Sulfonyl-Substituted Heteroleptic Cyclometalated Iridium(III) Complexes as Blue Emitters for Solution-Processable Phosphorescent Organic Light-Emitting Diodes

: The synthesis is reported of a series of blue-emitting heteroleptic iridium complexes with phenylpyridine (ppy) ligands substituted with sulfonyl, ﬂ uorine, and/or methoxy substituents on the phenyl ring and a picolinate (pic) ancillary ligand. Some derivatives are additionally substituted with a mesityl substituent on the pyridyl ring of ppy to increase solubility. Analogues with two ppy and one 2-(2 ′ -oxyphenyl)pyridyl (oppy) ancillary ligand were obtained by an unusual in situ nucleophilic displacement of a ﬂ uorine substituent on one of the ppy ligands by water followed by N^O chelation to iridium. The X-ray crystal structures of seven of the complexes are reported. The photophysical and electrochemical properties of the complexes are supported by density functional theory (DFT) and time-dependent DFT calculations. E ﬃ cient blue phosphorescent organic light-emitting devices (PhOLEDs) were fabricated using a selection of the complexes in a simple device architecture using a solution-processed single-emitting layer in the con ﬁ guration ITO/ PEDOT:PSS/PVK:OXD-7(35%):Ir complex(15%)/TPBi/LiF/Al. The addition of a sulfonyl substituent blue-shifts the electroluminescence by ca. 12 nm to λ maxEL 463 nm with CIE x,y coordinates (0.19, 0.29), compared to the benchmark complex FIrpic ( λ maxEL 475 nm, 0.19, 0.38) in directly comparable devices, con ﬁ rming the potential of the new complexes to serve as e ﬀ ective blue dopants in PhOLEDs. Replacing a ﬂ uorine by a methoxy group in these complexes red shifts the PL and EL λ max by ca. 4 − 6 nm. The e ﬃ ciency of the blue PhOLEDs of the sulfonyl-substituted complexes is, in most cases, signi ﬁ cantly enhanced by the presence of a mesityl substituent on the pyridyl ring of the ppy ligands.


■ INTRODUCTION
Luminescent transition metal complexes are widely exploited in many areas, 1 such as biological labeling probes, ion sensors, water splitting, solar cells, and phosphorescent emitters for organic light-emitting diodes (OLEDs) 2−8 and for solid-state lighting. 9,10−13 Key features of the complexes include their synthetic versatility, good color tuneability, generally good chemical and photochemical stability, relatively short excited-state lifetimes on the microsecond time scale, and highly efficient emission from a mixture of triplet metal-to-ligand charge-transfer (MLCT) states and π−π* transitions of the ligands. 14Tris-homoleptic complexes [Ir(C^N) 3 ], where C^N is a monoanionic bidentate ligand, have been extensively studied, for example, the prototype green emitter fac-Ir(ppy) 3 (ppy = 2-phenylpyridine). 2The highest occupied molecular orbital (HOMO) is primarily localized on the phenyl π and iridium d orbitals, while the lowest unoccupied molecular orbital (LUMO) resides mainly on the pyridine ring.Electron-withdrawing substituents, typically fluorine, are attached to the phenyl ring of the ppy ligands to blue-shift the emission by lowering the HOMO energy without significantly changing the LUMO energy. 15,16or heteroleptic complexes [Ir(C^N) 2 L] the ancillary ligand L can also be used to tune the emission color.−22 Other structural modifications that give blue emission involve different ancillary ligands such as N-heterocyclic carbenes, 23 pyridylazolate derivatives, 24 fluorine-substituted bipyridine chelates, 25 and dicyclometalated phosphate tripod ligands. 26he motivation for the present work was provided by the high level of ongoing research to develop new blue-emitting phosphors. 27,28his article describes the synthesis and X-ray structural and optoelectronic characterization of a systematic series of analogues of FIrpic 1 in which the phenyl ring bears a sulfonyl substituent at the C3 position to further lower the HOMO energy and blue-shift the emission.Additionally, we report derivatives with one or both of the fluorine atoms at C2 and C4 replaced by a methoxy group.The rationale here is that substituting one fluorine with an alkoxy group is known to have a minimal effect on the emission color while increasing the chemical stability of the complex to unwanted nucleophilic substitution reactions 29,30 during, or after, complex formation. 31t is also known that di/multifluorinated ligands can degrade during OLED operation. 32In some of the complexes a mesityl substituent is attached to C4 of the pyridyl ring, as it has been reported that this functionalization leads to enhanced PhOLED efficiency, 33 which is also the case for several complexes in the present work.The experimental photophysical and electrochemical data are supported by time-dependent density functional theory (TD-DFT) calculations.Blue PhOLEDs were fabricated with selected complexes as the emitter.

■ RESULTS AND DISCUSSION
It has been established previously that functionalizing the phenyl ring of FIrpic 1 with an electron-withdrawing group at C3 (i.e., between the two F atoms) induces a blue shift in emission: examples are cyano, 34 sulfonyl, 35e phosphoryl, 35e and perfluoroalkyl carbonyl substituents.24b The present work focuses on sulfonyl groups, which have only rarely been incorporated as substituents on C^N cyclometallating ligands in either neutral 35 or ionic 36 complexes.With the exception of complex 11, which has been reported previously by Fan et al. using a different method to introduce the p-toluenesulfonyl substituent, 35e to our knowledge the sulfone-containing complexes studied in this work have not been synthesized previously.The synthetic routes to complexes 10−15, 23−25, and 31 are shown in Schemes 1−3.All the complexes are characterized by NMR spectroscopy, mass spectrometry, and elemental analysis, with additional single-crystal X-ray crystallographic data for seven of the complexes.
The key step to introduce the sulfone substituent involves lithiation of 2-(2,4-difluorophenyl)pyridine 2 37 or the mesityl analogue 3 with lithium diisopropylamide (LDA), followed by iodination to yield 4 and 5 and then reaction with either sodium p-toluenesulfinate or sodium methylsulfinate in the presence of stoichiometric copper(I) iodide, to afford ligands 6−8 in 15−26% yields (Scheme 1).These yields were highly reproducible in several reactions and could not be improved by ultilising a copper(II) bis(arylsulfinate) reagent, 38 a CuI/Lproline system, 39 or ionic liquid additives. 40On one occasion, when using a batch of 4 that had not been rigorously purified (and was probably contaminated with iodine from the previous reaction) the di(methanesulfonyl) derivative 9 was also isolated in 11% yield alongside product 6.When the reaction was repeated with rigorously purified 4 and the addition of a trace of iodine, compound 9 was again obtained in very low yield.Exploring the mechanism of this reaction is outside the scope of this work.However, it can be postulated: (i) that iodine coordinates to the pyridyl N atom of ligand 4 or 6 to form a more electron-withdrawing pyridinium unit, thereby further increasing the reactivity (by increased electron deficiency) of the phenyl ring, or (ii) that a fluorine in ligand 4 or 6 is replaced by iodine prior to the reaction with the methylsulfinate anion.There is precedent for a combination of fluoro and iodo substituents on a phenyl ring facilitating S N Ar reactions at a C− F bond. 41 The synthesis of methoxy analogues 23−25 followed a similar protocol starting from phenylpyridine derivatives 16 and 17, which were readily obtained by standard Suzuki reactions of the commercially available 2-fluoro-4-methoxyphenylboronic acid (Scheme 2).Notably, the conversions of 18 and 19 to yield methanesulfonyl derivatives 20 and 21 proceeded in significantly higher yields (45 and 53%, respectively) than the comparable reactions of the difluoro analogues 4 and 5 in Scheme 1.However, the yield of p-tolyl derivative 22 is still low (21%).
To explore the effect of replacing both fluorine atoms with methoxy substituents, complex 31 was synthesized.To obtain the dimethoxy ligand 30, a different synthetic approach was used starting from the known compound 26 42 (Scheme 3).In this case, the sulfone functionality was obtained by oxidation of the methylthio group of 29 with meta-chloroperbenzoic acid in 83% yield.
To obtain the picolinate complexes 10−13, 23−25, and 31 standard conditions were employed, 43  containing methoxy substituents using IrCl 3 •3H 2 O have been reported previously.30b Attempts to form the corresponding tris-cyclometalated homoleptic [Ir(C^N) 3 ] complexes from 6 and 7 using established conditions, 31d by adding more C^N chelate to the [Ir(L) 2 Cl] 2 species, gave the unexpected and unusual [Ir-(C^N) 2 (N^O)] complexes 14 and 15 in 29% and 23% yields, respectively.There was no sign of the homoleptic complexes after purification of the product mixture.The structures of 14 and 15 were tentatively assigned based on mass spectra combined with 19 F NMR spectra, which showed the presence of only five fluorine atoms, together with 1 H NMR and COSY spectra that established that a fluorine ortho to a pyridyl group had been replaced on one of the ligands.The structures were subsequently unambiguously confirmed by X-ray crystallography of 15 (see below).The phenyl ring of ligands 6 and 7 is highly electron-deficient due to the presence of pyridyl, fluorine, and sulfone substituents; therefore, an S N Ar reaction at a C−F bond is entirely reasonable, for example, reaction with residual water in the ethylene glycol sovent, with subsequent N^O chelation as a 2-(2′-oxyphenyl)pyridyl (oppy) ancillary ligand.We are not aware of a previous example of an Ir complex of an oxyphenylpyridyl ligand.Indeed, six-membered ring chelation to Ir with a bidentate N^O ancillary ring appears  The N atoms of the two C^N ligands are coordinated trans to each other.The picolinate ligand, as well as oppy ligand in 15, forms a longer Ir−N bond, due to trans-effect of the σbonded C atom.In molecule 12, the methylsulfonyl and mesityl groups of one C^N ligand are both disordered between two orientations in a 0.55:0.45ratio (the Mes by a flipping motion).The Mes groups are nearly perpendicular to the adjacent Py rings, the dihedral angles being 86°for the ordered Mes group, 82°and 73°for the disordered Mes groups.In structure 13, one whole C^N ligand is disordered between two orientations differing by a ca.10°libration around the Ir atom, with occupancies refined to 0.684(5) and 0.316 (5), respectively.All the chelating ligands show small but significant nonplanarity.In the picolinato ligand, the dihedral angle between the Py and carboxylate planes varies from 0.9°in 13 to 10.6°in 12.In the C^N ligands, the twists between the phenyl and pyridine rings are the highest in complex 23 (12.7°and 14.7°), in other complexes they vary from 2°to 11°.The oxyphenyl-pyridine moiety in 15 is much more twisted with the two aromatic rings forming an angle of 37°.
Thermal, Photophysical, and Electrochemical Properties.The absorption and emission spectra of the complexes in dichloromethane solution are shown in Figures 3 and S7−S10, and the data are listed in Table 2.The complexes show strong absorption bands in the 250−300 nm region derived from ligand-centered π−π* transitions. 17Absorption bands in the range of 350-450 nm with lower extinction coefficients are ascribed to singlet and triplet metal-to-ligand charge-transfer ( 1 MLCT and 3 MLCT) states, following literature precedents 17 and the calculations of Hay. 46Emission of the complexes is in the blue or blue-green region.The following trends in this series of complexes can be seen.
(i) Addition of the sulfone moiety results in a blue shift in emission λ max relative to FIrpic of ∼10 nm.The identity of the R group in −SO 2 R (tolyl or methyl) has little effect on the optoelectronic properties.(ii) Exchanging a fluoro for a methoxy substituent results in a red shift of ∼5 nm.Therefore, all of the monofluoro− monomethoxy complexes (23−25) have bluer emission than FIrpic, while the dimethoxy−sulfone complex 31 has a λ max and emission profile comparable to FIrpic.(iii) Addition of a mesityl group results in little change to λ max , as observed previously. 33 (iv) Introducing a second sulfone group (complex 13) results in a red shift in emission of 50 nm and a change in spectral profile, compared to FIrpic.There is literature precedent for a sulfone group inducing a red shift when introduced at the 4-position of the phenyl ring of ppy in a neutral iridium complex.35b (v) Emission from the N^O bonded complexes 14 and 15 is red-shifted by ca. 10 nm relative to their pic analogues (10 and 11).The photoluminescence quantum yield (PLQY) and lifetime data are stated in Table 2.The observed lifetimes (τ P = 1.49− 3.53 μs) and the spectral profiles are indicative of phosphorescence from a mixture of ligand-centered and MLCT excited states. 17,47The PL spectra of representative complex 24 obtained at 80, 120, 160, 200, and 240 K in degassed 2-MeTHF (Figure S11) show the expected sharpening of the emission bands and a blue shift on cooling due to reduction in vibrations.This is consistent with a significant 3 LC contribution to the excited state. 47he thermal stabilities of the iridium complexes were evaluated using thermogravimetric analysis under a nitrogen atmosphere.The 5% weight loss temperatures (T d ), are above 360 °C for all complexes and are comparable to FIrpic (370 °C), suggesting the complexes should be thermally stable under device operation (Table S2).
The electrochemical properties of the complexes were examined by cyclic voltammetry (CV) in a 0.1 M NBu 4 PF 6 acetonitrile solution.The observed reversible oxidation waves are assigned to the Ir(III)/Ir(IV) couple (Table 3, Figures 4  and S12).All difluoro−sulfone pic complexes (10−12) show an increase in oxidation potential of ∼0.28 V relative to that of FIrpic, indicating the sulfone group is effective at lowering the HOMO.Exchanging a fluorine for a methoxy group (23−25) results in a smaller increase in oxidation potential of ∼0.13 V relative to FIrpic, and the dimethoxy−sulfone complex 31 has an oxidation potential comparable to that of FIrpic.If the LUMO energies are assumed to be very similar in these complexes, the relative HOMO energies from their CV data are consistent with the λ max values obtained from the emission spectra.The oxidation potential of 13 is +1.25 V, suggesting that addition of the second sulfone moiety lowers the HOMO; however, it must also lower the LUMO significantly, as the emission spectrum is strongly red-shifted.Complexes 14 and 15 both show two irreversible oxidation waves; if the scan is reversed before the second oxidation is reached the first oxidation wave is still irreversible.This suggests a chemical reaction takes place upon oxidation and could be attributed to the unusual connectivity around the Ir center.No additional ligand-centered oxidation waves were detected on scanning to +2.0 V for the pic complexes, and no reduction waves were observed on scanning between 0.0 and −2.0 V.
Density Functional Theory Calculations.Electronic structure calculations were performed on the 11 iridium complexes studied here (including FIrpic 1) to explore the frontier orbitals and understand the nature of the transitions involved in the absorption and emission bands observed (Tables S3−S13).The full geometries of the iridium complexes were optimized at B3LYP/LANL2DZ:3-21G*, and the optimized geometries are denoted as 1′, 10′−15′, 23′−25′, and 31′ to identify them as computed models and distinguish the predicted data from experimental data.Comparison between optimized and X-ray determined geometries reveal good agreement for the seven Ir bond distances with differences below 0.04 Å (Table S14).The complexes 10′−12′, 23′−25′, and 31′ have orbital contributions broadly similar to FIrpic 1′ with the HOMO on iridium and phenyl (ppy) and the LUMO on the pyridyl of the picolinate ligand (Figure 5).
The lower HOMO energies on the complexes containing the sulfonyl groups, compared to FIrpic, reflect the electronwithdrawing properties of the former.Frontier orbital energies are listed in Table 3 for direct comparison with observed CV data.The observed and computed HOMO energies are in very good agreement.The electron-donating methoxy groups in 23′−25′ increase the HOMO energies in 23′−25′ relative to 10′−12′.b The values reported are the cathodic peak potentials observed on the first scan.sulfonyl groups; thus, 31′ has remarkably similar orbital energies and distributions to FIrpic.The LUMOs in 10′−12′, however, do have substantial contributions from the pyridyl groups of the ppy ligands, which lower their orbital energies compared to FIrpic.Nevertheless, the effect of the sulfonyl groups on the HOMO energies is larger than on the LUMO energies in complexes 10′−12′ and 23′−25′ compared to the FIrpic frontier orbitals.Therefore, the HOMO−LUMO energy gaps (HLGs) in these complexes are larger than the HLG in FIrpic, and their emissions would be more blue-shifted than FIrpic as observed experimentally.
Compound 13′ with four sulfonyl groups has a similar HOMO as the other iridium complexes discussed above, but its LUMO is located on the ppy ligand(s) not on the picolinate ligand.The different LUMO is due to the four electronwithdrawing sulfonyl groups on the ppy ligands, which would be more easily reduced than the picolinate ligand.In 13′, the effect of the sulfonyl groups on the LUMO energy increases; thus, its HLG energy is closer to that of FIrpic than that of complexes 10′−12′.While the HOMO−LUMO gap for 13 is larger than expected, the TD-DFT prediction for 13′ (see below) is 490 nm compared to 1′ at 471 nm.Thus, this supports the experimental observation that the low-energy absorption of 13 is red-shifted compared to 1.
The HOMOs for the oppy-containing complexes, 14′ and 15′, are located at the oppy group bonded to iridium rather than on iridium and phenyl (ppy), as is the case for the HOMOs of the other complexes here.The different HOMOs result in higher HOMO energies compared to 10′−12′, and therefore 14′ and 15′ have the smallest HLG energies of all iridium complexes calculated here.The relatively high HOMO energies of 14′ and 15′ and different HOMOs to the other iridium complexes are reflected in the irreversible CV waves observed at lower oxidation potentials here.The LUMOs for 14′ and 15′ are on the ppy ligands.
TD-DFT computations were performed on the S 0 -optimized geometries of 1′, 10′−15′, 23′−25′, and 31′ to predict emission wavelengths of these iridium complexes.The initial excitation is assumed to give the lowest-energy singlet excited-state S 1 , which in the presence of the iridium results in intersystem crossing (ISC) to form the triplet excited state T 1 , and phosphorescence is observed from the S 0 ←T 1 process.The small Stokes shifts observed for most complexes suggest that the T 1 geometry is similar to the corresponding S 0 geometry.The reverse process S 0 ←T 1 is thus considered to have the same nature as the computed S 0 →T 1 process with the predicted emission wavelength adjusted to take into account the Stokes shift.
Table 4 compares the predicted S 0 ←T 1 wavelengths with observed emission wavelengths for all iridium complexes, where the agreement is excellent except for complexes 13′−15′.The emissions of complexes 13−15 are likely to be from different triplet excited-state geometries as shown in their subtly lower PLQYs and longer lifetimes than the other iridium complexes,  a Values from TD-DFT data on S 0 -optimized geometries with scaling energy factor of 0.945 based on dichloromethane at 298 K. b Observed highest-energy band from emission spectra (Table 2).c Observed emission is unlikely via ISC from S 1 to T 1 , as this ISC pathway may not be promoted due to a smaller iridium contribution in the HOMO.Different triplet states with higher energies are assumed to be responsible for emissions of 14 and 15.
1, 10−12, 23−25, and 31.In the case of 13, the larger Stokes shift observed suggests substantial rearrangement of the geometry in the T 1 excited state.The emissions from 14 and 15 may be from different triplet excited states of higher energies via ISC, as the HOMOs of 14′ and 15′ contain less iridium character (Table 4), and their S 1 excited states may not result in ISC to the predicted T 1 excited states.The emission intensities for FIrpic and 15 were recorded at different excitation wavelengths (10 nm intervals between 380 and 430 nm, Figures S8 and S9) to examine why the predicted T 1 excited-state energies for 14 and 15 differ from the observed emission data.These measurements reveal a significant decrease in the emission intensities for 15 compared to those for FIrpic as the excitation wavelengths lengthen.The observed decrease implies a different ISC pathway in 15 compared to FIrpic and thus a different triplet excited state in 15 to the predicted T 1 state.
Optimized geometries at the T 1 excited state for 10′, 13′, and 14′ were performed to examine the differences in the T 1 geometries compared to their corresponding optimized S 0 geometries.From T 1 vertical energies on both S 0 and T 1 geometries, the energy differences are 0.27, 0.30, and 0.28 eV  for 10′, 13′, and 14′, respectively, meaning that the T 1 geometry differs (i.e., rearranges) from its S 0 geometry in 13′ more than in 10′ and 14′.These energy values support the larger Stokes shift energy observed (Table 2) for 13 compared to the measured Stokes shift energies for all other iridium complexes.
The spin densities at the iridium atom in the optimized T 1 geometries for 10′, 13′, and 14′ are 0.36, 0.46, and 0.13, respectively.The low spin-density value of 0.13 for 14′ suggests that a different triplet excited state with higher energy is responsible for the phosphorescence in 14 and, by implication, 15.The emission wavelengths in 14 and 15 would therefore be shorter than the predicted S 0 ←T 1 emission wavelengths, as observed experimentally.The higher spin density at the iridium atom in the optimized T 1 geometry for 13′ compared to that for 10′ may explain the less vibronic structure emission observed in 13 compared to that in 10.An increased vibronic structure in the observed emission usually means an increased ligand contribution (and decreased metal character) to the excited state.
Phosphorescent Organic Light-Emitting Devices.PhOLEDs were fabricated using FIrpic 1 and complexes 10− 12 and 23−25 as the dopant phosphor.Complexes 13−15 and 31 were not studied in devices as their PL spectra are redshifted compared to the other analogues, and our focus is on obtaining blue devices.The device architecture comprised a simple single-emissive-layer, which was a blend of poly-(vinylcarbazole) (PVK) as the host material, OXD-7 (an electron-transporting material), and the Ir complex to give the architecture: ITO/PEDOT:PSS (45 nm)/PVK:OXD-7(35%)/ Ir complex(15%)(60 nm)/TPBi (30 nm)/LiF/Al.The emissive layer was spin-coated from chlorobenzene solution to avoid potential degradation of the complexes that can occur during thermal evaporation, especially for complexes with fluorinated ligands. 32An additional electron-transporting layer of TPBi was incorporated adjacent to the cathode to optimize charge balance and confine excitons in the emitter layer. 49The efficiency and luminance data of the devices are summarized in Table 5.The electroluminescence (EL) spectra for FIrpic 1 and complexes 10−12 and 23−25 are shown in Figure 6a.Several trends are apparent.
(1) For all the complexes the EL and PL spectra are similar, indicating good exciton confinement on the emissive molecules.
(2) The devices of 11, 12, 24, and 25 performed better than those of 10 and 23 both in terms of external quantum efficiency (EQE) and brightness (Figure 6c,e).For 12, 24, and 25 this can be attributed to the presence of the mesityl group on the ppy ligands, which has been shown previously to enhance device performance. 33The increased solubility of 11 (p-tolylsulfonyl substituent) compared to 10 (methylsulfonyl analogue) gave rise to better film quality for 11, which could explain its enhanced performance.(3) The FIrpic devices are the most efficient in this series.
This can be explained by a superior trapping efficiency for the FIrpic device compared to the other complexes; the HOMO level of the PVK host is −5.8 eV, which aligns well with FIrpic (−5.69 eV, observed; −5.49eV, calculated: Table 3).However, the HOMO levels of complexes 10−12 and 23−25 are equal to that of PVK or lower (Table 3), resulting in less efficient hole trapping.The lower efficiency for 10−12 and 23−25 is also consistent with the known trend that a rapid decline in efficiency is induced upon blue-shifting the emission color. 50(4) We also note that the extent of the efficiency roll-off is directly proportional to the current density and brightness for most of the devices.Efficiency roll-off in OLED devices has been comprehensively discussed. 51he higher efficiency, which is reproducible for complex 11 at ca. 1 mA/cm 2 (Figure 5d), is typical of some OLEDs at the device turn-on due to a slight change in material morphological defects caused by the initial current passing through the device. 52

■ CONCLUSIONS
In summary, a series of heteroleptic Ir complexes [Ir-(ppy) 2 (pic)] has been synthesized with sulfonyl substituents attached to C3 of the phenyl ring of the ppy ligands.Compared to the benchmark sky-blue emitter FIrpic, the additional electron-withdrawing sulfonyl substituent results in a blue shift of the λ max of PL and EL by up to ca. 12 nm as a result of the increased HOMO−LUMO gap.Replacing a fluorine atom by a methoxy group red shifts the λ max of PL and EL by ca.4−6 nm.Two complexes with an unusual 2-(2′-oxyphenyl)pyridyl ancillary ligand have been obtained by an in situ nucleophilic displacement of a fluorine substituent on a ppy ligand.(TD-)DFT calculations are in excellent agreement with the observed photophysical and electrochemical properties of the complexes.PhOLEDs have been fabricated using a selection of the complexes in a simple device architecture with a solutionprocessed single-emitting layer in the configuration ITO/ PEDOT:PSS/PVK:OXD-7(35%)/Ir complex(15%)/TPBi/ LiF/Al.The EL data confirm the potential of the new complexes to serve as blue dopants in functional devices with a notably lower CIE C y coordinate compared to FIrpic.The work provides further examples where the efficiency of blue PhOLEDs is enhanced by the presence of a mesityl substituent on the pyridyl ring of the ppy ligands.

■ EXPERIMENTAL SECTION
Materials, Synthesis, and Characterization.All commercially available chemicals were used without further purification.Reactions requiring an inert atmosphere were performed under a blanket of argon gas, which was dried over a phosphorus pentoxide column.Anhydrous solvents were dried through an HPLC column on an Innovative Technology Inc. solvent-purification system.Column chromatography was performed using 40−60 μm mesh silica gel.Analytical thin-layer chromatography (TLC) was performed on plates precoated with silica gel (Merck, silica gel 60F 254 ) and visualized using UV light (254, 315, 365 nm).NMR spectra (Figures S13−S34) were recorded on Bruker Avance 400 MHz, Varian Mercury 200 and 400 MHz, Varian Inova 500 MHz, or Varian VNMRS 600 and 700 MHz spectrometers.Chemical shifts are referenced to tetramethylsilane [Si(CH 3 ) 4 ] at 0.00 ppm.Melting points were determined in openended capillaries using a Stuart Scientific SMP3 melting point apparatus at a ramping rate of 1 °C/min.They are recorded to the nearest 0.1 °C.Electrospray ionization (ESI) and matrix-assisted laser desorption/ionization mass spectra were recorded on a Thermo-Finnigan LTQ FT (7.0 T magnet) spectrometer.Atmospheric solids analysis probe mass spectra were recorded on a Waters Xevo QTOF spectrometer.Gas chromatography mass spectrometry (GCMS) spectra were recorded on a Thermo-Finnigan Trace GCMS (EI and CI ion sources).Elemental analyses were obtained on an Exeter Analytical Inc. CE-440 elemental analyzer.HPLC was performed on an analytical/semiprep Varian LC system equipped with UV−vis and ES + mass detectors.
Solution Electrochemistry and Photophysics.Cyclic voltammetry experiments were recorded using a BAS CV50W electrochemical analyzer fitted with a three-electrode system consisting of a Pt disk (Ø = 1.8 mm) as the working electrode, a Pt wire as an auxiliary electrode, and a Ag/AgNO 3 (0.1 M[NEt 4 ][ClO 4 ] in CH 3 CN) system as the reference electrode.Experiments were conducted in dry acetonitrile solution with n-Bu 4 NPF 6 (0.1 M) as the supporting electrolyte at a scan rate of 100 mV/s.The reference electrode was assumed to be stable and was referenced externally to ferrocene (Cp 2 Fe) and decamethylferrocene (Cp* 2 Fe), which displayed potentials (E 1/2 ) of −0.41 V and +0.10 V, respectively, versus Ag/AgNO 3 under these conditions.It was not possible to use Cp 2 Fe or Cp* 2 Fe as an internal reference because their addition to the solutions of the complexes resulted in a distortion of the redox waves of the complexes.
Solution-state photophysical data were obtained using freshly prepared solutions of the complexes in dichloromethane (DCM).Emission and lifetime measurements used thoroughly degassed solutions achieved by three freeze−pump−thaw cycles and quartz cuvettes with a path length of 1 cm.The solutions had absorbance below 0.10 to minimize inner filter effects.UV−vis absorption measurements were recorded using a Unicam UV2−100 spectrometer operated with Unicam Vision(ver.3.50) software.Baseline correction was achieved by reference to pure solvent in the same cuvette.Absorption measurements were obtained using quartz cuvettes with a path length of 2 cm.
Excitation and emission spectra were recorded on a Jobin−Yvon− Horiba SpexFluoromax 3 Spectrometer.Solutions of the complexes in degassed DCM [<1 × 10 −5 M] were used for decay measurements.Samples were excited with a pulsed nitrogen laser emitting at 337.1 nm.Emission was focused onto a spectrograph and detected on a sensitive gated iCCD camera (Stanford Computer Optics) with subnanosecond resolution.Solution PLQYs were recorded in degassed solvent and determined using the relative method, with Ir(ppy) 3 (Φ PL = 0.46 in degassed dichloromethane, determined in house, vs Φ PL = 0.546 in 0.5 M H 2 SO 4 ) as the reference.The PLQYs were calculated according to the following equation: where subscripts "x" and "ref" denote the material being measured and the reference, respectively.Φ represents the PLQY, Grad is the gradient from the plot of integrated fluorescence intensity versus absorbance, and η is the refractive index of the solvent.Computational Studies.All calculations were performed with the Gaussian 09 package. 53All optimized S 0 geometries were performed using B3LYP 54 with the pseudopotential (LANL2DZ) 55 for iridium and 3-21G* basis set for all other atoms. 56This model chemistry was selected on the basis of related computational studies on [Ir-(ppy) 2 Cl] 2 57 and Ir(ppy) 3 . 58All S 0 -optimized geometries of the most stable conformers were true minima based on no imaginary frequencies found.Electronic structure and TD-DFT calculations were also from the optimized geometries at B3LYP/LANL2DZ:3-21G*. The MO diagrams and orbital contributions were generated with the aid of Gabedit 59 and GaussSum 60 packages, respectively.The TD-DFT method does not include spin−orbit couplings; thus, no oscillation strengths for the triplet excited states are given and also does not compute mixed singlet−triplet excited states.Optimized triplet excited-state (T 1 ) geometries were also performed on 10′, 13′, and 14′ to examine significant changes between their S 0 and T 1 geometries to explain the differences in observed and computed emission values for 13 and 14.Optimized triplet excited-state (T 1 ) geometries by TD-DFT can be considered unreliable, 58 so our interpretations are based on the assumption that these geometries are accurate.
Phosphorescent Organic Light-Emitting Devices.PhOLEDs were fabricated on indium tin oxide (ITO)-coated glass substrates of thickness 125 nm and possessing a sheet resistance of 15 Ω/x.
Poly(3,4-ethylenedioxythiophene) doped with high work function hole injection layer poly(styrenesulfonic acid) (PEDOT:PSS) (HIL1.3),from CLEVIOS, was spin-coated at 2500 rpm for 60 s to produce an ∼50 nm thick hole-injecting/transporting layer (HTL).The PEDOT:PSS layer was annealed at ca. 200 °C for 5 min to remove any residual water.A chlorobenzene solution of 25 mg/mL of poly(vinylcarbazole) (PVK; Mw = 90 000) was doped with 30% w/w of (1,3-phenylene)bis[5-(4-tert-butylphenyl)-1,3,4-oxadiazole] (OXD-7).Blended devices were made by mixing 15% w/w of the Ir complexes.The prepared mixtures were filtered with a 0.45 μm pore filter and spin-coated at 2500 rpm for 1 min on the top of the PEDOT:PSS layer and baked for 10 min at 120 °C.Each sample was shadow masked to produce four identical devices of area 4 × 5 mm; the samples were then introduced into a nitrogen glovebox, where 30 nm of 2,2′,2″-(1,3,5-benzenetriyl)tris-[1-phenyl-1H-benzimidazole] (TPBi) was evaporated as an electron injection/hole blocking layer at a rate of ∼1 Å/s under vacuum at a pressure of ca. 1 × 10 −6 Torr, followed by 0.8 nm LiF and a 100 nm capping layer of aluminum under the same evaporation conditions.Therefore, the device configuration for all complexes was: ITO/PEDOT:PSS (50 nm)/ PVK:OXD-7 (35%)/Ir complex (15%)/TPBi (30 nm)/LiF(0.8nm)/ Al(100 nm).All samples were encapsulated inside a glovebox using DELO UV cured epoxy (KATIOBOND) and capped with a 1.2 × 1.2 cm microscope glass slide then exposed to UV light for 3 min.Current−voltage data, device efficiency, brightness, and electroluminescence spectra were measured in a calibrated Labsphere LMS-100 integrating sphere.A home-written NI LabVIEW program was used to control an Agilent 6632B DC power supply, and the emission properties of the device were measured using an Ocean Optics USB4000 CCD fiber optic spectrometer.Thicknesses of the various layers in the device were measured with a J. A. Woolam VASE Ellipsometer using thin films that had been spin-coated on Si/SiO 2 substrates under the same conditions as the device films.
General Procedure for ortho Lithation Followed by Iodination.n BuLi (1.2 equiv) was added to a stirred solution of N,Ndiisopropylamine (DIPA; 1.3 equiv) in dry tetrahydrofuran (THF) at 0 °C, and the solution was stirred for 30 min.The temperature was then lowered to −78 °C, and a solution of the substrate (1 equiv) in dry THF was added in portions via cannula; the mixture was stirred for a further 1 h.A solution of I 2 (1.1−1.3 equiv) in dry THF was added to the flask dropwise.The solution was stirred at −78 °C and allowed to warm to room temperature (RT) overnight.The solution was then quenched with an aqueous solution of NaS 2 O 3 to remove unreacted LDA and I 2 .The solution was then extracted with DCM (3 × 75 mL), and the organic phases were combined, dried over MgSO 4 , and filtered; the solvent was reduced in vacuo.The residue was passed through a silica plug (eluent: DCM/ethyl acetate) to remove baseline impurities revealed by TLC.The solvent was removed, and the crude product was either triturated with cold hexane or purified by column chromatography, to remove any unreacted starting material.The product was then dried on a high vacuum line to leave the pure product.
General Procedure for Sulfone Synthesis.The iodo compound (1 equiv), the sodium sulfinate derivative (1.6 equiv) and CuI (1.5 equiv) were dissolved in dry dimethylformamide (DMF) and stirred under argon.The mixture was heated at 110 °C overnight.The mixture was cooled to RT, and water (20 mL) and DCM (20 mL) were added.The solid was filtered off and washed with more DCM.The organic layer was separated and dried over MgSO 4 , and the solvent was removed in vacuo.The residue was purified by column chromatography to give pure product, unless otherwise specified.This is in accordance with the method described in the literature. 38eneral Procedure for Synthesis of Iridium pic Complexes.Method A. IrCl 3 •3H 2 O (1 equiv) was added to a stirred solution of the cyclometalating ligand (2.1−2.2 equiv) in 2-ethoxyethanol.The mixture was heated to reflux at 130 °C under argon overnight.The mixture was then left to cool, and the solvent was removed in vacuo.The yellow solid, presumed to be the intermediate bis(μ-Cl) dimer complex, was washed with cold hexane to remove any free ligand and dried.Where possible, the bis(μ-Cl) dimer was characterized by 1 H NMR and mass spectrometry, although attempted purification usually led to decomposition.The dimer was then suspended in 2ethoxyethanol, and picolinic acid (<3 equiv) was added.The mixture was stirred and heated to reflux (130 °C) under argon overnight.The mixture was allowed to cool to RT, and the solvent was removed.The product was then purified by column chromatography to give pure product, unless otherwise specified.
Method B. The reaction was conducted via a one-pot method, and the intermediate bis(μ-Cl) dimer was not isolated.

* S Supporting Information
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.inorgchem.6b01179.
Thermal data; photophysical data; cyclic voltammograms; computational data; copies of NMR spectra (PDF) X-ray crystallographic data including files in CIF format with CCDC Nos.1413092−1413097 and 1474568 (ZIP)

Notes
The authors declare no competing financial interest.

Figure 2 .
Figure 2. (a).X-ray molecular structures of 11−13 and 15.Minor disorder components, solvent of crystallization, and all H atoms are omitted; thermal ellipsoids are drawn at the 50% probability level.

Table 2 .
Photophysical Data of the Iridium Complexes Data obtained in less than 10 μM dichloromethane solutions at 20 °C.b Data obtained in degassed dichloromethane solution with excitation wavelength at 380 nm.c Energy difference between lowest energy excitation maxima and highest energy excitation maxima.d Measured relative to Ir(ppy) 3 Φ PL = 0.46 in degassed dichloromethane at 20 °C; estimated error ±5%.

Table 3 .
Electrochemical and Computed Molecular Orbital Energy Data of Iridium Complexes

Table 4 .
Iridium Atom Contributions in HOMOs and Predicted Emission Wavelengths of Iridium Complexes

Table 5 .
Summary of Device Luminescence and Efficiency Data a Measured at a brightness of 10 cd/m 2 .b Measured at 12 V.