Bis(arylimidazole) Iridium Picolinate Emitters and Preferential Dipole Orientation in Films.

The straightforward synthesis and photophysical properties of a new series of heteroleptic iridium(III) bis(2-arylimidazole) picolinate complexes are reported. Each complex has been characterized by nuclear magnetic resonance, UV-vis, cyclic voltammetry, and photoluminescent angle dependency, and the emissive properties of each are described. The preferred orientation of transition dipoles in emitter/host thin films indicated more preferred orientation than homoleptic complex Ir(ppy)3.

Introduction

In the search for efficient light-emitting devices, organic light-emitting diodes (OLEDs) offer a full-spectrum inexpensive option because of the development of heavy atom emitters. Spin–orbit coupling (SOC) is strong in third-row elements and allows for theoretical device efficiency to be near unity by utilizing both singlet and triplet excitons for light generation.[1,2] Ir complexes have been successful thus far, thanks to room-temperature (rt) emission, strong SOC that allows efficient phosphorescence from its[3] MLCT state, μs time range excited state lifetime, and high photoluminescence quantum yield (PLQY).[3−6]

Ir(III) emitters exhibiting emission of every color, especially green and red, have been synthesized to date. Concerning green OLEDs, emitters such as Ir(ppy)2(acac) (Figure ) have delivered up to 54% external quantum efficiency (EQE) in devices.[7] Complexes such as Ir(MDQ)2(acac) (Figure ) (MDQ = methyldibenzo[f,h]quinoxaline) can provide >20% EQE in red-emitting OLED devices.[8] Quinoline-based ligands are better electron acceptors than pyridine because of the resonance stabilization energy of the phenyl moiety.[9] Furthermore, the additional π electrons and reduced donation power of the quinoline moiety in comparison to pyridine contribute to the stabilization of the emissive triplet state.

Figure 1

Structures of state-of-the-art iridium complexes and iridium complexes investigated in this study.

Emitters in the blue region, however, have been rare because of the high-energy excited state.[5] The current state- of- the- art blue emitters have structures such as (dfppy)2Ir(pic), called FIrPic) [dfppy = 2-(2,4-difluorophenyl)pyridine and pic = picolinate],[10,11] Ir(1-Aryl-NHC)3,[12] [NHC = N-heterocyclic carbene], or Ir(2-arylimidazole)3, as in Ir816 (Figure )].[13,14] Each class of emitters has its drawbacks, such as low color purity, poor stability, or poor tunability. With FirPic, for example, the strongly electron-withdrawing difluorophenyl moiety stabilizes the molecular ground state (with electron density primarily on the cyclometallated phenyl ring) more than the excited state (with electron density primarily on the pyridine ring).[15,16] Such energy level tuning widens the highest occupied molecular orbital–lowest unoccupied molecular orbital (HOMO–LUMO) gap, thus increasing emission energy to the blue region. Selective functionalization of either cyclometallating phenyl or picolinate ligand has been shown to modify emission energy, PLQY, excited state lifetime, and so forth.

In homoleptic (1-Aryl-NHC) iridium complexes, sufficiently withdrawing NHC moieties can harbor a high-energy excited state, leading to blue emission. However, in homoleptic (2-arylimidazole) iridium complexes, the excited state lies on the wingtip aryl moiety, leading to blue emission.[17] Both these types of complexes are synthetically more challenging to modify than arylpyridine complexes. Accordingly, many fewer studies have investigated the chemistry of these high-performing ligands,[18,19] especially in heteroleptic complexes.[20,21] The goal of this investigation was to develop a deeper understanding of emitter structure–property relationships in green- to red-emitting heteroleptic 1-arylimidazole iridium complexes. We report the synthesis and characterization of a series of functionalized heteroleptic bis(2-arylimidazole) iridium(III) picolinate complexes and indicate photophysical ramifications for ligand functionalization. This manuscript has been previously deposited on ChemRxiv.[22]

Results and Discussion

Synthesis of Iridium Complexes

To experimentally establish a structure–function relationship for this class of ligands, we envisioned a series of complexes with functionalization on both wingtip and cyclometallating phenyl moieties (Scheme ). Wingtip-modified ligands were reached by first synthesizing the appropriate imidazole with reproducible yields of nearly 50% and then coupling the appropriate aryl bromide through a key Negishi coupling reaction. This procedure reproducibly yielded each of the imidazole ligands 3–5 with good yields of 65–87%.

Scheme 1

Synthesis of Ir Complexes 6–14

After forming chloride-bridged iridium dimers with each ligand at high temperatures in acceptable yields of 52–80%, heteroleptic picolinate complexes were obtained after ligation with commercially available picolinate ligands to form complexes 9–12 with yields between 15 and 92%.

After synthesis and chromatography, X-ray quality single crystals of complexes 11, 12, and 14 were grown from a concentrated solution of CDCl3, and the solved structures are shown in Figure . Details of single-crystal X-ray diffraction (SC-XRD) experiments are given in the Supporting Information, along with CIF files. In the crystal structure of each complex, the Ir atom adopts the expected distorted octahedral geometry. Planar 2-arylimidazole ligands are perpendicular with chelating nitrogen atoms opposite to each other, which is also observed in (ppy)2Ir(pic) complexes.[23,24] Bond angles and distances were typical of other Ir(III) structures in a distorted octahedral, with no significant steric interactions.

Figure 2

ORTEP plot of the crystallographically determined structures of complexes 11 (left), 12 (center), and 14 (right) (thermal probability ellipsoids are drawn at the 50% probability level). Hydrogen atoms and solvated molecules are omitted for clarity.

Optoelectronic Characterization

Once the complexes were synthesized, we characterized them by cyclic voltammetry to better understand the structure–function relationship (Table , Figure ). Complex 10, with the least-accepting cyclometallating ligand, had the most destabilized ground state oxidation potential (E(S+/S)) value, while complex 11, with the least-accepting wingtip functionality, showed no change compared to complex 9 with the unmodified ligand.

Table 1

Electrochemical and Optical Properties of Ir Complexes 9–14

complex	εa (104 mol L–1 cm–1)	λmaxa (emission) (nm)	E(T1)b(eV)	E(S+/S)c (V vs NHE)	EHOMOd(eV vs vacuum)	ELUMOeste (eV vs vacuum)
9	316 (1.32), 336 (1.30), 376 (0.75), 402 (0.53)	571	2.59	1.13	–5.73	–3.14
10	336 (1.54), 377, (0.98), 400 (sh, 0.74), 463 (0.025)	577	2.43	1.05	–5.65	–3.22
11	314 (1.11), 342 (0.94), 371 (sh, 0.57), 402 (0.33)	571	2.59	1.13	–5.73	–3.14
12	318 (1.64), 335 (sh, 1.49), 372 (0.90), 399 (0.61)	555	2.65	1.12	–5.72	–3.07
13	338 (1.28), 368 (1.13), 397 (sh,), 462 (0.12)	654	2.30	1.18	–5.78	–3.48
14	317 (1.50), 337 (sh, 1.27), 372 (0.79), 402 (0.54)	555	2.70	1.11	–5.71	–3.01

Measured in degassed MeCN.

Value obtained by taking the high-energy onset of the emission spectrum and converting to eV using the formula E(T) = 1240 (eV/nm)/λonset (nm).

Measured in a 0.1 M Bu4NPF6 in MeCN solution with the glassy carbon working electrode, Pt reference electrode, and Pt counter electrode with ferrocene as an internal standard. Values estimated by taking Epa and are reported versus NHE.

Value obtained from the equation EHOMO = −4.6 eV – E(S+/S).

Value estimated by adding EHOMO and the E(T1).

Figure 3

Measured in a 0.1 M Bu4NPF6 in MeCN solution with a glassy carbon working electrode, Pt reference electrode, and Pt counter electrode with ferrocene as an internal standard. Values are reported vs NHE using the conversion Fc/Fc+ = 0.64 V vs NHE in MeCN.[25]

This indicated that modification of the wingtip aryl group does not affect the molecular ground state. Complex 14, with the most accepting picolinate ancillary ligand, displayed the most stabilized E(S+/S) at 1.11 V versus NHE. Decreasing the electron-accepting power of the picolinate complex through substitution with more inductive and resonance donating functional groups as in complexes 9, 12, and 13 led to successively most destabilized E(S+/S) values (from 1.12 for 12, 1.13 for 9, and 1.18 for 13), which indicated that the picolinate directly affected the molecular ground state energy, as is the case for FIrPic and its derivatives.

To understand how structural changes affected emission, we measured the UV–vis and photoluminescence of each complex dissolved in MeCN (Table , Figure ). Overall, the UV–vis absorbance spectra looked qualitatively similar for each of the complexes, as was expected. Looking more closely at the spectra, subtle differences are apparent.

Figure 4

UV–vis and normalized photoluminescence spectra of complexes 9–14 dissolved in Ar-saturated MeCN.

For instance, compared to complex 9 (unmodified ligand), decreasing the electron-accepting power of the cyclometallating ligand (complex 10) increased absorptivity, whereas modifying the imidazole wingtip (complex 11) decreased absorptivity. Modifying the picolinate ligand with methoxy functionality largely had no effect on absorptivity, whereas trifluoromethyl functionalization increased absorptivity and introduced a pronounced band at 460 nm.

The steady-state photoluminescence spectra of complexes 9–14 in MeCN were measured to understand the changes to the emissive state upon ligand modification. Adding electron density to the cyclometallating ligand led to a slight red shift in the emission spectra λmax value (577 nm for 10 vs 571 nm for 9), whereas modifying the imidazole wingtip had no effect on λmax (complex 11 vs 9).

Modifying the picolinate ligand was found to have a much more drastic effect, as increasing picolinate-accepting power led to an 83 nm red shift (654 nm for complex 13 vs 571 nm for 9). Decreasing accepting power through the addition of methoxy or methyl led to a less-pronounced blue shift (17 nm) for both complexes (complexes 12 and 14 vs complex 9). From the photoluminescence data, it is clear that picolinate ligand modifications affect emission more than absorption, and that the emission energy can be easily tuned by modifying the picolinate electron donating and accepting strengths. Because of its intrinsic accepting ability, the LUMO likely has high electron density on the picolinate ligand, as was previously described for very similar complexes of (ppy)2Ir(pic) derivatives.[24] Photoluminescence data support this hypothesis because emission energy should be lowered by increasing the accepting power of the LUMO-bearing moiety.[16]

Thus, from the electrochemical measurements, it appeared that the molecular HOMO (E(S+/S)) was likely localized on the Ir and the cyclometallating moiety of the imidazole ligand (as has been observed in calculations of homoleptic complexes).[17] The optical measurements indicate that the emissive triplet state (E(T1)) is likely composed of a mixed MLCT state, with very little ligand-centered triplet state character in the emissive state, as evidenced by the featureless charge transfer band present for each complex.[11]

The energy levels of complexes 9–14 are compiled in Figure for comparison. From the graph, the ramifications of ligand modification are easier to visualize, especially in the context of band gap energy. For instance, adding electron density to the cyclometallating ligand in complex 10 destabilized the E(S+/S) energy level more than E(LUMOest), lowering the band gap energy. Conversely, in complexes 12 and 13, the E(S+/S) level was slightly destabilized, but E(LUMOest) was destabilized to a greater degree, leading to overall band gap widening. In complex 13, both E(S+/S) and E(LUMOest) were stabilized, with much more stabilization occurring at the E(LUMOest) energy level.

Figure 5

Energy-level diagram for complexes 9–14. The E(S+/S) energy level was measured by cyclic voltammetry, and the value was converted to the vacuum scale by the equation E(HOMO) (eV) = −4.6 – E(S+/S) (V vs NHE). The ELUMOest energy level estimated as the high-energy onset of emission in degassed MeCN. The optical band gap value is the difference between the EHOMO and ELUMOest energy levels.

Once we had measured the relevant energy levels for each of the emitters, we then turned to gaining a better understanding of photophysical properties of the complexes in solution and in a variety of matrices. Our first set of measurements involved PLQY (Φ) data for the complexes (Table ). While not an IUPAC standard, we chose to reference our PLQY values in solution to a state-of-the-art emitter we had readily available, FIrPic (PLQY = 92% in dichloroethane, DCE).[26,27]

Table 2

PLQY Measurements of Complexes 9–14

complex	9	10	11	12	13	14	Ir816
PLQY (Φsolution, %)a	11	14	9	24	6	10	50b
PLQY (Φsolid, %)c				32	10	34
PLQY (Φsolid, %)d				23	7	16
PLQY (Φsolid, %)e				28	10	28	59
PLQY (Φsolid, %)f				19	3	18

Values taken from measurements made in DCE. Measured relative to FIrPic.[26]

Literature value.[31]

Measurement made in the solid state, 26DCZ(ppy) matrix.

Measurement made in the solid state, TcTa + OXD-7 matrix.

Measurement made in the solid state, XTM014 matrix.

Measurement made in the solid state, PMMA matrix.

As has been described for a series of recently published heteroleptic phenylpyridine acetylene-substituted picolinates, the observed PLQY values in solution were much lower than those for FIrPic but similar to those for Ir(ppy)2(pic) (14.7%).[24] In the reported Ir(ppy)2(pic) complexes, solution excited state lifetime measurements were not indicative of the pure radiative lifetime (τ0), and so, we chose to focus on PLQY in solution. Complex 12 exhibited the highest PLQY (24%), and complex 13 exhibited the lowest (6%). Complex 10 displayed higher PLQY than complexes 9, 11, and 14 (14% for 10 compared to 11% for 9, 9% for 11, and 10% for 14). Complexes 9–14 also exhibited extremely weak oxygen-sensitive phosphorescence in MeCN, as has been observed for other complexes.[28,29] Thus, we concluded that substitution on the wingtip and the cyclometallating ring had little effect on PLQY. Similarly, while methyl substitution on the ancillary ligand greatly enhanced PLQY, more electron-rich methoxy and strongly accepting trifluoromethyl both decreased PLQY in comparison to complex 9.

Because their PLQY values in solution exhibited the most variation, we chose to study the PLQY of emitters 12–14 when dissolved in a variety of host matrices (Table and Supporting Information). The PLQY values for Ir816 were much higher in solution and in XTM014 than for complexes 12–14. Contrary to what was observed in solution, complexes 12 and 14 exhibited similar PLQY values in each of the matrices. Complex 13 was a much less efficient emitter in each trial, perhaps largely because of the energy gap law. PLQY values for each complex in 26DCZ(ppy) were highest overall and lowest in poly(methyl methacrylate-co-butyl methacrylate) (PMMA). These results indicate that the local environment in PMMA favors nonradiative pathways, especially in comparison to 26DCZ(ppy).[30] Thus, it was observed that even though PLQY in PMMA (and solution) was low, judicious host selection could boost PLQY values.

Angle-Dependent Photoluminescence

The average orientation of transition dipole moments in thin films of emitters embedded in a host matrix is a useful and interesting aspect to analyze.[32] In the past years, it became more clear that orienting the transition dipole moments of the emitters horizontally, in plane with the substrate, can improve the light outcoupling efficiency of OLEDs and therefore their EQE.[33] Several factors can influence the orientation of the emitters, such as the deposition technique,[34] host material,[35] particular processing conditions,[36] and molecular shape of the emitter itself.[37] We analyzed the orientation of the average transition dipole moments of compounds 9, 10, and 11 in two different host matrices, poly(9-vinylcarbazole) (PVK) and poly(4-butylphenyl-diphenyl-amine) (poly-TPD). After preparing the solutions (9:1 host/guest ratio), we deposited them on quartz substrates, and we measured the transition dipole moments of the emitters from thin films. The experimental angular dependencies of the p-polarized photoluminescence intensities were used as the input for optical simulations using the software OrientExpress.[38] The angular emission profile of the vertical and horizontal radiating dipoles is simulated based on the thickness and (anisotropic) refractive indices of the layers and the substrate. Hereby, an even distribution of the emitters across the emitting layer is assumed, and the intensities of the contributing dipoles are added accordingly.

The relative weights of the contributions of horizontal and vertical dipoles to the emission profile are then obtained from the measured one by simple linear regression. The comparison of the experimental data with the optical simulations allows one to quantify the degree of the orientation of the transition dipole moments of the emitters in the host. The average orientation of the transition dipole moments can be represented by Θ, the ratio of the horizontal component of the transition dipole moments to the total transition dipole moments. For a fully random distribution of transition dipole moments, Θ = 0.66.

As a reference, we use the homoleptic Ir(ppy)3, which is randomly oriented in evaporated layers.[39] As for 9, 10, and 11, we mix the isotropic emitter in both PVK and poly-TPD (9:1 host/guest ratio) for comparison.

When films have the same thickness and the same host material, PL angle-dependent graphs can be roughly interpreted by comparing the PL at large (55°) angles with the PL at small angles (0°).

Horizontal dipoles emit more at small angles (as the dipole oscillating in a plane emits preferentially in directions perpendicular to it), while vertical dipoles emit more at high angles. Therefore, the layers with higher PL at 55° have average transition dipole moments less horizontally oriented. Figure shows the PL angle dependency of the emitters analyzed in PVK, both experimental measurements and simulated fit (used to quantify the orientation). The transition dipole moments of Ir(ppy)3, the isotropic reference material, are the most vertically oriented, followed by 10. 9 and 11 show the same behavior, and their transition dipole moments are the most horizontally oriented. A similar behavior is observed for the emitters in poly-TPD, apart from the fact that in poly-TPD, complex 10 has transition dipole moments slightly more vertically oriented than Ir(ppy)3. Table shows the orientation parameters quantified by combining the experimental data with optical simulation. Here, we can see that all the emitters perform better in poly-TPD rather than in PVK (they have higher Θ), and that 9 and 11 are more favorably oriented compared to the reference isotropic emitter and to 10. However, all the values for Θ are lower than 0.66, which indicates a vertical orientation of the average transition dipole moment.

Figure 6

PL angle dependency of Ir(ppy)3, 9, 10, and 11 blended in PVK (9:1 host/guest ratio). Experimental data (symbols) and simulated fit (lines).

Table 3

Orientation of Transition Dipole Moments of Ir(ppy)3, 9, 10, and 11, Blended in a Host Matrix of PVK or Poly-TPD

	PVK (Θ)	PolyTPD (Θ)
Ir(ppy)3	0.55	0.6
9	0.59	0.62
10	0.57	0.59
11	0.59	0.63

The apparent preferential vertical orientation of transition dipole moments of Ir(ppy)3 in the polymer hosts is surprising: for this homoleptic complex, one would expect fully random orientation because of its almost isotropic shape. The preferential vertical orientation of Ir(ppy)3 has also been reported for another unoriented, solution-deposited host material.[34] We note that the preference for the vertical orientation is the largest when using the PVK host material. Below, we briefly discuss a possible contribution to the apparent vertical orientation of the iridium compounds stemming purely from the optical anisotropy of the PVK host.

Solution-deposited layers of PVK have an unusual, positive uniaxial anisotropy (ne > no)[40] that may be rationalized in terms of a preferential horizontal orientation of the polymer backbone[41] with the highly polarizable carbazole side groups orienting such that the long axis of the carbazole moiety is perpendicular to the main chain. Taking into account all possible orientations of the main chain and the carbazole side groups, one then expects the film to be most polarizable in the direction perpendicular to the plane of the film, explaining why ne > no. For molecules in a dielectric medium, it is well-known that the polarizability of the medium can couple with the oscillating transition charge density on a dye molecule to produce an effective transition dipole moment μeff that can be larger than the transition dipole moment μbare of the bare molecule. The so-called empty cavity model[42] predicts:

The empty cavity model has been found to give an accurate description of the changes in the radiative rate constant for small metal organic luminophores[33] induced by the polarizability of the environment. If we now take into consideration a uniaxial optically anisotropic medium (no ≠ ne), then based on eq , one may expect an apparent preferential orientation of the effective transition dipole moment of an isotropically oriented dye in the direction corresponding to the largest value of the refractive index. In a pragmatic approach, one predicts preferential orientations of the effective transition dipole moment that deviate from the isotropic value by about a percent by simply inserting values for ne and no into 1. Assuming a slightly simpler relation between the effective transition dipole moment and the refractive index[43] (μeff ∞ n), one predicts slightly larger deviations from the isotropic values for the orientation, but the differences are still smaller than 4%. In summary, it is conceivable that the positive optical anisotropy of the PVK host contributes to the apparent vertical orientation of the Ir(ppy)3 emitter, but on the basis of the bulk dielectric constant, we cannot quantitatively account for the observed deviations. This may indicate the importance of specific molecular interactions between the emitter and the host.

Compared to the homoleptic Ir(ppy)3 emitter with its virtually isotropic shape, the heteroleptic emitters 9 and 11 show an inclination toward more horizontal preferential orientation of transition dipole moments in both polymeric hosts investigated (see Table ). Yet, on an absolute scale, the values of Θ obtained are indicative of a largely random orientation of the transition dipole moment of these luminophores.

Conclusions

Finding new, more efficient emitters for OLEDs is an important challenge in chemistry because lighting devices consume a large amount of energy worldwide. Toward this end, emitting molecules must be rationally designed to fit in to high-performance device configurations. Toward the aim of a deeper understanding of the ramifications of emitter functionalization, we synthesized a series of heteroleptic bis(2-arylimidazole) iridum(III) picolinates with different steric and electronic modifications, from simple alkyl substitution on the cyclometallating ring or the imidazole wingtip group to electronic modification of the picolinate ligand. The complexes were characterized by nuclear magnetic resonance (NMR), UV–visible absorption spectroscopy, photoluminescence spectroscopy, and mass spectrometry, and three complexes were characterized by SC-XRD. Three of the complexes were blended in two different host matrices, and the obtained thin films were characterized by photoluminescent angle dependency to measure the orientation of their average transition dipole moment. The other three complexes were used as dopants in host matrices, and the PLQY of these films was measured. It was determined that ligand substitution played a major role in the relative position of molecular energy levels (and thus emission wavelength) and the observed PLQY. Establishing the structure–property relationship in various phosphorescent emitters for OLEDs is important, as better design rules will come from a larger pool of known complexes. Further studies into structure–property relationships in Ir(III) emitters are currently underway.

Experimental Section

General Information

All commercially obtained reagents were used as received. Unless otherwise noted, all reactions were performed under an N2 atmosphere. Thin-layer chromatography (TLC) was conducted with Sigma T-6145-precoated TLC silica gel 60 F254 aluminum sheets and/or visualized with UV and potassium permanganate staining. Flash column chromatography was performed as described by still using Silicycle P60, 40–63 μm (230–400 mesh).[44]1H NMR spectra were recorded on Bruker AVIII-HD (400 or 200 MHz), and they are reported in ppm using solvent as an internal standard (CDCl3 at 7.26 ppm). Data are reported as follows: s = singlet, d = doublet, t = triplet, q = quartet, p = pentet, m = multiplet, b = broad, ap = apparent; coupling constant(s) in Hz; integration. UV–vis spectra were measured with an LS-55 spectrometer. Cyclic voltammetry was measured with a Bio-Logic S-200 potentiostat. Mass spectra were recorded on a Bruker autoflex MALDI-TOF or ESI mass spectrometer. 2,4,6-Triisopropylaniline was prepared according to the literature procedure.[45] SC-XRD measurements were conducted on a Bruker VENTURE SCXRD instrument. CCDC 1583689 (11), 1584183 (12), 1583690 (13) contain the supplementary crystallographic data for this paper. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre. For the analysis of the orientation of transition dipole moments, three of the Ir complexes synthesized were mixed each with two different host materials, PVK and poly(4-butylphenyl-diphenyl-amine) (poly-TPD). The host and guest materials were blended in a 9:1 ratio in chlorobenzene (0.5 wt %) and deposited on quartz substrates by spin-coating at 600 or 800 rpm, forming 18 to 33 nm thick films. The angle-dependent emission of the emitters in the host/guest system was measured in thin films by using a setup described previously.[38,46] Both layer preparation and angle-dependent measurements were performed in an inert N2 atmosphere (H2O and O2 < 1 ppm) to avoid photochemical degradation and photoluminescence quenching induced by combination of oxygen and UV light. To quantify the orientation of transition dipole moments of the emitters, the experimental measurements were combined with optical simulation by using the OrientExpress software, described in detail in the Supporting Information of ref (38). 2,6-Bis(3-(9H-carbazol-9-yl)phenyl)pyridine (26DCzppy), 4,4′,4″-tris(carbazol-9-yl)triphenylamine, and 1,3-bis[2-(4-tert-butylphenyl)-1,3,4-oxadiazo-5-yl]benzene (OXD-7) were purchased from lumtec. PMMA and chlorobenzene were purchased from Sigma-Aldrich.

The optical constants (refractive index, n, and extinction coefficient, k) of the pure host materials, PVK and poly-TPD, were determined experimentally by spectroscopic ellipsometry at various angles of incidence and transmission intensity measurements at 0° (perpendicular incidence). Optical constants for the host material also containing the molecular emitter (10 wt %) were assumed to be the same as the pure host. Optical constants were determined using WVASE 32 software from Woollam. Uniaxial optical anisotropy was taken into consideration, and ordinary and extraordinary optical constants were extracted from ellipsometry data of the pure host polymers. These data were used as the input for the optical simulations to quantify the orientation of transition dipole moments.

Photoluminescence Characterization

Each complex (10 wt %) was mixed in different host matrixes. The films were deposited from 10 mg mL–1 solution of complexes in chlorobenzene. Prior to deposition, all solutions were filtered with a 0.2 μm pore size filter and spin-coated at 1500 rpm for 60 s in inert atmosphere onto cleaned quartz substrates. The thin-film photoluminescence spectra and quantum yields were measured in air with a Hamamatsu C9920-02 absolute PLQY measurement system.

General Procedure for Imidazole Formation

1-(2′,6′-Diisopropylphenyl)imidazole (1)[47]

A 1 L round-bottom flask equipped with a stirbar was charged with 750 mL MeOH and sparged vigorously with N2. After 20 min, 2,6-diispropylaniline (20 mL, 18.8 g, 106 mmol) and glyoxal (14.4 mL, 106 mmol, 40% in H2O) were added and stirred at rt. After 16 h, large yellow crystals were observed. To this suspension, NH4Cl (5.79 g, 212 mmol) was added, a reflux condenser was attached, and the solution was warmed to reflux. Once the crystals had dissolved, formaldehyde (16 mL, 212 mmol, 37.5% in H2O) was added, and heating continued. After 1.5 h, 85% H3PO4 (25 mL) was added, and the reaction was further refluxed and monitored by TLC. After 6 h, the flask was cooled down to rt, and then to 0 °C, and pH was adjusted to 9 with 15% KOH. The organic phase was extracted with Et2O (3 × 250 mL), washed with H2O (3 × 500 mL), and concentrated on a rotary evaporator after drying with MgSO4. It was submitted to column chromatography using 650 mL SiO2 and gradient hexanes to 30% EtOAc/hexanes. The product spot was concentrated to yield an off-white solid (12.4 g, 51% yield). Analytical data matched literature values. 1H NMR (400 MHz, CDCl3): δ 7.53 (s, 1H), 7.49 (d, J = 8 Hz, 1H), 7.30 (d, 7.1 Hz, 3H), 6.99 (s, 1H), 2.45 (m, J = 4.8 Hz, 2H), 1.19 (d, J = 4.8 Hz, 12H).

1-(2′,4′,6′-Triisopropylphenyl)imidazole (2)

Reaction performed as for compound 1, yield: 5.0 g, 52%. 1H NMR (400 MHz, CDCl3): δ 7.52 (s, 1H), 7.49 (s, 1H), 7.27 (s, 2H), 6.77 (s, 1H), 2.85 (m, J = 4.8 Hz, 1H), 2.55 (m, J = 6.9 Hz, 2H), 1.28 (d, J = 6.9 Hz, 6H), 1.13 (d, J = 7.0 Hz, 6H), 1.10 (d, J = 4.8 Hz, 6H). 13C NMR (100 MHz, CDCl3): δ 150.5, 146.3, 138.8, 130.8, 129.4, 121.8, 35.5, 28.3, 24.6, 24.5, 24.2.

General Procedure for Negishi Coupling

1-(2′,6′-Diisopropylphenyl)-2-phenylimidazole (3)[47]

A flame-dried Schlenk flask equipped with a stirbar was charged with compound 1 (1.0 g, 4.34 mmol) and dry THF (5 mL) and allowed to cool to 0 °C with stirring. After 15 min, n-BuLi (1.74 mL, 2.5 M in hexanes) was added via a syringe and stirred at 0 °C. After 30 min, ZnCl2 (3.2 mL, 1.9 M in 2Me-THF) was added via a syringe and stirred at 0 °C. After 30 min, solution was allowed to warm to rt, and then, it was warmed to 40 °C. The solution was concentrated to a total volume of 4 mL, and to that were added dry toluene (4 mL), iodobenzene (0.84 g, 0.46 mL, 4.1 mmol), and Pd(PPh3)4 (0.20 g, 0.17 mmol, 4% catalyst loading). The solution was warmed to 90 °C with stirring and was monitored by TLC. After 21 h, the reaction was cooled to rt, and H2O (5 mL) was added. It was washed with 10% HCl (25 mL), dried with MgSO4, concentrated on a rotary evaporator, and submitted to column chromatography using 400 mL SiO2 and gradient hexanes to 20% EtOAC/hexanes. Product fractions were concentrated to yield a beige solid (1.1 g, 87%). Analytical data matched literature values. 1H NMR (400 MHz, C6D6): δ 7.82 (d, J = 7.1 Hz, 2H), 7.45 (d, J = 0.9 Hz, 1H), 7.18 (t, J = 7.8 Hz, 1H), 7.00 (d, J = 7.9 Hz, 2H), 6.95 (at, J = 6.9 Hz, 2H), 6.91 (ad, J = 7.0 Hz, 1H), 6.67 (d, J = 1.1 Hz, 1H), 2.58 (m, J = 6.9 Hz, 2H), 1.06 (s, 9H), 0.97 (d, J = 6.9 Hz, 6H), 0.83 (d, J = 6.9 Hz, 6H).

1-(2′,4′,6′-Triisopropylphenyl)-2-phenylimidazole (4)

Reaction performed as for compound 3, yield: 1.3 g, 65%. 1H NMR (400 MHz, CDCl3): δ 7.42–7.39 (m, 2H), 7.29 (s, 1H), 7.18–7.16 (m, 3H), 7.07 (s, 2H), 6.93 (s, 1H), 2.96 (m, J = 6.9 Hz, 1H), 2.44 (m, J = 6.9 Hz, 2H), 1.30 (d, J = 7.0 Hz, 6H), 1.10 (d, J = 6. Hz, 6H), 0.88 (d, J = 7.0 Hz, 6H). 13C NMR (100 MHz, CDCl3): δ 150.7, 147.3, 145.7, 132.2, 130.7, 129.0, 128.4, 128.3, 127.4, 124.0, 122.4, 34.6, 31.9, 28.5, 25.4, 24.3, 23.1, 22.9.

1-(2,6-Diisopropylphenyl)-2-(4″-tert-butylphenyl)imidazole (5)

Reaction performed as for compound 3, yield: 1.2 g, 81%. 1H NMR (400 MHz, C6D6): δ 7.79 (d, J = 7.8 Hz, 2H), 7.45 (d, J = 1.1 Hz, 1H), 7.23 (t, J = 7.8 Hz, 1H), 7.08 (at, J = 8.7 Hz, 4H), 6.71 (d, J = 1.1 Hz, 1H), 2.58 (m, J = 6.8 Hz, 2H), 1.06 (s, 9H), 0.97 (d, J = 6.9 Hz, 6H), 0.83 (d, J = 6.9 Hz, 6H). 13C NMR (100 MHz, C6D6): δ 150.6, 146.8, 146.0, 135.1, 129.6, 129.4, 128.5, 127.8, 127.7, 127.5, 127.4, 126.8, 124.9, 124.11, 123.0, 34.1, 30.8, 28.1, 24.6, 22.5.

General Procedure for Chloro-Bridged Iridium Dimer Formation

Di-μ-chlorobis(1-(2′,6′-diisopropylphenyl)-2-phen-2″-ylimidazol-3-yl)diiridium (6)

A sealable pressure tube equipped with a stirbar was charged with compound 3 (0.63 g, 2.05 mmol) and tridecane (5 mL) and sparged with N2. After 35 min, [Ir(COD)Cl]2 (0.32 g, 0.47 mmol) was added, and the tube was sealed under N2. The tube was allowed to be warmed to 240 °C with stirring and was monitored by TLC. After 15 h, yellow ppt was observed. A crude reaction was submitted directly to column chromatography using 350 mL SiO2 and a gradient of hexanes to 20% EtOAc/hexanes. The product spot was concentrated to yield a bright yellow solid (0.60 g, 77%). 1H NMR (400 MHz, CDCl3): δ 7.76 (s, 4H), 7.57 (t, J = 7.6 Hz, 4H), 7.45–7.38 (m, 8H), 6.51–6.32 (m, 8H), 6.92 (s, 4H), 6.52 (t, J = 7.2 Hz, 4H), 6.34 (m, 8H), 6.06 (d, J = 7.5 Hz, 4H), 3.0–2.83 (m, 8H), 1.37–1.21 (m, 42H), 0.94 (d, J = 6.8 Hz, 16H). 13C NMR (100 MHz, CDCl3): δ 158.9, 146.8, 146.7, 135.0, 135.1, 133.7, 131.9, 130.6, 127.9, 124.8, 124.6, 121.7, 120.2, 29.0, 28.7, 25.2, 24.6, 24.0, 23.4.

Di-μ-chlorobis(1-(2′,4′,6′-triisopropylphenyl)-2-phen-2″-ylimidazol-3-yl)diiridium (7)

Reaction was performed as for compound 7, yield: 0.265 g, 52%. 1H NMR (400 MHz, C6D6): δ 8.17 (d, J = 1.5 Hz, 4H), 7.22 (t, J = 7.7 Hz, 4H), 7.13 (dd, J = 7.4 Hz, 1.5 Hz, 4H), 7.06 (dd, J = 7.7, 1.4 Hz, 4H), 6.89 (d, J = 1.5 Hz, 4H), 6.86 (dd, J = 7.8, 1.1 Hz, 4H), 6.69 (dt, J = 6.6, 1.5 Hz, 4H), 6.51 (dt, J = 7.3, 1.2 Hz, 4H), 6.42 (dd, J = 7.7, 1.5 Hz, 4H), 3.01 (m, J = 6.9 Hz, 8H), 1.13 (at, J = Hz, 24H), 1.02 (d, J = 6.9 Hz, 12H), 0.77 (d, J = 6.9 Hz, 12H). 13C NMR (100 MHz, CDCl3): δ 158.9, 151.1, 146.7, 146.4, 146.2, 145.1, 135.3, 131.9, 131.5, 129.0, 128.9, 127.8, 127.7, 122.6, 122.5, 122.4, 121.7, 120.6, 120.4, 119.7, 35.0, 34.7, 31.9, 29.1, 28.8, 27.2, 25.6, 25.2, 25.1, 24.7, 24.4, 24.3, 24.1, 23.5, 22.9. HRMS (ESI-MS): Theo. for C96H117ClIr2N8, [M – Cl]+, 1802.8353; found for C96H117ClIr2N8, [M – Cl]+, 1802.8354.

Di-μ-chlorobis(1-(2′,6′-diisopropylphenyl)-2-(4″-tertbutylphen-2″-yl)imidazol-3-yl)diiridium (8)

Reaction was performed as for compound 7, yield: 0.235 g, 80%. 1H NMR (200 MHz, CDCl3): δ 7.83 (m, 2H), 7.63–7.26 (m, 18H), 6.89 (bs, 4H), 6.51–6.32 (m, 8H), 5.91 (bs, 4H), 2.85 (bs, 4H), 1.54 (s, 24H), 1.30 (s, 24H), 1.04 (s, 45H). 13C NMR (100 MHz, CDCl3): δ 159.3, 150.0, 149.9, 147.4, 147.3, 146.9, 146.3, 133.8, 131.9, 130.5, 129.1, 128.7, 128.4, 127.4, 124.6, 124.6, 121.1, 119.9, 117.0, 34.5, 34.5, 31.7, 31.6, 31.6, 28.9, 28.8, 23.9, 23.8, 23.8, 25.1.

General Procedure for Heteroleptic Ir(III) Picolinate Synthesis

Bis(1-(2′,6′-diisopropylphenyl)-2-phen-2″-ylimidazol-3-yl)iridium(III) Picolinate (9)

A 25 mL round-bottom flask equipped with a stirbar and a reflux condenser was charged with dichloromethane (15 mL) and compound 6 (0.25 g, 0.015 mmol) and sparged with N2. After 10 min, 2-picolinic acid (0.006 g, 0.045 mmol) and NEt3 (0.006 g, 0.011 mL, 0.075 mmol) were added, and the reaction was warmed to reflux with stirring and monitored by TLC. After 18 h, the yellow solution was concentrated and passed through an SiO2 plug, first using 20% EtOAc/hexanes and then using acetone. Product fractions were concentrated to yield a bright yellow solid (0.060 g, 65%, based on picolinic acid). 1H NMR (400 MHz, CDCl3): δ 8.34 (d, J = 7.5 Hz, 1H), 7.91–7.83 (m, 2H), 7.53 (td, J = 2.5 Hz, 7.2 Hz, 4H), 7.38–7.24 (m, 8H), 6.88 (d, J = 1.6 Hz, 1H), 6.81 (d, J = 1.7 Hz, 1H), 6.64–6.58 (m, 5H), 6.47–6.38 (m, 2H), 6.34 (d, J = 1.7 Hz, 1H), 6.1 (d, J = 7.8 Hz, 1H), 6.05 (d, J = 7.8 Hz, 1H), 2.81–2.56 (m, 4H), 1.35–0.9 (m, 24H). 13C NMR (100 MHz, CDCl3): δ 173.9, 159.6, 158.2, 153.4, 149.1, 148.1, 147.1, 146.6, 146.4, 146.3, 145.3, 137.0, 135.5, 135.4, 133.7, 133.2, 133.1, 133.0, 130.8, 130.7, 128.3, 128.2, 127.7, 127.1, 126.0, 125.3, 124.9, 124.8, 124.8, 124.7, 124.6, 122.8, 122.3, 121.8, 121.6, 120.5, 120.1, 28.8, 28.7, 28.5, 25.2, 24.5, 24.0, 23.8, 23.6, 23.5.

Bis(1-(2′,6′-triisopropylphenyl)-2-(4″-tert-butylphen-2″-yl)imidazol-3-yl)iridium(III) Picolinate (10)

Reaction was performed as for compound 9 with the exception of 2 equiv picolinate and NEt3 used, yield: 0.020 g, 15%. 1H NMR (400 MHz, CDCl3): δ 8.29 (d, J = 7.9 Hz, 1H), 7.82 (t, J = 7.7 Hz, 1H), 7.64 (d, J = 5.2 Hz, 1H), 7.52 (t, J = 7.7 Hz, 2H), 7.38–7.26 (m, 6H), 7.21 (t, J = 6.4 Hz, 1H), 6.83 (s, 1H), 6.81 (s, 1H), 6.77 (s, 1H), 6.54 (d, J = 8.2 Hz, 1H), 6.47 (d, J = 8.2 Hz, 1H), 6.39 (s, 1H), 6.03 (d, J = 8.2 Hz, 1H), 5.99 (d, J = 8.2 Hz, 1H), 2.85 (m, 1H), 2.73 (m, 1H), 2.49 (m, 1H), 1.85 (m, 1H), 2.05 (m, 4H), 1.26–0.8 (m, 68H). 13C NMR (100 MHz, CDCl3): δ 173.9, 159.8, 158.4, 153.7, 151.1, 150.3, 148.4, 147.6, 147.5, 146.7, 146.6, 146.5, 145.0, 136.6, 133.1, 133.1, 132.7, 132.6, 131.0, 130.7, 127.7, 126.8, 125.9, 125.1, 124.9, 124.7, 124.4, 122.3, 121.8, 121.3, 121.2, 117.9, 117.3, 34.5, 31.7, 31.7, 31.6, 30.0, 28.9, 28.8, 28.6, 28.4, 25.3, 25.2, 25.1, 24.5, 23.8, 23.7, 23.7. HRMS (ESI-MS): Theo. for C56H67IrN5O2, [M + H]+, 1034.4919; found for C56H67IrN5O2, [M + H]+, 1034.4930.

Bis(1-(2′,4′,6′-triisopropylphenyl)-2-phen-2″-ylimidazol-3-yl)iridium(III) Picolinate (11)

Reaction was performed as for compound 9 with the exception of 11 equiv picolinate and 17 equiv Net3, yield: 0.050 g, 92%. 1H NMR (400 MHz, CDCl3): δ 8.26 (d, J = 7.7 Hz, 1H), 7.83–7.76 (m, 2H), 7.20–7.05 (m, 6H), 6.79 (s, 1H), 6.72 (s, 1H), 6.81 (d, J = 1.5 Hz, 1H), 6.54–6.30 (m, 7H), 6.25 (s, 1H), 6.06–5.94 (m, 2H), 2.94 (m, 3H), 2.05 (m, 4H), 1.25–0.7 (m, 36H). 13C NMR (100 MHz, CDCl3): δ 173.9, 159.6, 158.3, 153.5, 151.4, 151.2, 149.0, 148.1, 146.6, 146.1, 146.0, 145.9, 145.4, 136.9, 135.7, 135.5, 133.7, 133.2, 130.7, 128.1, 127.7, 127.0, 125.8, 125.2, 122.8, 122.7, 122.6, 122.5, 122.4, 122.3, 121.9, 121.8, 120.4, 120.0, 34.6, 29.9, 28.8, 28.8, 28.6, 25.2, 24.7, 24.7, 24.5, 24.3, 24.1, 23.9, 23.7, 23.6. HRMS (ESI-MS): Theo. for C54H63IrN5O2, [M + H]+, 1006.4606; found for C54H63IrN5O2, [M + H]+, 1006.4619.

Bis(1-(2′,6′-triisopropylphenyl)-2-(phen-2″-yl)imidazol-3-yl)iridium(III) 4-Methylpicolinate (12)

Reaction was performed as for compound 9 with the exception of 10 equiv picolinate and NEt3 used, yield: 0.048 g, 43%. 1H NMR (400 MHz, CDCl3): δ 8.19 (s, 1H), 7.75 (d, J = 5.5 Hz, 1H), 7.54 (m, 2H), 7.36 (m, 5H), 7.08 (dd, J = 5.5, 2.0 Hz, 1H), 6.90 (d, J = 1.5 Hz, 1H), 6.82 (d, J = 1.5 Hz, 1H), 6.58 (m, 4H), 6.47 (t, J = 6.8 Hz, 1H), 6.42 (d, J = 6.8 Hz, 1H), 6.37 (d, J = 1.6 Hz, 1H), 6.12 (d, J = 8.2 Hz, 1H), 6.06 (d, J = 8.2 Hz, 1H), 2.79 (m, 1H), 2.73 (m, 1H), 2.65 (m, 1H), 2.46 (s, 3H), 2.20 (m, 1H), 1.24 (d, J = 6.9 Hz, 3H), 1.22 (d, J = 6.9 Hz, 3H), 1.17 (d, J = 7.0 Hz, 3H), 1.11 (d, J = 6.9 Hz, 3H), 1.03 (d, J = 7.1 Hz, 3H), 0.95 (m, 9H). 13C NMR (100 MHz, CDCl3): δ 174.1, 159.6, 158.3, 152.8, 148.9, 148.6, 148.4, 147.1, 146.6, 146.5, 146.3, 145.6, 135.5, 135.4, 133.7, 133.2, 133.1, 133.1, 130.8, 130.7, 128.6, 128.2, 128.1, 127.8, 126.0, 125.3, 124.8, 124.8, 124.6, 124.5, 122.7, 122.2, 121.6, 121.6, 121.5, 120.3, 120.0, 28.8, 28.8, 28.7, 28.5, 25.2, 24.7, 24.6, 24.5, 24.0, 23.8, 23.6, 21.5. HRMS (ESI-MS): Theo. for C49H53IrN5O2, [M + H]+, 936.3823; found for C49H53IrN5O2, [M + H]+, 936.3842.

Bis(1-(2′,6′-triisopropylphenyl)-2-(phen-2″-yl)imidazol-3-yl)iridium(III) 4-Trifluoromethylpicolinate (13)

Reaction was performed as for compound 9 with the exception of 10 equiv picolinate and NEt3 used, yield: 0.047 g, 40%. 1H NMR (400 MHz, CDCl3): δ 8.60 (s, 1H), 8.10 (d, J = 5.5 Hz, 1H), 7.55 (m, 2H), 7.42 (dd, J = 5.1, 2.1 Hz, 1H), 7.38–7.30 (m, 5H), 6.92 (d, J = 1.5 Hz, 1H), 6.84 (d, J = 1.5 Hz, 1H), 6.62 (m, 2H), 6.54–6.41 (m, 4H), 6.32 (d, J = 1.7 Hz, 1H), 6.11 (d, J = 8.0 Hz, 1H), 6.07 (d, J = 8.2 Hz, 1H), 2.79 (m, 1H), 2.80–2.56 (m, 3H), 2.15 (m, 1H), 1.21 (at, J = 6.9 Hz, 6H), 1.15 (d, J = 6.9 Hz, 3H), 1.10 (d, J = 7.0 Hz, 3H), 1.05 (d, J = 6.9 Hz, 3H), 1.03 (d, J = 7.1 Hz, 3H), 0.93 (m, 9H). 13C NMR (100 MHz, CDCl3): δ 172.4, 159.4, 158.2, 155.4, 150.1, 147.1, 147.0, 146.6, 146.4, 144.0, 138.4 (q, J = 34.9 Hz), 135.4, 135.2, 133.4, 133.2, 133.0, 132.9, 131.0, 130.8, 128.5, 128.3, 125.9, 125.2, 125.0, 124.9, 124.8, 124.6, 124.1, 122.9, 122.4, 122.1, 121.9, 121.3, 120.9, 120.6, 28.8, 28.8, 28.8, 28.8, 28.6, 28.6, 25.2, 24.7, 24.4, 24.0, 23.8, 23.7, 23.5. HRMS (ESI-MS): Theo. for C49H50IrN5O2F3, [M + H]+, 990.3540; found for C49H50IrN5O2F3, [M + H]+, 990.3562.

Bis(1-(2′,6′-triisopropylphenyl)-2-(phen-2″-yl)imidazol-3-yl)iridium(III) 4-Methoxypicolinate (14)

Reaction was performed as for compound 9 with the exception of 10 equiv picolinate and NEt3 used, yield: 0.066 g, 58%. 1H NMR (400 MHz, CDCl3): δ 7.91 (d, J = 2.9 Hz, 1H), 7.64 (d, J = 6.1 Hz, 1H), 7.52 (m, 2H), 7.38–7.29 (m, 5H), 6.89 (d, J = 1.6 Hz, 1H), 6.81 (d, J = 1.5 Hz, 1H), 6.76 (dd, J = 6.1, 1.7 Hz, 1H), 6.61–6.51 (m, 4H), 6.46–6.37 (m, 3H), 6.09 (d, J = 8.0 Hz, 1H), 6.03 (d, J = 8.2 Hz, 1H), 3.93, (s, 3H), 2.75–2.55 (m, 3H), 2.2 (m, 1H), 1.21 (at, J = 6.9 Hz, 6H), 1.15 (d, J = 6.9 Hz, 3H), 1.09 (d, J = 7.0 Hz, 3H), 1.03 (d, J = 6.8 Hz, 3H), 1.03 (d, J = 7.1 Hz, 3H), 0.93 (m, 9H). 13C NMR (100 MHz, CDCl3): δ 173.8, 166.6, 159.7, 155.3, 149.5, 148.3, 147.1, 146.7, 146.5, 146.4, 145.6, 135.5, 133.7, 133.2, 133.1, 133.2, 130.8, 130.7, 128.2, 128.1, 126.0, 125.4, 124.9, 124.5, 122.7, 122.3, 121.6, 120.3, 119.9, 114.7, 112.0, 56.18, 28.8, 28.7, 28.5, 25.2, 24.7, 24.5, 24.0, 23.9, 23.7, 23.5. HRMS (ESI-MS): Theo. for C49H53IrN5O3, [M + H]+, 952.3772; found for C49H53IrN5O3, [M + H]+, 952.3796.