Electronic Spectra of Tris(2,2′-bipyridine) - ACS Publications

Article pubs.acs.org/IC

Electronic Spectra of Tris(2,2′-bipyridine)-M(II) Complex Ions in Vacuo (M = Fe and Os) Shuang Xu,† James E. T. Smith,‡ Samer Gozem,§ Anna I. Krylov,§ and J. Mathias Weber*,‡ †

JILA and Department of Physics and ‡JILA and Department of Chemistry and Biochemistry, University of Colorado, Boulder, Colorado 80309-0440, United States § Department of Chemistry, University of Southern California, Los Angeles, California 90089-0482, United States S Supporting Information *

ABSTRACT: We measured the electronic spectra of massselected [M(bpy)3]2+ (M = Fe and Os, bpy = 2,2′-bipyridine) ions in vacuo by photodissociation spectroscopy of their N2 adducts, [M(bpy)3]2+·N2. Extensive band systems in the visible (predominantly charge transfer) and near-ultraviolet (ππ*) spectral regions are reported. The [M(bpy)3]2+·N2 target ions were prepared by condensing N2 onto electrosprayed ions in a cryogenic ion trap at ca. 25 K and then mass-selected by timeof-flight mass spectrometry. The electronic photodissociation spectra of the cold, gas-phase ions closely reflect their intrinsic properties, i.e., without perturbation by solvent effects. The spectra are interpreted using time-dependent density functional theory calculations both with and without accounting for relativistic effects.

1. INTRODUCTION Tris(2,2′-bipyridine)-metal complex ions, [M(bpy)3]n+, constitute an important group of metal−organic complexes due to their photophysical properties.1−30 In particular, the dicationic complexes [M(bpy)3]2+ with d6 metals, i.e., M = Fe, Ru, or Os, are known for their high absorption cross sections of visible light.31,32 The transitions responsible for the absorption in this spectral region have predominantly metal-to-ligand charge transfer (MLCT) character.8,33 The electronic states involved in such transitions are of significant interest, since these charge transfer states have led to applications of [M(bpy)3]2+ as chemical photosensitizers and photocatalysts.34−37 In addition, [M(bpy)3]2+ complexes are prototypes from which many other metal-polypyridyl complexes for specific tasks in several areas of chemistry are derived. The electronic spectra of [M(bpy)3]2+ ions based on d6 metals have previously been measured in solution and in diluted crystals,1−3,5,7,8,13,16 and the properties of their excited states have been studied extensively in the condensed phase.6,15,18,20,25,27,38−40 However, interaction with surrounding solvent molecules or crystal lattices leads to solvatochromic shifts and changes in excited-state lifetimes, and therefore alters the absorption spectrum and photophysics of the solute, particularly for ionic species. As a result, spectra obtained in the condensed phase do not reflect the intrinsic photophysical properties of the ions under study. By measuring the electronic spectrum of mass-selected ions in vacuo, the electronic states of an ionic species can be probed in the absence of a chemical environment. © 2017 American Chemical Society

Typically, the number of mass-selected ions in a given experimental cycle is too small to cause detectable attenuation of light by the sample. Therefore, photodissociation spectroscopy is often used instead, monitoring the destruction of the target ions after photon absorption as a function of the irradiation wavelength. The only significant fragmentation channel in the photodissociation of [M(bpy)3]2+ ions is the loss of a bpy ligand.24,28,30,41 However, in contrast to many other [M(bpy)3]2+ complexes,30,42 density functional theory (DFT) predicts the dissociation energy threshold to be 333 kJ/ mol for M = Fe and 424 kJ/mol for M = Os (see Methods section). Consequently, single-photon absorption in the visible and near UV spectral range does not provide sufficient energy to overcome the dissociation threshold. To circumvent this problem, we use the messenger tagging method, which has been applied to similar complexes before.29,30,43−45 By forming weakly bound (ca. 5 kJ/mol) N2 adducts, [M(bpy)3]2+·N2, it is possible to perform photodissociation spectroscopy by monitoring the loss of N2 from the [M(bpy)3]2+·N2 ion. This method requires that the additional N2 molecule does not significantly change the properties of the [M(bpy)3]2+ ion. While this may not always be the case, the innocence of an N2 tag molecule has been shown from its negligible influence on the MLCT bands for M = Ru and on the ππ* UV bands for M = Fe.29,30 In this work we present the electronic spectra of two cryogenically prepared complex ions in vacuo, [Fe(bpy)3]2+ and Received: March 8, 2017 Published: June 6, 2017 7029

DOI: 10.1021/acs.inorgchem.7b00620 Inorg. Chem. 2017, 56, 7029−7037

Article

Inorganic Chemistry [Os(bpy)3]2+, by applying photodissociation spectroscopy to their N2 adducts, [M(bpy)3]2+·N2.

The ion packets were produced at 10 Hz repetition rate, while the laser pulses interrogated the target ions at 5 Hz, allowing alternating measurement of the fragment signal with and without laser irradiation to subtract background signals resulting from the decay of the target ions due to collisions with background gas. The laser intensity was measured by a pyroelectric joulemeter. Background subtraction and photon flux normalization were performed to yield the final spectra shown in this work. A full spectrum is composed of several concatenated segments corresponding to different crystal tuning ranges. The spectra presented here are averages of 10−30 scans for each spectral range taken on different days to increase the signal-tonoise ratio and ensure reproducibility. Computational. We use several different approaches to compute the excitation energies and oscillator strengths for the [M(bpy)3]2+ complexes. All approaches employ density functional theory46 (DFT) with the B3LYP functional.47,48 In one case we use a purely nonrelativistic, all-electron basis set (DZP) for both M = Fe and M = Os systems (abbreviated B3LYP/DZP). In a second approach (denoted B3LYP/DZP+SOC), we account for spin−orbit coupling (SOC) effects perturbatively with the one-electron Breit−Pauli (BP) Hamiltonian:49−51

2. METHODS Sample Preparation. Solutions of [Fe(bpy)3]2+ were prepared by dissolving FeCl2 (Alfa Aesar, 99%) and a 3-fold excess of 2,2′bipyridine (Sigma-Aldrich, ≥99%) in water with the concentration of Fe2+ at 5 mM. A red solution formed, which was then diluted with methanol (HPLC grade, EMD Millipore) to about 0.5 mM [Fe(bpy)3]2+, and used for electrospray without further purification. [Os(bpy)3]2+ was synthesized following literature procedures,6 modified to prepare solutions suitable for electrospray ionization. In a typical synthesis, 0.5 g of K2OsCl6 (Alfa Aesar, Os 38.7%) and 0.64 g of 2,2′-bipyridine were mixed in 25 mL of glycerol (Alfa Aesar, 99%+) and refluxed at 240 °C for 1 h. The resulting green-colored mixture was cooled, diluted with water,and then mixed with an excess of NH4PF6 (Alfa Aesar, 99.5%) until the supernatant was colorless. The resulting mixture was filtered to collect the dark green precipitate, which mainly consisted of Os(bpy)3(PF6)2. The precipitate was then washed with water, but even after washing contained residual glycerol, which can impair the performance of electrospray ionization. To remove the residual glycerol, the solid was dissolved in acetonitrile, and the resulting green solution was concentrated by slow evaporation in a fume hood until most of the solvent (acetonitrile) had evaporated, leaving a large fraction of the glycerol on the inner wall of the vial. The small amount of solution at the bottom of the vial was extracted using a pipet and diluted again with acetonitrile. This procedure was repeated three times before the solution was finally diluted with acetonitrile for electrospray without further purification. CAUTION: Many osmium compounds are toxic. K2OsCl6 is harmful with acute toxicity (oral, dermal, inhalation) and causes serious eye irritation or damage (GHS07, H302, H312, H332, H315, H319, H335). Appropriate precautions (dust mask, eye shields, gloves) should be taken to avoid inhalation and skin contact while handling. Photodissociation Spectroscopy. Our experimental setup has been described in detail in previous publications,29,42 and an overview is presented in the Supporting Information. In brief, gas-phase [M(bpy)3]2+ ions (M = Fe and Os) were produced by electrospraying solutions containing the corresponding ions (see above). Upon electrospraying, ions were accumulated to form ion packets in an octupole ion guide/trap combination where pulsed, biased end-caps are used to accumulate and release ions. The ion packets were then guided by a sequence of octupole ion guides through several differential pumping stages before entering a 3D quadrupole ion trap. The quadrupole ion trap is mounted on a closed-cycle helium cryostat, which was held at 25 K for the present experiments. N2 was used as the buffer gas and injected into the trap through a pulsed valve to form [M(bpy)3]2+·N2 adducts. Ions were ejected from the trap 95 ms after injection and then mass-separated in a Wiley−McLaren timeof-flight mass spectrometer. The [M(bpy)3]2+·N2 target ions were subsequently irradiated by the pulsed output of a tunable optical parametric converter system (220−710 nm). Upon photon absorption by an ion in vacuo, the electronic excitation energy is converted into vibrational energy, generating a hot ion. Consequently, the weakly bound (ca. 5 kJ/mol) N2 tag will be lost from the complex by unimolecular dissociation. Photodissociation spectra were measured by monitoring the intensity of [M(bpy)3]2+ daughter ions coming from [M(bpy)3]2+·N2 parent ions via the reaction

[M(bpy)3 ]2 + ·N2 + ℏω → [M(bpy)3 ]2 + + N2

SO Ĥ = −

a02 2

∑ i,A

ZA (riA × pi)·Si riA3

(2)

where i denotes the electrons, A denotes the nuclei, ZA is the charge of the nuclei, and a0 is the fine-structure constant. While the two-electron BP Hamiltonian term neglected here can be significant in molecules with only light atoms,52,53 the one-electron term is expected to dominate due to the presence of heavier atoms (Fe and Os) in the [M(bpy)3]2+ systems studied here, especially for M = Os. This results from the ZA term, which is not only explicitly included in the oneelectron term but also influences the riA term (larger nuclei result in smaller nuclear-electron distances in core electrons). The SOC can be expressed as a sum of its x, y, and z components:

⟨H SO⟩ = ⟨HxSO⟩ + ⟨H ySO⟩ + ⟨HzSO⟩

(3)

where SO

⟨HαSO⟩ = ⟨Ψ(s , ms)|Ĥ α |Ψ′(s′, ms′)⟩, α = x , y , z

(4)

The perturbative approach used here is expected to yield quantitatively accurate SOCs (depending on the accuracy of the electronic structure approach employed) for molecules containing atoms up to the fourth row elements of the Periodic Table (i.e., for Fe), but has not been tested for heavier elements such as Os.54 Final energies accounting for SOC were obtained by rediagonalizing the Hamiltonian matrix in the basis of all spin-components of the ground and excited singlet and triplet states after including the singlet−triplet and triplet−triplet off-diagonal SOC terms. Diagonalization of this Hamiltonian leads to perturbed energies and wave functions that have mixed singlet−triplet character and therefore shared oscillator strengths, resulting in a different computed spectral profile. The third approach (B3LYP/def2-TZVP and B3LYP/def2TZVPD) uses the def2-TZVP55 and def2-TZVPD56,57 basis sets for all atoms, respectively. These basis sets do not include an effective core potential (ECP) for the Fe atom, but they do for Os. We note that since ECPs account for some relativistic effects, including SOC, for the core electrons, calculations using ECP should not be combined with the explicit calculation of SOC as done above for the all-electron basis set. Finally, to make the calculations comparable between Fe and Os, and to explore basis set effects, additional calculations (B3LYP/ECP10-MDF) were performed with the ECP-10-MDF basis set for the Fe atom, which includes an ECP for Fe. All calculations in this work employ the hybrid B3LYP functional. Our choice stems from studies indicating that B3LYP provides accurate geometries58 and reasonably good energetics for Fe(II) complexes.59,60 The B3LYP functional with spin−orbit interaction was also used by Heully et al. for the [Ru(bpy)3]2+ complex21 and yielded good agreement with the experimental spectra. Although it might be

(1)

as a function of the photon energy. Parent and daughter ions were separated in a second mass spectrometry step using a reflectron, which the ions enter ca. 15 μs after irradiation, leaving sufficient time for relaxation of electronic excitation energy into vibrational energy and thereby causing N2 loss.41 As discussed in detail in a previous publication,29 the energetics of the evaporation process allow the photodissociation spectrum to be viewed as identical to the intrinsic photoabsorption spectrum of the [M(bpy)3]2+ chromophore ions. 7030

DOI: 10.1021/acs.inorgchem.7b00620 Inorg. Chem. 2017, 56, 7029−7037

Article

Inorganic Chemistry expected that long-range corrected functionals would be better suited for describing charge transfer (e.g., MLCT) states,61 a recent benchmark study on a number of Pt- and Ir-based complexes found that B3LYP yielded more accurate energetics and that long-range corrected functionals did not yield a clear advantage.62 They attributed that to the fact that MLCT states involve a transition out of strongly hybridized metal d-orbitals that already had some partial charge transfer character to the ligands. This may explain why B3LYP does not fail for those types of MLCT states. We note that exploratory calculations with range-corrected functionals (CAM-B3LYP and LRCwPBEPBE, both using DZP basis sets) for the systems studied here yielded no better (for M = Os) or significantly poorer agreement (for M = Fe) with the experiment (see Supporting Information). In all calculations, we assume singlet ground states of D3 symmetry in accord with the literature.9,14 Charge distributions of the groundstate structures were calculated with B3LYP/def2-TZVP, using natural population analysis.63 Excitation energies were calculated by timedependent DFT (TDDFT64−66). Geometries were obtained with B3LYP/DZP (for both calculations with and without SOC), and with B3LYP/def2-TZVP (for B3LYP/def2-TZVP, B3LYP/def2-TZVPD, and B3LYP/ECP-10-MDF). B3LYP/def2-TZVPD and B3LYP/DZP calculations were performed in Q-Chem. 67 The B3LYP/DZP calculations with SOC were carried out using the SOC implementation for TDDFT by Subotnik and co-workers,68 while B3LYP/def2TZVP and B3LYP/ECP-10-MDF calculations were performed using the Turbomole program suite.69

3. RESULTS AND DISCUSSION Figure 1 shows the photodissociation spectra of [M(bpy)3]2+· N2 (M = Fe, Ru,29 Os). All spectra exhibit an extensive region of charge transfer transitions (16200−33000 cm−1) in the visible and near UV, and bands predominantly of ππ* character in the deeper UV (33000−38400 cm−1). The overall sequence of the band positions of the three species in spectra taken in solution5,8,15,27,39 and in the gas phase is very similar. The individual electronic bands in all spectra are very broad, consistent with ultrafast intersystem crossing, which has been found in all three species.15,27,39 The bands observed in vacuo shift to lower wavenumbers upon solvation, consistent with the expectation that the stabilization of the excited states by solvation is more pronounced than for the ground state, particularly for the charge transfer bands. Since the electronic spectrum for M = Ru has been reported and analyzed in detail in a previous publication,29 we will focus the discussion on M = Fe and Os and include a short discussion for M = Ru only for completeness. M = Ru.29 The spectrum has an onset at 18 700 cm−1. The main absorption in the visible region appears as two resolved peaks at 23 160 and 24 660 cm−1. The UV band of [Ru(bpy)3]2+ is found to be a single broad feature at 36 100 cm−1. The bands observed in the experiment can be assigned on the basis of TDDFT calculations. However, while nonrelativistic TDDFT calculations are able to interpret the most prominent features, SOC needs to be taken into account to correctly describe the onset of the spectrum.21,29 We note that the weak interaction of N2 with [Ru(bpy)3]2+ (ca. 5 kJ/ mol) results in very small spectroscopic shifts (ca. 100 cm−1) of the electronic bands, as shown both experimentally and computationally.29 The similarity of the species probed here, and the similar binding energies of N2 adducts imply that the spectra in the present work reflect their intrinsic photophysics. M = Fe. The photodissociation spectrum of [Fe(bpy)3]2+ (see Figures 1 and 2) is very similar to that of [Ru(bpy)3]2+, but the features appear at lower energies for M = Fe. The onset of the spectrum is at 19 300 cm−1 (feature I) and rises to an

Figure 1. Photodissociation spectra of [M(bpy)3]2+·N2 (M = Fe, Ru, Os) at 25 K trap temperature. The points are raw data, and the full lines are 20-point (40 for M = Os) sliding averages to guide the eye. Data for M = Ru were taken from ref 29 with permission of AIP Publishing Copyright 2016.

intense peak at 19 970 cm−1 (feature II). Another strong absorption with slightly lower intensity is found at 21 570 cm−1 (feature III). There are extensive features (IV−VI) in the blue and near-UV region. The most prominent band is in the UV (34 360 cm−1) and shows some substructure (features VII and VIII), reminiscent of the UV spectra of other [M(bpy)3]2+ ions.29,30 The upper panel of Figure 2 compares the spectrum of [Fe(bpy)3]2+ in aqueous solution (room temperature, adapted from ref 27) and the photodissociation spectrum (trap temperature at 25 K) of [Fe(bpy)3]2+·N2. The overall envelopes of the spectra are quite similar; generally the features in the spectrum measured in vacuo are better resolved than in the room temperature solution spectrum, but despite the low temperature in the ion trap, the individual electronic bands are still broad. This suggests that the residual width in the spectra reported here is due to lifetime broadening, consistent with ultrafast experiments that show excited-state lifetimes of less than 50 fs.27 Our TDDFT calculations can be used to qualitatively assign the experimental data. The lowest-energy transition (1E) 7031

DOI: 10.1021/acs.inorgchem.7b00620 Inorg. Chem. 2017, 56, 7029−7037

Article

Inorganic Chemistry

experimental feature. It follows that features III and IV can be assigned to the 4E and 5E transitions. In the UV region, the dominant peak (feature VII) can be assigned unambiguously to the 8A2 transition, since this is the only prominent transition predicted in this region. Features VII and VIII are due to a single electronic band of ππ* character, where feature VII contains the band origin, while feature VIII is part of the Franck−Condon envelope, as seen in the spectra of several other [M(bpy)3]2+ ions.29,30 Bands VII and VIII are somewhat better resolved than in our previous study on bare [Fe(bpy)3]2+,30 but the peak of the band origin of A2 transition at 34 360 cm−1 (feature VII) is at the same position in both spectra. The assignment of all the features observed in the experiment is summarized in Table 1. The shift of the TDDFT energies and the varying quality of the agreement of the calculated and experimental pattern across the spectrum highlight the fact that the calculations only qualitatively describe the experiment. Including diffuse functions into the basis set (def2-TZVPD) improves the agreement with the experimental pattern, but the transition energies still need to be shifted significantly. To analyze the character of the transitions, we examined the molecular orbitals in the output of the calculations (see Supporting Information for details). Since the def2-TZVP basis sets use effective core potentials for larger metal atoms such as Ru and Os but not to Fe, we performed calculations with a def2-TZVP basis for all atoms, and another set of calculations with the ECP-10-MDF basis set for Fe, keeping def2-TZVP basis sets for all other atoms. Changing the basis set used for the central metal atom results in only small variations in the character of the orbitals involved in the calculated transitions. Their designations with respect to the highest occupied and lowest unoccupied molecular orbitals (HOMO and LUMO) are consistent, and the overall spectral patterns are very similar for the two sets of calculated spectra. The optimal shift for comparison of the ECP-10-MDF calculations with the experiment is −4520 cm−1, but even with this shift, the agreement in the UV region is not as good as for the def2TZVP or the def2-TZVPD calculations. The bottom panel of Figure 2 shows the computed electronic spectra for [Fe(bpy)3]2+ for B3LYP/DZP both with and without accounting for SOC. The inclusion of SOC produces a small splitting in the energies of some excited states (as expected, see also refs 21 and 29) and also results in a small shift in the spectrum due to SOC between the singlet ground state with at least one low-lying triplet state. However, beyond such small effects, the influence of SOC on the spectrum of [Fe(bpy)3]2+ is negligible. This is expected for Fe, which is a fourth row element with a relatively small atomic number. M = Os. The signal-to-noise ratio for M = Os is generally lower than for the other spectra in this series, because the best achievable parent ion intensity was much lower (only about 25% of that for M = Fe). The spectrum of [Os(bpy)3]2+ can be divided into four groups of bands with increasing energies and intensities (see Figures 1 and 3). We observe the lowest-energy absorption peak at 16 140 cm−1 (feature I in Figure 3), followed by a broader feature with a similar peak intensity at 17 740 cm−1. The second, more intense group of bands centers around 23 000 cm−1 and consists of features III−V. The third group has no resolved substructure and peaks at ca. 28 000 cm−1. The highest energy group of bands is in the UV region and has some more resolved features (VI, VII) around 34 000 cm−1. A solution spectrum of [Os(bpy)3]2+ measured by Shaw et al.39 in

Figure 2. Upper panel: Photodissociation (in vacuo) and absorption (aqueous solution) spectrum of [Fe(bpy)3]2+·N2 at 25 K trap temperature. Black points are raw data; the red curve is a 20-point sliding average to guide the eye. The gray curve is the absorption spectrum of [Fe(bpy)3]2+ in aqueous solution (room temperature, adapted from ref 27). Lower panels: Calculated electronic spectrum of [Fe(bpy)3]2+ using the B3LYP functional. The basis sets are noted in the panels. In the def2-TZVPD and def2-TZVP panels, blue and red bars represent the transitions of E and A2 symmetry, respectively. The bottom panel shows calculations using DZP basis sets for all atoms with (black) and without (orange) SOC. The def2-TZVPD calculation is offset by −3000 cm−1 in this figure, the def2-TZVP calculation by −3390 cm−1, and the DZP calculations by −3600 cm−1 to allow better comparison with the experimental data. All oscillator strengths below 24 × 103 cm−1 (i.e., to the left of the dashed line) have been multiplied by 5 for clarity.

predicted by theory (def2-TZVP) is at 13 930 cm−1, which is below the range of the spectra reported here, while the next transition (2E) is more than 8000 cm−1 higher in energy. Given the good overall match of the calculated and experimental spectral envelopes, the onset of the experimental spectrum shown here is unlikely to be due to the 1E transition, which has a small computed oscillator strength. In order to better compare the computational and experimental spectral patterns, we shift the calculated spectra in Figure 2 to match the 2E transition to the onset and the 3E transition to the first intense 7032

DOI: 10.1021/acs.inorgchem.7b00620 Inorg. Chem. 2017, 56, 7029−7037

Article

Inorganic Chemistry Table 1. Calculated Transition Energies E, Oscillator Strengths f, and Transition Types for Selected Excited States of [Fe(bpy)3]2+ Compared to Experimentally Observed Features B3LYP/def2-TZVP state 2E 3E 4E 5E 4A2 6E 11E 14E 8A2 a

type MLCT/dd MLCT MLCT dd/MLCT MLCT MLCT MLCT ππ*/MLCT ππ*/MLCT

−1

E [cm ] 22153 23437 24634 25774 29343 30044 34063 35846 36749

−1

f 1.6 3.6 7.5 1.9 2.5 1.5 8.2 3.6 6.4

× × × × × × × × ×

B3LYP/DZP + SOCa

B3LYP/ECP-10-MDF E [cm ] −3

10 10−2 10−2 10−2 10−3 10−2 10−2 10−1 10−1

E [cm−1]

f

23353 24548 25792 27539 30476 31143 35053 36181 37045

2.3 4.0 7.3 1.1 9.7 1.6 1.9 2.8 5.8

× × × × × × × × ×

−3

10 10−2 10−2 10−2 10−4 10−2 10−1 10−1 10−1

21834 23592 24808 27357 29257 30346 33162 37089 38057

experimental

f 1.5 1.5 4.9 6.9 7.0 7.1 1.0 2.4 7.0

× × × × × × × × ×

−4

10 10−2 10−2 10−3 10−4 10−3 10−2 10−1 10−1

E [cm−1]

featureb

19300 19970 21570 23230 26480

I II III IV V

29310 34360

VI VII

State label and type of transition assigned on the basis of calculations without SOC. bSee Figure 2.

Figure 3. Photodissociation spectrum of [Os(bpy)3]2+·N2 at 25 K trap temperature throughout the full spectral range of the present experiment. Black points are raw data; the green curve is a 40-point sliding average to guide the eye. Filled blue and red bars represent calculated transitions of E and A2 symmetry, respectively, as calculated with the B3LYP/def2-TZVP approach. The analogous states calculated using the def2-TZVPD basis are shown as empty columns. In contrast to M = Fe, the computed spectra are shown without shifts.

acetonitrile solution in the visible range shows considerably less resolved structure than the spectrum in the present work (see Figure 4). Our TDDFT B3LYP/def2-TZVP calculations capture the overall shape of the spectrum, with the exception of the onset. Owing to its high nuclear charge, relativistic effects are significant in Os, and SOC should therefore be taken into account in [Os(bpy)3]2+. In particular, we expect significant mixing of singlet and triplet transitions leading to a redistribution of oscillator strength and, consequently, noticeable changes in the spectral shape.5,8,39 Similar to M = Ru, the onset of the spectrum is not well described by calculations without SOC. We note that strong SOC can also provide fast relaxation channels for the excited states, resulting in short intersystem crossing lifetimes and broader spectral features, consistent with the experiment. Unlike M = Fe, addition of diffuse functions in the basis set does not improve the overall agreement between theory and experiment (see Figures 3 and 4), as the low- and mid-energy range calculated states are at very similar energy positions, but the calculations with diffuse functions overestimate the energies in the higher energy range.

Figure 4. Upper panel: MLCT region of the spectrum shown in Figure 3. The gray curve is the absorption spectrum of [Os(bpy)3]2+ in acetonitrile solution (room temperature, adapted from ref 39). The second panel shows our B3LYP/def2-TZVP (black columns) and B3LYP/def2-TZVPD (magenta columns) TDDFT calculations. The third panel shows the work of Ronca et al.26 taking SOC into account (note that Ronca et al. only give the main transitions up to 23 000 cm−1 in ref 26). The bottom panel shows our B3LYP/DZP calculations with (black) and without (orange) SOC, shifted by −800 cm−1 to improve pattern comparison with the other data shown here. For clarity, oscillator strengths for transitions below 19 × 103 cm−1 (i.e., to the left of the dashed line) have been multiplied by the factors indicated.

7033

DOI: 10.1021/acs.inorgchem.7b00620 Inorg. Chem. 2017, 56, 7029−7037

Article

Inorganic Chemistry

Table 2. Transition Energies E and Oscillator Strengths f for Selected Calculated Excited States of [Os(bpy)3]2+ (TDDFT26 with and without Spin-Orbit Interaction) Compared to Experimentally Observed Features computational statea Ronca et al.

a

this work - B3LYP/def2-TZVP

ST2,3: 1E 3E ST22,23: 8E ST25,26: 9E ST34,35: 12E 1A2 2E 3E 3A2 8E 12E 7A2

experimental (this work) E [cm−1]

f

16091 16962 20527 20785 22672 18015 20268 22094 24610 29838 36110 36709

1.5 2.1 6.0 5.8 1.7 2.7 4.2 2.1 9.3 2.3 3.8 7.3

× × × × × × × × × × × ×

E [cm−1]

type −2

10 10−2 10−2 10−2 10−1 10−3 10−2 10−1 10−3 10−1 10−1 10−1

MLCT MLCT MLCT MLCT MLCT ππ* ππ*

featureb

16140 17740 20700

I II III

22310 17740 20700 22310 24160 27890 36530

IV II III IV V VI VIII

a

The state designation follows that given by Ronca et al. in ref.26 The authors do not list higher energy states than those shown here. bSee Figures 3 and 4.

Ronca et al.26 incorporated relativistic effects in their calculations of [Os(bpy)3]2+ by employing a two-component zero-order regular approximation (ZORA) approach that includes both scalar relativistic (SR) effects and SOC effects within the TDDFT framework.70−72 They find that SR effects result in a red shift of the peaks corresponding to features III and IV, compared to calculations employing only SOC. Since we do not include SR effects in any of our calculations, this may explain some of the discrepancy between our computed spectra and those of Ronca et al., which are closer to the experiment. By introducing SOC effects, Ronca et al. show that the transitions in the singlet manifold split, similar to M = Ru.21,29 In addition, they found transitions containing significant triplet character at lower energies, enabled by SOC, which are in remarkably good agreement with our experimental results (see Figure 4). Consequently, we are able to assign many of the features observed in our experiment as shown in Table 2. Analysis of our calculations suggests that the lower energy features (I−V) are predominantly of MLCT character, while the highest energy features are of ππ* character. Our B3LYP/DZP calculations with the perturbative SOC correction reproduce some of the results observed by Ronca et al. In particular, both our calculations and those by Ronca et al. show that SOC leads to (1) a blue-shift of peaks of features III and IV relative to calculations without SOC, (2) splitting of the peak in feature III, and (3) some loss in intensity of singlet transitions due to intensity borrowing by triplet transitions. However, our B3LYP/DZP+SOC calculations do not reproduce the peaks of feature I. This is most likely due to limitations of the perturbative approach used here, which can lead to larger errors for heavy metals (Os is a sixth row transition metal). Solvatochromic Behavior. The overall similarity of the spectra in solution and in vacuo for all complexes studied here allows extraction of solvatochromic shifts for many of the electronic bands. All observed electronic bands shift to lower wavenumbers in solutions and crystals. Solvatochromic shifts for M = Fe in aqueous solution are in the range of −580 cm−1 to −1200 cm−1 (see Table 3). For M = Ru, the shifts range from −1060 cm−1 to −1460 cm−1.29 For M = Os, the largely unresolved features in acetonitrile solution reported by Shaw et al.39 only allow determination of solvatochromic shifts for the bands in the group with features III−V in our spectrum, which shift by about 930 cm−1 to the

Table 3. Experimental and Calculated (B3LYP/DZP with Polarizable Continuum Solvent Model) Solvatochromic Shifts in cm−1 experimental metal

feature

Fea

II III V VI VII IV V IX III IV V

Rub

Osc

d

water −1170 −1200 −580 −740 −860 −1060 −1460 −1140

calculated

acetonitrile

water

acetonitrile

−1040 −1075 −1080 −930 −930 −930

−536 −660 −441 −599 −1440 −460 −637 −1501 −496 −803 −647

−524 −647 −436 −586 −1435 −452 −629 −1498 −485 −795 −637

a

Comparison with room temperature aqueous solution data in ref 27. Nomenclature taken from ref 29. Features IV and V are the two most intense features in the MLCT region, and feature IX is the peak in the ππ* region. Room temperature aqueous solution data were also taken from ref 29; room temperature acetonitrile solution data are from ref 24. cComparison with room temperature acetonitrile solution data in ref 26. dSpectroscopic features not listed have no sufficiently well resolved features in solution. b

red in acetonitrile solution. Shifts in crystal matrices5 are also generally to the red and of similar magnitude as for solutions. Note that solvatochromic shifts are not the same for each transition, as they depend on the changes in charge distribution in the solute upon excitation. For MLCT transitions, the red shifts can be qualitatively understood with the charge transfer character of the bands. Solvation stabilizes both the ground state and the excited state, but the stabilization for the MLCT excited state will be larger in magnitude, resulting in a red shift. It is interesting that this effect is also observed with similar magnitude in the intense UV bands, although they have mainly ππ* character, with only small amounts of charge transfer mixed in (see Table 1 and Supporting Information). Calculations with a polarizable continuum solvent model (PCM) qualitatively capture the solvatochromic red shift, but they cannot provide good quantitative shifts. At lower photon 7034

DOI: 10.1021/acs.inorgchem.7b00620 Inorg. Chem. 2017, 56, 7029−7037

Article

Inorganic Chemistry energies, solvatochromic shifts are underestimated by the PCM calculations, while they are overestimated at higher energies. The deviations suggest that purely dielectric solvent effects only provide a partial description of solvatochromic effects and that explicit water molecules need to be taken into account to better recover solvatochromic shifts computationally. The magnitude of observed deviations is in line with a recent benchmark study of the accuracy of PCM models for excitation energies.73 A more accurate description of solvatochromic shifts requires using explicit polarizable solvent models, such as the effective fragment potential method.74−76 The solvent calculations and their comparison with experimental data are summarized in Table 3.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

J. Mathias Weber: 0000-0002-5493-5886 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We gratefully acknowledge the National Science Foundation for support of this work under Grant CHE-1361814 (J.M.W.). A.I.K. acknowledges support by the United States Department of Energy, Basic Energy Sciences through Grant DE-FG02101205ER15685.

4. SUMMARY AND CONCLUSIONS We measured and characterized the spectra of cryogenically prepared tris(2,2′-bipyridine)-metal complex ions in vacuo via messenger spectroscopy of their nitrogen adducts, [M(bpy)3]2+·N2 (M = Fe, Os). The absence of perturbations by solvent molecules means that the spectra represent the intrinsic electronic spectra of these species, unperturbed by effects arising from interactions with a chemical environment. This allows direct comparison of the experimental spectra to computed gas-phase spectra. For [Fe(bpy)3]2+, singlet transitions calculated using TDDFT are sufficient to describe the spectral features seen in the experiment. The incorporation of SOC effects in [Fe(bpy)3]2+ has a limited effect on the computed energies and oscillator strengths. For [Os(bpy)3]2+, while nonrelativistic TDDFT calculations can satisfactorily reproduce transitions in the high-energy range, SR effects and SOC are important for reproducing the onset (low-energy region) of the spectrum, as demonstrated by Ronca et al.26 using ZORA TDDFT calculations. We find that accounting for SOC effects using a perturbative approach reproduces some features of the ZORA calculations, such as peak shifts, splitting, and intensity borrowing from some singlet to triplet transitions, but does not recover the intensity borrowing of low-lying triplet transitions near the onset that result in absorption at the red edge of the visible spectrum. Therefore, a quantitative agreement of calculations with experiment in the case of [Os(bpy)3]2+ requires accounting rigorously for both SR and SOC effects. Finally, ultrafast intersystem crossing is present for both M = Fe and Os species, and their spectra remain broad at 25 K trap temperature. Comparison with UV/vis absorption spectra from the literature shows that all observed bands shift to lower wavenumbers upon solvation, and polarizable continuum solvent models qualitatively capture this behavior.

■

(bpy)3]2+ and [Os(bpy)3]2+ with and without long-range corrected functionals (PDF)

■

REFERENCES

(1) Palmer, R. A.; Piper, T. S. 2,2′-Bipyridine Complexes. I. Polarized Crystal Spectra of Tris(2,2′-Bipyridine)Copper(II) -Nickel -Cobalt(II) -Iron(II) and -Ruthenium(II). Inorg. Chem. 1966, 5, 864−878. (2) Lytle, F. E.; Hercules, D. M. The Luminescence of Tri(2,2′Bipyridine) Ruthenium (II) Dichloride. J. Am. Chem. Soc. 1969, 91, 253−257. (3) McWhinnie, W. R.; Miller, J. D. The Chemistry of Complexes Containing 2,2′-Bipyridyl, 1,10-Phenanthroline, or 2,2′,6′,2″-Terpyridyl as Ligands. Adv. Inorg. Chem. Radiochem. 1970, 12, 135−215. (4) Adamson, A. W.; Demas, J. N. A New Photosensitizer. Tris(2,2′Bipyridine)Ruthenium(II) Chloride. J. Am. Chem. Soc. 1971, 93, 1800−1801. (5) Felix, F.; Ferguson, J.; Güdel, H. U.; Ludi, A. Electronic-Spectra of M(Bipy)32+ Complex-Ions (M = Fe, Ru and Os). Chem. Phys. Lett. 1979, 62, 153−157. (6) Creutz, C.; Chou, M.; Netzel, T. L.; Okumura, M.; Sutin, N. Lifetimes, Spectra, and Quenching of the Excited-States of Polypyridine Complexes of Iron(II), Ruthenium(II), and Osmium(II). J. Am. Chem. Soc. 1980, 102, 1309−1319. (7) Decurtins, S.; Felix, F.; Ferguson, J.; Gudel, H. U.; Ludi, A. The Electronic-Spectrum of Fe(bpy)32+ and Os(bpy)32+. J. Am. Chem. Soc. 1980, 102, 4102−4106. (8) Kober, E. M.; Meyer, T. J. Concerning the Absorption-Spectra of the Ions M(bpy)32+ (M = Fe, Ru, Os, bpy = 2,2′-Bipyridine). Inorg. Chem. 1982, 21, 3967−3977. (9) Ferguson, J.; Herren, F. A Model for the Interpretation of the Electronic Spectra of the Complex Ions M(bpy)32+ (M = Fe, Ru, Os) in D3 and C2 Sites. Chem. Phys. 1983, 76, 45−59. (10) Ferguson, J.; Herren, F. The Electronic Structure of the Metalto-Ligand Charge-Transfer States of M(bpy)32+ (M= Fe, Ru, Os). Chem. Phys. Lett. 1982, 89, 371−375. (11) Ferguson, J.; Herren, F.; Krausz, E. R.; Maeder, M.; Vrbancich, J. Electronic Spectroscopy of Fe(bpy)32+, Ru(bpy)32+, Os(bpy)32+, Cr(bpy)32+ and Related Compunds. Coord. Chem. Rev. 1985, 64, 21−39. (12) Ohsawa, Y.; Dearmond, M. K.; Hanck, K. W.; Moreland, C. G. A 1h Nmr Study of Reduced Paramagnetic Tris(2,2′-Bipyridine) Complexes of Iron(II), Ruthenium (II), and Osmium(II). J. Am. Chem. Soc. 1985, 107, 5383−5386. (13) Constable, E. C. Homoleptic Complexes of 2,2′-Bipyridine. Adv. Inorg. Chem. 1989, 34, 1−63. (14) Daul, C.; Baerends, E. J.; Vernooijs, P. A Density Functional Study of the MLCT States of [Ru(bpy)3]2+ in D3 Symmetry. Inorg. Chem. 1994, 33, 3538−3543. (15) Damrauer, N. H.; Cerullo, G.; Yeh, A.; Boussie, T. R.; Shank, C. V.; McCusker, J. K. Femtosecond Dynamics of Excited-State Evolution in Ru(bpy)32+. Science 1997, 275, 54−57.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.inorgchem.7b00620. Schematic overview of the experimental apparatus; calculated electronic transitions of [Fe(bpy)3]2+ in the spectral region of the experiment; calculated orbitals involved in the transitions of [Fe(bpy)3]2+; calculated electronic transitions of [Os(bpy)3]2+ in the spectral region of the experiment; calculated orbitals involved in the transitions of [Os(bpy)3]2+; calculated electronic spectrum of [Fe(bpy)3]2+ and [Fe(bpy)3]2+·N2; comparison of experimental and calculated spectra of [Fe7035

DOI: 10.1021/acs.inorgchem.7b00620 Inorg. Chem. 2017, 56, 7029−7037

Article

Inorganic Chemistry

Making Oxygen with Ruthenium Complexes. Acc. Chem. Res. 2009, 42, 1954−1965. (37) Blakemore, J. D.; Crabtree, R. H.; Brudvig, G. W. Molecular Catalysts for Water Oxidation. Chem. Rev. 2015, 115, 12974−13005. (38) Cannizzo, A.; van Mourik, F.; Gawelda, W.; Zgrablic, G.; Bressler, C.; Chergui, M. Broadband Femtosecond Fluorescence Spectroscopy of Ru(bpy)32+. Angew. Chem., Int. Ed. 2006, 45, 3174− 3176. (39) Shaw, G. B.; Styers-Barnett, D. J.; Gannon, E. Z.; Granger, J. C.; Papanikolas, J. M. Interligand Electron Transfer Dynamics in Os(bpy)32+: Exploring the Excited State Potential Surfaces with Femtosecond Spectroscopy. J. Phys. Chem. A 2004, 108, 4998−5006. (40) Yoon, S.; Kukura, P.; Stuart, C. M.; Mathies, R. A. Direct Observation of the Ultrafast Intersystem Crossing in Tris(2,2 ′-Bipyridine) Ruthenium(II) Using Femtosecond Stimulated Raman Spectroscopy. Mol. Phys. 2006, 104, 1275−1282. (41) Xu, S.; Weber, J. M. Absorption Spectrum of a Ru(II)-Aquo Complex in Vacuo: Resolving Individual Charge-Transfer Transitions. J. Phys. Chem. A 2015, 119, 11509−11513. (42) Xu, S.; Gozem, S.; Krylov, A. I.; Christopher, C. R.; Mathias Weber, J. Ligand Influence on the Electronic Spectra of Monocationic Copper-Bipyridine Complexes. Phys. Chem. Chem. Phys. 2015, 17, 31938−31946. (43) Garand, E.; Fournier, J. A.; Kamrath, M. Z.; Schley, N. D.; Crabtree, R. H.; Johnson, M. A. Characterization of an Activated Iridium Water Splitting Catalyst Using Infrared Photodissociation of H2 Tagged Ions. Phys. Chem. Chem. Phys. 2012, 14, 10109−10113. (44) Wolk, A. B.; Leavitt, C. M.; Garand, E.; Johnson, M. A. Cryogenic Ion Chemistry and Spectroscopy. Acc. Chem. Res. 2014, 47, 202−210. (45) Leavitt, C. M.; Wolk, A. B.; Fournier, J. A.; Kamrath, M. Z.; Garand, E.; Van Stipdonk, M. J.; Johnson, M. A. Isomer-Specific IR-IR Double. Resonance Spectroscopy of D2-Tagged Protonated Dipeptides Prepared in a Cryogenic Ion Trap. J. Phys. Chem. Lett. 2012, 3, 1099−1105. (46) Parr, R. G.; Yang, W. Density-Functional Theory of Atoms and Molecules; Oxford University Press: New York, 1989. (47) Lee, C. T.; Yang, W. T.; Parr, R. G. Development of the ColleSalvetti Correlation-Energy Formula into a Functional of the ElectronDensity. Phys. Rev. B: Condens. Matter Mater. Phys. 1988, 37, 785−789. (48) Becke, A. D. Density-Functional Exchange-Energy Approximation with Correct Asymptotic-Behavior. Phys. Rev. A: At., Mol., Opt. Phys. 1988, 38, 3098−3100. (49) Abegg, P. W. Abinitio Calculation of Spin-Orbit-Coupling Constants for Gaussian Lobe and Gaussian-Type Wave-Functions. Mol. Phys. 1975, 30, 579−596. (50) Bethe, H. A.; Salpeter, E. E. Quantum Mechanics of One and Two Electron Atoms; Plenum Publishing Corp.: New York, 1977. (51) Schwerdtfeger, P., Ed.; Relativistic Electronic Structure Theory; Elsevier: Amsterdam, 2002. (52) Marian, C. M. Spin-Orbit Coupling and Intersystem Crossing in Molecules. Wiley Interdisc. Rev. - Comput. Mol. Sci. 2012, 2, 187−203. (53) Ågren, H.; Vahtras, O.; Minaev, B. Response Theory and Calculations of Spin-Orbit Coupling Phenomena in Molecules. Adv. Quantum Chem. 1996, 27, 71−162. (54) Epifanovsky, E.; Klein, K.; Stopkowicz, S.; Gauss, J.; Krylov, A. I. Spin-Orbit Couplings within the Equation-of-Motion Coupled-Cluster Framework: Theory, Implementation, and Benchmark Calculations. J. Chem. Phys. 2015, 143, 064102. (55) Weigend, F.; Ahlrichs, R. Balanced Basis Sets of Split Valence, Triple Zeta Valence and Quadruple Zeta Valence Quality for H to Rn: Design and Assessment of Accuracy. Phys. Chem. Chem. Phys. 2005, 7, 3297−3305. (56) Rappoport, D.; Furche, F. Property-Optimized Gaussian Basis Sets for Molecular Response Calculations. J. Chem. Phys. 2010, 133, 134105. (57) Andrae, D.; Häussermann, U.; Dolg, M.; Stoll, H.; Preuss, H. Energy-Adjusted Ab Initio Pseudopotentials for the 2nd and 3rd Row Transition-Elements. Theor. Chimica Acta 1990, 77, 123−141.

(16) Krausz, E.; Riesen, H. Developments in Laser Selective Spectroscopy and Photophysics of d6 Metal-(Diimine) Complexes. Coord. Chem. Rev. 1997, 159, 9−40. (17) Spence, T. G.; Burns, T. D.; Guckenberger, G. B.; Posey, L. A. Wavelength-Dependent Photodissociation of [Fe(bpy) 3 · (CH3OH)N]2+ Clusters, N = 2−6, Triggered by Excitation of the Metal-to-Ligand Charge-Transfer Transition. J. Phys. Chem. A 1997, 101, 1081−1092. (18) Gawelda, W.; Cannizzo, A.; Pham, V. T.; van Mourik, F.; Bressler, C.; Chergui, M. Ultrafast Nonadiabatic Dynamics of FeII(bpy)32+ in Solution. J. Am. Chem. Soc. 2007, 129, 8199−8206. (19) Yeh, A. T.; Shank, C. V.; McCusker, J. K. Ultrafast Electron Localization Dynamics Following Photo-Induced Charge Transfer. Science 2000, 289, 935−938. (20) Bhasikuttan, A. C.; Suzuki, M.; Nakashima, S.; Okada, T. Ultrafast Fluorescence Detection in Tris(2,2 ′-Bipyridine)Ruthenium(II) Complex in Solution: Relaxation Dynamics Involving Higher Excited States. J. Am. Chem. Soc. 2002, 124, 8398−8405. (21) Heully, J.-L.; Alary, F.; Boggio-Pasqua, M. Spin-Orbit Effects on the Photophysical Properties of Ru(bpy)32+. J. Chem. Phys. 2009, 131, 184308. (22) Fan, J.; Autschbach, J.; Ziegler, T. Electronic Structure and Circular Dichroism of Tris(Bipyridyl) Metal Complexes within Density Functional Theory. Inorg. Chem. 2010, 49, 1355−1362. (23) Johansson, E. M. J.; Odelius, M.; Plogmaker, S.; Gorgoi, M.; Svensson, S.; Siegbahn, H.; Rensmo, H. Spin-Orbit Coupling and Metal-Ligand Interactions in Fe(II), Ru(II), and Os(II) Complexes. J. Phys. Chem. C 2010, 114, 10314−10322. (24) Kirketerp, M. B. S.; Nielsen, S. B. Absorption Spectrum of Isolated Tris(2,2′-Bipyridine)Ruthenium(II) Dications in Vacuo. Int. J. Mass Spectrom. 2010, 297, 63−66. (25) Sousa, C.; de Graaf, C.; Rudavskyi, A.; Broer, R.; Tatchen, J.; Etinski, M.; Marian, C. M. Ultrafast Deactivation Mechanism of the Excited Singlet in the Light-Induced Spin Crossover of Fe(2,2 ′-Bipyridine)32+. Chem. - Eur. J. 2013, 19, 17541−17551. (26) Ronca, E.; De Angelis, F.; Fantacci, S. Time-Dependent Density Functional Theory Modeling of Spin-Orbit Coupling in Ruthenium and Osmium Solar Cell Sensitizers. J. Phys. Chem. C 2014, 118, 17067−17078. (27) Aubö ck, G.; Chergui, M. Sub-50-Fs Photoinduced Spin Crossover in Fe(bpy)32+. Nat. Chem. 2015, 7, 629−633. (28) Stockett, M. H.; Nielsen, S. B. Does a Single Ch3cn Molecule Attached to Ru(Bipy)32+ Affect Its Absorption Spectrum? J. Chem. Phys. 2015, 142, 171102. (29) Xu, S.; Smith, J. E. T.; Weber, J. M. The Electronic Spectrum of Cryogenic Ruthenium-Tris-Bipyridine Dications in Vacuo. J. Chem. Phys. 2016, 145, 024304. (30) Xu, S.; Smith, J. E. T.; Weber, J. M. Uv Spectra of Tris(2,2′Bipyridine)-M(II) Complex Ions in Vacuo (M = Mn, Fe, Co, Ni, Cu, Zn). Inorg. Chem. 2016, 55, 11937−11943. (31) Kalyanasundaram, K. Photophysics, Photochemistry and Solar Energy Conversion with Tris(Bipyridyl)Ruthenium(II) and Its Analogues. Coord. Chem. Rev. 1982, 46, 159−244. (32) Hanazaki, I.; Nagakura, S. Electronic Structure of Tris(α,α′Dipyridyl)Iron(II) Ion. Inorg. Chem. 1969, 8, 648−654. (33) Kober, E. M.; Sullivan, B. P.; Meyer, T. J. Solvent Dependence of Metal-to-Ligand Charge-Transfer Transitions - Evidence for Initial Electron Localization in Mlct Excited States of 2,2′-Bipyridine Complexes of Ruthenium(II) and Osmium(II). Inorg. Chem. 1984, 23, 2098−2104. (34) Juris, A.; Balzani, V.; Barigelletti, F.; Campagna, S.; Belser, P.; von Zelewsky, A. Ru(II) Polypyridine Complexes - Photophysics, Photochemistry, Electrochemistry, and Chemi-Luminescence. Coord. Chem. Rev. 1988, 84, 85−277. (35) Yagi, M.; Kaneko, M. Molecular Catalysts for Water Oxidation. Chem. Rev. 2001, 101, 21−35. (36) Concepcion, J. J.; Jurss, J. W.; Brennaman, M. K.; Hoertz, P. G.; Patrocinio, A. O. T.; Iha, N. Y. M.; Templeton, J. L.; Meyer, T. J. 7036

DOI: 10.1021/acs.inorgchem.7b00620 Inorg. Chem. 2017, 56, 7029−7037

Article

Inorganic Chemistry (58) Bowman, D. N.; Jakubikova, E. Low-Spin Versus High-Spin Ground State in Pseudo-Octahedral Iron Complexes. Inorg. Chem. 2012, 51, 6011−6019. (59) Jakubikova, E.; Bowman, D. N. Fe(II)-Polypyridines as Chromophores in Dye-Sensitized Solar Cells: A Computational Perspective. Acc. Chem. Res. 2015, 48, 1441−1449. (60) Laurent, A. D.; Jacquemin, D. TD-DFT Benchmarks: A Review. Int. J. Quantum Chem. 2013, 113, 2019−2039. (61) Dreuw, A.; Weisman, J. L.; Head-Gordon, M. Long-Range Charge-Transfer Excited States in Time-Dependent Density Functional Theory Require Non-Local Exchange. J. Chem. Phys. 2003, 119, 2943−2946. (62) Niehaus, T. A.; Hofbeck, T.; Yersin, H. Charge-Transfer Excited States in Phosphorescent Organo-Transition Metal Compounds: A Difficult Case for Time Dependent Density Functional Theory? RSC Adv. 2015, 5, 63318−63329. (63) Reed, A. E.; Weinstock, R. B.; Weinhold, F. Natural-Population Analysis. J. Chem. Phys. 1985, 83, 735−746. (64) Furche, F. On the Density Matrix Based Approach to TimeDependent Density Functional Response Theory. J. Chem. Phys. 2001, 114, 5982−5992. (65) Bauernschmitt, R.; Ahlrichs, R. Treatment of Electronic Excitations within the Adiabatic Approximation of Time Dependent Density Functional Theory. Chem. Phys. Lett. 1996, 256, 454−464. (66) Bauernschmitt, R.; Häser, M.; Treutler, O.; Ahlrichs, R. Calculation of Excitation Energies within Time-Dependent Density Functional Theory Using Auxiliary Basis Set Expansions. Chem. Phys. Lett. 1997, 264, 573−578. (67) Shao, Y.; Gan, Z.; Epifanovsky, E.; Gilbert, A. T. B.; Wormit, M.; Kussmann, J.; Lange, A. W.; Behn, A.; Deng, J.; Feng, X.; Ghosh, D.; Goldey, M.; Horn, P. R.; Jacobson, L. D.; Kaliman, I.; Khaliullin, R. Z.; Kus, T.; Landau, A.; Liu, J.; Proynov, E. I.; Rhee, Y. M.; Richard, R. M.; Rohrdanz, M. A.; Steele, R. P.; Sundstrom, E. J.; Woodcock, H. L., III; Zimmerman, P. M.; Zuev, D.; Albrecht, B.; Alguire, E.; Austin, B.; Beran, G. J. O.; Bernard, Y. A.; Berquist, E.; Brandhorst, K.; Bravaya, K. B.; Brown, S. T.; Casanova, D.; Chang, C.-M.; Chen, Y.; Chien, S. H.; Closser, K. D.; Crittenden, D. L.; Diedenhofen, M.; DiStasio, R. A., Jr.; Do, H.; Dutoi, A. D.; Edgar, R. G.; Fatehi, S.; Fusti-Molnar, L.; Ghysels, A.; Golubeva-Zadorozhnaya, A.; Gomes, J.; Hanson-Heine, M. W. D.; Harbach, P. H. P.; Hauser, A. W.; Hohenstein, E. G.; Holden, Z. C.; Jagau, T.-C.; Ji, H.; Kaduk, B.; Khistyaev, K.; Kim, J.; Kim, J.; King, R. A.; Klunzinger, P.; Kosenkov, D.; Kowalczyk, T.; Krauter, C. M.; Lao, K. U.; Laurent, A. D.; Lawler, K. V.; Levchenko, S. V.; Lin, C. Y.; Liu, F.; Livshits, E.; Lochan, R. C.; Luenser, A.; Manohar, P.; Manzer, S. F.; Mao, S.-P.; Mardirossian, N.; Marenich, A. V.; Maurer, S. A.; Mayhall, N. J.; Neuscamman, E.; Oana, C. M.; Olivares-Amaya, R.; O’Neill, D. P.; Parkhill, J. A.; Perrine, T. M.; Peverati, R.; Prociuk, A.; Rehn, D. R.; Rosta, E.; Russ, N. J.; Sharada, S. M.; Sharma, S.; Small, D. W.; Sodt, A.; Stein, T.; Stueck, D.; Su, Y.-C.; Thom, A. J. W.; Tsuchimochi, T.; Vanovschi, V.; Vogt, L.; Vydrov, O.; Wang, T.; Watson, M. A.; Wenzel, J.; White, A.; Williams, C. F.; Yang, J.; Yeganeh, S.; Yost, S. R.; You, Z.-Q.; Zhang, I. Y.; Zhang, X.; Zhao, Y.; Brooks, B. R.; Chan, G. K. L.; Chipman, D. M.; Cramer, C. J.; Goddard, W. A., III; Gordon, M. S.; Hehre, W. J.; Klamt, A.; Schaefer, H. F., III; Schmidt, M. W.; Sherrill, C. D.; Truhlar, D. G.; Warshel, A.; Xu, X.; Aspuru-Guzik, A.; Baer, R.; Bell, A. T.; Besley, N. A.; Chai, J.D.; Dreuw, A.; Dunietz, B. D.; Furlani, T. R.; Gwaltney, S. R.; Hsu, C.P.; Jung, Y.; Kong, J.; Lambrecht, D. S.; Liang, W.; Ochsenfeld, C.; Rassolov, V. A.; Slipchenko, L. V.; Subotnik, J. E.; Van Voorhis, T.; Herbert, J. M.; Krylov, A. I.; Gill, P. M. W.; Head-Gordon, M. Advances in Molecular Quantum Chemistry Contained in the QChem 4 Program Package. Mol. Phys. 2015, 113, 184−215. (68) Ou, Q.; Subotnik, J. E. Electronic Relaxation in Benzaldehyde Evaluated Via Td-Dft and Localized Diabatization: Intersystem Crossings, Conical Intersections, and Phosphorescence. J. Phys. Chem. C 2013, 117, 19839−19849. (69) Ahlrichs, R.; Bär, M.; Häser, M.; Horn, H.; Kölmel, C. Electronic-Structure Calculations on Workstation Computers - the Program System Turbomole. Chem. Phys. Lett. 1989, 162, 165−169.

(70) Lenthe, E. v.; Baerends, E. J.; Snijders, J. G. Relativistic Regular Two-Component Hamiltonians. J. Chem. Phys. 1993, 99, 4597−4610. (71) van Lenthe, E. v.; Snijders, J. G.; Baerends, E. J. The Zero-Order Regular Approximation for Relativistic Effects: The Effect of Spin− Orbit Coupling in Closed Shell Molecules. J. Chem. Phys. 1996, 105, 6505−6516. (72) Wang, F.; Ziegler, T.; van Lenthe, E.; van Gisbergen, S.; Baerends, E. J. The Calculation of Excitation Energies Based on the Relativistic Two-Component Zeroth-Order Regular Approximation and Time-Dependent Density-Functional with Full Use of Symmetry. J. Chem. Phys. 2005, 122.20410310.1063/1.1899143 (73) Mewes, J.-M.; Herbert, J. M.; Dreuw, A. On the Accuracy of the General, State-Specific Polarizable-Continuum Model for the Description of Correlated Ground- and Excited States in Solution. Phys. Chem. Chem. Phys. 2017, 19, 1644−1654. (74) Gordon, M. S.; Smith, Q. A.; Xu, P.; Slipchenko, L. V. Accurate First Principles Model Potentials for Intermolecular Interactions. Annu. Rev. Phys. Chem. 2013, 64, 553−578. (75) Kosenkov, D.; Slipchenko, L. V. Solvent Effects on the Electronic Transitions of P-Nitroaniline: A Qm/Efp Study. J. Phys. Chem. A 2011, 115, 392−401. (76) Gurunathan, P. K.; Acharya, A.; Ghosh, D.; Kosenkov, D.; Kaliman, I.; Shao, Y.; Krylov, A. I.; Slipchenko, L. V. Extension of the Effective Fragment Potential Method to Macromolecules. J. Phys. Chem. B 2016, 120, 6562−6574.

7037

DOI: 10.1021/acs.inorgchem.7b00620 Inorg. Chem. 2017, 56, 7029−7037