Tuning of Photochemical and Photophysical Properties of [ RuII ( 2 , 2 ’-bipyridine ) 2 Lx ] Complexes using Nonchromophoric Ligand Variations

Cálculos de orbitais moleculares para o complexo cis-[Ru(bpy) 2 L x ](PF 6 ) 2 , onde bpy é 2,2’-bipiridina e L é 3-aminopiridina (complexo 1 com x = 2), foram realizados com o programa Gaussian 03 usando o método DFT. A estrutura eletrônica e as propriedades moleculares de 1 foram caracterizadas no vácuo e em solução com acetonitrila e comparadas com os resultados obtidos para o complexo com L = 5,6-bis(3-amidopiridina)-7-oxanorborneno (complexo 2 com x = 1). Os espectros eletrônicos dos complexos 1 e 2 foram investigados por TD-DFT. Os dados experimentais de voltametria cíclica, UV-vis, fotoquímica e fotofísica foram comparados com dados teóricos de maneira a estabelecer a influência de L nas transições eletrônicas e interpretar as diferenças entre os comportamentos fotoquímicos desses complexos.


Introduction
Previous studies conducted in our laboratories demonstrated that visible light photolysis of complex 1, cis-[Ru(bpy) 2 L 2 ](PF 6 ) 2 , where bpy is 2,2'-bipyridine and L = 3-aminopyridine (3Amnpy), in CH 3 CN solution, leads to ligand photosubstitution. 1 On the other hand, the similar complex 2, cis-[Ru(bpy) 2 L](PF 6 ) 2 , where L = 5,6-bis(3amidopyridine)-7-oxanorbornene (3Amdpy 2 oxaNBE), shows photophysical properties with emission in the visible region (l em = 600 nm; t = 650 ns) and undergoes an electron transfer process using methylviologen (MV 2+ ) as quenching agent. 1 The ligand 3Amdpy 2 oxaNBE contains a cyclic olefin connected to two pyridine rings which are coordinated to the {Ru II (bpy) 2 } moiety in complex 2 (Scheme 1). 1 This bidentate monomer-ligand was obtained from a reaction between 3Amnpy and 5,6-bis-carboxylate-7- oxanorbornene monomer, so that, in the Ru II complexes, the influence of the substitute groups in the meta-positions of the pyridine rings could be compared. 1n view of the different photobehaviors of 1 and 2, the primary photoprocesses and the characterization of the electronic transitions of these complexes need to be understood.Thus, a theoretical investigation of 1 based on DFT calculations was carried out to gain further insight into the influence of the nonchromophoric ligand on the transition between the MLCT and MC states, which are closely related to the photochemical properties.In addition to what has already been calculated for complex 2, 1 a more detailed theoretical investigation was performed with 2 to improve the discussion.Investigations of the photoreactivity of 1 in various solvents and in different irradiation wavelengths were also performed. 4][5][6][7][8][9][10][11]

Materials and procedures
Complexes 1 and 2 were prepared as previously described. 1All solvents used were of HPLC grade.The optical spectra were recorded on a Varian spectrophotometer model Cary 500 NIR, using 1.00 cm path length quartz cells.
The NMR data were acquired using a Bruker DRX-400 spectrometer.The samples were prepared under argon and analyzed at room temperature using CD 3 CN.The chemical shifts (d) are given with reference to tetramethylsilane (TMS).
Monochromatic irradiations at 330, 440 and 500 nm were generated either using a 200 W xenon lamp in an Oriel model 68805 Universal Arc Lamp source selected with an appropriate interference filter (Oriel) or a RMR-600 model Rayonet Photochemical reactor using RMR-4200 lamps.The experiments were carried out at room temperature in 1.00 cm path length quartz cells with 4 polished sides capped with rubber septa.The magnetically stirred solutions (initial complex concentration of ca. 10 -5 -10 -2 mol L -1 ) were deoxygenated with argon.Potassium (tris-oxalate) ferrate(III) was used in actinometry. 12,13The progress of the photoreactions was monitored either by UV-vis or 1 H NMR spectra.
The emission spectra at 25.0 o C and 77 K were recorded on an Aminco-Bowman spectrofluorometer model J4-8960A with a high-pressure xenon lamp and an IP 28 type photomultiplier.

Computational methods
The calculations were made using the Gaussian 03 package. 14The starting molecular geometries were obtained at the HF/3-21G level of theory.The final molecular geometry optimizations were performed using the Kohn-Sham density functional theory (DFT). 15The Becke three-parameter hybrid exchange-correlation function (B3LYP) 16 was used with the pseudo-potential basis set LanL2DZ. 17 No symmetry condition was imposed.Vibrational frequencies were calculated from the analytic second derivatives to check the minimum on the potential energy surface.][20] The fragments {Ru II (bpy) 2 }, {3Amnpy} and {3Amdpy 2 oxaNBE} were used to investigate the metalligand interaction energies and the composition of the orbitals.The molecular orbitals are expanded in the converged molecular or atomic orbital of these fragments.The Mulliken population of a fragment orbital in a molecular orbital was used to denote the percentage of the fragment orbital character of that molecular orbital.The differences between the one-electron energies of the appropriate virtual and occupied molecular orbitals were used as a first approximation for excitation energies.
Molecular orbital (MO) compositions and the overlap populations between molecular fragments were calculated using the AOMix program. 21,22The atomic charges were calculated using NBO analyses.
Time-dependent density functional theory (TDDFT) was used to calculate the energies and intensities of the electronic transitions. 23,24The electronic transitions were transformed into simulated spectra using the Swizard program 25 and Gaussian functions with half-widths of 25,000 cm -1 .

Results and Discussion
A general view of the structures for complexes 1 and 2 in CH 3 CN is shown in Figure 1.Selected bond lengths and angles for the optimized geometry are given in Table 1.The X-ray data for [Ru(bpy) 3 ](PF 6 ) 2 are also listed. 26oth complexes 1 and 2 have very similar geometrical arrangements.Each complex exhibits the Ru atom bounded to two bipyridyl ligands (bpy1 and bpy2) in cis configuration along with the L (3Amnpy or 3Amdpy 2 oxaNBE) ligands.The bpy1 molecule (characterized by eq N1 and eq N7 pyridinic atoms), one pyridine N atom from bpy2 ( eq N19), one pyridine N atom of L ( eq N31) and the Ru atom lie roughly in the equatorial plane.The other N-pyridine atom ( ax N25) of L is trans-positioned to a pyridine unit of the bpy2 ( ax N13) located in the axial position.The complexes are characterized by C1 symmetry.
There is a satisfactory agreement between the theoretical data of complexes 1 and 2 concerning the bond length of each N-pyridinic atom and the central atoms (Table 1).The largest difference between complexes 1 and 2 concerns the bond lengths when one bpy ligand in [Ru(bpy) 3 ](PF 6 ) 2 is replaced by a different L (3Amnpy or 3Amdpy 2 oxaNBE).The experimental bond lengths of the bpy N atoms for the tris-bpy complex are rather short in relation to complexes 1 and 2. The bond lengths Ru-N(L) for L = 3Amnpy (ca.2.16 Å) and 3Amdpy 2 oxaNBE (ca.2.20 Å) are longer when compared with the Ru-N bond lengths observed for the bpy ligand, which is 2.05 Å.This could be caused by a diminished back bonding in complexes 1 and 2.
The polyhedral coordination of the complexes corresponds to an octahedral arrangement of the ligands in the coordination sphere of the metal.For complex 2, the polyhedron is slightly distorted, with the trans eq N1(bpy1)-Ru-eq N31(3Amdpy 2 oxaNBE) angle equal to 171 o , for instance.
Considering that [Ru(bpy) 3 ](PF 6 ) 2 differs from 1 and 2 by replacing one bpy molecule for L, while the coordinated bpy molecules show similar N(bpy)-Ru-N(bpy) biting angles of 78 o C as expected, the N(L)-Ru-N(L) angles in complexes 1 and 2 increase by 14-17 degrees.In addition, while the angle ax N25(L)-Ru-eq N31(L) for complex 2 (96.5°) is larger than that found for complex 1, 92.6°, the eq N1(bpy1)-Ru-eq N31(L) angle decreases ca.4.0 o .The large bite ring could explain why the two py rings of the 3Amdpy 2 oxaNBE are not orthogonal, contrary to the 3Amnpy molecules.
Additional information about complex binding was obtained by NBO charge analyses and charge decomposition analyses (CDA) using the AOMix program. 21,22able 2 shows the NBO charges for complexes 1 and 2. The calculated charge distributions show that Ru atoms carry similar positive charges, while the N-pyridinic atoms bound to Ru (N1, N7, N13, N19, N25, N31) and the N-substituted (N37, N38, N47) atoms in the L ligands are negatively charged.The highest negative charges are in the N-substituted atoms of each complex (N37, N38, N47) and the highest ones among them are located in complex 1 (N37 and N38).Investigation of the NBO charges in free bpy, 3Amnpy and 3Amdpy 2 oxaNBE shows that no noticeable global charge transfer occurs between the pyridine units from either bpy or L (3Amnpy or 3Amdpy 2 oxaNBE) and the Ru centre after L coordination, considering the fact that the NBO charge values in the N-pyridinic atoms are similar in each case.In counterpart, the negative charges in the N-substituted atoms differ 0.15-0.17units between the 3Amnpy and 3Amdpy 2 oxaNBE.This probably occurs because of the greater electronegativity of O in relation to N. As a consequence, the electron densities from the N37/ N47 atoms migrate towards the amide-carbonyl moiety.The relative values of these transfers can indeed be deduced from the NBO orbital occupancies.Table 3 shows the donation and back-donation charge transfer between different fragments for complexes 1 and 2. It can be observed that, while the back-donation from the {Ru II (bpy) 2 } moiety to L is insignificant, the donation from L to the {Ru II (bpy) 2 } moiety involves 0.56 electron.It is interesting to note the CDA between the {Ru II (bpy) 2 (3Amdpy 2 -)} and {(-oxaNBE)} fragments, suggesting that the oxaNBE moiety has a large contribution in the relative charge donation to the {Ru II (bpy) 2 } moiety.

Molecular orbital compositions
Further understanding of the nature of the coordination bonds and the electronic structure in the studied complexes can be provided by analysis of the valence molecular orbital composition.The energies and composition of the frontier molecular orbitals of complexes 1 and 2 are given in Table 4.The frontier molecular orbital representations are presented as Electronic Supplementary Information (Tables S1 and S2).
The theoretical data for complex 1 show that the HOMO orbital has a high percentage of Ru non-bonding d-orbital (86%), as occurs with 2. 1 In fact, the Ru orbitals are present in all HOMOs of 1 shown in the selected frontier orbitals, but it is not the case for 2. However, the type of L defines differences between the LUMOs from 1 and 2. The LUMO in 2 is exclusively located on the 3Amdpy 2 oxaNBE ligand, whereas the bpy orbital predominates in the LUMO+1 and LUMO+2. 1 For complex 1, the p* orbitals of the bpy predominate in the LUMO, LUMO+1, LUMO+2 and LUMO+3 with a few Ru antibonding participation.Another substantial difference, considering the analyses of the frontier orbitals from 1 and 2, is the LUMO+12 situated 6 eV above the HOMO, which has a large Ru antibonding character (82%).The HOMO-1 in complex 1 comes basically from the Ru nonbonding orbitals, while the HOMOs -2, -3 and -4 have a sizeable contribution from the 3Amnpy.
The HOMO-LUMO gaps are 3.4 and 3.2 eV for 1 and 2 respectively, whereas the back-donation towards the L ligand is weak in both cases, as revealed by the charge donation analysis (Table 3).Considering a correlation between the HOMO-LUMO energy gap and the degree of charge delocalization, 27 it can be deduced that the Ru-bpy covalent interaction involves charge-donation from the L ligands to the {Ru II (bpy) 2 } fragment.Similar values were observed in the case of L = 3Amdpy 2 oxaNBE (complex 2). 1 The presence of the 3Amdpy 2 oxaNBE ligand in 2 decreases the oxidation potential of the metal complex to a less positive value compared to the 3Amnpy and bpy derivative complexes (E 1/2 ([Ru(bpy) 3 ] 2+ ) = +1.29,E 1/2 (1) = +1.11and E 1/2 (2) = +0.77V vs. Ag/AgCl in CH 3 CN).This occurs in agreement with the large Ru-L bond lengths in relation to Ru-bpy. 1 Since the HOMO-LUMO gap energies are unaffected by the different L, the net effect of L = 3Amdpy 2 oxaNBE in 2 is to lower the energy of the 3 MLCT states, as previously discussed. 1Thus, when comparing complex 2 to the parent complex 1, this effect helps to decouple the 3 MLCT from the higher energy 3 MC states, resulting in an emissive 3 MLCT, which does not occur in 1.

Theoretical and experimental electronic spectrum studies
The experimental and theoretical absorption spectra of complexes 1 and 2 in CH 3 CN are characterized by two rather strong bands (Table 5; Figure 2).The relative positions and intensities are dependent on the nonchromophoric ligands.According to the experimental data (Table 5), the lowest energy absorption bands of complex 2 are red-shifted approximately 30 nm in relation to those of complex 1.
When the oxaNBE monomer-group is present in the substituted pyridine ligand, an extended conjugation is observed (Table 5), lowering the transition energy while increasing the oscillator strength.Similar tendencies are observed for the oxidation of Ru II to Ru III (Table 6).The higher metal oxidation potential of complex 1, compared to that of 2, reflects the stabilization of the +2 oxidation state by the 3Amnpy, the less s donor ligand.Figure 3 points out that more electron-withdrawing ligands decrease the electron density at the metal center for the related series of complexes and lead to more positive potentials than the complexes with more electron-donating ligands.
For complex 2, the lowest energy calculated absorption could be assigned as a HOMO-2→LUMO+1 electronic transition with an MLCT nature (Table 5).There are four electronic transitions at higher energies: HOMO-1→ LUMO+4, HOMO-2→LUMO+4, HOMO→LUMO+6 and HOMO→LUMO+7.The energy differences among these transitions are not significant, ca.0.1 eV, although the differences in the calculated intensities are substantial (f = 0.015-0.038).In this context, for complex 2, it is noteworthy that the experimental absorption at 3.62 eV could be originated from the highest filled Ru d p orbitals to the p* orbital of 3Amdpy 2 oxaNBE.Furthermore, the calculations do not predict any MC transitions in the region studied for complex 2 (up to LUMO+20).These features correlate with the electrochemical behavior and structural features, such as the photophysical and photochemical results described in the next section.
Apart from the MLCT transitions, complex 1 has an MC transition at 3.78 eV (328 nm), on account of the HOMO→LUMO+12 transition.Because of the low intensity (f = 0.0060), this transition is obscured by the intense MLCT transition at 3.69 eV (f = 0.0683).The ground state calculation predicts 0.09 eV (726 cm -1 ) as the energy gap between the first MLCT and the MC states.
Replacement of bpy by 3Amdpy 2 oxaNBE does not have any significant influence on the position of the highest MLCT absorption at ca. 340 nm (3.64 eV) of the {Ru II (bpy) 2 } moiety (Table 6), although the e value increases significantly on account of the contribution of the Ru→3Amdpy 2 oxaNBE transition.In contrast, the 3Amnpy ligand shifts this band to 328 nm (3.78 eV), which matches the MC transition as attributed by the simulated spectrum.For a series of related [Ru II (bpy) 2 L x )] q+ complexes (Table 6), the experimental energy of this transition decreases in the order 3Amnpy (3.78 eV) > py (3.67 eV) > 3Amdpy 2 oxaNBE (3.62 eV) > bpy (3.59 eV) > 4Amnpy (3.56 eV) > Cl -(3.26 eV).
Table 7 presents the lowest energy bands for complexes 1 and 2 in different media.The absorption bands are solvent sensitive in agreement with the MLCT assignments.

Photochemistry studies
As reported in the earlier paper, complex 2 was photochemically unreactive. 1 On the other hand, the photolysis of complex 1 in CH 3 CN caused photoreaction with a quantum yield of 0.26 at l exc = 440 nm, producing the cis-[Ru(bpy) 2 (CH 3 CN) 2 ] 2+ ion. 1 In order to better understand this photochemical process, the changes in the UV-vis spectra of complex 1 in a CH 3 CN solution during photolysis with 440 nm light (I 0 = 1×10 -9 einstein s -1 ) were analyzed (Figure 4a).A progressive depletion of the absorption band is noted at 464 nm, with concomitant blue shifts to 440 nm (Figure 4b) and then to 428 nm (Figure 4c).This implies that the formation of the cis-[Ru(bpy) 2 (3Amnpy)(CH 3 CN)] 2+ ion with l max = 440 nm occurs, followed by a second coordination of the p acceptor CH 3 CN ligand, leaving a complex with a much smaller extinction coefficient when compared to complex 1.The l max of the mono-solvent complex coincides with the irradiation wavelength.Then, after 20 measurements, a depletion of the absorption of the first photoproduct species can be observed due to secondary photolysis (Figure 4b-c).The new band profile is similar to that shown by the cis-[Ru(bpy) 2 (CH 3 CN) 2 ] 2+ complex ion (Table 6).From these UV-vis spectrophotometer time scale measurements, these reactions occur quite smoothly.Isosbestic points are preserved even at a very high conversion.The quantum yield  for the 3Amnpy photosubstitution at 440 nm was dependent on the irradiation time up to at least 3% decomposition of complex 1 due to the secondary photolysis.
The presences of the mono and bis-acetonitrile complexes during photolysis of 1 were confirmed by 1 H NMR measurements.A CD 3 CN solution of cis-[Ru(bpy) 2 (3Amnpy) 2 ] 2+ was irradiated at room temperature with 420 nm light (I 0 = 1x10 -7 einstein s -1 ).The 1 H NMR spectra were recorded with 30 min time intervals (Figure 5).Simple intensity arguments tell us that the peaks with chemical shifts between 6.9 and 9.5 ppm are from the bipyridine and Amnpy hydrogens. 29During irradiation, decreasing and increasing resonance peaks were observed, indicating photorelease of 3Amnpy for 2 h.In particular, the signal at 9.0 ppm decreases and a new signal at 9.4 ppm appears.This indicates the replacement of 3Amnpy by acetonitrile in the metal coordination sphere.The shift of the bipyridine hydrogen signal at 9.0 ppm to a higher frequency indicates the change of the s-donor nature of the trans ligand (3Amnpy versus acetonitrile).It is interesting to note that the shift of the 3Amnpy NH 2 signal from 4.42 to 4.52 ppm also indicates that the 3Amnpy is released.After four irradiation cycles, a well-resolved 1 H NMR spectrum is observed and assigned to the bis-acetonitrile complex.It differs from the spectrum of the sample irradiated for 30 min, attributed to the cis-[Ru(bpy) 2 (3Amnpy)(CH 3 CN)] 2+ complex ion.The final 1 H NMR spectrum profile was identical to the 1 H NMR spectrum of the cis-[Ru(bpy) 2 (CH 3 CN) 2 ] (PF 6 ) 2 species synthesized thermally, confirming the attribution.
Considering the observed stepwise photosubstitution of 1 in CH 3 CN, photolyses in different solvents were carried out as a function of the irradiation wavelength (Figure 6).While thermal reactions were not observed up to 10 h at 25.0 °C, in all the irradiated solutions changes in the UV-vis spectra with blue shifted final spectra were observed.Table 7 shows the initial quantum yield data calculated for wavelengths where the e value ratios between the starting complex 1 and the products were the largest.
Photolysis at 330 nm provides large quantum yields (Table 7).It is relevant that this wavelength is close to the calculated l max of the lowest-energy spin-allowed MC absorption band (Table 5).
The quantum yield data in Table 7 involve two simultaneous variables: l max (MLCT) and l irr .The variation of l irr should not influence the quantum yields if the deactivation from upper to lower energy excited states were 100% efficient.However, it is clear from Table 7 that photosubstitution quantum yields are quite dependent on l irr for all the studied cases.Furthermore, the interconversions to common states, presumably the lowest in energies, do not occur with 100% efficiency.
The significant decrease in the f values and the large dependence on l irr suggest a change in the nature of the lowest excited state from 3 MC to 3 MLCT.These results, combined with the absorption experiments and DFT calculations, confirm that the lowest energy reactive excited state is a 3 MC for 1.
The tendency in the photochemical properties of complexes 1 and 2 parallels the changes in the photophysical properties found in the emission spectra.
In sharp contrast with the photochemical results, substitution of bpy in [Ru(bpy) 3 ](PF 6 ) 2 by 3Amdpy 2 oxaNBE shifts the MLCT emission maximum to 565 nm, while the 3Amnpy ligand shifts this band to 590 nm.The excited state energy (0-0 energy) can be calculated according to the equation: E 0-0 = n max + 1.29 Dn 1/2 . 30Although proposed for a Gaussian band shape, this relationship provides a qualitative basis for comparing the spectra of closely analogous complexes.For 2, [Ru(bpy) 3 ](PF 6 ) 2 and 1, the n max and Dn 1/2 values, defined as the full width at half maximum, are 2.20, 2.16, 2.10 and 0.16, 0.11, 0.11 eV, respectively.Thus E 0-0 values are 2.42, 2.29 and 2.24 eV according to the assumption of a Gaussian band shape for 2, [Ru(bpy) 3 ](PF 6 ) 2 and 1 respectively.From these results, the 3 MLCT spectroscopic energy gap between complexes  hν hν 1 and 2 is 0.18 eV.The difference seems to be reasonable on account of the change in the electron densities about the metal centers after the attachment of the oxaNBE group to the substituted pyridine ligand.Concerning the excitation of complex 2 at 450 nm at room temperature, the position of the emission maximum does not change when compared to the corresponding [Ru(bpy) 3 ](PF 6 ) 2 emission (577 nm).Complex 1 shows an emission at 592 nm with intensity much lower than that of complex 2 (Figure 7).
The observation of a weaker MLCT luminescence at the longer wavelength and the high sensitivity of f for l irr suggest that complex 1 has moved from the reactive to the unreactive category at room temperature, although still showing some intermediate nature (Table 7).
In the same way, going from 3Amnpy to 3Amdpy 2 oxaNBE modifies the excited state order, where a reactive MC state is below or comparable in energy to the lowest MLCT state.These results are consistent with the observation of the extended p-conjugation over the 3Amdpy 2 oxaNBE ligand.It increases the opportunity for electron delocalization at the triplet level.This delocalization should lower the nuclear reorganization energy that accompanies nonradiative decay of the 3 MLCT, stabilizing the triplet states.

Figure 1 .
Figure 1.Optimized molecular structures obtained at the B3LYP/ LanL2DZ level for complexes 1 and 2.

Figure 3 .
Figure 3. Correlation between MLCT l abs and E 1/2 values for complexes 1, 2 and related complexes in CH 3 CN.

Figure 4 .
Figure 4. Changes in the absorption spectra resulting from the continuous photolysis of complex 1 in CH 3 CN using 440 nm light.

Table 2 .
NBO charge analyses calculated at the B3LYP/LanL2DZ level for complexes 1 and 2 a N-pyridinic atoms in the pyridine rings; b N-substituted atoms (amine in 1 or amide in 2) in the position 3 of the L pyridine rings.

Table 3 .
Charge decomposition analyses (CDA) calculated at the B3LYP/ LanL2DZ level for fragments from complexes 1 and 2

Table 4 .
Molecular orbital characters and energies for complexes 1 and 2 obtained from calculations at the B3LYP/LanL2DZ level

Table 5 .
1xperimental and simulated (TDDFT)1MLCT absorption spectrum data for complexes 1 and 2 a MLCT is Metal to Ligand Charge Transfer; LLCT is Ligand to Ligand Charge Transfer; MC is Metal-Centered.

Table 7 .
Spectroscopic (l max ; e max ) and quantum yield (f) data in various solvents