Effect of gold(I) in the Room Temperature Phosphorescence of ethynylphenanthrene

The synthesis of two series of gold(I) complexes containing the general formulae PR 3 -Au-C≡C-phenanthrene (PR 3 = PPh 3 ( 1a / 2a ), PMe 3 ( 1b / 2b ), PNaph 3 ( 1c / 2c )) or (diphos)(Au-C≡C-phenanthrene) 2 (diphos = 1,1- bis (diphenylphosphino)methane, dppm ( 1d / 2d ); 1,4- bis (diphenylphosphino)butane, dppb ( 1e / 2e )) have been synthesized. The two series differ on the position of the alkynyl substituent on the phenanthrene chromophore, being at the 9-position (9-ethynylphenanthrene) for the L1 -series and at the 2-position (2-ethynylphenanthrene) for the L2 -series. The compounds have been fully characterized by 1 H and 31 P NMR and IR spectroscopy, mass spectrometry and single crystal X-ray diffraction resolution in the case of compounds 1a , 1e , 2a , and 2c . The emissive properties of the uncoordinated ligands and corresponding complexes have been studied in solution and within organic matrixes of different polarity (PMMA and Zeonex). We have observed room temperature phosphorescence (RTP) for all gold(I) complexes while only fluorescence can be detected for the pure organic chromophore. In particular, the L2 -series present better luminescent properties regarding intensity of emission, quantum yields and RTP effect. Additionally, while the inclusion of all the compounds in organic matrixes induces an enhancement of the observed RTP due to the decrease in non-radiative deactivation, only the L2 -series completely supress fluorescence giving rise to pure phosphorescent materials.


Introduction
Room-temperature phosphorescence (RTP) has become important for a variety of applications in different fields such as photodynamic therapy, [1] bioimaging, [2] optical limiting, [3] photon upconversion, [4] oxygen sensing, [5] and light-emitting diodes (OLEDs). [6] This phenomenon is difficult to be successfully achieved through pure organic molecules under normal conditions (room temperature and normal pressure) because of the weak spinorbital coupling (SOC) between excited singlet and triplet states as well as the fast nonradiative deactivation of triplet excitons. Heavy-atom induced phosphorescence of organic chromophores that originates from SOC is a studied mechanism for RTP emitting materials due to the well-known heavy atom effect. [7][8][9][10][11][12][13] For these reasons, the phenomenon of RTP has become more popular for inorganic and organometallic complexes, easily containing heavy metal atoms, that exhibit strong RTP, mostly originating from the 3 MLCT transition. [14] The more favoured SOC process in the presence of heavy atoms is relevant for triplet state harvesting as a well-known procedure to achieve high intensities in light emitting devices, LEDs. [15] It must be considered that enhancing SOC by heavy atom effects affords an increase not only in the efficiency of triplet formation, ST, and the radiative rate constant of the triplet, kr T , but also in the non-radiative rate constant for triplet deactivation knr T .
Nevertheless, the important point is that the increase in kr T and ST must surpass the increase of knr T to achieve a net gain in the emission RTP from the chromophores (Scheme 1).

Scheme 1.
Schematic representation of the photophysical pathways involved in these systems.
The chemistry of organogold(I) compounds is receiving much attention among heavy metal atom complexes. In comparison with the surrounding metals of the periodic table, gold displays a maximum relativistic effect which affects the resulting photophysical properties. [16][17][18][19] The formation of a gold(I)-carbon σ bond can significantly modify the electronic states of an organic chromophore by enhancing spin-orbit interactions, thus increasing the rate of intersystem crossing between singlet and triplet states with respect to the free aromatic counterpart. [20,21] In this way, while only fluorescence emission can be detected for specific organic luminophores, their corresponding organogold(I) compounds can exhibit dual luminescence (fluorescence and phosphorescence) at room temperature with emission yields that depend sensitively on the position of metal coordination. [17] RTP is observed to be weakly recorded in solution. For this reason, some feasible strategies have been analysed for enhancing this RTP, such as crystal engineering, [22] polymerization, [23] aggregation, [24] or immobilization within organic matrixes [5] . All these mechanisms are focused to supress the non-radiative decay knr and the quenching kq rate constants. In particular, the fabrication of thin films or matrix supports containing luminophores are highly desirable outside from academia for practical applications such as smart devices and luminescent sensors. This is due to their easily tune shape, size, flexibility or rigidity and to the fact that the resulting materials are much easier and simpler to handle in comparison with solution or powdered forms [25] .
In this work, we present the synthesis and structural characterization of two series of ethynyl phenanthrene gold(I) compounds (L1-series and L2-series) that differ on the position of coordination of the gold(I) atom on the chromophore, changing the location of the alkynyl moiety. Their luminescent properties and, in particular, the resulting RTP behaviour have been carefully studied both in solution and when the samples are immobilized within organic matrixes with different polarity. The nuclearity of the compounds, steric hindrance, alkynyl position and polarity of the organic matrix have been evaluated as key parameters for obtaining RTP materials with increasing phosphorescence quantum yields and very long lifetimes in the order of hundreds of microseconds.

Synthesis and characterization
A series of mono-and dinuclear gold(I) compounds containing, on one side, ethynylphenanthrene as chromophore and mono-or diphosphanes at the second coordination position have been synthesized following the reaction pathway shown below. The compounds have been obtained with two different isomers of ethynylphenanthrene, with the alkynyl group at 2-and 9-position. The 9-ethynylphenanthrene (L1) was commercially available while the 2-ethynylphenanthrene compound (L2) was synthesized by a Sonogashira coupling reaction from 2-bromophenanthrene and trimethylsilylacetylene in order to get the TMS-protected compound, L2', and subsequent deprotection with K2CO3 to get the desired product, L2, in good yields (Scheme S1 and Figures S1-S2).
The reaction of L1 or L2 and AuCl(tht) in the presence of sodium acetate as a base gave rise to the formation of the insoluble polymers [Au-C≡C-phenanthrene]n (1 and 2 respectively, from L1 and L2) based on the procedure reported for similar gold polymers.
The formation of the polymer has been evidenced by the disappearance of the C≡C-H vibration of the ethynylphenanthrene and the shift to shorter wavenumbers of the C≡C vibration in the corresponding IR spectra.
Single-crystals suitable for single X-ray diffraction were successfully obtained for 1a, 1e, 2a and 2c through slow diffusion of dichloromethane/hexane solutions of the compounds at room temperature. The corresponding molecular structures are presented in Figure 1 and the selected bond distances and angles are summarized in Table 1.  The Au-C and C≡C distances and P-Au-C and Au-C≡C angles (Table 1) are in the usual ranges for Au(I) alkynyl complexes. [28,[31][32][33][34][35][36][37][38][39][40][41]  (1e) and 3.133 (2c) Å respectively. These intermolecular interactions could be considered as hydrogen bonds that could have a very significant role in the packaging of these kind of molecules as it has been reported previously. [33,34,[43][44][45][46][47] It seems that the large aromaticity of the phenanthrene group favours these kind of intermolecular contacts with respect to possible aurophilic interactions, usual in gold(I) complexes. These contacts may be also hindered by the bulkiness of the phenyl groups unless the two Au(I) atoms may be located at the adequate distance such as in the case of dppm derivative (see below photophysical behaviour).

Photophysical characterization
The absorption and emission spectra of all the complexes 1a-e and 2a-e and of the free ligands L1 and L2 were recorded in 1x10 -5 M acetonitrile solutions (except 1c, which was recorded in THF due to its low solubility in acetonitrile) at room temperature and the obtained data are summarized in Table 2.
The electronic absorption spectra of all the compounds ( Figure 2) show an intense band with vibronic resolution at 310-330 nm. These bands display the same profile as the corresponding spectrum of the L1 and L2 ligands, but are red-shifted due to coordination to the metal atom and thus they are assigned to metal perturbed * intraligand transitions. [14,[48][49][50] Additionally, all the absorption bands are 10-20 nm blue-shifted for all L2 and corresponding complexes compared with the analogous L1-compounds.
An additional absorption band centered at 300 nm appears for 1c/2c, that contain the trinaphthylphosphane (PNaph3) ligand, which is ascribed to the π-π*(Naph3) transition. [31] In the case of 1d/2d, a lower-energy absorption band is observed as a tail which has been assigned to the absorption of aggregates and arises from σ*(A ···Au)-π* transitions, as an indication of the possible aurophilic contacts in solution for these complexes. [51]  The emission spectra of all the compounds were recorded in the presence and absence of O2 (  (see Table 2). This is also evidenced in the corresponding phosphorescence quantum yields, P, which are one order of magnitude larger than the corresponding fluorescence data, F, in all the L2-gold(I) complexes except for 2e (the dppb derivative) with a comparatively lower difference though still 5 times larger. Moderately large P values (2-3-fold with respect to F) were measured for the L1-compounds (see Table 3).
The generally large values of room temperature phosphorescence in the L2-derivatives might be related either to the lower steric hindrance of the resulting complexes derived from 2-ethynylphenanthrene with higher linearity in comparison to the 9-ethynylphenanthrene data where the chromophore was in a perpendicular disposition with the alkynyl edge, or other electronic aspects that deserve future investigation.
Phosphorescence emission cannot be detected for the free ligands L1 and L2, providing direct evidence of the role of gold(I) heavy atom in the intersystem crossing and phosphorescence emission at room temperature. Table 3. Luminescent quantum yields in N2-saturated samples.
In order to minimize the non-radiative processes and maximize the RTP, the samples were immobilized within two different organic matrixes, polymethylmetacrylate (PMMA) and Zeonex 480. As found previously in solution, RTP is more favoured for the L2-series where the almost complete disappearance of fluorescence in all gold complexes was observed, giving rise to effectively pure phosphorescent materials subsequent to removing oxygen (Figures 3 and 4). RTP is also favoured to some extend for the L1-series when the compounds are immobilized in these matrixes ( Figure S40), in particular for the dppm derivative, 1d. This may be due to the scalability of the intersystem crossing due to Au(I)···Au(I) intramolecular contacts.   Luminescence quantum yields were also measured in all cases for the prepared dopedmatrix samples (Table 4). In this way, the improvement of the RTP can be quantitatively . [52] Additionally, calculation of the corresponding radiative and non-radiative rate constants, from the  and  values, demonstrated that, as expected, the non-radiative relaxation is less favored in PMMA ( Table 6). The largest kr calculated value for 2d (1877.7 s -1 ) is ca. 4 orders of magnitude larger than that previously reported for phenanthrene (0.26 s -1 ). [14]

Theoretical calculations
TD-DFT calculations were carried out to determine the excited states and their involved molecular orbitals in acetonitrile solution (see Table S7 and Figure S41). For each series, the same compounds were studied containing the different phosphanes: PPh3, PMe3, PNaph3, and PMePh2 (this can provide results those will be compared with bimetallic species in the absence of metal···metal interactions). For the L1 series (9-ethynylphenanthrene), two bands are calculated at 3.5 and 4.2 eV corresponding to transitions between  orbitals of the phenanthrene rings (Table S8). The first band is always most intense and it is assigned to HOMO→LUMO transition (f ≈ 1.0), being these orbitals centred in the phenanthrene rings (~70 %) with an important contribution of C≡C fragment. The second one is found for HOMO-1→LUMO transition (f ≈ 0.34), being the HOMO-1 is only a phenanthrene orbital (> 95 %). Those related triplet states are located at 2.3 and 3.8 eV, respectively. Since that this description is found for PPh3 (1a), PMe3 (1b) and PMePh2 derivatives, one can expect identically values of photochemical properties for dppb compound (1e) where metal···metal interaction is absence (optimized distance about 9 Å), and probably for dppm compound (1d, ~ 3.7 Å). However, more complex situation is found for PNaph3 (1c) because several transitions are expected between these two bands involving the naphthyl groups of the phosphane.
For the L2 series (2-ethynylphenanthrene), three bands are now calculated between 3.7 and 4.2 eV between two occupied and two empty  orbitals (with the triplet states between 2.6 and 3.4 eV, Table S9). The first band is newly most intense of three (f ≈ 1.4, 0.9 and 0.3, respectively), and it corresponds to HOMO→LUMO transition as previously is shown for In the case of diphosphanes, the calculations show the presence of two independent chromophores for dppb derivatives in agreement to large Au···Au distance in the most favourable conformation (about 9 Å). Nevertheless, similar situation is found for dppm having a splitting of less than 4 nm (Au···Au distances are calculated at 4.4 and 3.7 Å for 1d and 2d, respectively), but their conformation can easily modified and the presence Au···Au contacts would provide major changes.
Interestingly, we can observe that the triplet state closer in energy to the S1 appears slightly at higher energy in the L1 series. On the contrary, the corresponding triplet state in L2-series is lower in energy to S1 (see Figure 5). This means that the population of the triplet state in L2 complexes is much more favoured than in L1 and for this, RTP process is more efficient.

Conclusions
The position of the alkynyl moiety at the phenanthrene chromophore determines the resulting The presence of an organic environment such as the PMMA matrix has been observed to be a very good way to improve RTP.
Luminescent lifetimes in the order of hundreds of microseconds are perfectly in agreement with the phosphorescence emission and indicate the successful synthesis of very promising phosphorescent materials. Calculation of radiative and non-radiative rate constants indicate that the main reason for the lower phosphorescence recorded in Zeonex compared with that recorded in PMMA may be due to more favored deactivation processes in this matrix.
TD-DFT theoretical calculations let us to identify the energies of the lowest energy singlet excited state, S1 and the corresponding closer triplet state. The population of the triplet state is expected to be much more favored in L2-series in agreement with the more favored nonradiative deactivation process and the lower RTP effect.

Experimental section General procedures
All manipulations have been performed under prepurified N2 using standard Schlenk techniques. All solvents have been distilled from appropriated drying agents. Commercial reagents PPh3, PMe3, P(1-Naph)3, dppm, dppb and PMMA were purchased from Aldrich, and 2-bromophenanthrene was purchased from Fluorochem.

Crystal data
The crystal data and experimental details for the data collection are given in Tables S1-S4.
The single crystal data for 1a, 1e and 2a were measured using a Bruker-Nonius KappaCCD diffractometer with an APEX-II detector with graphite-monochromatized Mo-Kα (λ = 0.71073 Å) radiation. Data collection and reduction were performed using the program COLLECT [53] and HKL DENZO AND SCALEPACK [54] respectively, and the intensities were corrected for absorption using SADABS. [55] The single crystal data for 2c was collected using a Rigaku-Oxford Diffraction SuperNova dual-source diffractometer with an Atlas CCD detector using mirror-monochromated Cu-Kα (λ = 1.54184 Å) radiation. The data collection and reduction were performed using the program CrysAlisPro and Gaussian face index absorption correction method was applied. The structures were all solved by intrinsic phasing (SHELXT) [56] and refined by full-matrix least squares on F 2 using the OLEX2 software, [57] which utilizes the SHELXL-2014 module. [58] CCDC-2019549 to CCDC-2019552 contain the supplementary crystallographic data for these structures. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.

Physical measurements
Infrared spectra have been recorded on a FT-IR 520 Nicolet Spectrophotometer. 1

Theoretical Calculations.
Density functional calculations were carried out using the GAUSSIAN package. [59] The hybrid density function method known as B3LYP was applied. [60,61] Effective core potentials (ECP) were used to represent the innermost electrons of the gold atom and the basis set of valence triple- quality with an extra d-polarization function. [62] A similar description was used for all main group elements. [63] Atomic charges and populations analysis have been confirmed from analysis of Natural Bond Order. [64] Solvent effects of acetonitrile were taken into account by PCM calculations, [65] keeping the optimized geometries for the gas phase without symmetry restrictions. Excited states and absorption spectra were obtained from the time-depending algorithm implemented in Gaussian09. [66]

Synthesis and characterization
Hydrogen atoms assignment is described in Scheme 1.

Synthesis of 2b
The synthesis of complex 2b was performed following the same procedure of 2a by

Synthesis of 2c
The synthesis of complex 2c was performed following the same procedure for 2a but using