Potent Bis-Cyclometalated Iridium Photoreductants with β

Article pubs.acs.org/IC

Cite This: Inorg. Chem. XXXX, XXX, XXX−XXX

Potent Bis-Cyclometalated Iridium Photoreductants with β‑Diketiminate Ancillary Ligands Jong-Hwa Shon and Thomas S. Teets* Department of Chemistry, University of Houston, 3585 Cullen Blvd., Room 112, Houston, Texas 77204-5003, United States S Supporting Information *

ABSTRACT: In this work, we outline a strategy to prepare a class of improved visible-light photosensitizers. Bis-cyclometalated iridium complexes with electron-rich β-diketiminate (NacNac) ancillary ligands are demonstrated to be potent excited-state electron donors. Evaluation of the photophysical and electrochemical properties establishes the excited-state redox potentials of the complexes, and Stern−Volmer quenching experiments inform on the kinetics of photoinduced electron transfer to the model substrates methyl viologen (MV2+) and benzophenone (BP). Compared to fac-Ir(ppy)3 (ppy = 2-phenylpyridine), widely regarded as a state-of-the-art photoreductant, the complexes we describe have excitedstate redox potentials that are more potent by 300−400 mV and rates for photoinduced electron transfer that are accelerated by as much as a factor of 3. These complexes emerge as promising targets for application in photocatalytic reactions and other photochemical processes.

■

INTRODUCTION Photosensitizers, which transfer electrons or holes following absorption of light, are important components in a number of chemical processes. Photocatalysis, which has emerged as a prominent strategy in fuel-forming transformations,1−3 organic synthesis,4−6 and biomass degradation,7 often relies on visiblelight photosensitizers to more effectively absorb light and separate charge. In addition, photosensitizers have long been an invaluable tool for studying electron transfer processes in biomolecules where phototriggering the charge transport event permits accurate measurement of the kinetics.8−11 The substrate scope, rate, and efficiency of a photocatalytic reaction and the range of moieties that can be photooxidized or photoreduced in a biomolecule are governed by the choice of the photosensitizer and its excited-state redox potential. Perhaps the best-known photosensitizer is [Ru(bpy)3]2+ (bpy = 2,2′-bipyridine), which can act as both an oxidant and a reductant in its triplet excited state. Excited-state redox potentials can be calculated using modified Latimer diagrams,12 and for [Ru(bpy)3]2+ values of E(RuIII/RuII*) = −1.2 V and E(RuII*/RuIII) = 0.4 V can be estimated (all quoted potentials are vs the ferrocenium/ferrocene couple). Although [Ru(bpy)3]2+ has been employed as a photosensitizer in numerous visible-light photoredox schemes, recent research has borne out the notion that more potent photosensitizers can lead to improved catalytic reactivity. A well-known example is fac-Ir(ppy)3 (ppy = 2-phenylpyridine),13 which has an excited-state potential (E(IrIV/IrIII*)) of −2.1 V4. (Note: All excited-state potentials are quoted herein as reduction potentials; more negative values correlate with more reducing excited states.) Reductive transformations often involve photosensitized C−X (X = halogen, O) bond © XXXX American Chemical Society

cleavage, and work from the Stephenson group has shown that the stronger photoreductant fac-Ir(ppy)3 can lead to an enhanced substrate scope for hydrodehalogenation reactions, allowing unactivated alkyl, alkenyl, and aryl iodide substrates to be tolerated.14 In work from the same group, photocatalytic deoxygenation reactions, relevant to lignin depolymerization, were also unveiled using fac-Ir(ppy)3 as the photocatalyst.15,16 Other groups have employed fac-Ir(ppy)3 as a photocatalyst in C−H functionalization reactions,17 arylation of thiols,18 radical polymerization reactions,19 and hydrotrifluoromethylation reactions20 to give a few representative and recent examples. In spite of these recent advances, there are still existing limitations of the substrate scope for visible-light photoredox transformations and challenges associated with phototriggered oxidation or reduction of certain redox-active amino acids. There are currently no reports of visible-light-promoted transformations involving unactivated alkyl chloride or alkyl bromide substrates, and for degradation of lignin model substrates, preoxidation of a benzyl alcohol to a ketone proximal to the ether linkage was required prior to photocatalytic Cα−O bond cleavage.16 Fundamental advances in the design of transition-metal complexes with more potent excitedstate redox potentials could lead to important breakthroughs in these areas, but there is considerably less ongoing research on developing improved photosensitizers compared to the vast amount of research using existing photosensitizers to discover new photochemical reactions and biochemical pathways. Recent developments include a zirconium-based photosensitizer with redox-active ligands, which was shown to be active for Received: November 8, 2017

A

DOI: 10.1021/acs.inorgchem.7b02859 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry a few reductive transformations.21 Work from the Gray group has shown that homoleptic tungsten isocyanide complexes W(CNAr)6, where Ar is a substituted aryl substituent, are potent excited-state reductants with long-lived excited states capable of stoichiometric photoinduced electron transfer to a number of substrates.22−24 To date, no catalytic reactions involving these photosensitizers have been disclosed. Our group has previously shown that nitrogen-containing βketoiminate (acNac) and β-diketiminate (NacNac) ancillary ligands have profound effects on the ground-state redox properties and triplet-state dynamics of bis-cyclometalated iridium complexes when compared to isoelectronic facIr(C^N)3 (C^N is a cyclometalating ligand) or Ir(C^N)2(acac) (acac = acetylacetonate) complexes.25 In this work, we demonstrate that bis-cyclometalated iridium complexes with NacNac ancillary ligands are potent excited-state reductants. A series of complexes of the type Ir(ppy)2(NacNacR) (R = Me, NMe2, OEt) is described, where R is the substituent at the 2and 4-positions of the NacNac backbone. The R group on the NacNac influences the excited-state energy and the groundstate potential, both of which contribute to the excited-state potential. The rates of photoinduced electron transfer to the model substrates methyl viologen (MV2+) and benzophenone (BP) were determined by Stern−Volmer quenching studies and track well with the driving-force dependence predicted by Marcus theory.26 Complexes with R = Me or NMe2 outperform state-of-the-art complex fac-Ir(ppy)3 as photoreductants. Their excited-state potentials for electron transfer are ca. 300−400 mV more potent than those of fac-Ir(ppy)3, and this increase in driving force leads to small but significant enhancements in photoinduced electron-transfer rates when the substrate is MV2+, with much larger enhancements with BP. In total, this study reveals a new class of potent photoreductants with many potential applications and unveils design strategies that could lead to further improvements in future generations of photosensitizers based on this platform.

Both of the new complexes were characterized by X-ray crystallography; the structures are collected in Figure 1, and a

Figure 1. X-ray crystal structures of complexes 2 and 3. Ellipsoids are shown at the 50% probability level with hydrogen atoms and solvent molecules omitted.

summary of crystallographic parameters can be found in the Supporting Information, Table S1. A comparison of the structures of the new compounds described here with previously reported structures of Ir(bt)2(NacNacMe) (bt = 2phenylbenzothiazole)25 and compounds akin to 1 with CF3substituted NacNacMe derivatives27 reveals two different conformations of the NacNacR ancillary ligand. In most cases, a planar NacNac core is observed, with the two Ir−N, N−C, and C−C distances of the six-membered chelate ring nearly identical. In this typical arrangement, a close approach and parallel arrangement of the N-phenyl ring and the cyclometalated phenyl ring are observed in what can be described as a π-stacking arrangement. In contrast, for the structure of Ir(ppy)2(NacNacNMe2) (2), a puckered conformation of the NacNac core is observed with inequivalent bond distances and no close approach of the N-phenyl and cyclometalated phenyl rings. We do not believe this is a steric effect, as previously characterized complexes from our group with backbone CF3 groups do not exhibit this puckering,27 and there are structurally characterized complexes of t-butyl substituted NacNac ligands that exhibit a planar conformation. That said, the solution NMR spectra of 2 do not bear any abnormalities, suggesting that this conformation is only a solid-state phenomenon. Electrochemistry. Ground-state redox properties of the complexes were evaluated by cyclic voltammetry. Voltammograms for the new complexes 2 and 3 are shown in the Supporting Information, Figure S1. Sweeping cathodically, compounds 1−3 are difficult to reduce, and only irreversible features are observed beyond −2.5 V vs the ferrocenium/ ferrocene (Fc+/Fc) couple. As opposed to the reduction features, which are ill-defined and depend minimally on the identity of the NacNac ligand, the potential of the reversible oxidation couple is sensitive to the identity of the R group on the NacNac backbone (see Table 1). For complex 1 (R = Me), E(IrIV/IrIII) occurs at −0.07 V.25 Replacing the backbone Me groups with NMe2 results in a significant cathodic shift in the redox potential with E(IrIV/IrIII) = −0.26 V for complex 2. Addition of OEt groups to the NacNac backbone has the opposite effect, anodically shifting the oxidation potential to +0.03 V for complex 3. Photophysical Properties. The UV−vis absorption spectra do not show any remarkable features, and like most other cyclometalated iridium complexes, the spectra are dominated by intense 1(π → π*) transitions in the UV (λ
5 × 103 M−1 cm−1 at λ = 400 nm and discernible absorption out to λ ∼ 550 nm, suggesting that these complexes are viable candidates for visible-light photosensitization. Room temperature emission spectra for the complexes are overlaid in Figure 3 with data summarized in Table 1. The emission

Figure 4. Overlaid emission spectra of complexes 1−3 recorded at 293 K in MeCN and 77 K in toluene glass. Samples were excited at λ = 420 nm.

shifts upon cooling samples of 1 to 77 K in rigid toluene glass, and we see the same phenomenon with the new complexes. Whereas room-temperature spectra are broad and featureless, spectra at 77 K exhibit distinct vibronic structure with maxima that are blue-shifted by >1900 cm−1 relative to the roomtemperature spectra. As we25,27 and many others28,29 have documented, these large rigidochromic shifts are indicative of significant charge-transfer character in the excited state. Photoinduced Electron Transfer. Excited state potentials, E(IrIV/IrIII*), can be estimated via eq 1:12

Figure 3. Overlaid emission spectra of complexes 1−3 recorded in MeCN at 293 K. For emission spectra, samples were excited at λ = 420 nm.

spectra are broad in each case with poorly resolved vibronic structure, but the substitution of the NacNac ligand influences the emission wavelengths. In MeCN at room temperature, the complex Ir(ppy)2(NacNacMe) (1) has an emission maximum of 595 nm, which red-shifts by 1034 cm−1 in NacNacNMe2 complex 2 (λ = 634 nm) and blue shifts by 706 cm−1 in NacNacOEt

E(Ir IV /Ir III*) = E(Ir IV /Ir III) − E ES IV

(1)

III

The ground-state E(Ir /Ir ) values were determined from cyclic voltammograms recorded in MeCN. The excited-state energy (EES) values were approximated from the first (highenergy) maximum of the low-temperature emission spectra, C

DOI: 10.1021/acs.inorgchem.7b02859 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry recorded in toluene (Table 1). Toluene was used as a solvent for low-temperature measurements because it forms a transparent glass at 77 K. Acknowledging the difference in polarity between toluene and MeCN at 77 K in a rigid solvent environment, we do not expect solvent polarity to have a large influence on the observed emission maxima. Values for the excited-state potentials are summarized in Table 2 along with Table 2. Summary of Excited-State Electron-Transfer Parameters Measured in MeCN Solution kq × 10−9/M−1 s−1

a

complex

E(Ir /Ir *) /V

MV2+

BP

fac-Ir(ppy)3 1 2 3

−2.1 −2.4 −2.5 −2.4

12(1) 17(2) 16(2) 9.4(9)

1.9(2) 5.3(7) 5.6(6) 3.8(4)

IV

III

a

Referenced to Fc+/Fc.

kinetic data for photoinduced electron-transfer reactions with model substrates (see below). Using eq 1, state-of-the-art photosensitizer fac-Ir(ppy)3 has an excited-state potential of −2.1 V vs Fc+/Fc.30 In the complex Ir(ppy)2(NacNacMe) (1), the excited-state energy is diminished relative to that of facIr(ppy)3, and the ground-state IrIV/IrIII potential is cathodically shifted; the latter is more sensitive to the ligand substitution, resulting in an excited-state potential of −2.4 V for complex 1. In complex 2, with electron-donating NMe2 groups on the NacNac backbone, the excited-state endures a slight bathochromic shift which is more than counterbalanced by the decrease in E(IrIV/IrIII), resulting in a value of −2.5 V for E(IrIV/IrIII*). Adding OEt groups to the NacNac backbone has the opposite effect: the excited-state energy in 3 is slightly higher than 1 (by ca. 0.03 eV), but the anodic shift in the redox potential counterbalances this shift and leads to a predicted E(IrIV/IrIII*) of −2.4 V. Stern−Volmer quenching experiments31 were used to determine the rates of photoinduced electron transfer by facIr(ppy)3 and complexes 1−3 to the substrates methyl viologen (MV2+; E(MV2+/MV•+) = −1.1 V32) and benzophenone (E(BP/BP•−) = −2.2 V33,34). Upon excitation, the complexes transfer an electron to the quencher (Q), and in the absence of a sacrificial reductant back electron transfer regenerates the quencher and the photosensitizer’s ground state, as summarized in Scheme 1. Stern−Volmer data for the complexes 1−3 with the two acceptors (MV2+ and BP) is plotted in Figure 5. Good

Figure 5. Stern−Volmer plots for complexes 1−3 with MV2+ (top) and BP (bottom) as quenchers, including best-fit lines. Data were recorded at 293 K in MeCN solvent with λex = 420 nm for steady-state measurements and 453 nm for time-resolved measurements.

agreement between steady-state and time-resolved quenching data is observed, suggesting a dynamic quenching mechanism, as summarized in Scheme 1. The quenching rate constants (kq) summarized in Table 2 were determined by dividing the slope of the Stern−Volmer plot by the emission lifetime in the absence of quencher.31 The reported numbers represent an average of the values obtained from the steady-state and timeresolved experiments. The quenching rate constant for facIr(ppy)3 with MV2+ was determined to be 1.2(1) × 1010 M−1 s−1 (see data in Figure S2). Compared to fac-Ir(ppy)3 a slight but statistically significant increase in kq is observed for complexes 1 and 2, for which kq is 1.7(2) × 1010 M−1 s−1 and 1.6(2) × 1010 M−1 s−1, respectively. For complex 3, a slight decrease in kq is observed relative to fac- Ir(ppy)3, with a value of 9.4(9) × 109 M−1 s−1 for the complex with the NacNacOEt ancillary ligand. For the substrate BP, a much more prominent increase in the rate of photoinduced electron transfer is observed in the NacNacR complexes. The quenching rate for fac-Ir(ppy)3 with BP in THF was previously measured in THF,35 and a value of 4.2 × 107 M−1 s−1 was obtained. We reproduced this value and also found a significant rate increase when the solvent was changed to MeCN, with kq = 1.9(2) × 109 M−1 s−1 in this solvent (Figure S3). The rate constants are much larger for the NacNacR complexes 1−3; again, complexes 1 and 2 give nearly identical values, 5.3(7) × 109 M−1 s−1 (1) and 5.6(6) × 109 M−1 s−1 (2), with a slightly smaller rate constant of 3.8(4) × 109 M−1 s−1 for complex 3. To confirm that excited-state electron transfer (Scheme 1) is the operative quenching mechanism, and not energy transfer,

Scheme 1. Summary of Photoinduced Electron-Transfer Pathway

D

DOI: 10.1021/acs.inorgchem.7b02859 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry we conducted a photolysis experiment with Ir(ppy)2(NacNacNMe2) (2) in the presence of MV2+ and the sacrificial reductant triethanolamine (TEOA). As shown in Figures S4−S6, complexes 1−3 are minimally quenched by TEOA alone with observed quenching rates (kq) that are 3−4 orders of magnitude smaller than the quenching rate constants for MV2+. When complex 2 is irradiated together with MV2+ and TEOA in MeCN, the highly colored MV•+ radical gradually appears, as determined by the growth of UV−vis absorption features with λmax = 395 and 605 nm,36 as well as a distinct blue color in the solution (Figure 6).

Figure 7. Plot of the ground-state oxidation potential, E(IrIV/IrIII), vs the excited-state energy, EES, for the compounds fac-Ir(ppy)3 and 1−3.

and vice versa. That said, we do observe that the value of E(IrIV/IrIII) is more strongly perturbed than the excited-state energy when electron-rich NacNacR ancillary ligands are incorporated, such that complexes 1−3 all are more potent excited-state electron donors than fac-Ir(ppy)3 by ca. 300−400 mV. This difference in excited-state potentials is significant, as it introduces the possibility of using these photosensitizers on substrates which do not react with the excited-state of facIr(ppy)3. For example, existing photocatalytic reactions can operate on activated organic halide substrates or unactivated organic iodide substrates,4,14 but there are no examples involving alkyl, vinyl, or aryl chloride/bromide substrates, which are much more inexpensive than their iodide congeners. Although redox potentials of organic halides are difficult to quantify and depend strongly on solvent a electrolyte conditions, owing to the irreversibility of heterogeneous electron transfer with these compounds, a general trend emerges in electrochemical studies of organic halides.37−39 The reduction potentials for organic bromide compounds are ca. 100 mV more negative than analogous iodide compounds, and the potentials for chloride substrates are shifted by an additional 100−250 mV relative to the bromides. As such, the 300−400 mV gain in potential for complexes 1−3 relative to fac-Ir(ppy)3 could allow an expanded substrate scope for photocatalytic radical reactions involving halide substrates. Similarly, previous work on C−O bond cleavage in lignin substrates required preoxidation of the substrate prior to ether bond cleavage.16 The reduction potentials for Csp3−O bonds are estimated to be −2.6 V vs Fc+/Fc,40,41 and the −2.5 V excited-state potential for complex 2 is very close to being able to do direct C−O bond cleavage and, under the right conditions, may operate on such substrates. In addition to establishing the excited-state redox potential, it is also important to measure the kinetics and mechanism of photoinduced electron transfer and compare with known photosensitizers. Emission quenching is observed when the complexes fac-Ir(ppy)3 and 1−3 are irradiated in the presence of the electron acceptors MV2+ and BP. The poor spectral overlap between the excited states of the iridium complexes and the acceptors, coupled with the observation of the MV•+ radical when a sacrificial reductant is included (see Figure 6), suggest that electron transfer, and not excited-state energy transfer, is the origin of the quenching.42 Excited-state electron transfer between [Ru(bpy)3]2+ and MV2+is one of the best-studied electron transfer reactions, with kq = 2.8 × 109 M−1 cm−1 in MeCN.43 With iridium-based photosensitizers, photoinduced

Figure 6. Evolution of the UV−vis absorption spectra when a solution of 2 (55 μM) is irradiated at 420 nm in the presence of MV2+ (2.9 × 10−4 M) and TEOA (1.79 × 10−2 M) for 45 min. Spectra were recorded in 5 min intervals. The inset shows photographs of the cuvette at the start and end of photolysis.

■

DISCUSSION We demonstrated previously that incorporating electron-rich, strongly π-donating β-diketiminate (NacNac) ligands destabilizes the Ir-centered HOMO in bis-cyclometalated iridium complexes, and that the HOMO likely involves considerable contribution from NacNac-centered π-orbitals in these complexes.25,27 Here, we show that this perturbation of the orbital energies also results in complexes that are more potent excited-state electron donors than fac-Ir(ppy)3, which has emerged as the state-of-the-art photosensitizer for transformations initiated by photoinduced electron transfer.4 As visible-light photoredox catalysis continues to grow as a prominent strategy in reactions related to the synthesis of renewable fuels, commodity chemicals, and fine chemicals, improved photosensitizers could potentially benefit a number of these areas by leading to improved yields and efficiencies for existing processes, expanding substrate scopes, and/or giving rise to new reactions altogether. The compounds we describe here are certainly promising targets for incorporation into photocatalytic reaction schemes. The first criterion with which to evaluate a photosensitizer is its excited-state reduction potential, which defines the thermodynamic limit for which substrates can be acted upon in a photocatalytic reaction. As shown by eq 1, there are two parameters which determine the value of E(IrIV/IrIII*): the ground-state IrIV/IrIII potential and the excited-state energy. For the family of compounds studied here, there is a trade-off between excited-state energy and ground-state potential, as shown in Figure 7, such that complexes with higher excitedstate energies tend to have more positive oxidation potentials, E

DOI: 10.1021/acs.inorgchem.7b02859 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry electron transfer to MV2+ involves a large driving force of >1 V, leading to efficient quenching with kq values almost 1 order of magnitude larger than [Ru(bpy)3]2+ and approaching the diffusion limit of ca. 2−4 × 1010 in MeCN.44,45 That said, the rates of electron transfer to MV2+ are not entirely diffusion limited, and in the cases of complexes 1 and 2, the increase in electron-transfer driving force is accompanied by a modest increase in kq. Complexes 1 and 2 outperform fac-Ir(ppy)3 by a factor of 1.3−1.4, a small but statistically relevant amount. In contrast, benzophenone is much more difficult to reduce, with a reduction potential of ca. −2.1 V, such that the driving forces for photoinduced electron transfer are much smaller. As such, kq values are significantly smaller than observed for MV2+ but with a much greater dependence on the ancillary ligand. For fac-Ir(ppy)3, an electron-transfer rate constant of 1.9(1) × 109 M−1 s−1 is observed with benzophenone as the acceptor. This value is doubled in complex 3 and nearly tripled in complexes 1 and 2. This dependence of the electron-transfer rate on the driving force is predicted by Marcus electron-transfer theory,26 and Figure 8 shows the dependence of log(kq) on the electron-

Figure 9. Comparison of NacNac chelate ring to the o, m, and p positions of a six-member aryl ring.

IrIV/IrIII potentials and blue-shifted emission bands (provided the nature of the excited-state did not change altogether) when these electron-withdrawing groups were added either to the backbone of the NacNac or to the N-phenyl groups. The increase in excited-state energy and anodic shift in oxidation potential in NacNacOEt complex 3, when compared with 1 and 2, are possible only if the ethoxy groups function as electronwithdrawing substituents. This is all consistent with the 2- and 4-positions of the backbone behaving like meta substituents (the σm Hammett constant for an ethoxy group is +0.10).46 With this in mind, and the observation that complexes 1 and 2 are superior to 3 as photoreductants, a way forward is to incorporate additional electron-donating substituents into the NacNac, crafting an even more electron-rich version. There are very few meta substituents that are net electron donating, Me (σm = −0.07) and NMe2 (σm = −0.16) being rare exceptions, but electron-donating substituents at the ortho and para positions (i.e., on the nitrogen donors or the 3-position of the ligand backbone) could be a way to perturb the E(IrIV/ IrIII*) potentials to even more negative values. It may be synthetically challenging to incorporate heteroatom electrondonating groups like NMe2 or OR at the 3-position of the backbone because that carbon typically reacts as a nucleophile,47,48 although it is possible to add modestly electrondonating alkyl groups to that position.48 Replacing the Nphenyl groups with either alkyl groups or more electron-rich aryl groups (e.g., 4-NMe2C6H4) should also be synthetically tractable and is a strategy we will explore in future work.

Figure 8. Dependence of log(kq) on the electron-transfer driving force, ΔG, plotted with a best-fit line as a guide. The data point for [Ru(bpy)3]2+ with MV2+ was generated from literature values.12,43

■

CONCLUSIONS A series of cyclometalated iridium photosensitizers with enhanced excited-state redox reactivity is described here. Incorporation of electron-rich β-diketiminate ligands greatly perturbs the orbital energies and results in complexes with excited-state redox potentials more potent than that of facIr(ppy)3. Excited-state electron transfer reactions with model substrates MV2+ and BP are also presented, and for the complexes introduced here, the rates of photoinduced electron transfer exceed those of existing state-of-the-art photosensitizers. Analysis of structure−reactivity relationships shows that complexes with electron-donating substituents on the NacNac are superior, suggesting future designs involving variants with additional electron-donating groups. The bis-cyclometalated NacNac complexes described here and their future derivatives are promising candidates for application as photocatalysts, and in future efforts, we will investigate these photosensitizers in the context of a number of photocatalytic reactions related to renewable fuels and organic synthesis.

transfer driving force, estimated by the difference between the excited-state potential of the photosensitizer and the reduction potential of the acceptor. Noting some uncertainty in the value of the excited-state redox potential, caused by difficulty in accurately determining the excited-state energy, we do see an approximately linear dependence in Figure 8, suggesting these excited-state electron transfer reactions lie in the Marcus normal region. More importantly, the data in Table 2 and Figure 8 show that the increase in excited-state potential brought on by addition of electron-rich NacNac ligands correlates with an increase in excited-state electron transfer rate, further supporting the efficacy of 1−3 as photocatalyst targets. One final insight revealed by this study are design criteria for preparing even more potent photosensitizers based on biscyclometalated iridium NacNac complexes. The six-membered NacNac chelate ring can be thought of analogously to a sixmembered aryl ring, where the two bound nitrogen atoms are akin to the ortho positions, the 2- and 4-positions of the backbone are meta, and the 3-position of the backbone is the para position, as indicated in Figure 9. The trends in excitedstate energy and redox potentials for complexes 1−3 are consistent with this notion. In our previous work on CF3substituted NacNac complexes, we showed anodic shifts in the

■

EXPERIMENTAL SECTION

Materials. All reactions were executed in a nitrogen-filled glovebox operating at