Effect of Ligands with Extended π-System on the Photophysical

Jun 14, 2010 - University of Houston. .... Effect of pH on the Photophysical and Redox Properties of a .... Photophysical properties of metal complexe...
0 downloads 0 Views 950KB Size
14664

J. Phys. Chem. B 2010, 114, 14664–14670

Effect of Ligands with Extended π-System on the Photophysical Properties of Ru(II) Complexes† Yujie Sun,† Maya El Ojaimi,‡ Richard Hammitt,‡ Randolph P. Thummel,‡,* and Claudia Turro†,* Department of Chemistry, The Ohio State UniVersity, Columbus, Ohio 43210 and Department of Chemistry, UniVersity of Houston, Houston, Texas 77204 ReceiVed: March 23, 2010; ReVised Manuscript ReceiVed: May 10, 2010

Density functional theory calculations were performed on a series of six ruthenium complexes possessing tridentate ligands: [Ru(tpy)2]2+ (1; tpy ) [2,2′;6′,2′′]-terpyridine), [Ru(tpy)(pydppx)]2+ (2; pydppx ) 3-(pyrid2′-yl)-11,12-dimethyldipyrido[3,2-a: 2′,3′-c]phenazine), [Ru(pydppx)2]2+ (3), [Ru(tpy)(pydppn)]2+ (4; pydppn ) 3-(pyrid-2′-yl)-4,5,9,16-tetraazadibenzo[a,c]naphthacene), [Ru(pydppn)2]2+ (5), and [Ru(tpy)(pydbn)]+ (6; pyHdbn ) 3-pyrid-2′-yl-4,9,16-triazadibenzo[a,c]naphthacene). The calculations were compared to experimental data, including electrochemistry and electronic absorption spectra. The theoretical results reveal that the lowestlying singlet and triplet states in 4 and 5 are pydppn-based ππ* in character, which are remarkably different from the lowest-lying metal-to-ligand charge transfer (MLCT) states in 1-3. The calculated lowest triplet states in 4 and 5 are consistent with the 3ππ* states observed experimentally. However, although the extended π-system of pydbn- is similar to that of pydppn, the HOMO of 6 lies above those of 4 and 5, resulting in strikingly different spectroscopic properties. Calculations show that the lowest triplet excited state of 6 is a combination of 3MLCT and 3ππ*. This work demonstrates that the electronic structure of the tridentate ligand has a pronounced effect on the photophysical properties of ruthenium(II) complexes and that DFT and TDDFT methods are a useful tool that can be used to predict photophysical and redox properties of transition metal complexes. Introduction Molecular assemblies derived from ruthenium polypyridyl complexes continue to attract intense interest in the fields of dye-sensitized solar cells,1,2 artificial photosynthesis, and photocatalysis,3,4 as well as potential agents in biomedical diagnostics and as pharmaceuticals.5,6 Their versatile use originates from their favorable electrochemical and photophysical properties, along with the ease of structural modification.7 A considerable amount of work has focused on ruthenium complexes with bidentate ligands related to 2,2′-bipyridine (bpy), including the prototype complex [Ru(bpy)3]2+, which exhibits a high emission quantum yield from the lowest-energy triplet metal-to-ligand charge transfer (3MLCT) excited state and a long lifetime at ambient temperature.7 In contrast to the highly emissive [Ru(bpy)3]2+ (Φ ) 0.062 and τ ) 820 ns in CH3CN at 298 K), [Ru(tpy)2]2+ (1; tpy ) [2,2′;6′,2′′]-terpyridine, Figure 1a) exhibits very weak emission (Φ < 5 × 10-6) and a short-lived excited state (τ ) 120 ps) in CH3CN at room temperature, making it less useful for photochemical applications.8-10 However, ruthenium complexes with tpy-type tridentate ligands have been generating increasing interest owing to the ease of substitution at the 4′ position of the central pyridine ring of the ligand, making it possible to construct linear supramolecular assemblies.8-10 It is well known that metal-centered ligand-field (LF) states close in energy to the lowest 3MLCT state in 1 result in the †

Part of the “Michael R. Wasielewski Festschrift”. * To whom correspondence should be addressed. E-mails: (C.T.) chemistry.ohio-state.edu, (R.P.T.) [email protected]. † The Ohio State University. ‡ University of Houston.

efficient deactivation of the latter by the former, leading to the short excited state lifetime and weak emission of the complex.7-10 Structural modifications, including substitution of the tpy ligand and addition of polymethylene bridges between the pyridine rings to increase the bite angle have been proposed to increase the energy gap between the LF and 3MLCT states in these complexes, thus extending the 3MLCT lifetime.8-11 In general, research involving complexes related to 1 has focused on increasing the LF-3MLCT gap to extend the excited state lifetime.8-11 In the present work, ligands with extended π-systems were designed to lower the energy of the ligand-centered (LC) 3ππ* state below that of the 3MLCT state. The photophysical properties of a series of four new ruthenium complexes, [Ru(tpy)(pydppx)]2+ (2; pydppx ) 3(pyrid-2′-yl)-11,12-dimethyldipyrido[3,2-a:2′,3′-c]phenazine), [Ru(pydppx)2]2+ (3), [Ru(tpy)(pydppn)]2+ (4; pydppn ) 3-(pyrid-2′-yl)-4,5,9,16-tetraazadibenzo[a,c]naphthacene), and [Ru(pydppn)2]2+ (5) were recently reported by us,6b and the molecular structures of the ligands are shown in Figure 1a. Complexes 2 and 3 are weakly emissive in CH3CN at 298 K, similar to [Ru(tpy)n(pydppz)2-n]2+ (n ) 0, 1; pydppz )3-(pyrid-2′-yl)dipyrido[3,2-a:2′,3′-c]phenazine, Figure 1a). It should be noted that the luminescence quantum yields of the pydppz and pydppx complexes are greater than that of 1 and exhibit longer 3MLCT lifetimes (τ ) 1-4 ns).12 In contrast, the excited state lifetimes of 4 and 5 are dramatically longer (τ ) 20-25 µs).6b Comparison of the transient absorption spectra of 4 and 5 to that of the free pydppn ligand indicates that the lowest excited states in these complexes are pydppn-based 3ππ*, similar to that of the related complex [Ru(bpy)2(dppn)]2+ (dppn ) benzo[i]dipyrido[3,2-a:2′,3′-

10.1021/jp102613n  2010 American Chemical Society Published on Web 06/14/2010

Photophysical Properties of Ru(II) Complexes

J. Phys. Chem. B, Vol. 114, No. 45, 2010 14665

Figure 1. (a) Molecular structures of ligands and (b) ORTEP plot of the cation [Ru(tpy)(pydbn)]+.

c]phenazine).13 These long-lived 3ππ* states are able to photosensitize 1O2 with near unit quantum yield in 4 and 5, resulting in efficient DNA photocleavage under irradiation with visible light.6,12 Theoretical work on ruthenium complexes with bidentate ligands is plentiful in the literature,14-19 however, calculations on complexes with tridentate ligands related to tpy are scarce.20,21 The present work utilizes density functional theory (DFT) to explore the electronic structures of the ground and excited states of 1-5 and a cyclometalated complex, [Ru(tpy)(pydbn)]+ (6), in which pydbn- represents the deprotonated form of pyHdbn (pyHdbn ) 3-pyrid-2′-yl-4,9,16-triazadibenzo[a,c]naphthacene, Figure 1a). The calculated results for 1-6 are compared with experimental data, including those from electrochemistry, steady-state spectroscopy, and transient absorption. Particularly, the lowest energy triplet states of 4-6 are discussed in detail. Experimental Section Materials. The synthesis and characterization of complexes 1-5 were previously reported.6b All reagents and solvents were purchased from commercial sources and were used as received. 2-(Pyrid-2′-yl)benzo[h]quinoline was prepared according to a reported procedure.22 2-(Pyrid-2′-yl)benzo[h]quinoline-5,6-dione. To a suspension of 2-(pyrid-2′-yl)benzo[h]quinoline (80 mg, 0.31 mmol) in acetic acid (2 mL) was added I2O5 (110 mg, 0.33 mmol). The mixture was heated at 120 °C for 2 h, and the reddish solution was allowed to cool to room temperature. An aqueous solution of Na2S2O3 (10%, 20 mL) and EtOAc (20 mL) were added, and the residue was stirred at room temperature for 1.5 h. The organic phase was separated and washed with water and brine, dried over MgSO4 and concentrated to generate the product as yellow flakes (56 mg, 62%). 1H NMR (DMSO-d6): δ 8.84 (d, 1H, J ) 7.45 Hz), 8.75 (dd, 1H, J ) 4.45, 1.75 Hz), 8.72 (d, 1H, J ) 7.4 Hz), 8.47 (dd, 2H, J ) 7.85 Hz), 8.05 (m, 2H), 7.88 (td, 2H, J ) 7.4, 1.15 Hz), 7.67 (td, 1H, J ) 7.45 Hz, 1.15 Hz), 7.56 (ddd, 1H, J ) 7.7, 4.6, 1.15 Hz); IR 1678 cm-1. 3-Pyrid-2′-yl-4,9,16-triazadibenzo[a,c]naphthacene (pyHdbn). A suspension of 2,3-diaminonaphthalene (198 mg, 1.25 mmol) and 2-(pyrid-2′-yl)benzo[h]quinoline-5,6-dione (88 mg, 0.307 mmol) in ethanol (20 mL) was heated at reflux for 2 h. The residue was allowed to cool to room temperature, and the

precipitate was collected, washed with EtOH and acetone, and dried under vacuum to provide pyHdbn as a yellow solid (105 mg, 84%). [Ru(tpy)(pydbn)](PF6). A suspension of pyHdbn (105 mg, 0.257 mmol) and Ru(tpy)Cl3 (113 mg, 0.257 mmol) in EtOH/ water (3:1, 48 mL) was stirred at reflux under Ar for 22 h. The mixture was concentrated and was added dropwisely to a solution of NH4PF6 (300 mg, 1.80 mmol) in water (7 mL). The precipitate was filtered and washed with water. The solid was purified by chromatography on alumina, eluting with CH2Cl2/ acetone (1:1). The crude product was dissolved in a minimum amount of acetone and precipitated with ether (10 mL). The precipitate was collected and washed with ether, providing [Ru(tpy)(pydbn)](PF6) (112 mg, 49%) as a purple solid. 1H NMR (acetone-d6): δ 9.51 (d, 1H, J ) 8.6 Hz), 9.08 (s, 1H), 9.02 (d, 1H, J ) 8.5 Hz), 8.95 (s, 1H), 8.92 (d, 2H, J ) 8.0 Hz), 8.90 (d, 1H, J ) 8.2 Hz), 8.68 (d, 2H, J ) 8.0 Hz), 8.42 (d, 2H, J ) 8.0, 1.15 Hz), 8.36 (m, 1H), 8.31 (m, 1H), 8.25 (t, 1H, J ) 8.0 Hz), 8.08 (td, 1H, J ) 7.45, 1.7 Hz), 7.86 (d, 1H, J ) 6.3 Hz), 7.83 (t, 2H, J ) 8.0 Hz), 7.70 (m, 2H), 7.66 (d, 2H, J ) 5.7 Hz), 7.30 (dd, 1H, J ) 5.7, 1.1 Hz), 7.08 (m, 2H), 6.99 (t, 1H, J ) 7.45 Hz), 6.25 (dd, 1H, J ) 6.3, 1.1 Hz); MS (MALDI-TOF): m/z 742.15 [M - PF6]+. X-ray Determination of [Ru(tpy)(pydbn)](PF6). All measurements were made with a Siemens SMART platform diffractometer equipped with a 4K CCD APEX II detector. A hemisphere of data (1271 frames at 6 cm detector distance) was collected using a narrow-frame algorithm with scan widths of 0.30% in omega and an exposure time of 40 s/frame. The data were integrated using the Bruker-Nonius SAINT program, with the intensities corrected for Lorentz factor, polarization, air absorption, and absorption due to variation in the path length through the detector faceplate. A psi scan absorption correction was applied on the basis of the entire data set. Redundant reflections were averaged. Final cell constants were refined using 6162 reflections having I > 10σ(I), and these, along with other information pertinent to data collection and refinement, are listed in Table S1 of the Supporting Information. The Laue symmetry was determined to be -1, and the space group was shown to be either P1 or P-1. The asymmetric unit consists of one cation, one anion, and one molecule of methylene chloride solvent, all in general positions. The anion was found to be disordered 70:

14666

J. Phys. Chem. B, Vol. 114, No. 45, 2010

30 over two different orientations, and this was treated by refinement of ideal rigid bodies with occupancy factors estimated by comparison of the isotropic displacement parameters. An ORTEP plot of the cation is shown in Figure 1b, and selected bond lengths, bond angles, and torsion angles are collected in Table S2 of the Supporting Information. Instrumentation. The 1H NMR spectra were recorded at room temperature on a JEOL ECX-400 spectrometer at 400 MHz. Chemical shifts are referenced to the residual solvent peak and are reported in parts per million (ppm) with J values in (0.5 Hz. Infrared spectra were measured on a Thermo-Nicolet 370 FT-IR spectrometer. Mass spectra were obtained on a Voyager-DE-STR MALDI-TOF mass spectrometer. Electrochemical studies were carried out on a CV-50W voltammetric analyzer in a three-electrode cell with a glassy carbon working electrode, a Pt wire auxiliary electrode, and a saturated calomel electrode (SCE) as the reference electrode; the latter was separated from the bulk solution by a glass frit. The measurements were conducted at a scan rate of 100 mV/s in deoxygenated CH3CN or CH2Cl2 containing 0.1 M tetra-n-butylammonium hexafluorophosphate as the supporting electrolyte. The oxidation potential of ferrocene (0.42 V vs SCE) was measured under identical conditions and was used as the internal reference. Steady-state absorption spectra were recorded on a HP diode array spectrometer (HP8453) with HP8453 Win System software. Corrected steady-state emission spectra were measured on an SPEX Fluoromax-2 spectrometer with a 90° optical geometry equipped with a 150 W Xe arc lamp as the source. Transient absorption spectra were collected using a home-built instrument described in detail previously.23 Computational Methods. All calculations were performed with the Gaussian 0324 program package employing the DFT method with Becke’s three-parameter hybrid functional and Lee-Yang-Parr’s gradient corrected correlation functional (B3LYP).25 The Stuttgart/Dresden (SDD) basis set and effective core potential were used for the Ru atom,26 and the 6-31G* basis set was applied for H, C, and N.27 The geometries of the singlet ground states and the lowest energy triplet states of 1-6 were optimized in CH3CN using the conductive polarizable continuum model (CPCM). The local minimum on each potential energy surface was confirmed by frequency analysis. Time-dependent DFT calculations produced the singlet and triplet excited states of each complex starting from the optimized geometry of the corresponding singlet ground state, using the CPCM method with CH3CN as the solvent. The dipole moment of each complex was calculated using the corresponding optimized geometries of the ground states and the lowest triplet states. The calculated absorption spectra, electronic transition contributions, and electron density difference maps were generated by GaussSum 1.0.28 Results and Discussion Ground State Geometry, Orbital Analysis, and Electrochemistry. Figure 1b displays the crystal structure of 6, and various crystallographic parameters are listed in Tables S1 and S2. A comparison of the experimental crystal structure and that from the calculated minimized ground state of 6 is provided in Table S3, highlighting the similarities in the bond lengths and angles between the two methods. Table S4 lists selected geometric features of the optimized singlet ground state of 1, which are consistent with the reported crystal structure and the results of other theoretical calculations.20,21,29 The calculated geometric parameters for 2-5 are also included in the Supporting Information, which parallel those of 1 and 6 (Table S5).

Sun et al.

Figure 2. MO diagrams comparing the relative energies of the frontier orbitals of 1-6 (red indicates Ru(d) orbital and blue indicates LUMO).

The calculated energy diagram of the frontier molecular orbitals of 1-6 is shown in Figure 2. Complexes 1-3 were reported to display an oxidation potential in the range of +1.31 to +1.38 V vs SCE, assigned to the Ru(III/II) couple.6b These values are similar to those of the related [Ru(tpy)n(pydppz)2-n]2+ (n ) 0, 1) complexes.12 Table S6 shows that the HOMO, HOMO-1, and HOMO-2 of 1-3 are localized on the Ru(d) orbitals, similar to those of [Ru(tpy)n(pydppz)2-n]2+ (n ) 0, 1).12 With the further extended π-delocalization of the pydppn ligand, 4 exhibits a mainly pydppn-centered HOMO with a small contribution from the Ru(d) orbitals, which lies very close in energy (0.16 eV) to the Ru(d)-centered HOMO-1. Similarly, the Ru(d)-based HOMO-2 in 5 also lies at slightly lower energy (0.30 eV) than its degenerate HOMO and HOMO-1, which are primarily localized on pydppn with a small contribution from the Ru(d) orbitals. Selected occupied molecular orbitals of 4-6 are shown in Figure 3. The cyclic and square-wave voltammograms of the free pydppn ligand in deoxygenated CH2Cl2 show an irreversible oxidation wave at +1.57 V vs SCE (Figure S2). This oxidation potential is significantly lower than those of other ligands, showing that pydppn is easier to oxidize than pydppz and pydppx.6,7,12 It is necessary to mention that DFT calculations previously predicted a dppn-based (dppn ) 4,5,9,16-tetraazadibenzo[a,c]naphthacene) HOMO for dppn-containing Ru(II) complexes using DFT methods, in which the dppn ligand possesses the same extended π-system as pydppn.13,30 It was also reported that the dppn ligand in [Ru(bpy)2(dppn)]2+ can be oxidized by O2 under photolysis conditions, while the oxidation state of the Ru(II) center remained unchanged.31 Substitution of pydppn with the cyclometalating pydbnligand in 6 alters the electron density and energy of the molecular orbitals significantly. The HOMO of 6 is primarily localized on the Ru(d) orbitals, but with a large contribution from pydbn- (Figure 3). Furthermore, the HOMO of 6 lies at higher energy than those of 1-5 by 0.70-0.98 eV. Indeed, as shown in Figure S3, the cyclic and square-wave voltammograms of 6 in CH3CN exhibit a reversible oxidation peak at +0.67 V vs SCE attributed to the Ru(III/II) couple, shifted to more negative potentials as compared with those of 1-5 (0.64-0.74 V). This shift matches the calculations, and the value is consistent with reported oxidation potentials of Ru(II) coordinated by a cyclometalating ligand.32 The HOMO-1 and HOMO-2 of 6 are also primarily Ru(d) in character, with some contribution from the ligands (Table S6). It should be noted, however, that the HOMO-3 of 6 is centered mainly on the distal

Photophysical Properties of Ru(II) Complexes

J. Phys. Chem. B, Vol. 114, No. 45, 2010 14667

Figure 3. Selected occupied molecular orbitals (a) 4, (b) 5, and (c) 6 (isovalue ) 0.04).

TABLE 1: First Reduction Potentials and Calculated LUMO Energies of 1-6 complexes

red E1/2 /Va

∆Eexp/Vb

ELUMO/eVc

∆ELUMO/eVd

1 2 3 4 5 6

-1.21 -0.99 -0.96 -0.70 -0.66 -0.95e

0.22 0.25 0.51 0.55 0.26

-2.59 -2.97 -2.97 -3.10 -3.21 -2.80

0.38 0.38 0.51 0.62 0.21

a The first reduction potential of 1-5 from ref 6d; in CH3CN vs red SCE. b Difference of the first reduction potentials (E1/2 ) of 2-6 relative to that of 1. c Calculated LUMO energy. d ∆ELUMO ) ELUMO(1) - ELUMO(2-6). e This work, in CH3CN vs SCE.

portion of the pydbn- ligand, with only a small contribution from the Ru(d) orbitals. This orbital lies at a slightly higher energy than the HOMOs of 1-5 (Figure 2). As expected, the calculated LUMOs of 1-6 are localized on the ligands (Table S6), typical of ruthenium polypyridyl complexes.7 The LUMOs of the heteroleptic complexes 2, 4, and 6 are localized on the ligand with the extended π-system, pydppx, pydppn, and pydbn-, respectively, whereas those of the homoleptic complexes 1, 3, and 5 have electron density distributed over the two ligands. The LUMOs of 4 and 5 exhibit greater electron density on the portion of pydppn distal to the metal, whereas those of 2 and 3 show a relatively high density on the proximal part of the pydppx ligand (Table S6). The differences in the energies of the LUMOs of 2-6 relative to that of 1 correspond well to the red difference in the experimental first reduction potentials, E1/2 red (Table 1). The E1/2 value measured for 1, -1.21 V vs SCE, shifts to -0.99 and -0.96 V vs SCE in 2 and 3, respectively, both of which possess the pydppx ligand with an extended aromatic structure. The calculated LUMOs of 2 and 3 lie 0.38 eV below that of 1, which qualitatively matches the experimental trend in electrochemical shift (Table 1). The largest aromatic ligand of the series, pydppn in 4 and 5, red values of -0.70 and -0.66 V vs SCE, at 0.51 results in E1/2 and 0.55 V more positive potentials than that of 1, respectively. The energy differences of the calculated LUMOs of 4 and 5 compared with that of 1, 0.51 and 0.62 eV, are also consistent with the experimental differences (Table 1). As shown in Figure S3, three reversible reduction potentials at -0.96, -1.43, and -1.75 V vs SCE were recorded for 6 in CH3CN. The calculated gap between the LUMOs of 1 and 6 is 0.21 eV, consistent with the 0.26 V difference between the first reduction potentials of these complexes. Complex 6 is easier to oxidize than 1-5; this shift can be attributed to the additional electron density on the metal center upon cyclometalation by pydbn-.32 Selected MOs of 1-6 are shown in the Supporting Information (Table S6).

Singlet Excited States and Absorption Spectra. TD-DFT calculations were performed starting from the optimized geometries of the ground states of 1-6 in CH3CN, resulting in the vertical singlet electronic transitions (λ > 300 nm and f > 0.01) listed in Tables S7-S12. A comparison between the calculated singlet electronic transitions and the experimental absorption spectra of 1-3 shows a qualitative match between them, with MLCT bands predicted in the 400-500 nm range.6b The intense peaks of 1-3 in the ultraviolet (uv) region can be assigned as predominantly ligand-centered, in agreement with the known spectral profiles of tpy and pydppx.6b Different from 1-3, the lowest singlet electronic transitions of 4 and 5 are pydppn-based 1ππ* at 545 nm (f ) 0.0317) and 552 nm (f ) 0.0585), respectively (Tables S10 and S11). In the 450-500 nm range, the calculated results of 4 and 5 reproduce the broad MLCT band, showing strong transitions at 446 nm (f ) 0.2295) and 448 nm (f ) 0.4984), respectively, blue-shifted ∼30 nm as compared with the experimental maxima at 473 and 479 nm.6b Overestimation of the absorption maxima of ruthenium polypyridyl complexes of similar magnitude with the TD-DFT method has been previously reported.33 However, the pydppn-based 1ππ* transitions at 432 nm (f ) 0.0643) and 402 nm (f ) 0.0001) are comparable to the LC band at 414 nm in the experimental absorption spectrum of 4.6b Similarly, an intense pydppnbased transition is predicted for 5 at 411 nm (f ) 0.2048), corresponding to the absorption feature of pydppn at 415 nm.6b In the ultraviolet region (300-400 nm), the calculated absorption spectra of 4 and 5 also reflect the experimental spectra, including intensities, where the pydppn-centered 1 ππ* transitions are more intense in 5 than in 4 (Tables S10 and S11).6b The calculated and experimental absorption spectra of 4 are compared in Figure 4a. Due to the introduction of the cyclometalating ligand pydbn-, the experimental and calculated absorption spectra of 6 are dramatically different from those of 1-5. The absorption spectrum of 6 in CH3CN shows a broad band with a maximum at 499 nm (ε ) 14.3 × 103 M-1 cm-1) and a shoulder at 577 nm (ε ) 9.6 × 103 M-1cm-1) (Figure 4b), which is qualitatively reproduced by the calculations (Table S12 and Figure 4b). The calculated singlet excited states exhibit several transitions at 660 nm (f ) 0.0609), 496 nm (f ) 0.0635), and 470 nm (f ) 0.2819), with the lowestenergy peak blue-shifted by 29 nm relative to the experimental absorption maximum (499 nm), similar to those of 4-5. The lowest-energy MLCT state corresponds to a transition where the electron is localized on pydbn-. At higher energy, several LC bands are observed for 6 at 419 nm (ε ) 17.7 × 103 M-1 cm-1), 396 nm (ε ) 21.5 × 103 M-1 cm-1),

14668

J. Phys. Chem. B, Vol. 114, No. 45, 2010

Figure 4. Absorption spectra and calculated singlet electronic transitions of (a) 4 and (b) 6 in CH3CN.

TABLE 2: Orbital Contributions of the Lowest Triplet Excited States of 4-6 in CH3CNa complex

λcalc/nm

4

888.2

5

896.5 896.0

6

840.9

major transitions (contribution) H f L (0.770), H f L+1 (-0.131), H f L+5 (-0.144), H f L+9 (0.131) H-1 f L (0.267), H-1 f L+1 (0.480), H f L (0.485), H f L+1 (0.242) H-1 f L (0.480), H-1 f L+1 (-0.250), H f L (-0.259), H f L+1 (0.485) H-3 f L (0.702), H-2 f L (0.278)b

H ) HOMO, L ) LUMO. b Additional transitions H-9 f L (0.10794), H-3 f L+2 (-0.18208), H-3 f L+4 (-0.10608), H-3 f L+9 (0.11050), H-2 f L+2 (-0.12265), H f L (0.15933). a

347 nm (ε ) 51.3 × 103 M-1 cm-1), 313 nm (ε ) 65.3 × 103 M-1 cm-1), and 298 nm (ε ) 60.4 × 103 M-1 cm-1), which are also qualitatively reflected by the calculated singlet transitions. Triplet States. To further elucidate the identity of the lowest excited states of 4 and 5, the lowest-energy triplet state and higher-energy triplet excited states were also calculated using TD-DFT. As shown in Table 2, the lowest

Sun et al. triplet excited state of 4 in CH3CN, T1, corresponds to a HOMO f LUMO vertical transition at 888 nm (1.40 eV), consistent with the estimated 3ππ* energy of ∼1.50 eV from the transient absorption quenching experiments with various organic chromophores.6b The electron density difference map (EDDM) of the lowest triplet excited state of 4 clearly illustrates the pydppn-localized 3ππ*, showing the promotion of the electron density from the terminal naphthalene portion of the ligand to the region of pydppn proximal to the metal (Figure 5). As expected and also listed in Table 2, 5 exhibits behavior similar to that of 4, with two nearly degenerate calculated pydppn-based 3ππ* transitions at 896 nm (1.38 eV).6b The EDDMs of these two states of 5 are also included in Figure 5. In addition, transient absorption spectra of 4 and 5 have been reported and compared with that of the free ligand pydppn, from which lifetimes of 20 and 24 µs were obtained for 4 and 5, respectively.6b These values are significantly longer than those of 2 and 3 (1-4 ns) whose lowest triplet excited states are 3MLCT in character.6b Investigation of the calculated geometries of the lowest energy singlet (S) and triplet (T) states of 2-5 can also be used to gain insight into the identity of the lowest energy triplet state. As listed in Table S5, the calculated Ru-N bond lengths between the optimized geometries of the S and T states of 2 display a relatively large deviation, with the largest increase of Ru-N2 bond length from 2.011 Å (S) to 2.050 Å (T). However, only a small change is calculated for S and T of 4 and 5, with differences in Ru-N bond lengths of