (Fig. 11). According to this scheme the FC geometry of the 1(M+X)LCT excited molecule should be maintained in order to populate the bright 3(M+X)LCT state. Otherwise distortions namely Cu2I2 ring deformation, act as a non-radiative path at the Conical Intersection point which corresponds to photochemical funnel between S2 and dark S1. . Figure 11. Radiative transitions (red arrows) and non-radiative transitions (curvy arrows) involved in the suggested mechanism of phosphorescence. The suggested mechanism can explain the photophysical behaviour in deaerated solution where emission quenching is independent of oxygen concentration. While at room temperature none of the two complexes is emissive, in frozen solution at 77 K the spectral behaviour of both 1c and 3c is almost identical to those of other rigid media (Fig. S12 in SI). So the quenching of phosphorescence in these complexes can be attributed to their geometrical flexibility in solution rather than specific interactions with solvent molecules. Catalytic Studies. A survey on the catalytic properties of complexes 1-5 in the CuAAC reaction between benzyl azide and phenyl acetylene, the model reaction, was accomplished. The trials were conducted using the reaction conditions optimized by Díez and co-workers29 for the catalyst CuBr(PPh3)3, namely, using water as the solvent and 0.5 mol% [Cu] catalyst loading under air. As shown in Table 3, all catalysts tested proved to be active in the model system under these conditions. However, those with the less bulky phosphine ligands (L1 and L3) displayed the best activities (entries 1-3 and 6-8) in the shortest reaction times. Furthermore, within each series of catalysts no significant effect of the nature of the halide was observed. Finally, less than 20% yield of triazole was obtained in blank experiments carried out with copper halide salts CuX (X = Cl, Br and I) as catalysts in the absence of any phosphine ligand (entries 10-12). It is important to note that no sign of oxidation or disproportionation of any of the Cu(I) complexes was observed during the catalysts screening, even though the reaction were performed in aqueous media and under aerobic conditions. This observation underlines the effective protection imparted by terphenyl phosphine ligands to the Cu(I) centre during the catalysis. Table 3. Catalytic Performance of Complexes 1-5 in the [3+2] Cycloaddition of Benzyl Azide and Phenyl Acetylene.a entry [Cu] time (h) Yield (%)b 1 1a 6 92 2 1b 6 96 3 1c 6 96 4 2a 20 58 5 2b 20 55 6 3a 6 92 7 3b 6 97 8 3c 6 95 9 5a 20 62 10 CuCl 6 20 11 CuBr 6 16 12 CuI 6 16 aReaction conditions: benzyl azide (1mmol), phenyl acetylene (1 mmol), Cu catalyst (0.005 mmol of Cu), H2O (3 mL) under air. bI- solated yields (average of two runs). To study the scope of the reaction complexes 1c and 3c were selected as the catalysts, since the these two Cu(I) iodide derivatives exhibited the best long-term air stability of each series. As shown in Table 4, a variety of azides and alkynes were smoothly converted into the corresponding triazoles in excellent yields under the reaction conditions. Benzyl, cynnamyl and phenyl azides reacted in short reaction times (5-6 h), whereas the alkyl-substituted n-octyl azide required longer reaction time (12 h) to reach completion (14o-14r). Conversely, the outcome was largely unaffected by the substituents on the terminal alkynes. Thus, aryl, alkyl, heteroaryl (pyridine), carboxylate and hydroxyl groups were equally well tolerated (see for example 14a-14g). Organic azides, particularly those of low molecular weight or with high nitrogen content, are potentially explosive.46 Therefore, various protocols that make use of in situ generated azides for the preparation of triazoles have been reported in the literature.47 Given that most of the azides employed in this work were prepared from the reaction of alkyl halides with NaN3, the activity of catalysts 1c and 3c were examined in the three-component reaction of benzyl bromide, NaN3 and phenyl acetylene using the same reaction conditions as those described above for the reactions with the preformed azides. Under these conditions, only catalyst 3c provided the triazole in quantitative yield, although in extended reaction time when compared with the reaction carried out with the isolated benzyl azide. The versatility of catalyst 3c for the synthesis of triazoles by the three-component reaction of alkyl halides, NaN3 and 1-alkynes is illustrated in Table 5. In all instances, the triazoles were obtained in good to high yields under otherwise unaltered reaction conditions. It is interesting to note that only very few Cu-based catalyst systems accomplished the one-pot synthesis of triazoles with such a low catalyst loading.29c,48 Table 4. Scope of Cycloaddition Reactions of Azides and Alkynes Catalyzed by Complexes 1c and 3c.a,b aReaction conditions: azide (1mmol), 1-alkyne (1 mmol), Cu catalyst (0.5 mol% Cu), H2O (3 mL) under air. bIsolated yield base on azide (average of two runs) Table 5. Scope of Three-Component Reactions of Aryl Halides, NaN3 and 1-Alkynes Catalyzed by Complex 3c.a,b aReaction conditions: alkyl halide (1 mmol), NaN3 (1.3 mmol), 1- alkyne (1 mmol), catalyst 3c (0.5 mol% Cu), H2O (3 mL). bIsolated yields (average of two runs). cWith MeI 5-yodo-1,2,3-triazoles are versatile intermediates for further functionalization via the halogen substituent.49 However, there are still a limited number of efficient Cu-based catalyst systems for the CuAAC reactions between organic azides and 1-iodo alkynes.29c,48,50 This prompted us to investigate the behavior of complexes 1c and 3c in this challenging transformation. Benzyl azide and iodoethynylbenzene were chosen as the model system. All reactions were conducted in water at room temperature under air and with 0.5 mol% [Cu]. Neither of the two Cu(I) catalyst was active under these conditions (Table 6, entries 1 and 2). Changing the solvent to THF or increasing the catalyst loading up to 5 mol% did not produced any significant improvement in the reaction yield (entries 3 and 5). The addition of nitrogen bases has been reported to be beneficial in these reactions.48,50 Thus, by performing the cycloaddition in the presence of 4 mol% of 2,6- lutidine high yields of the 5-iodotriazole was obtained with both catalyst (entries 6 and 7). However, catalyst 3c was less selective since small amounts of dehalogenated 5-H-triazol were also observed. No change either in yield or in selectivity was observed when the reactions were carried out under neat conditions (entries 8 and 9). Table 6. Screening of reaction conditions for the cycloaddition between benzyl azide and iodoethynylbenzene catalyzed by complexes 1c and 3c.a entry catalyst cat. loading (mol% Cu) solvent additive Yield (%)b 1 1a 0.5 H2O - Trace 2 3c 0.5 H2O - Trace 3 1a 0.5 THF - Trace 4 3c 0.5 THF - Trace 5 3c 5 THF - 20% 6 1a 1 H2O 2,6- lutidine >95 7 3c 1 H2O 2,6- lutidine 90c 8 1a 1 neat 2,6- lutidine >95 9 3c 1 neat 2,6- lutidine 90c aReaction conditions: benzyl azide (0.5 mmol), iodoethynylbenzene (0.5 mmol), solvent (3 mL), 24 h. bDetermined by 1H NMR with an internal standard. c5-10% of 5H-triazole was formed. The optimized reaction conditions were then applied to different azides and 1-iodoalkynes giving the products 1,4-disubstituted 5- iodo-1,2,3-triazoles with the yields in the range of 45–95% (Table 7). n-octyl azide was the less reactive; consequently, reactions involving this azide were performed under neat conditions to obtain the products in moderate to good yields (15c and 15f). Only in the case of the cycloaddition of benzyl azide and iodoethynylcyclopropane, trace amounts of the dehalogenated triazol were detected (15d). Table 7. Scope of Cycloaddition Reaction s of Azides and 1- Iodoalkynes Catalyzed by 1c.a,b aReaction conditions: azide (0.5 mmol), iodoethynylbenzene (0.5 mmol), catalyst 1c (1 mol% Cu), H2O (3 mL), 24 h. bIsolated yields (average of two runs). cWith 2 mol% of Cu. dNeat. eTrace amounts of 5-H-triazole detected by 1H NMR. CONCLUSIONS A family of Cu(I) halide complexes stabilized by dialkyl terphenyl phosphines, PR2Ar’, has been prepared and characterized. X-ray diffraction studies disclosed a bimetallic dihalo-bridged structure for compounds [CuX(PR2Ar’)]2 with a distorted trigonal planar geometry at the metal centre and, in most cases, a C2h-like symmetry. The presence of a Cu-Cu bond was detected for two iodide derivatives with an intermetallic distance of ca. 2.7 Å. Compounds [CuI(PMe2ArXyl2)]2 1c and [CuI(PEt2ArMes2)]2 3c behave in the same way in degassed liquid solution at 300 K (none of the two complexes is emissive), in solid-solid solution, i.e. Zeonex films, in the solid state at 77 K and in frozen cyclohexane solution(both complexes are emissive). In microcrystalline powder samples with sufficient rigid structure and order, however, the radiative relaxation of excited states is highly conditioned by interactions with adjacent molecules at room temperature. In such a way that just by blocking very specific vibrational modes at dimeric core site it is possible to “turn on” the emission at room temperature. In this study the centrosymmetric structure of complexes allowed the identification of such non-radiative deactivation vibrational modes and showed how structural knowledge can influence the design of novel luminescent Cu(I) complexes in the future. Among the dinuclear Cu(I) complexes prepared, [CuI(PMe2ArXyl2)]2 1c and [CuI(PEt2ArMes2)]2 3c are efficient catalysts for [3+2] cycloaddition of azides and terminal alkynes. Catalytic cycloadditons are conducted in water under air using low catalysts loadings. Remarkably, these conditions are compatible with the use of 1-iodoalkynes, a substrate scarcely used in Cu- AAC reactions. EXPERIMENTAL SECTION All preparations and manipulations were carried out under oxygen-free nitrogen, using conventional Schlenk techniques. Solvents were rigorously dried and degassed before use. Ligands L1- L5,30a,31a organic azides51 and iodoalkynes50a were synthesized following previously reported procedures. Reagents were purchased from commercial suppliers and used without further purification. NMR spectra were recorded on Bruker Avance DPX-300, Avance DRX-400, Avance DRX-500, and 400 Ascend/R spectrometers. The 1H and 13C resonances of the solvent were used as the internal standard and the chemical shifts are reported relative to TMS while 31P was referenced to external H3PO4. Elemental analyses were performed by the Servicio de Microanálisis of Instituto de Investigaciones Químicas (IIQ). CITIUS. X-ray diffraction studies were accomplished at Centro de Investigación Tecnología e Innovación, CITIUS (Universidad de Sevilla) and Centro de Investigación en Química Sostenible, CIQSO (Universidad de Huelva). Diffuse reflectance spectra were acquired in a Shimadzu UV-2501PC equipped with an integrating sphere. The powdered samples were smashed between two quartz lamellae, accommodated on a BaSO4 filled support and the spectra run using an identical BaSO4 filled support as blank. The remission function, F(R), was calculated using the Kubelka–Munk equation for optically thick samples. UV–vis absorbance spectra were acquired on a UV–vis–NIR Varian Cary 5000 spectrophotometer and fluorescence spectra were recorded on a SPEX Fluorolog-3 Model FL3–22 spectrofluorimeters. Time-resolved emission spectra were acquired in the same apparatus, using a pulsed xenon lamp with a full-width of 3 µs. Spectra and decays were collected with a minimum 50 μs delay to remove any interference from the lamp. Luminescence quantum efficiencies were measured by the absolute method with an Integrated Sphere. Films for optical characterization were prepared in zeonex (10% in toluene) matrix by drop-casting onto a quartz substrate with an emitter concentration of 1% (m/m). The Raman analysis was carried out using a Labram 300 Jobin Yvon spectrometer, equipped with a HeNe laser 17mW operating at 633 nm. The laser beam was focused with a 50x Olympus objective lens. The laser power at the surface of the samples was varied with the aid of a set of neutral density filters (D 0.6). DFT calculations have been carried out with the program Gaussian 09 using the B3LYP functional, 6-31G** basis set for C, H and a standard double-ζ polarized basis set, namely, the LANL2DZ set for Cu, I and P augmented with polarization functions on P (a d orbital with exponent 0.364 for P, upon EMSL basis set exchange database) For Cu, P and I we applied also Effective Core Potentials. Geometry optimization and TD-DFT calculations were performed in cyclohexane with the PCM solvation method. Molecular Orbital visualisation for all the complexes was performed using Avogadro 1.2.0. Synthesis of [CuCl(PMe2ArXyl2)] 1a. A mixture of the ligand PMe2ArXyl, L1, (0.080 g, 0.231 mmol) and CuCl (0.023 g, 0.231 mmol) in CH2Cl2 (5 mL) was stirred at room temperature for 20 h. After filtration, the resulting solution was evaporated to dryness affording 1a as colorless solid. Yield: 0.098 g, 98%. 1H NMR (300 MHz, CDCl3):  7.57 (t, 3JHH = 7.6 Hz, 1H, p-C6H3), 7.30 (t, 2H, p-Xyl), 7.18 (d, 3JHH = 7.6 Hz, 4H, m-Xyl), 7.07 (dd, 3JHH = 7.6 Hz, 4JHP  1 Hz, 2H, m-C6H3), 2.05 (s, 12 H, CH3-Xyl), 1.14 (d, 2JHP = 6.8 Hz, 6H, P-CH3).13C NMR (75 MHz, CDCl3):  147.1 (br s, o-C6H3), 140.2 (br s, ipso-Xyl), 135.4 (o-Xyl), 131.8 (p-C6H3),