Chromophore-Functionalized Phenanthro-diimine Ligands and Their Re(I) Complexes.

A series of diimine ligands has been designed on the basis of 2-pyridyl-1 H-phenanthro[9,10- d]imidazole (L1, L2). Coupling the basic motif of L1 with anthracene-containing fragments affords the bichromophore compounds L3-L5, of which L4 and L5 adopt a donor-acceptor architecture. The latter allows intramolecular charge transfer with intense absorption bands in the visible spectrum (lowest λabs 464 nm (ε = 1.2 × 104 M-1 cm-1) and 490 nm (ε = 5.2 × 104 M-1 cm-1) in CH2Cl2 for L4 and L5, respectively). L1-L5 show strong fluorescence in a fluid medium (Φem = 22-92%, λem 370-602 nm in CH2Cl2); discernible emission solvatochromism is observed for L4 and L5. In addition, the presence of pyridyl (L1-L5) and dimethylaminophenyl (L5) groups enables reversible alteration of their optical properties by means of protonation. Ligands L1-L5 were used to synthesize the corresponding [Re(CO)3X(diimine)] (X = Cl, 1-5; X = CN, 1-CN) complexes. 1 and 2 exhibit unusual dual emission of singlet and triplet parentage, which originate from independently populated 1ππ* and 3MLCT excited states. In contrast to the majority of the reported Re(I) carbonyl luminophores, complexes 3-5 display moderately intense ligand-based fluorescence from an anthracene-containing secondary chromophore and complete quenching of emission from the 3MLCT state presumably due to the triplet-triplet energy transfer (3MLCT → 3ILCT).

Introduction

Multichromophore n class="Chemical">compounds, i.e. species combining two or more photofunctional units, offer wide possibilities to manipulate the energy of electronic transitions on the molecular level and consequently within the bulk materials. Depending on the properties of the constituting blocks and the interplay between them, such molecules can be utilized for a diversity of light-harvesting, light-emitting, and charge transport purposes.[1−4] The efficiency of the targeted photophysical processes is defined by the dynamics of the excited state, which can be chemically tuned by proper molecular design.

An appealing approach to systems demonstratn class="Chemical">ing unconventional photophysical behavior involves coupling of an organic chromophore with a transition-metal ion.[5] The presence of a d-block ion participating in the electronic transitions (e.g., metal to ligand or ligand to metal charge transfer, MLCT/LMCT) increases the number of accessible excited states. Furthermore, spin–orbit coupling induced by the heavy atom often activates fast intersystem crossing (ISC), leading to lower energy emissive states with triplet spin multiplicity.[6] Minimization of the electronic interaction between organic and metal-containing fragments in these molecules, e.g. by providing large spatial separation and/or by the lack of direct conjugation, may decouple the fluorescent and phosphorescent emitters to give dual singlet–triplet emission, suitable for ratiometric sensing[7−10] and panchromatic light generation.[11] In some cases even direct attachment of the metal center to the extended organic chromophore affords dual-emissive compounds[12−15] or complexes demonstrating prompt fluorescence[16] because of slow ISC rates. Nevertheless, the majority of the bi- or multichromophore metal–organic assemblies undergo efficient intramolecular energy transfer,[5,17−19] which makes these compounds fundamentally important for luminescent sensing,[20,21] singlet oxygen generation (photodynamic therapy),[22−24] and triplet–triplet annihilation upconversion.[25,26]

From a preparative viewpoint, one facile way to construct the metal–organic architectures relies on the coordination of the metal center to a presynthesized chromophore ligand. In this respect the rhenium(I) carbonyl diimine derivatives constitute an attractive and readily accessible class of photoactive complexes. These compounds play a noticeable role in bioimaging and photocatalysis[27−30] due to robust stereochemistry of the {Re(CO)(diimine)} (x = 2, 3) motif and its tailorability and rich photophysical behavior, dominated by the diimine-defined (MLCT or ligand-centered, LC) triplet excited state.[31]

Representative n class="Chemical">diimines containing extended aromatic systems or functionalized with an ancillary fluorophore, which are shown in Figure , have been successfully employed for the preparation of rhenium(I) carbonyl species. Extensive research efforts have been devoted to the compounds built on the family of dipyridophenazine ligands (A), the majority of which undergo formation of the dark (nonemissive) phenazine-localized triplet excited states of either ligand-centered (3LC) or intraligand charge transfer (3ILCT) nature.[32−36] On the other hand, Re(I) complexes based on phenanthroline-imidazole ligands (B) bearing electron donor (triphenylamine) groups display moderately intense MLCT/ILCT phosphorescence,[37] while for those with naphthalene and coumarine fragments luminescence is largely quenched due to the population of the low-lying 3LC(chromophore) states from the 3MLCT states.[38] A similar triplet–triplet energy transfer, which occurs in a “ping-pong” manner, has been found for phenanthroline (C)[39] and pyridyl-benzoimidazole (D)[40] ligands decorated with N-(1,10-phenanthroline)-4-(1-piperidinyl)naphthalene-1,8-dicarboximide and anthracene chromophores. In the case of phenanthrolines (C), bipyridines (E), and related ligands, anchoring the extended electron-rich groups results in intraligand charge transfer and a dominating triplet emission of ILCT and MLCT parentage.[41−45]

Figure 1

Representative chromophore diimine ligands used for the synthesis of rhenium(I) carbonyl complexes.

Representative chromophore n class="Chemical">diimine ligands used for the synthesis of rhenium(I) carbonyl complexes.

Recently, pyridyl-imidazoles fused with pyrene and phenanthrene motifs have been shown to serve as efficient chelating functions in Ru(II), Os(II), and Ir(III) complexes despite the steric hindrance introduced by the polyaromatic cores.[46,47] Inspired by this preparative success and rich photophysical behavior of Re(I) diimine carbonyl compounds, herein the coordinating pyridyl-phenanthroimidazole motif (Figure ) has been employed for the construction of donor–acceptor bichromophore dyes. We have synthesized the derivatives of 1-phenyl- and 1-(4-bromophenyl)-2-(pyridin-2-yl)-1H-phenanthro[9,10-d]imidazole ligands (L1 and L2) tailored to ethynyl-connected chromophores (anthracene, L3; diphenylamino-anthracene, L4; anthracene-ethynyl-dimethylaniline, L5), which were further used to generate the series of rhenium(I) chloro tricarbonyl complexes [Re(CO)3X(diimine)] (X = Cl, 1–5; X = CN, 1-CN).

Experimental Section

General Comments

The solution 1H and 1H–1H COSY NMR spectra were recorded on Bruker Avance 400 and AMX 400 spectrometers with chemical shifts referenced to residual solvent resonances. The infrared spectra were measured on a Shimadzu FTIR-8400S instrument. Mass spectra were recorded on a Bruker maXis II ESI-QTOF instrument in the ESI+ and ESI– (for 2–5) modes. Microanalyses were performed in the analytical laboratory of the University of Eastern Finland. The synthesis of the ligands (L1–L5) are provided in the Supporting Information.

Synthesis of Complexes Re(CO)3Cl(L1–L5) (1–5)

Pentacarbonyln class="Chemical">rhenium(I) chloride (100 mg, 0.28 mmol) and a stoichiometric amount of the corresponding ligand L1–L5 (0.29 mmol) were suspended in ethanol (30 mL) and degassed by purging nitrogen for 15 min with stirring. The reaction mixture was refluxed for 5 h under a nitrogen atmosphere to give a yellow (1–3), orange (4), or red (5) suspension. The precipitate was collected, washed with ethanol and diethyl ether, dried, and purified by recrystallization.

Re(CO)3Cl(L1) (1)

Recrystallized by slow evaporation of its toluene solution to give bright yellow crystalline material (180 mg, 96%). IR (CH2Cl2, ν(CO) cm–1): 2024s, 1921s, 1894s. 1H NMR (DMSO-d6, 298 K; δ): 9.34 (d, JHH = 7.9 Hz, 1H), 9.21 (d, JHH = 5.4 Hz, 1H), 9.01 (m, 2H), 8.14–7.83 (m, 7H), 7.72–7.61 (m, 3H), 7.42 (t, JHH = 7.9 Hz, 1H), 7.02 (d, JHH = 8.4 Hz, 1H), 6.80 (d, JHH = 8.4 Hz, 1H). ESI+ MS (m/z): 698.04 [M + Na]+ (calcd 698.04). Anal. Calcd for C29H17ClN3O3Re: C, 51.44; H, 2.53; N, 6.21. Found: C, 51.51; H, 2.58; N, 6.25.

Re(CO)3CN(L1) (1-CN)

Complex 1 (100 mg, 0.15 mmol) and silver cyanide (22 mg, 0.16 mmol) were suspended n class="Chemical">in acetonitrile (60 mL), and the mixture was refluxed for 3 h under a nitrogen atmosphere in the absence of light. The suspension was cooled to room temperature and filtered through Celite, and the solvent was removed under reduced pressure. The solid residue was recrystallized by a gas-phase diffusion of diethyl ether into a dichloromethane solution of 1-CN at room temperature to give a bright yellow crystalline material (90 mg, 90%). IR (CH2Cl2, ν(CO) cm–1): 2024s, 1926s, 1914s. 1H NMR (DMSO-d6, 298 K; δ): 9.33 (dd, JHH = 8.1 and 1.2 Hz, 1H), 9.23 (d, JHH = 5.2 Hz, 1H), 9.03 (dd, JHH = 8.1 and 5.2, 2H), 8.15 (m, 1H), 8.08 (dt, JHH = 8.1 and 1.2 Hz, 1H), 8.02–7.92 (m, 4H), 7.88 (dd, JHH = 7.7 and 1.2 Hz, 1H), 7.75–7.69 (m, 3H), 7.45 (dd, JHH = 7.7 and 0.8 Hz, 1H), 7.04 (d, JHH = 7.7 Hz, 1H), 6.83 (d, JHH = 8.1 Hz, 1H). ESI+ MS (m/z): 689.07 [M + Na]+ (calcd 689.07). Anal. Calcd for C30H17N4O3Re: C, 53.97; H, 2.57; N, 8.39. Found: C, 53.79; H, 2.65; N, 8.31.

Re(CO)3Cl(L2) (2)

Recrystallized by a gas-phase dn class="Chemical">iffusion of diethyl ether into a dichloromethane solution of 2 at room temperature to give a bright yellow crystalline material (200 mg, 95%). IR (CH2Cl2, ν(CO) cm–1): 2025s, 1922s, 1895s. 1H NMR (DMSO-d6, 298 K; δ): 9.33 (d, JHH = 8.3 Hz, 1H), 9.21 (d, JHH = 5.5 Hz, 1H), 9.02 (m, 2H), 8.22–8.10 (m, 4H), 7.96 (t, JHH = 7.8 Hz, 1H), 7.86 (t, JHH = 7.8 Hz, 1H), 7.75–7.67 (m, 2H), 7.58 (dd, JHH = 8.3 and 2.5 Hz, 1H), 7.51 (t, JHH = 7.8 Hz, 1H), 7.06 (d, JHH = 8.3 Hz, 1H), 6.91 (d, JHH = 8.3 Hz, 1H). ESI– MS (m/z): 787.93 [M + Cl]− (calcd 787.93), 797.96 [M + HCOO]− (calcd 797.96). Anal. Calcd for C29H16BrClN3O3Re: C, 46.07; H, 2.13; N, 5.56. Found: C, 46.08; H, 2.15; N, 5.62.

Re(CO)3Cl(L3) (3)

Recrystallized by a gas-phase dn class="Chemical">iffusion of diethyl ether into acetone/methanol solution of 3 at room temperature to give bright yellow crystalline material (236 mg, 96%). IR (CH2Cl2, ν(CO) cm–1): 2025s, 1922s, 1894s. 1H NMR (DMSO-d6, 298 K; δ): 9.36 (d, JHH = 8.1 Hz, 1H), 9.24 (d, JHH = 5.3 Hz, 1H), 9.04 (m, 2H), 8.81 (s, 1H), 8.75 (d, JHH = 8.1 Hz, 2H), 8.49 (dd, JHH = 8.1 and 1.8 Hz, 1H), 8.40 (dd, JHH = 8.1 and 1.8 Hz, 1H), 8.30 (dd, JHH = 8.1 and 1.8 Hz, 1H), 8.24 (d, JHH = 8.4 Hz, 2H), 8.16 (t, JHH = 8.1 Hz, 1H), 7.98 (t, JHH = 7.8 Hz, 1H), 7.88 (t, JHH = 7.8 Hz, 1H), 7.80–7.65 (m, 7H), 7.54 (t, JHH = 7.8 Hz, 1H), 7.21 (d, JHH = 8.1 Hz, 1H), 7.04 (d, JHH = 8.1 Hz, 1H). ESI– MS (m/z): 910.08 [M + Cl]− (calcd 910.08), 920.11 [M + HCOO]− (calcd 920.11). Anal. Calcd for C45H25ClN3O3Re: C, 61.60; H, 2.87; N, 4.79. Found: C, 61.80; H, 3.13; N, 4.87.

Re(CO)3Cl(L4) (4)

Recrystallized by slow evaporation of its ethanol/dichloromethane solution at room temperature to give a bright orange crystalline material (280 mg, 96%). IR (CH2Cl2, ν(CO) cm–1): 2025s, 1922s, 1894s. 1H NMR (DMSO-d6, 298 K; δ): 9.37 (d, JHH = 7.8 Hz, 1H), 9.24 (d, JHH = 4.6 Hz, 1H), 9.04 (m, 2H), 8.86 (d, JHH = 8.4 Hz, 2H), 8.52 (d, JHH = 8.2 Hz, 1H), 8.42 (d, JHH = 8.2 Hz, 1H), 8.31 (d, JHH = 8.2 Hz, 1H), 8.19–8.13 (m, 3H), 7.98 (t, JHH = 7.8 Hz, 1H), 7.88 (t, JHH = 7.8 Hz, 1H), 7.82–7.70 (m, 6H), 7.62 (m, 2H), 7.55 (t, JHH = 7.8 Hz, 1H), 7.26–7.21 (m, 5H), 7.06–6.99 (m, 4H), 6.94 (t, JHH = 7.4 Hz, 2H). ESI– MS (m/z): 1077.15 [M + Cl]− (calcd 1077.15), 1087.18 [M + HCOO]− (calcd 1087.18). Anal. Calcd for C57H34ClN4O3Re: C, 65.54; H, 3.28; N, 5.36. Found: C, 65.22; H, 3.39; N, 5.31.

Re(CO)3Cl(L5) (5)

Recrystallized by a gas-phase dn class="Chemical">iffusion of diethyl ether into a dichloromethane/ethanol solution of 5 at room temperature to give a bright red crystalline material (274 mg, 96%). IR (CH2Cl2, ν(CO) cm–1): 2025s, 1922s, 1894s. 1H NMR (DMSO-d6, 298 K; δ): 9.37 (d, JHH = 8.1 Hz, 1H), 9.24 (d, JHH = 5.3 Hz, 1H), 9.04 (m, 2H), 8.81 (m, 2H), 8.72 (m, 2H), 8.51 (m, 1H), 8.41 (m,1H), 8.30 (m, 1H), 8.16 (m, 1H), 7.98 (t, JHH = 7.8 Hz, 1H), 7.90–7.69 (m, 10H), 7.54 (t, JHH = 7.8 Hz, 1H), 7.22 (d, JHH = 8.4 Hz, 1H), 7.04 (d, JHH = 8.4 Hz, 1H), 6.84 (d, JHH = 8.8 Hz, 2H), 3.03 (s, 6H). ESI– MS (m/z): 1053.15 [M + Cl]− (calcd 1053.15), 1063.18 [M + HCOO]− (calcd 1063.18). Anal. Calcd for C55H34ClN4O3Re: C, 64.73; H, 3.36; N, 5.49. Found: C, 64.4; H, 3.48; N, 5.61.

X-ray Structure Determinations

The crystals of n class="Gene">L1, L5, 1, 1-CN, 2, and 5 were immersed in cryo-oil, mounted in a Nylon loop, and measured at a temperature of 150 or 100 K (L5). The structures were determined on Bruker Kappa Apex II and Agilent Technologies Xcalibur diffractometers using Mo Kα (λ = 0.71073 Å) and Cu Kα (λ = 1.54184 Å) radiation, respectively. The APEX2[48] and CrysAlisPro[49] program packages were used for cell refinements and data reductions. A semiempirical or numerical absorption correction (SADABS[50] or SCALE3 ABSPACK[49]) was applied to all data. The structures were solved by direct methods using the SHELXS-2014[51] program with the WinGX[52] graphical user interface. Structural refinements were carried out using SHELXL-2014.[51] The crystallization solvent molecules in 1-CN could not be resolved unambiguously; their contribution to the calculated structure factors was taken into account by using a SQUEEZE[53] routine of PLATON.[54] The missing solvent was not taken into account in the unit cell content.

All non-H atoms were anisotropn class="Chemical">ically refined, and all hydrogen atoms were positioned geometrically and constrained to ride on their respective parent atoms with C–H = 0.89–0.99 Å and Uiso = 1.2–1.5Ueq(parent atom). The crystallographic details are summarized in Table S1.

Photophysical Measurements

Freshly distn class="Chemical">illed solvents (cyclohexane, toluene, dioxane, chlorobenzene, o-dichlorobenzene, dichloromethane, dimethylformamide, and acetonitrile) were used for the solution experiments. For complexes 1–5, all solutions were carefully degassed before lifetime and quantum yield measurements by three “freeze–pump–thaw” cycles. UV–vis absorption spectra were recorded on a Shimadzu UV-1800 spectrophotometer. The excitation and emission spectra in solution were measured on a HORIBA FluoroMax-4 spectrofluorometer. Lifetimes were monitored on a HORIBA Scientific FluoroLog-3 spectrofluorometer. The emission quantum yields were determined by the comparative method[55] using coumarin 102 in ethanol (Φr = 0.764) as a standard with the refraction indexes of dichloromethane and ethanol equal to 1.42 and 1.36, respectively. The uncertainties of the quantum yield determinations were in the range of ±5% (an average of three measurements).

Computational Details

Electronic structure calculations were performed with the Gaussian 16 program package[56] within the framework of DFT/TD-DFT using the PBE0[57,58] hybrid density functional. A 6-311G+(d) basis set was used for the calculations involving pure ligands, while a def2-TZVPPD/6-311G(d) (Re/other atoms) basis set was used for the rhenium complexes.[59] The natures of the stationary points were confirmed by calculating the vibrational frequencies for the optimized S0, S1, and T1 geometries. Calculations were performed in dichloromethane solvent using the C-PCM method.[60,61] Ground state S0 optimizations were also performed under vacuum for comparison between X-ray structures.

Results and Discussion

Synthesis and Characterization

The famn class="Chemical">ily of chelating ligands based on a 2-pyridyl-1H-phenanthro[9,10-d]imidazole core was prepared according to the reaction sequence summarized in Scheme . The 1-phenyl (L1)[47] and 1-(4-bromophenyl) (L2) derivatives were synthesized in good yields (80% and 94%, respectively) by the Debus–Radziszewski method starting from phenanthrene-9,10-dione, 2-pyridinecarboxaldehyde, and the corresponding aniline (see the Supporting Information for the experimental details). The subsequent two-step Sonogashira cross-coupling allows for the conversion of Br-aryl intermediate L2 into the anthracene-functionalized compounds, which were isolated in moderate yields (38–70%) as yellow (L3, L4) and red (L5) crystalline solids, soluble in common organic solvents.

Scheme 1

Synthesis of Ligands L1–L5

Reagents and conditions: (a) NH4OAc, toluene/AcOH, 68 °C, N2, 12 h; (b) X = Br, HC2SiMe3, n-propylamine, Pd(PPh3)4, CuI, 57 °C, N2, 48 h, then K2CO3 in THF/MeOH for 1 h; (c) n-propylamine, Pd(PPh3)4, CuI, 57 °C, N2, 48 h.

Reagents and conditn class="Chemical">ions: (a) NH4OAc, toluene/AcOH, 68 °C, N2, 12 h; (b) X = Br, HC2SiMe3, n-propylamine, Pd(PPh3)4, CuI, 57 °C, N2, 48 h, then K2CO3 in THF/MeOH for 1 h; (c) n-propylamine, Pd(PPh3)4, CuI, 57 °C, N2, 48 h.

The ESn class="Chemical">I+-MS of L1–L5 display the dominating signals corresponding to the protonated molecular ions at m/z 372.15, 450.05, 572.21, 739.28, and 715.28, respectively (Figure S1, Supporting Information). The 1H NMR spectra of these compounds display well-resolved multiplets in the aromatic region (Figure S2), which were assigned on the basis of 2D 1H–1H COSY experiments; their multiplicites and relative intensities are completely compatible with the structural patterns shown in Scheme .

The structures of L1 and L5 were establn class="Chemical">ished by XRD analysis (Figure and Figure S3). Their molecular arrangements found in the solid state are in agreement with the spectroscopic data obtained in solution. The phenanthro-imidazole core in both ligands is nearly flat, as well as the anthracenyl motif in L5.

Figure 2

Molecular view of ligand L5 (thermal ellipsoids are shown at the 50% probability level).

Molecular view of ln class="Chemical">igand L5 (thermal ellipsoids are shown at the 50% probability level).

The planes of n class="Chemical">N(3)-bound phenyl(ene) rings in L1 and L5 are almost perpendicular to those of the imidazole fragment (the angles are 80 and 90° for L1 for two independent molecules and 85° for L5), indicating that the conjugation between the polyaromatic moieties in chromophore-functionalized ligands L3–L5 is most probably disrupted.

Despite tn class="Chemical">he fact that the NMR data suggest unrestricted rotation of the C6H4 ring around the N(3)–carbon bond in solution at room temperature, it is reasonable to propose that the preferable conformation in the fluid medium corresponds to that found in the crystal due to the minimized intramolecular H–H repulsion. This conclusion is also supported by the computational analysis of the optimized geometries (Supporting Information). The crystal packing of L5 reveals extensive intermolecular interactions (Figure S4), which evidently cause visible bending of the extended −C6H4–C2–An–C2–C6H4– fragment.

The coordn class="Chemical">inating ability of the pyridyl-imidazole chelating moiety, fused with sterically demanding polyaromatic systems (phenanthrene, pyrene), has been successfully demonstrated for some late-transition-metal ions (Ir(III), Os(II), Ru(II)).[46,47,62] Similarly, ligands L1–L5 can be employed for binding the {Re(CO)3Cl} fragment to give the complexes 1–5, which belong to the family of well-known [Re(CO)3Cl(diimine)] species (see Scheme ).

Scheme 2

Synthesis of Complexes 1–5

Reagents and conditions: (a) EtOH, reflux, 5 h, N2; (b) AgCN, MeCN, reflux, 3 h, N2.

Reagents and conditn class="Chemical">ions: (a) EtOH, reflux, 5 h, N2; (b) AgCN, MeCN, reflux, 3 h, N2.

Compounds 1–5 were syntn class="Chemical">hesized in high yields following the established procedures.[34,63] In the case of 1 bearing ligand L1, the chloride ion was substituted for the cyanide by reacting 1 with a stoichiometric amount of AgCN. Molecular structures of crystallographically characterized complexes 1, 1-CN, 2 and 5 are shown in Figure and Figures S5 and S6; the relevant experimental details and selected bond distances and angles given in Tables S1 and S2.

Figure 3

Molecular views of complexes 1-CN and 5 (thermal ellipsoids are shown at the 50% probability level).

Molecular views of complexes 1-Cn class="Chemical">N and 5 (thermal ellipsoids are shown at the 50% probability level).

The n class="Chemical">rhenium ions in these complexes adopt a pseudo-octahedral coordination geometry typical for the related tricarbonyl diimine species, with the N(1)–Re(1)–N(2) angles ranging from 73.95 to 74.25°. The steric hindrance introduced by the bulky phenanthrene moiety is expected to cause a strong repulsion between the equatorial C(2)O(2) ligand and the aromatic H–C group. This results in a significant displacement of the metal center from the plane of polycyclic NN′ motif (Figure ); similarly to compounds with benzoquinoline ligands,[64] the dihedral angles between the planes of the Re(1), C(1), and C(2) atoms and of the imidazolyl ring are equal to 34.3° (1), 41.6° (1-CN), 27.5° (2), and 34.8° (5). In addition, the observed distortions systematically lead to the elongation of the Re(1)–Nimi(1) bonds with respect to Re(1)–Npy(2) distances (Table S2), which is in contrast to congener Re(I) complexes with pyridyl-imidazole ligands.[65−67]

Not surprn class="Chemical">isingly, compounds 1, 1-CN, 2, and 5 display extensive π–π, π-C≡C. and CH-π intermolecular interactions (Figures S5 and S6), which essentially influence the crystal packing and might further increase the angle between the diimine backbones and the equatorial planes of the complexes, simultaneously inducing some curvature of the polyaromatic systems.

The spectroscopn class="Chemical">ic data recorded for complexes 1–5 and 1-CN in solution are compatible with the structures depicted in Scheme and the results of the XRD analysis. The IR spectra of chloro tricarbonyl species 1–5 in the CO stretching region are nearly identical and show three strong bands at ca. 2025, 1922, and 1895 cm–1, which correspond to fac-M(CO)3 fragments lacking 3-fold symmetry.[43,68] The lower energy CO vibrations for 1-CN are somewhat shifted to higher frequencies (2024, 1926, and 1914 cm–1), which can be attributed to the substitution of the Cl– for the CN– ligand with a stronger π-accepting character.

The n class="Disease">ESI-MS of 1–5 and 1-CN reveal the signals [1 + Na]+, [(1-CN) + Na]+ and [(2–5) + Cl]−/[(2–5) + HCOO]− ions with the isotopic patterns, which fit completely the calculated patterns (Figure S7). The 1H NMR spectra of these complexes contain well-resolved sets of resonances, which indicate a stereochemically rigid molecular arrangement under ambient conditions. The complete illustrative assignment, performed for 2 on the basis of the 1H–1H COSY and NOESY experiments (Figure S8), confirms that the solid-state structure remains unchanged in solution, including the preferential orientation of the imidazolyl-N-connected phenylene ring due to the restricted rotation around the N(3)–C6H4 bond.

Photophysical Properties and Theoretical Analysis

Ligands L1–L5

Compounds L1–L5 contan class="Chemical">in extended aromatic motifs (phenanthrene (phen) and secondary anthracenyl (an) chromophores), the electronic properties of which define their optical behavior and therefore substantially differentiate L1/L2, L3, and L4/L5. The UV–vis spectra for L1 and L2 in dichloromethane solution are very much alike and display absorption bands in the range 250–360 nm (Table and Figure S9), corresponding to the π–π*phen transitions in the aromatic systems, whereas in the case of L3 the presence of the anthracene moiety results in the appearance of additional longer wavelength absorptions (380–430 nm). Ligands L4 and L5, functionalized with electron-donating −NR2 groups, expectedly display broad bathochromically shifted bands (470 and 490 nm) attributed to the intramolecular charge transfer.

Table 1

Experimental and Calculated Electronic Absorption Data for L1 and L3–L5

	exptl		calcd
	λ, nm	ε, M–1 cm–1	λ, nm	f	MO configuration, % contribution
L1	362	13000	343	0.76	HOMO → LUMO, 94
	330	17000	319	0.04	HOMO → L+1, 83
	261	58000	284	0.09	H-1 → LUMO, 37; HOMO → L+2, 30
L3	425	20000	447	0.62	HOMO → LUMO, 99
	402	22000
	362	20000	343	0.75	H-1 → L+1, 95
L4	464	12000	529	0.44	HOMO → LUMO, 94
	409	1200	431	0.32	H-2 → LUMO, 94
	363	26000	343	0.73	H-1 → L+1, 95
L5	490	52000	561	1.44	HOMO → LUMO, 99
	363	20000	355	0.25	HOMO → L+2, 91

According to TD-DFT calculations (using PBE0 hybrid density functional taking into account dichloromethane solvent with the C-PCM method), the predicted lowest energy excitations S0 → S1 for L1 and L3–L5 are mainly HOMO → LUMO in character (Table ). For L1, it involves phenanthrene-localized π → π*phen transition mixed with phenanthrene → pyridyl charge transfer, LC/ILCT (Figure ). In L3 both the HOMO and LUMO are located on the ethynyl-anthracene fragment with some participation of the phenylene spacer (Figure S10). For L4 and L5 the HOMO is substantially delocalized over the entire anthracene-containing chromophore (Figure and Figure S10), which leads to a visible charge transfer (ILCT) as a result of the S0 → S1 transition. Computationally assessed absorptions (Table ) correlate rather well with experimental data for L1 and L3, which feature ππ* character of transitions. However, for CT compounds L4 and L5 the energies of S0 → S1 transitions are considerably underestimated.

Figure 4

Frontier molecular orbitals for L1 and L5.

Frontier molecular orbn class="Chemical">itals for L1 and L5.

Compounds n class="Gene">L1 and L2 exhibit nearly identical violet-blue photoluminescence in solution (Figure and Table ), which is insensitive to molecular oxygen (3O2) and has a short lifetime of 1.9 ns that indicates its singlet origin. The emission bands are vibronically structured (ν ca. 1250 cm–1), showing no effect of the bromide substituent in L2, which is in accordance with the 1ππ*phen nature of the excited state, dominated by the orbitals of the phenanthroimidazole system. Both the absorption and fluorescence spectra of L1 and L2 are very similar to those of 1,2-diphenyl-1H-phenanthro[9,10-d]imidazole and the derivatives,[69,70] which supports the given assignment.

Figure 5

Normalized excitation (dotted lines) and emission (solid lines) spectra of L1–L5 (CH2Cl2, 298 K).

Table 2

Emission Spectral Data for L1–L5 (298 K, Aerated CH2Cl2)

	λem, nma	Φ, %a	τ, ns	kr, s–1 b	knr, s–1 c
L1	370, 388	37	1.9 ± 0.1	1.9 × 108	3.3 × 108
L2	370, 388	22	1.9 ± 0.1	1.2 × 108	4.1 × 108
L3	433, 459, 486	65	3.3 ± 0.2	2.0 × 108	1.1 × 108
L4	446w, 475w, 566	52	2.1 ± 0.1 (446 nm); 1.4 ± 0.1 (566 nm)	3.7 × 108	3.4 × 108
L5	484w, 602	92	2.6 ± 0.1 (484 nm); 3.8 ± 0.2 (602 nm)	2.4 × 108	2.1 × 107

λex 260 nm (L2, L3), 360 nm (L3), 365 nm (L4, L5).

kr values were estimated by Φ/τobs.

knr values were estimated by (1 – Φ)/τobs.

Normaln class="Chemical">ized excitation (dotted lines) and emission (solid lines) spectra of L1–L5 (CH2Cl2, 298 K).

The emn class="Chemical">ission band of L3 is substantially red shifted and looks essentially the same as the fluorescence profiles of phenyl-enthynyl-anthracene and the related anthracene-based compounds.[71,72] This observation correlates with the results of a theoretical analysis (Figure S10), which points to a major contribution of the anthracene moiety into the lowest-lying S1 state of ππ*an nature. Notably, the congener dye bearing an anthracenyl unit at the N1-phenylene spacer of diphenyl-phenanthroimidazole[73] shows a different electronic structure with the HOMO and LUMO distributed over the phenanthrene and anthracene fragments, respectively.

Ligands L4 and L5 dn class="Chemical">isplay emissions with observed lifetimes of several nanoseconds and intensities being insensitive to the presence of molecular oxygen that corresponds to prompt fluorescence. In contrast to L1–L3, compounds L4 and L5 containing pendant amine groups display structureless emission bands of considerably lower energy (Table and Figure ). These variations in luminescence characteristics indicate crucial changes in the character of the electronic transitions, which are associated with ILCT arising from the donor–acceptor (R2N−π-imidazole) molecular structure of L4 and L5 (Figure and Figure S10). Consequently, these dyes show strong fluorescence solvatochromism (Table , Figure , and Figure S11). Increasing polarity of the solvent (cyclohexane < toluene < dichloromethane < dimethylformamide) results in the bathochromic shift of the major emission band maximum to 2845 cm–1 (L4) and 4853 cm–1 (L5). The lowest energy absorptions are visibly less influenced by the variation of the solvent, which corresponds to the trends typical for charge-transfer luminophores.[74,75] For both fluorophores the highest emission intensity was determined in dichloromethane solutions, with quantum yields Φem reaching 52% and 92% for L4 and L5, respectively. The optical and electronic properties of L5 are close to those of the Me2N–C6H4–C2–anthracenyl–C2–C6H4–X chromophore (X = electron-withdrawing group),[76,77] confirming a negligible participation of phenanthroimidazole in the observed electronic transitions for L5. A severe drop in Φem for L5 in a polar solvent (DMF, Φem = 7%) is comparable to that observed for this bipolar diethynyl-anthracene congener[76] and is in line with a larger Stokes shift that implies more explicit geometry distortions of the S1 state and, consequently, more efficient nonradiative relaxation processes.

Table 3

Photophysical Data for L4 and L5 in Different Solvents

	solvent	λabs, nm (ε, 104 M–1 cm–1)	λem, nm	Φ, %b	τ, ns (aer)
L4	cyclohexane	260 (11.2), 310 (3.6), 344 (2.6), 363 (2.6), 409 (0.8), 472 (0.9)	440w, 503	41	c
	toluene	314 (5.9), 345 (4.3), 364 (3.9), 409 (1.2), 470 (1.3)	443w, 531	47	1.4 (443 nm); 8.7 (531 nm)
	CH2Cl2	262 (13.5), 311 (5.6), 344 (3.6), 363 (2.6), 409 (1.2), 464 (1.2)	446w, 475w, 566	52	2.1 (446 nm); 1.4 (566 nm)
	DMF	312 (6.4), 343 (3.8), 362 (2.9), 411 (1.4), 465 (1.4)	444w, 472w, 587	38	3.9 (444 nm); 1.4 (587 nm)
L5	cyclohexanea	261 (17), 268 (15), 316 (6), 344 (3), 363 (3), 459 (5)	509, 542	64	c
	toluene	319 (5.8), 364 (2.7), 471 (4.9), 490 (4.9)	539	77	2.5 (539 nm)
	CH2Cl2	261 (12.4), 273 (9.8), 318 (5.9), 363 (2.0), 471 (4.9), 490 (5.2)	484w, 602	92	2.6 (484 nm); 3.8 (602 nm)
	DMF	319 (6.0), 362 (2.0), 473 (4.6), 495 (5.3)	484mw, 518w, 676	7	2.8 (484 nm); 3.2 (671 nm)

Extinction coefficients in cyclohexane are estimated approximately due to low solubility of L5.

λex 365 nm.

Could not be determined accurately due to low solubility.

Figure 6

Normalized emission spectra of L5 at 298 K: (A) in various solvents, with the inset showing their visual appearance under UV light (λ 365 nm); (B) in toluene–CH2Cl2 mixtures; (C) dependence of the emission energy on polarity (dielectric constant) of toluene–CH2Cl2 solutions, R2 = 0.996.

Normaln class="Chemical">ized emission spectra of L5 at 298 K: (A) in various solvents, with the inset showing their visual appearance under UV light (λ 365 nm); (B) in toluene–CH2Cl2 mixtures; (C) dependence of the emission energy on polarity (dielectric constant) of toluene–CH2Cl2 solutions, R2 = 0.996.

Extinctn class="Chemical">ion coefficients in cyclohexane are estimated approximately due to low solubility of L5.

Could not be determined accurately due to low solubn class="Chemical">ility.

The emn class="Chemical">ission energies of L4 and L5 show a linear dependence of the medium dielectric constant, which can be conventionally varied by using mixtures of two different solvents (e.g., toluene/CH2Cl2, Figure and Figure S11). This type of dependence is common for systems with photoinduced ILCT where stabilization of charge separated (i.e., polarized) excited state by means of dipole (solute)–dipole (solvent) interactions accounts for the red shift of emission in a polar environment.[34,74,76,78] A Lippert–Mataga analysis reveals satisfactory linear correlation of the Stokes shift with the solvent polarity (see Figure S12 and the accompanying calculations), from which differences in the electric dipole moment between the S0 and S1 states (Δμ) for L4 and L5 were found to be 29 ± 5 and 44 ± 8 D, respectively. These values correlate well with those obtained for donor–acceptor species with comparable linear dimensions.[75,78]

In addition to the main low-energy ILCT emission band, it is possible to distinguish weak high energy signals in the spectra of L4 and L5 (measured from multiply recrystallized samples to exclude the impurities), which are most clearly seen in DMF solutions (Figure and Figures S11 and S13). It should be noted that the increase in HE/LE intensity ratio for L5 in DMF is primarily ascribed to the decrease in LE charge-transfer emission. The difference in the excitation spectra, monitored at the HE and LE bands (Figure S13), suggests two independent emissive states, which are not exceptional for donor–acceptor compounds.[78−80] The structured profile of the HE emissions may be ascribed to anthracene-centered ππ* transitions, which lack charge-transfer character. The corresponding excitation patterns for HE bands (Figure S13) are reminiscent of those for L3 and the related compounds[72] and reveal a red shift typically found for diethynyl anthracene derivatives in comparison to the monosubstituted species. In accordance with the suggested assignment, the energies of HE bands demonstrate insignificant solvatochromic dependence due to an expectedly smaller change in the dipole moment upon ππ* excitation.

Response to Protonation

The electron-donatn class="Chemical">ing N-functions of L1–L5 can be easily protonated, which substantially affects the photophysical characteristics of these compounds. Adding an excess of trifluoroacetic acid to solutions of L1 and L2 causes an immediate appearance of a yellow-greenish color and changes the emission from deep blue (λem 370, 388 nm) to vivid green due to the emergence of two broad bands maximized at ca. 413 and 517 nm (Figure and Figure S14).

Figure 7

Effect of protonation on the absorption and emission spectra of L1 (c = 5.8 × 10–5 M, CH2Cl2, 298 K). The inset shows the visual appearance of L1 (left) and [L1 + H]+ (right) under UV light (λ 365 nm).

Effect of protonation on tn class="Chemical">he absorption and emission spectra of L1 (c = 5.8 × 10–5 M, CH2Cl2, 298 K). The inset shows the visual appearance of L1 (left) and [L1 + H]+ (right) under UV light (λ 365 nm).

The correspondn class="Chemical">ing 1H NMR spectra (Figure S15) indicate that protonation most probably occurs on the diimine (pyridyl-imidazole) part of the molecule, which is reflected by the large low-field shift of the pyridyl H(9) atom as a result of electronic deshielding, analogously to the coordination-induced shift observed for rhenium complexes (Figure S8). Protonation of the pyridinium fragment is expected to stabilize the LUMO level that activates the ILCTphen→py excited state and decreases the energies of both absorption and emission. Reminiscent variations in the optical characteristics due to an emission switch to the ILCT mode have been recently described for a phenanthro-imidazole probe for formaldehyde.[81] Treatment of L3 and L4 with the acid leads to a large decrease in fluorescence intensity without a considerable change in emission energy (Figure S14). This effect might be attributed to the increased electron-withdrawing ability of the diimine motif that facilitates photoinduced electron transfer from the anthracene-based donor. The basicity of the terminal Ph2N–C6H4– group in L4 is expected to be much lower than that of pyridyl and dimethylaminophenyl functions (in L5) (pKa values of conjugate acids of pyridine, dimethylaniline, and Ph3N in acetonitrile are 12.53, 11.43, and 1.3, respectively[82,83]); therefore, protonation of the diphenylamine group in L4 is not likely to occur.

In tn class="Chemical">he case of L5, however, the high proton affinity of the aniline Me2N–C6H4– function is evidently responsible for reversible switching of the optical properties. In acidic medium the long-wavelength absorption (492 nm) is gradually decreased, which is accompanied by the simultaneous growth of the adjacent blue-shifted band with a clearly distinguishable isosbestic point at 476 nm (Figure ). Accordingly, the low-energy broad charge-transfer emission (602 nm in CH2Cl2) is replaced by the high-energy structured band (487 nm) that is assigned to π → π*an anthracene-centered transitions. Addition of a base (e.g., triethylamine, DBU) restores the original orange ILCT fluorescence.

Figure 8

Effect of protonation on the absorption and emission spectra of L5 (c = 10–5 M, CH2Cl2, 298 K).

Effect of protonation on tn class="Chemical">he absorption and emission spectra of L5 (c = 10–5 M, CH2Cl2, 298 K).

Rhenium(I) Complexes

The photophysn class="Chemical">ical properties of compounds 1–5 and 1-CN are summarized in Table . The absorption spectra of 1, 1-CN, and 2 in dichloromethane solution (Figure S16) display high-energy (HE) bands at ca. 260 nm, which are primarily assigned to 1π → π*phen transitions within the aromatic system of the phenanthro-diimine fragments. The longer wavelength weaker absorptions with maxima at ca. 375–378 nm (ε = (2.1–2.3) × 104 M–1 cm–1) and tails up to 450 nm are not observed for the free ligands and apparently correspond to the spin-allowed 1ML′LCT (L′ = Cl/CN/CO) transitions (i.e., {Re(CO)3X} → NN ligand, Figure S17 and Table S3), which conforms with earlier theoretical and spectroscopic investigations of [Re(CO)3X(diimine)] compounds.[84−88]

Table 4

Photophysical Properties of Complexes 1–5 and 1-CN in Solution (Degassed CH2Cl2, 298 K)

	λabs, nm (ε, 104 M–1 cm–1)	λem, nm	Φ, %a	τobs, ns	kr, s–1 b	knr, s–1 c
1	259 (5.1), 377 (2.1)	369, 392, 408, 616	2 (616 nm)	2.6 ± 0.1 (392 nm); 186 ± 9 (616 nm)	1.1 × 105	5.3 × 106
1-CN	258 (5.8), 375 (2.2)	582	20	3550 ± 250	5.6 × 104	2.3 × 105
2	259 (6.1), 378 (2.3)	371, 392, 407, 624	2 (624 nm)	2.1 ± 0.1 (405 nm); 142 ± 7 (624 nm)	1.4 × 105	6.9 × 106
3	263 (14.1), 303 (2.9), 384 (3.4), 403 (3.4), 427 (2.5)	436, 460, 486(sh)	3	4.1 ± 0.2	7.3 × 106	2.4 × 108
4	262 (18.8), 384 (4.9), 470 (2.1)	572	10	4.5 ± 0.2	2.2 × 107	2.0 × 108
5	260 (15.2), 380 (3.9), 478 (6.5), 493 (6.4)	614	22	2.7 ± 0.1	8.1 × 107	2.9 × 108

λex 365 nm.

kr values were estimated by Φ/τobs.

knr values were estimated by (1 – Φ)/τobs.

The UV–vn class="Chemical">is absorption patterns of 3–5 are reminiscent of those obtained for free ligands L3–L5. In the low-energy region these complexes feature electronic transitions, which are determined by the anthracene-based chromophores (Figure S18 and Table S3). These bands are only slightly red shifted with respect to those of L3–L5, which contrasts with the coordination response of donor-functionalized dipyridophenazine[34,78] and pyridyl-triazole[19] ligands and points to essentially independent electronic behavior of phenathro-diimine and anthracene motifs. Nevertheless, the ILCT absorptions for 4 and 5 demonstrate a substantial increase in extinction coefficients to 21000 and 64000 M–1.cm–1, respectively (cf. 12000 and 52000 M–1.cm–1 for L4 and L5). Analogously to the aforementioned compounds 1 and 2, the ML′LCT band at ca. 380 nm is evidently present for 3–5 as well and is clearly visible in the spectrum of 5 (but is absent for L5; Figure S16).

Complexes 1–5 and 1-CN are lumn class="Chemical">inescent at room temperature both in solution and in the solid state (except 3) (see Figure and Figures S19 and S20). In fluid medium 1 and 2 demonstrate dual emission, which is possible to manipulate by varying the excitation wavelength. Selective irradiation into the long-wavelength absorption (λexc 375 nm) of 1 and 2 results in a predominantly low-energy structureless emission with a maximum at ca. 620 nm (Φem = 2%) and lifetimes of 0.19 (1)/0.14 (2) μs (Table ); the latter values are comparable to those of other Re(I) diimine tricarbonyl compounds.[37,38,67,89,90] The intensity of the LE band appreciably decreases in aerated solution, which suggests a triplet origin of the excited state (i.e., 3ML′LCT likely mixed with some 3LC). Higher energy excitation (λexc 330 nm) results in appearance of an HE band with a maximum at 392 nm, particularly pronounced for complex 2 (Figure B), in addition to low-energy 3CT emission. The vibronic structure (ν ca. 1440 cm–1) of the HE band, its position, and the corresponding excitation spectrum, which are close to those of L1 and L2, point to the ligand-centered 1LC nature of HE emissions (i.e., fluorescence) that is also evidenced by short lifetimes of 2.1–2.6 ns.

Figure 9

Normalized excitation (A) and emission (B) spectra of 1-CN and 2 (the corresponding spectra of L2 (filled) are shown for comparison) and (C) excitation (dotted lines) and emission spectra (solid lines) of 3–5 (degassed CH2Cl2, 298 K).

Normaln class="Chemical">ized excitation (A) and emission (B) spectra of 1-CN and 2 (the corresponding spectra of L2 (filled) are shown for comparison) and (C) excitation (dotted lines) and emission spectra (solid lines) of 3–5 (degassed CH2Cl2, 298 K).

In thn class="Chemical">is respect it should be noted that Re(I) diimine complexes, dually emissive in solution under ambient conditions, are rare. Thus, two 3MLCT emissive states with nearly the same energies arising from different conformers were proposed for [Re(Me2bipy)(CO)2(PR3)2]+ compounds.[91] Two clearly resolved emission bands assigned to 1ππ* (HE) and 3ππ* (LE) origins were observed for the family of [Re(NN)(CO)3Cl] species (NN = pyridylimidazo[1,5-a]pyridine ligands), though at a low intensity of less than 0.3%.[92,93] Ultimately, dual luminescence from ππ* (HE) and MLCT (LE) levels was suggested for dirhenium metallacycle[94] and [Re(NN)(CO)3Cl] compounds (NN = dipyridophenazine-based ligands),[95] while for [Re(NN)(CO)3(L)]+/0 species both 1ILCT and 3MLCT excited states were found to undergo radiative relaxation in solution.[44] In addition, a few complexes bearing pendant fluorophores exhibit singlet and triplet luminescence from two different spatially separated emissive centers.[39,43,96]

The unconventional behavior of 1 and 2 can be tentatively rationalized by the qualitative energy level diagram (Figure ). It is assumed that, due to the presence of a polyaromatic phenanthrene system, irradiation of these complexes with λexc 330 nm can optically populate both ligand-centered 1ππ*phen and 1ML′LCT states, which are normally not mixed due to the large energy separation[87] and thus can potentially relax following independent pathways. The singlet charge-transfer state expectedly undergoes fast intersystem crossing (ISC) to the lowest triplet 3CT state, responsible for the long-lived phosphorescence signal at ca. 620 nm. This assignment in general matches the excited 1CT state dynamics of [Re(CO)3X(bipy)] compounds in detail studied by ultrafast spectroscopy and theoretical studies involving spin–orbit coupling calculations.[97,98]

Figure 10

Qualitative energy level diagram for 1 and 2 (left) and 1-CN (right).

Qualitatn class="Chemical">ive energy level diagram for 1 and 2 (left) and 1-CN (right).

However, for the 1ππ*n class="Chemical">phen state the ISC process seems to be too slow to completely overcome prompt radiative relaxation S → S0 that generates a fluorescence band maximized at 392 nm. Slow rates of ISC for 1LC excited states have been claimed for some other second- and third-row metal complexes with the ligand emissive centers directly bound to the metal ions.[16,99−101] In the case of the proposed 1LC(ππ*) state the ISC is symmetry forbidden (due to the same origin of molecular orbitals involved in 1ππ* and 3ππ* states)[102] and might become allowed by mixing with higher energy MLCT states.[101] Changing the chloride for cyanide in 1 has a dramatic influence on the photophysical performance. The emission of 1-CN is independent of the excitation energy and reveals a poorly structured profile (Figure B). Together with a nearly 20-fold increase in lifetime (from 0.19 to 3.55 μs) and a visible hypsochromic shift of 948 cm–1 (34 nm) in comparison with 1, these results point out that the emissive excited state changed its character and now consists of 3ML′LCT and 3LC contributions. Despite the fact that the radiative rate constant for 1-CN is two times smaller than that for 1 (kr = 5.6 × 104 and 1.1 × 105 s–1, respectively), which also complies with the somewhat different lowest lying triplet state, the cyanide derivative shows a 1 order of magnitude higher quantum efficiency (Φem = 20%) due to largely suppressed nonradiative relaxation processes. It is known that the ancillary X ligand in [Re(CO)3X(NN)] compounds (X = pseudohalide or neutral L) can strongly affect the 3MLCT/3ML′LCT levels.[31,35,103] For the complexes with small energy separation between the ML′LCT and LC triplet states the variation in X nature (e.g., substitution of Cl for pyridine)[35] or its electronic characteristics (protonation or boronation of −CN)[90,104] can rise the energy of a triplet charge transfer level so that it becomes comparable to or exceeds that of the 3LC transition. Furthermore, a better π-accepting ability of CN– vs Cl– simultaneously to increasing the energy of the 3ML′LCT state is expected to stabilize the ground state of 1-CN, causing a hypsochromic shift in emission and a decrease in radiative rate constant. The absence of a HE fluorescence band in the spectrum of 1-CN under 330 nm excitation shows that the ISC (1ππ* → 3ππ*) rate is faster than S → S0 decay. However, the proposed largely oversimplified model does not allow us to rationalize the hypothesized facilitation of intradiimine ISC upon substitution of Cl for CN. One plausible explanation might originate from increased energies of higher ML′LCT excited states in 1-CN that could lead to their more effective mixing with the 1LC state.[101] Furthermore, it has been shown computationally that Re–X stretching modes (which are eventually different for 1 and 1-CN) can participate in vibronic coupling of the spin-mixed CT excited states[105−107] and therefore may also influence these second-order LC–ML′LCT interactions.

In tn class="Chemical">he solid state complexes 1, 1-CN, and 2 behave very much alike (Figure S20 and Table S4) and display vibronically structured luminescence bands. The values of observed lifetimes (3.9–4.7 μs) are close to that determined for 1-CN in solution, which indicates a certain similarity of their emissive excited states (i.e., significant mixing of 3CT and 3LC characters). Destabilization of the 3CT state due to a phase transition (solution → solid) that accounts for a change in the lowest lying excited state (or efficient 3IL–3CT coupling) for 1 and 2 is not uncommon for Re(I) phosphors[87,90] and confirms the small energy gap ΔE(3IL–3CT) suggested above.

The character of lumn class="Chemical">inescence in 3–5 is different, as no MLCT contribution was observed regardless of the excitation wavelength; only the short-lived singlet emission associated with the anthracene-based moiety has been detected (Figure C and Table ). For all of these complexes the excitation and emission profiles virtually coincide with the corresponding spectra of the parent ligands L3–L5. A small red shift (159–325 cm–1) of the emission maxima in 3–5 means that coordination of the metal center introduces only a slight perturbation to the frontier molecular orbitals. Completely in line with the behavior of uncoordinated organic compounds with ILCT features (L4 and L5) the emissions of the corresponding complexes 4 and 5 display strong dependence on the nature of solvents, media polarity, and the presence of the protonating agent (Figure S21 and Table S5). Furthermore, the ligand-originated fluorescence correlates with the results of theoretical analysis, according to which the lowest lying excited state S1 for 3–5 is delocalized over the ancillary chromophore (Table S3 and Figure S18).

Quantum yields for 3–5 are apprecn class="Chemical">iably lower with respect to those of the parent ligands L3–L5, which is reflected by much smaller radiative rate constants calculated for the metal complexes. Gradually decreasing the energy of the ligand-based lowest lying excited state (1ILCT) results in less drastic emission quenching (Φem = 10% and 22% for 4 and 5). The lifetimes for 3–5 (τobs = 4.1, 4.5, 2.7 ns) are on the same order as those of the ligands (3.3, 1.4, and 3.8 ns for L3–L5) meaning that the drop in fluorescence intensity upon complexation is not determined by the energy transfer from organic chromophores to the metal ion and other components of the coordination sphere.

Interestn class="Chemical">ingly, the excitation spectrum of 5 does not reveal an MLCT band around 380 nm, which is resolved in the absorption spectrum. This points to the lack of MLCT participation in the dynamic processes which lead to the emissive excited state, indicating no appreciable energy transfer 1MLCT → 1ILCT occurs. Efficient nonradiative relaxation of the 3MLCT state in Re compounds containing the anthracene motif has been previously rationalized by sensitization of the polyaromatic triplet state (i.e., 3MLCT → 3ππ*).[40,108] In the case of 4 and 5 a similar triplet–triplet energy transfer (3MLCT → 3ILCT) also seems to be a feasible pathway to quench the emission from the {Re(diimine)} fragment. An alternative possibility that cannot be ruled out on the basis of the available data involves electron transfer from the electron-rich chromophore to the photoexcited {ReII(diimine)} motif; such a process is not exceptional for rhenium(I) luminophores, and its probability depends on the properties of the diimine ligand.[43,108] The described photophysical behavior of 4 and 5 clearly contrasts with optical characteristics of the majority of rhenium(I) diimine compounds, which predominantly exhibit triplet emission or the formation of the intraligand dark triplet states.[32−36]

Conclusions

We have developed the famn class="Chemical">ily of readily accessible diimine ligands based on a coordinating pyridyl-phenanthroimidazole motif (L1, L2), which was functionalized with electron-rich anthracene-based units to give bichromophoric molecules L3–L5. The optical properties of ligands L1 and L2 are defined by ππ*/CT (phenanthrene → pyridyl) electronic transitions, whereas for L3–L5 they are localized in the anthracene motif. The donor–acceptor architecture of L4 and L5 enables efficient low-energy intramolecular charge-transfer transitions (maximum λabs 490 nm with tails below 550 nm for L5 in CH2Cl2). The title compounds are luminescent in solution, showing moderate to high quantum efficiencies, and cover a wide range of emission wavelengths from violet to orange-red (Φem = 22–92%, λem 370–602 nm in CH2Cl2). The charge-transfer character of emission for L4 and L5 determines their pronounced fluorescence solvatochromism. The presence of electron donor groups (pyridyl and dimethylaminophenyl) susceptible to protonation allows for reversible switching of emission parameters by affecting the charge-transfer processes. Ligands L1–L5 easily get coordinated to the rhenium(I) center, forming the corresponding complexes of the general composition [Re(CO)3X(diimine)] (X = Cl, 1–5; X = CN, 1-CN). Compounds 1 and 2 demonstrate dual fluorescence–phosphorescence arising from the same organic fragment that is a rare behavior for rhenium(I) luminophores. This phenomenon is tentatively attributed to the excitation-dependent population of poorly coupled 1ππ* and 3CT excited states, which undergo independent radiative relaxation leading to high- and low-energy emission bands. Changing the chloride for cyanide has a pronounced effect on luminescence intensity that reveals 10-fold increase in 1-CN in comparison with 1.

Photoemission of complexes 3–5 is dominated by the intraligand transitions, only slightly perturbed by the presence of the metal ion. The lack of phosphorescence originating from 3ML′LCT or 3IL(phenanthrene) might be explained by the triplet energy transfer to the nonemissive state of the anthracene moiety. Moderately intense intraligand fluorescence detected for 4 and 5 is drastically different from the properties of the major part of the known rhenium(I) luminophores.

The reported ln class="Chemical">igands, showing intense absorption and emission bands in a wide range of the visible spectrum, are thought to be suitable for the preparation of a variety of photofunctional coordination metal complexes.