Blue-to-Green Emitting Neutral Ir(III) Complexes Bearing Pentafluorosulfanyl Groups: A Combined Experimental and Theoretical Study

A structure–property relationship study of neutral heteroleptic (1 and 2, [Ir(C∧N)2(L∧X)]) and homoleptic (3 and 4, fac-[Ir(C∧N)3]) Ir(III) complexes (where L∧X = anionic 2,2,6,6-tetramethylheptane-3,5-dionato-κO3,κO6 (thd) and C∧N = a cyclometalating ligand bearing a pentafluorosulfanyl (−SF5) electron-withdrawing group (EWG) at the C4 (HL1) and C3 (HL2) positions of the phenyl moiety) is presented. These complexes have been fully structurally characterized, including by single-crystal X-ray diffraction, and their electrochemical and optical properties have also been extensively studied. While complexes 1 ([Ir(L1)2(thd)]), 3 (Ir(L1)3), and 4 (Ir(L2)3) exhibit irreversible first reduction waves based on the pentafluorosulfanyl substituent in the range of −1.71 to −1.88 V (vs SCE), complex 2 ([Ir(L2)2(thd)]) exhibits a quasi-reversible pyridineC∧N-based first reduction wave that is anodically shifted at −1.38 V. The metal + C∧N ligand oxidation waves are all quasi-reversible in the range of 1.08–1.54 V (vs SCE). The optical gap, determined from the lowest energy absorption maxima, decreases from 4 to 2 to 3 to 1, and this trend is consistent with the Hammett behavior (σm/σp with respect to the metal–carbon bond) of the −SF5 EWG. In degassed acetonitrile, for complexes 2–4, introduction of the −SF5 group produced a blue-shifted emission (λem 484–506 nm) in comparison to reference complexes [Ir(ppy)2(acac)] (R1, where acac = acetylacetonato) (λem 528 nm in MeCN), [Ir(CF3-ppy) (acac)] (R3, where CF3-ppyH = 2-(4-(trifluoromethyl)phenyl)pyridine) (λem 522 nm in DCM), and [Ir(CF3-ppy)3] (R8) (λem 507 nm in MeCN). The emission of complex 1, in contrast, was modestly red shifted (λem 534 nm). Complexes 2 and 4, where the −SF5 EWG is substituted para to the Ir–CC∧N bond, are efficient phosphorescent emitters, with high photoluminescence quantum yields (ΦPL = 58–79% in degassed MeCN solution) and microsecond emission lifetimes (τε = 1.35–3.02 μs). Theoretical and experimental observations point toward excited states that are principally ligand centered (3LC) in nature, but with a minor metal-to-ligand charge-transfer (3MLCT) transition component, as a function of the regiochemistry of the pentafluorosulfanyl group. The 3LC character is predominant over the mixed 3CT character for complexes 1, 2, and 4, while in complex 3, there is exclusive 3LC character as demonstrated by unrestricted density functional theory (DFT) calculations. The short emission lifetimes and reasonable ΦPL values in doped thin film (5 wt % in PMMA), particularly for 4, suggest that these neutral complexes would be attractive candidate emitters in organic light-emitting diodes.


■ INTRODUCTION
Phosphorescent iridium complexes bearing arylpyridine cyclometalating (C ∧ N) and ancillary ligands (either anionic (L ∧ X) or neutral (L ∧ L)) have gained widespread interest among researchers because of their remarkable optoelectronic properties: e.g., good color tunability, high photoluminescence quantum yields (Φ PL ), short emission lifetimes (τ e ), and high photo-and thermostability. 1−3 This confluence of properties render these complexes as attractive candidates as emitters for solid-state electroluminescent devices, the most common of which are organic light-emitting diodes (OLEDs) or light-emitting electrochemical cells (LEECs), 4−7 bioimaging agents, 8,9 and sensing applications. 10 With respect to their use in OLEDs, neutral Ir(III) complexes are generally more desirable in comparison to cationic Ir(III) complexes, as they can be easily vacuum deposited. High-efficiency white flat-panel displays require combined emission from red, green, and blue (RGB) emitters. While the color purity and efficiency of red and green Ir(III) emitters are satisfactory, the dearth of highefficiency blue Ir(III) emitters remains an issue.
Unlike cationic iridium complexes, where color tuning is facile due to independent modulation of the electronics of both the C ∧ N and L ∧ L ligands, for neutral iridium complexes, bearing a nonchromophoric L ∧ X ligand, the emission color is governed by the electronics of the C ∧ N ligands. Substitution of the C ∧ N ligands with EWGs renders the HOMO to be more stabilized than the LUMO. This fact leads to an increased HOMO−LUMO gap, and thus blue emission is achieved. While the near-universal strategy of blue-shifting the emission by incorporation of one or two fluorine atoms in the C ∧ N ligand, such as 2-(4-fluorophenyl)pyridine (FppyH) or 2-(4,6difluorophenyl)pyridine (dFppyH), is popular, 11−14 the issue of emitter degradation via defluorination negatively affects its incorporation into the emitter design. 15 Apart from fluorine atoms, other examples of EWGs used with a view to blueshifting the emission of Ir(III) complexes consist of sulfonyl (−SO 2 R), 16−18 trifluoromethyl (−CF 3 ), 19−23 trifluoromethoxy (−OCF 3 ), 24,25 and pentafluorosulfanyl (−SF 5 ). 26 These C ∧ N ligands are often used in conjunction with nonchromophoric ancillary ligands such as acac, 27 thd, 28 picolinate (pic), 29 and 3oxo-1,3-diphenylprop-1-en-1-olate (dbm). 30 Although impressive performances have been achieved in solution, translating these performances into devices with good stability and efficiency metrics is still a challenge. 17 Thus, the design and syntheses of new blue-emitting phosphors for OLEDs and LEECs are essential.
Relatively strong intermolecular π−π interactions are generally observed for planar, nonhindered C ∧ N ligands, leading to the formation of small Ir(III) complex crystallites, which are responsible for unfavorable self-quenching. 31 Incorporation of bulky hydrophobic and chemically inert groups within the C ∧ N ligands leads to the prevention of such aggregate formation, while it also improves the photostability of these complexes in the amorphous phase. Introduction of bulky and strongly electron-withdrawing −SF 5 groups (σ p = 0.68) 32 instead of a trifluoromethyl (−CF 3 ) group (σ p = 0.54) should lead to significantly blueshifted emission, concomitant with a blue-shifted absorption, due to stabilization of the C ∧ N-based HOMO. The −SF 5 group is strongly electron-withdrawing and has been shown to be lipophilic, thermally and chemically stable, and also biologically active. 33−35 However, despite these favorable properties, as yet it is an underexplored moiety in the field of organic semiconductor materials. 26 Nevertheless, these properties make the −SF 5 group an attractive candidate for replacing the commonly used C(aryl)−F motif, serving the same purpose of modulating the HOMO energy but without affecting the stability of the emitter.
The significance of the regiochemistry of the substituent on the C ∧ N ligand in tuning the emission wavelength has been  demonstrated by molecular orbital analyses using density functional theory (DFT) calculations. 36,37 For example, the absorption and emission spectra of an Ir(III) complex bearing dFppy C ∧ N ligands (dFppyH = 2-(2,4-difluorophenyl)pyridine) are more blue-shifted in comparison to an analogous Ir complex with ppy C ∧ N ligands due to greater HOMO stabilization in comparison to LUMO stabilization. When the fluorine atom is located at the 5-position on the C ∧ N ligands (para to the Ir−C C ∧ N bond), its inductive electron-withdrawing effect is counterbalanced by weak π donation, thereby raising the HOMO and reducing the band gap. 38 In a recent study, we investigated the optoelectronic properties of cationic Ir(III) complexes bearing an −SF 5 EWG on ppy or phenylpyrazole (ppz) C ∧ N ligands, 26 with the substituent position of the EWG varied so as to adopt either a para or meta relationship with respect to the Ir−C C ∧ N bond. In this work, a family of neutral emissive Ir(III) complexes with the −SF 5 group attached at the 4-(para with respect to the Ir− C C ∧ N bond) or 5-position (meta with respect to the Ir−C C ∧ N bond) of the C ∧ N ligands (HL1 and HL2; Chart 1) is reported. To ensure neutrality of these complexes, three anionic ligands were employed: complexes 1 and 2 adopt two anionic C ∧ N ligands and 2,2,6,6-tetramethylheptane-3,5-dionato-κO 3    (thd) as the ancillary ligand, while complexes 3 and 4 are fachomoleptic complexes containing three C ∧ N ligands (Chart 1). The effect of the meta/para position of the −SF 5 EWG with respect to the Ir−C C ∧ N bond on the optoelectronic properties of these complexes is discussed and corroborated on the basis of DFT calculations, with the results compared to several benchmark complexes (complexes R1−R10; Chart 2).

■ RESULTS AND DISCUSSION
Synthesis. The syntheses of the C ∧ N ligands and the iridium complexes 1−4 are shown in Scheme 1. As we reported previously, the C ∧ N ligands were synthesized under Stille 39 palladium-catalyzed cross-coupling conditions in good yields. 40 Ligands HL1 and HL2 were reacted with IrCl 3 ·3H 2 O, and the resulting iridium dimers [Ir(L1) 2 (μ-Cl)] 2 (D1) and [Ir-(L2) 2 (μ-Cl)] 2 (D2) were obtained in good yield and used directly in the next synthetic step. 41 Complexes 1 and 2 were isolated in good yield through cleavage of D1 and D2 with the thd ligand under basic conditions. The homoleptic complexes 3 and 4 were synthesized upon reaction of 3.1 equiv of the C ∧ N ligands with 1 equiv of Ir(acac) 3 . 27 A long reaction time (72 h) at high temperature (200°C) favors the formation of the thermodynamically stable facial ( fac) isomer in comparison to the kinetically stable meridional (mer) isomer, 27 as observed by 1 H NMR spectroscopy. All the neutral complexes were purified by column chromatography. The successful syntheses of complexes 1-4 confirm the stability of the −SF 5 group toward strong bases and high temperatures in organic alcoholic solvents. For both 3 and 4, 19 F NMR indicated the presence of a small impurity (∼2%), which, in light of the satisfactory microanalysis, was inferred to be the mer isomer. This impurity could not be removed either by chromatography or by repeated recrystallization. In fact, the formation of an inseparable trace amount of the kinetically stable mer isomer during the synthesis of the fac isomer is already well documented. 42,43 Recrystallization of 3 and 4 on a small scale provided single crystals, which were found to be the expected fac isomer by X-ray crystallography (see Figure 1). Complexes 1−4 are stable in the presence of air and moisture and are soluble in common organic solvents such as acetonitrile and dichloromethane.
All ligands, dimers and complexes were characterized by 1 H, 19 F and 13 C NMR spectroscopy (Figures S1−S8 and S9−S28 in the Supporting Information), ESI-HRMS, melting point determination, and elemental analyses. The structures of complexes 1−4 were unequivocally determined by single crystal X-ray diffraction and corroborated the C 2 (1 and 2) and C 3 (3 and 4) symmetry assignments ascribed to the complexes on the basis of the solution-state 1 H and 19 F NMR. The downfield shift of the proton ortho to the cyclometalating carbon atom points toward the electron-withdrawing nature of the −SF 5 group on the phenyl ring ( Figure S6). A similar downfield shift was also found for the proton ortho to the carbon atom that is involved in the C ph −C pyridine bond. The 19 F NMR spectra exhibit a doublet for the equatorial fluorine atoms and a pentet for the axial fluorine for the −SF 5 group in an intensity ratio of 4:1 as an AB 4 system ( Figure S7). 44 The HRMS analyses for 1−4 showed the indicative peak of the cation [M + H] + .
Crystal Structures. Crystals of 1−4 suitable for X-ray analysis were grown by slow diffusion of an antisolvent (1, ethanol; 2, methanol; 3, diethyl ether; 4, hexane) into concentrated solutions of the complexes in dichloromethane ( Figure 1). Table S1 in the Supporting Information contains relevant crystallographic parameters, and Table S2 in the Supporting Information compares selected bond distances and angles observed in the crystal to those predicted by DFT calculations. In all cases, the metal ion exhibits a pseudooctahedral coordination geometry. In the case of the heteroleptic complexes 1 and 2, the pyridyl nitrogen atoms of the C ∧ N ligands are in a mutually trans relationship with respect to each other, as is common for many [Ir(C ∧ N) 2 (L ∧ X)] complexes, such as R1. 29 In the case of the homoleptic complexes, all of the nitrogen atoms are in a cis relationship. The average Ir−C C ∧ N (1.995 Å) and Ir−C C ∧ N (2.031 Å) bond

Inorganic Chemistry
Article distances in 1 and 2 are similar to those in R1 (Ir−C C ∧ N , 1.991 Å; Ir−C C ∧ N , 2.037 Å) (Table S2). A similar structural picture was found in [Ir(L1) 2 (dtBubpy)][PF 6 ] in our previous study, 26 where dtBubpy is 4,4′-di-tert-butyl-2,2′-bipyridine (average Ir− C C ∧ N distance 2.053 Å and average Ir−C C ∧ N distance 2.018 Å). In 1 and 2 the average Ir−O distance (2.123 Å) was found to be longer in comparison to the Ir−C C  (Table S2). In all of the complexes, the steric bulk of the tert-butyl (complexes 1 and 2) and −SF 5 groups prohibit the formation of any significant π−π stacking interactions. Electrochemical Properties. In order to assess the effect of the −SF 5 group on the ground-state electronics of complexes 1−4, cyclic (CV) and differential pulse voltammetry (DPV) measurements were undertaken. Degassed MeCN was employed as the solvent, and the redox potentials are referenced with respect to SCE (Fc/Fc + = 0.38 V in MeCN). 45 The relevant electrochemical data can be found in Table 1 and Table S3 in the Supporting Information, while the CV and DPV traces are shown in Figure 2 (the full set of redox potentials is detailed in Table S3).
At positive potential, complexes 1−4 each display oxidation waves that are quasi-reversible and single electron in nature, in the range of 1.08−1.54 V. DFT calculations indicate that incorporation of the −SF 5 EWG results in a stabilization of the HOMOs of 1−4 in comparison to reference complex R2 and that the HOMOs of these complexes are comprised of almost equal contributions from the metal center and the C ∧ N ligands ( Figure 3). On the basis of the DFT calculations, and in line with literature precedent of related neutral Ir(III) complexes, 29,42 the first oxidation of complexes 1−4 is assigned to abstraction of an electron from the metal-based orbitals (the Ir III /Ir IV redox couple) as well as some contribution from the C ∧ N phenyl orbitals. The lower calculated HOMO energies for 1 (E HOMO = −5.57 eV), 2 (E HOMO = −5.64 eV), 3 (E HOMO = −5.66 eV), and 4 (E HOMO = −5.72 eV) in comparison to that of R2 (E HOMO = −5.17 eV) are consistent with the expected stabilization of the C ∧ N phenyl based orbitals by the −SF 5 group in comparison to R2 (Table 1). A significant anodic shift of the oxidation potential of 2 by 460 mV was observed in comparison to that of 1, which is due to the increasing electronwithdrawing nature of the −SF 5 group when it is moved from a meta position to a para position with respect to the Ir−C bond of the C ∧ N ligands. This observation is also in agreement with the increasing Hammett parameter of the −SF 5 group when it is positioned regiospecifically (σ m = 0.61, σ p = 0.68); this behavior is less pronounced for the fac-homoleptic complexes 3 and 4. Assuming the oxidation potentials are invariant with respect to solvent, the oxidation potentials of 1 and 2 were found to be more positive in comparison to those of R1−R3 and R5, implying that −SF 5 is a stronger EWG in comparison to −F and −CF 3 , coincident with the smaller Hammett parameters of these substituents (Table 1). 32 Assuming the fact that thd is a slightly better donor in comparison to acac, as implied by the slight cathodic shift of the oxidation potential of R2 in comparison to that of R1, complex R6 has an oxidation potential more positive than that of 1 but a value of E 1/2 ox lower than that of 2, suggesting that the electron-withdrawing ability follows the order L2 > dFppy > L1. Complexes 3 and 4 are harder to oxidize in comparison to the fac-homoleptic reference complexes R7−R9, which is in line with the Hammett parameters but are 0.03 V (for 4) to 0.13 V (for 3) easier to oxidize in comparison to R10.
At negative potentials, multiple ligand-based multielectronic reductions can be observed for 1−4 ( Figure 2). While for 1, 3 and 4 the first reductions are irreversible, for 2 this reduction is found to be quasi-reversible and monoelectronic; the second reduction in 2 mirrors the behavior of the first reduction waves for 1, 3, and 4. For 1, 3, and 4, DFT calculations point to a LUMO that has predominant C ∧ N character with significant contribution from the −SF 5 group, whereas for 2, the LUMO remains localized on the C ∧ N ligand but without the contribution from the −SF 5 group. Therefore, and in line with our previous results for cationic iridium complexes, 26 the first reduction wave is plausibly assigned to direct reduction of the −SF 5 moiety for 1, 3, and 4, where the −SF 5 group, upon accepting an electron, may release a fluoride ion, thus rendering the reduction irreversible. Due to the strong electronwithdrawing nature of the −SF 5 group, all of the first reduction potentials of 1−4 are anodically shifted by between 270 and 770 mV in comparison to that of R1. There is a noticeable anodic shift of 0.45 V of the first reduction potential of complex 2 in comparison to that of complex 1. This reduction wave in 2 is also distinctively reversible and as a consequence does not involve the −SF 5 group and in fact represents reduction of the pyridine ring of the C ∧ N ligand. The DFT prediction of the LUMO energy actually aligns well with the second irreversible reduction wave (E red2 = −1.86 (irr)). Further, the trend  Figure 4, and the data are summarized in Table S4 in the Supporting Information. Figure S29 in the Supporting Information compares the experimentally determined absorption spectra for each of the complexes with the transitions predicted by TD-DFT. The absorption spectra of complexes 1, 3, and 4 show two highly absorbing bands between 210 and 300 nm and additional, less absorptive bands beyond 300 nm. For complexes 1 and 2, prominent spin-allowed 1 π → π* ligandcentered ( 1 LC) transitions localized on the C ∧ N ligand and ligand-to-ligand charge transfer ( 1 LLCT) transitions from the ancillary ligand to the C ∧ N ligands, as predicted by TD-DFT calculations, correlate with the high-energy bands. For complexes 3 and 4, these bands are 1 π → π* ligand-centered ( 1 LC) transitions as predicted by TD-DFT. The lower energy hypochromic bands between 270 and 300 nm are assigned to a 1 LC transition for complexes 3 and 4, whereas for complexes 1 and 2 these 1 LC transitions are mixed with 1 MLCT transition contributions (Tables S5−S8 in the Supporting Information). The nature of the transitions between 300 and 400 nm becomes more complex with bands consisting of an admixture of 1 LC and 1 MLCT transitions, with varying but more significant 1 MLCT content along with ligand-to-ligand transitions ( 1 LLCT) in the case specifically for 1 and 2; the 1 LLCT transitions are evidently absent for 3 and 4. For all of the complexes the band located between 412 and 456 nm is assigned as the HOMO → LUMO transition, which is principally 1 MLCT (Ir(dπ) → L1/L2(π*)) in nature but mixed with 1 LC ( 1 π → π*) character (Tables S5−S8). Complexes 1−4 display a shoulder at λ >450 nm, albeit with very low molar absorptivity, which is a feature also observed for R1 at 487 nm and R2 at 468 nm (Table S4). 29,47 These bands are assigned as spin-forbidden 3 MLCT and 3 LLCT by direct population of the triplet state, which gains intensity by mixing with the higher lying 1 MLCT through the strong spin−orbit coupling of the Ir metal center. 48 The spectra observed for 1−4 are very similar to those of the corresponding pentafluorosulfanyl-substituted cationic complex [Ir(L1) 2 (dtBubpy)]-[PF 6 ], suggesting that the dominant absorptions in these complexes are due to the "(C ∧ N) 2 Ir" fragment. 26

Article
The emission spectra in deaerated MeCN solution at room temperature for complexes 1−4 are shown in Figure 5a, while Figure 5b shows the doped film (5 wt % in PMMA) emission spectra at room temperature. Table 2 summarizes the relevant solution-phase photophysical data of 1−4 as well as the reference complexes R1−R10. The solid-state photoluminescence data are shown in Table 3. In MeCN solution, the emission of complexes 1−4 varies from sky blue to green, with the emission maxima in the range of 484−537 nm. The emission profiles are structured, which is an indication of an excited state that has 3 LC character. Spin-unrestricted DFT calculations predict that the spin density is principally localized on a combination of the C ∧ N ligands and the metal center ( Figure 6), pointing toward a mixed 3 LC and 3 MLCT excited state. These predictions are in line with a variety of features that are characteristic of 3 LC or 3 MLCT emission: the structured vibronic features in the phosphorescence spectra and short (τ e < 1.5 μs for 1, 3, and 4; τ e = 3.02 μs for 2) radiative lifetimes.
Complex 2 (where the −SF 5 EWG is positioned para to the Ir−C C ∧ N bond) exhibits an emission maximum that is blueshifted in comparison to 1 (where the −SF 5 EWG is positioned meta to the Ir−C C ∧ N bond), which fits with the magnitude of the Hammett meta and para parameters of the −SF 5 group. An analogous observation can be made for complexes 3 and 4. In line with the red shift of the absorption onset from 4 to 2 to 3 to 1 (see inset magnified spectra in Figure 4), the emission maxima are also red-shifted accordingly ( Table 2) and this trend also follows the gradually decreasing HOMO−LUMO  a Absorption data are in solvents as mentioned in Table S4   gap from 4 to 2 to 3 to 1, as calculated by DFT (Table 1 and Figure 3). The predicted emission maxima, E AE = E(T 1 ) − E(S 0 ) at the T 1 optimized geometries (adiabatic electronic emission) obtained by DFT calculations 36 for 1−4, are respectively, at 566, 521, 539, and 443 nm and match closely those observed experimentally and reproduce the observed trend of red-shifted emission maxima from complex 4 to 2 to 3 to 1.
Incorporation of the −SF 5 EWG on the cyclometalating phenyl rings helps to stabilize the frontier molecular orbitals (Figure 3), leading to a blue shift in the observed emission color for complexes 2−4 in comparison to reference complexes R1 and R2 (Table 2). For 1, slight red shifts in its emission maximum in comparison to those of R1 (6 nm, 213 cm −1 ) and R2 (9 nm, 321 cm −1 ) are observed due to a more stabilized emissive triplet state in 1 in comparison to those in R1 and R2. Considering the fact that 3 LC emission is insensitive to solvent polarity, 28 Table 2 for λ em of 1−4). Both the λ em,0−0 and λ em,0−1 peaks of [Ir(L2) 2 (dtBubpy)][PF 6 ] are, however, more blueshifted in comparison to those of 1−4 (see Table 2 for λ em of 1−4).
The Φ PL values vary widely between 15 and 79% ( Table 2). The observed τ e values are single component, pointing toward emission from a single species. Complexes 2 and 4, where the −SF 5 EWG is positioned para to the Ir−C C ∧ N bond, were found to be more brightly emissive with longer τ e in comparison to those of complexes 1 and 3, where the −SF 5 EWG is positioned meta to the Ir−C C ∧ N bond. The calculated radiative (k r ) and nonradiative (k nr ) decay constants (k r = Φ PL / τ ε and k nr = (k r /Φ PL ) − k r ) are in the range of (2.31−4.29) × 10 5 and (0.69−13.09) × 10 5 s −1 , respectively. The higher Φ PL values of 2 vs 1 and 4 vs 3 are supported by the decrease in nonradiative decay by 19 and 4 times, respectively. The low Φ PL values and high k nr rates observed for complexes 1 and 3 may be explained by considering the deactivation of the emissive excited state through vibrational modes of the S−F bonds in the −SF 5 group that is meta to the Ir−C C ∧ N bond, as we 49 had previously observed in a cationic iridium complex containing a −CF 3 -substituted guanidylpyridine ancillary ligand. Frequency calculations of complexes 1 and 2 suggest that there is a strong coupling between the wagging mode of the C ∧ N ligand with the wagging mode of the equatorial S−F bonds of the −SF 5 group in complex 1 (vibrational modes 46 and 47; E v46 = 322.59 cm −1 and E v47 = 324.23 cm −1 ; cf. Table  S9 in the Supporting Information for other minor contributing vibrational modes that couple with the spin density). These couplings are found to be very weakly present in complex 2, as the −SF 5 group is in the para position in this complex and is therefore farther away from the C C ∧ N −Ir bond. A similar observation is found for the homoleptic complexes 3 and 4. The principal vibrational modes of deactivation of the excited state of complex 3 are 90 (E v90 = 647.11 cm −1 ) and 116 (E v116 = 865.29 cm −1 ), where the asymmetric stretching modes of the C C ∧ N −C C ∧ N bonds and Ir−N C ∧ N bonds couple with the rocking mode of the equatorial S−F bonds and the equatorial out-ofplane vibration of the S atom of the −SF 5 group (cf. Table S10 in the Supporting Information for other minor contributing vibrational modes that couple with the spin density). Similar to what was observed for 2, these deactivation modes are less pronounced in complex 4, where the −SF 5 group is para to the C C ∧ N −Ir bond.
In order to assess their potential as emitters in OLEDs, the PL properties of 1−4 were investigated as doped PMMA thin films (PMMA = poly(methyl methacrylate)). The sample films were fabricated by spin-coating 5 wt % of the emitter in PMMA in chlorobenzene solutions in air. In doped thin films, the complexes exhibit phosphorescence behavior similar to that in solution with Φ PL values of 23−51% (Figure 5b and Table 3). The emission maxima of 1−4 in doped films are similar to those in solution. In the case of 1−4, the C ∧ N ligands mainly contribute to the T 1 states, as shown in Figure 6, and thus the changes of the molecular dipole orientation are relatively small upon photoexcitation. As the PMMA molecules around the iridium complexes do not change their dipoles in the solid state, the nature of the T 1 states remains unchanged, unlike the positive rigidochromic effect observed by Ikawa et al. 28 for iridium complexes containing an O ∧ O-based aromatic ancillary

■ EXPERIMENTAL SECTION
General Synthetic Procedures. Commercial chemicals were used without further purification. Ligands HL1 and HL2 were synthesized using a literature procedure. 26 All reactions were performed using standard Schlenk techniques under an inert (N 2 ) atmosphere with reagent grade solvents. Flash column chromatography was performed using silica gel (60 Å, 40−63 μm). Silica plates with aluminum backings (250 μm with indicator F-254) were used for analytical thin layer chromatography (TLC). Compounds were visualized under UV irradiation. 1 H (for ligands and dimers), 13 C, and 19 F NMR spectra were recorded on a Bruker Avance spectrometer at 400, 125, and 376 MHz, respectively. The following abbreviations have been used for multiplicity assignments: "s" for singlet, "d" for doublet, "t" for triplet, "p" for pentet, "m" for multiplet, and "br" for broad. Deuterated chloroform (CDCl 3 ) and deuterated dichloromethane (CD 2 Cl 2 ) were used as the solvents of record. 1 H and 13 C NMR spectra were referenced with respect to the NMR solvent peaks. An Electrothermal melting point apparatus was used to record melting points (mps). Mps were recorded in open-ended capillaries and are uncorrected. Thermogravimetric analysis (TGA) data were collected on a TA Instruments SDT 2960 apparatus. High-resolution mass spectra (HRMS) were recorded at the EPSRC UK National Mass Spectrometry Facility at Swansea University on a quadrupole timeof-flight (Q-TOF) instrument using a Model ABSciex 5600 Triple TOF in positive electrospray ionization (pESI) mode, and spectra were recorded using sodium formate solution as the calibrant. Elemental analyses were performed by Mr. Stephen Boyer, London Metropolitan University.

* S Supporting Information
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.inorgchem.7b01075. The research data supporting this work can be accessed at: http://dx.doi.org/10.17630/6da56570-6be6-4931-9a72-63144c352b0a.
NMR spectra of all C ∧ N ligands, dimers, and complexes and supplementary optoelectronic and DFT data of complexes 1−4 (PDF) for the gift of materials. We thank Dr. Nail Shavaleev for the synthesis of the complexes 1 and 2 in this study.