Tuning the Photophysical Properties of Homoleptic Tris-Cyclometalated Ir(III) Complexes by Facile Modification of the Imidazo-Phenanthridine and Their Application to Phosphorescent Organic Light-Emitting Diodes.

To explore the excited-state electronic structure of the blue-emitting Ir(dmp)3 dopant material (dmp = 3-(2,6-dimethylphenyl)-7-methylimidazo[1,2-f]phenanthridine), which is notable for durable blue phosphorescent organic light-emitting diode (PhOLED), a series of homoleptic dmp-based Ir(III) complexes (DMP-R, tris[3-(2,6-dimethylphenyl)-7-R-imidazo[1,2-f]phenanthridin-12-yl-κC 12,κN 1]iridium, R = H, CH3, F, and CF3) were prepared by introducing an electron-donating group (EDG; -CH3) or an electron-withdrawing group (EWG; -F and -CF3) at the 7-position of the imidazo-phenanthridine ligand. The photophysical analysis demonstrated that the alteration from EDG to EWGs led to redshifted structureless emission profiles, which were correlated with variations in the 3MLCT/3ILCT ratio in the T1 excited state. From electrochemical studies and density functional theory calculations, it turned out that the excited-state nature of the dmp-based Ir(III) complexes was significantly affected by the inductive effect of the 7-substituent of the cyclometalating dmp ligand. As a result of the lowest unoccupied molecular orbital energy stabilization by the EWGs that suppressed the non-radiative pathway from the emissive triplet excited state to the 3 d-d state, the F- and CF3-modified Ir(dmp)3 complexes (DMP-F and DMP-CF 3 ) showed quantum yields of 27-30% in the solution state, which were at least 4- or 5-fold higher than those shown by DMP-H and DMP-CH 3 . A PhOLED device based on DMP-CF 3 [CIE chromaticity (0.17, 0.39)], which demonstrated a distinct 3MLCT characteristic, exhibited better electroluminescent efficiencies with an external quantum efficiency of 13.5% than that based on DMP-CH 3 .

Introduction

Cyclometalated Ir(III) complexes are widely applied to diverse applications such as molecular sensors, optoelectronics, biosensors, photocatalysts, and so forth.[1−9] Particularly, the organometallic Ir(III) complexes are the best fit for triplet emitters in phosphorescent organic light-emitting diodes (PhOLEDs) owing to their ability to harvest all electro- and photogenerating singlets (25%) and triplets (75%), which means a theoretical internal quantum efficiency of nearly 100%.[10−14] To realize high-efficiency OLEDs, there have been considerable efforts in a wide range from materials science including molecular engineering to electronic device technology.[15−20] Recently, many studies have been conducted focusing on improving the photoelectric properties of the Ir(III) complex through changing the chelating ligand, consequently enhancing the performance of PhOLEDs. These molecular design strategies provide insights into understanding the triplet excited state origin, dipole orientation, and excimer formation of the Ir(III) emitter in the emitting layer.[21−23] Selecting appropriate chelating ligands of Ir(III) complexes that can allow tunable emission via the involvement of electron-donating (EDG) or -withdrawing (EWG) groups on cyclometalated ligands is a crucial task in modeling and synthesis of Ir(III) complexes.[24−27] Compared to phenylpyridine (ppy) ligands, which are typically used as dopants for red or green OLEDs, cyclometalated phenylimidazole (pmi) ligands tend to form higher triplet energies (pmi: 3π–π = 380 nm vs. ppy: 3π–π = 508 nm),[28−30] leading to destabilization of the lowest unoccupied molecular orbital (LUMO) energy, as well as are suitable as dopants for blue OLEDs.[31−35] In general, the phosphorescence color purity and efficiency of Ir(III) complexes coordinated with pmi chelates would be fine-tuned by the (i) introduction of substituents with different electronic effects, such as EWG and EDG moieties or (ii) expanding π-conjugation between chelating moieties (imidazole and phenyl moiety) to increase the rigidity of the Ir(III) complex.[26,36−38] For example, the introduction of a fluoro moiety to the parent pmi ligand was found to dramatically improve the emission efficiency and electroluminescence in a blue PhOLED, resulting in an external quantum efficiency (EQE) of 18.3% at L0 = 1,000 cd m–2 with Commission International de L’Eclairage (CIE) coordinates of [0.17, 0.30].[39] More recently, Thompson and Forrest et al. have confirmed that the fused dmp moiety (dmp = 3-(2,6-dimethyl-phenyl)-7-methylimidazo[1,2-f]phenanthridine) is an effective C^N chelating ligand for achieving an efficient blue electroluminescence in a PhOLED (CIE values of [0.16, 0.31] and an EQE of 8.5% at L0 = 1,000 cd m–2) producing a long operational lifetime (T80 = 334 ± 5 h at L0 = 1,000 cd m–2).[37] The fused feature of the imidazo-phenanthridine ligand can help stabilize the higher excited state of the complex; otherwise, it would readily undergo a known metal-centered (MC, 3d–d) deactivation process.[40] The efficiency and durability of the Ir(dmp)3 emitter could be further improved by modulating the electronic properties of the substituents.

However, despite their potential, the fundamental structure–property relationship of Ir(dmp)3 complexes has been less investigated for blue PhOLED application based on the organometallic Ir(III) dopant.[36,37,40,41] In efforts to investigate the relationship between the electronic structure and emission property of the Ir(dmp)3 dopant, we herein prepared four electronically tuned Ir(dmp)3 complexes, that is, tris[3-(2,6-dimethylphenyl)-7-R-imidazo[1,2-f]phenanthridin-12-yl-κC12,κN1]Ir (DMP−R; R = H, CH3, F, and CF3) (Scheme ). From photophysical/electrochemical analyses and density functional theory (DFT) calculations, it was found that the introduction of an EDG (−CH3) or EWGs (−F and −CF3) at the 7-position of the imidazo-phenanthridine ligand leads to a significant variation in excited-state properties of Ir(dmp)3 complexes with an elucidation of the electronic structure and radiative process. The device performance of PhOLEDs based on the synthesized Ir(dmp)3 dopants was also examined, and the results indicated that the participation of the −CF3 group in the cyclometalated dmp ligand is relatively effective.

Scheme 1

Synthetic Route and Chemical Structures Leading to the Cyclometalated Ir(III) Complexes Prepared in This Work

Experimental Section

General Information

The synthetic scheme of all compounds is depicted in Scheme and the detailed experimental procedures are given in the Supporting Information. The Ir(III) complex tris[3-(2,6-dimethylphenyl)-7-methylimidazo[1,2-f]phenanthridin-12-yl-κC12,κN1]iridium(III) (DMP–CH) as well as the intermediates 6-aminophenanthridine, 6-amino-2-methylphenanthridine, 6-amino-2-fluorophenanthridine, 6-amino-2-(trifluoromethyl)phenanthridine, and 3-(2,6-dimethylphenyl)-7-methylimidazo[1,2-f]phenanthridine ligand were facilely synthesized as described previously.[40,42−44] All ligands were allowed to react with Ir(COD)2BF4 in 1,2-propanediol to form the desired products, that is, homoleptic Ir(III) complexes. The ligands and imidazo-phenanthridine-based Ir(III) complexes were identified by 1H and 13C NMR spectroscopy, mass spectrometry (HR-MS), and elemental analysis (Figures S1–S15 and Table S1). 1H and 13C NMR spectra were measured on a 400 MHz spectrometer (AS400, Agilent, USA) at the KBSI Ochang Center. Details of the X-ray analysis and equipment used in this study for steady-state absorption/emission experiments, decay lifetime measurements, electrochemical measurements, DFT calculations, and thermal stability, among others, are presented in the Supporting Information.

Synthesis of 3-(2,6-Dimethylphenyl)imidazo[1,2-f]phenanthridine (L-H)

Phenanthridin-6-amine (11.5 mmol) and 2-bromo-2-(2,6-dimethylphenyl)acetaldehyde (12.7 mmol) were dissolved in isopropanol (45 mL) and heated at 85 °C overnight. Then, sodium bicarbonate (23.0 mmol) was added to the mixture slowly and refluxed for an additional 1 day. The resultant was cooled to ambient temperature. The isopropanol was removed by evaporation, and the residue was partitioned between water and dichloromethane (DCM). The organic layer was treated with sodium sulfate, and column chromatography was performed on silica gel using n-hexane/ethyl acetate/DCM (4:1:1, v/v/v) as an eluent to give L-H (2.8 g, 75%). 1H NMR (CDCl3, 400 MHz, δ): 9.07 (s, 1 H), 8.53 (d, J = 8.0 Hz, 1 H), 8.45 (m, 1 H), 7.78 (t, J = 3.6 Hz, 2 H), 7.51 (m, 2 H), 7.41 (t, J = 7.60 Hz, 1 H), 7.32 (t, J = 9.8 Hz, 2 H), 7.27 (s, 1 H), 7.25 (s, 1 H), 2.08 (s, 6 H). ESI–MS (m/z): calcd. for 323.1543; found [M + H]+, 323.1547. Anal. Calcd for C23H18N2: C, 85.68; H, 5.63; N, 8.69. Found: C, 85.32; H, 5.56; N, 8.57.

Synthesis of 3-(2,6-Dimethylphenyl)-7-fluoroimidazo[1,2-f]phenanthridine (L-F)

A procedure analogous to that described earlier to obtain L-H was used to synthesize L-F. Yield: 2.2 g (56%). 1H NMR (CDCl3, 400 MHz, δ): 8.85 (d, J = 8.0 Hz, 1 H), 8.31 (d, J = 8.0 Hz, 1 H), 8.13 (d, J = 8.0 Hz, 1 H), 7.78 (m, 2 H), 7.42 (m, 2 H), 7.29 (m, 3 H), 6.99 (t, J = 8 Hz, 1 H), 2.08 (s, 6 H). ESI–MS (m/z): calcd for 341.1449; found [M + H]+, 341.1454. Anal. Calcd for C23H17FN2: C, 81.16; H, 5.03; N, 8.23. Found: C, 80.76; H, 5.09; N, 8.34.

Synthesis of 3-(2,6-Dimethylphenyl)-7-(trifluoromethyl)imidazo[1,2-f]phenanthridine (L-CF3)

A procedure analogous to that described earlier to obtain L-H was used to synthesize L-CF from 2-(trifluoromethyl)phenanthridin-6-amine (11.5 mmol). Yield: 3.1 g (69%). 1H NMR (CDCl3, 400 MHz, δ): 8.93 (d, J = 7.20 Hz, 1 H), 8.74 (s, 1 H), 8.45 (dd, J = 7.00, 2.20 Hz, 1 H), 7.78 (m, 2 H), 7.50 (m, 2 H), 7.41 (t, J = 7.60 Hz, 2 H), 7.27 (s, 1 H), 7.25 (s, 1 H), 2.07 (s, 6 H). ESI–MS (m/z): calcd for 391.1417; found [M + H]+, 391.1423. Anal. Calcd for C24H17FN2: C, 73.84; H, 4.39; N, 7.18. Found: C, 74.56; H, 4.43; N, 7.26.

Synthesis of Tris[3-(2,6-dimethylphenyl)imidazo[1,2-f]phenanthridin-12-yl-κC12,κN1]iridium. (DMP–H)

L-H (6.84 mmol) and bis(1,5-cyclooctadiene)iridium(I) tetrafluoroborate (Ir(COD)2BF4) (1.95 mmol) were added to a round-bottomed two-necked flask under a N2 atmosphere without light. Deaerated 1,2-propanediol (250 mL) was placed in this flask and mildly heated to 50 °C for 30 min. For an additional 30 min, the temperature was slowly increased to 180 °C. The resultant was cooled, and then all volatiles were evaporated. The residues were isolated by column purification, which used DCM/n-hexane (1:1, v/v) as an eluent to give DMP–H (0.7 g, 31%). 1H NMR (CDCl3, 400 MHz, δ): 8.46 (s, 3 H), 7.68 (d, J = 7.7 Hz, 3 H), 7.37 (td, J = 7.7, 1.4 Hz, 3 H), 7.30 (t, J = 7.7 Hz, 3 H), 7.16 (m, 15 H), 7.09 (d, J = 8.40 Hz, 3 H), 6.81 (s, 3 H), 2.08 (s, 9 H), 1.88 (s, 9 H). 13C NMR (CDCl3, 176.1 MHz, δ): 155.16, 149.63, 140.02, 139.68, 136.52, 133.88, 132.80, 130.94, 129.55, 128.85, 128.05, 127.81, 127.58, 127.11, 125.78, 124.84, 124.63, 115.65, 112.29, 20.90. ESI–MS (m/z): calcd. for 1156.3804; found [M], 1156.3802. Anal. Calcd for C69H51IrN6: C, 71.67; H, 4.45; N, 7.27. Found: C, 70.95; H, 4.42; N, 7.20.

Synthesis of Tris[3-(2,6-dimethylphenyl)-7-fluoroimidazo[1,2-f]phenanthridin-12-yl-κC12,κN1] iridium. (DMP–F)

A procedure analogous to that described earlier to obtain DMP–H was used to obtain DMP–F from L-F (6.84 mmol). Yield: 0.94 g (40%). 1H NMR (CDCl3, 400 MHz, δ): 8.70 (d, J = 10.2 Hz, 3 H), 7.59 (d, J = 7.6 Hz, 3 H), 7.31 (t, J = 7.6 Hz, 3 H), 7.16 (m, 12 H), 7.05 (dd, J = 9.2, 4.8 Hz, 3 H), 6.87 (m, 3 H), 6.80 (s, 3 H), 2.07 (s, 9 H), 1.85 (s, 9 H). 13C NMR (CDCl3, 176.1 MHz, δ): 160.62, 159.24, 154.85, 149.41, 139.95, 139.59, 137.10, 132.81, 130.54, 130.29, 129.79, 129.71, 128.13, 127.97, 127.77, 124.65, 117.07, 114.99, 114.86, 112.55, 110.74, 110.61, 20.85. ESI–MS (m/z): calcd for 1210.3522; found [M], 1210.3524. Anal. Calcd for C69H48F3IrN6: C, 68.47; H, 4.00; N, 6.94. Found: C, 67.79; H, 4.04; N, 7.01.

Synthesis of Tris[3-(2,6-dimethylphenyl)-7-(trifluoromethyl)imidazo[1,2-f]phenanthridin-12-yl-κC12,κN1] Iridium. (DMP–CF)

A procedure analogous to that described earlier to obtain DMP–H was used to obtain DMP–CF from L-CF (6.84 mmol). Yield: 0.95 g (36%). 1H NMR (CDCl3, 400 MHz, δ): 8.70 (s, 3 H), 7.73 (d, J = 8.0 Hz, 3 H), 7.40 (dd, J = 9.0, 1.8 Hz, 3 H), 7.33 (t, J = 7.6 Hz, 3 H), 7.20 (m, 9 H), 7.15 (d, J = 7.20 Hz, 3 H), 7.10 (d, J = 7.60 Hz, 3 H), 6.84 (s, 3 H), 2.09 (s, 9 H), 1.85 (s, 9 H). 13C NMR (CDCl3, 176.1 MHz, δ): 155.47, 139.94, 139.51, 137.31, 135.68, 132.67, 130.16, 130.09, 130.04, 128.27, 128.23, 128.07, 127.12, 126.94, 126.60, 126.08, 125.16, 125.07, 124.19, 123.52, 122.26, 116.14, 112.71, 20.88. ESI–MS (m/z): calcd for 1360.3426; found [M], 1360.3423. Anal. Calcd for C72H48F9IrN6: C, 63.57; H, 3.56; N, 6.18. Found: C, 62.94; H, 3.52; N, 6.12.

Results and Discussion

Molecular Structure

The X-ray structures and crystallographic properties of DMP–CH and DMP–F are provided in Figure and in the Supporting Information (Tables S2–S6). These crystals of DMP–CH and DMP–F were cultivated by slow evaporation in DCM solution. DMP–CH and DMP–F are crystallized in the monoclinic crystal system (space groups: P21/c for DMP–CH (R1 = 0.0461) and P21/n for DMP–F (R1 = 0.0552)). As shown in Figure and Table , the Ir(III) complexes exhibit octahedral geometries, and the coordination parameters of the complexes, such as their C–Ir–N bite angles and Ir–C/Ir–N bond distances, are closely akin to those of other previously reported phenylimidazole-based Ir(III) complexes.[26,28,45]

Figure 1

(a) Perspective views of DMP–CH and DMP–F. Hydrogen atoms were omitted for clarity. (b) Intermolecular π–π interactions of DMP–CH and DMP–F.

Table 1

Selected Single-Crystal Parameters for DMP–CH and DMP–Fa

		DMP–CH3		DMP–F
parameters		selected value	average	selected value	average
bond length (Å)	Ir–C	2.034(5)	2.026	2.011(8)	2.022
		2.015(4)		2.124(9)
		2.028(5)		2.031(9)
	Ir–N	2.186(4)	2.189	2.164(8)	2.169
		2.201(4)		2.164(7)
		2.180(4)		2.178(7)
angle (deg)	∠C–Ir–N	81.7(2)	81.13	81.0(3)	81.52
		80.2(2)		81.4(3)
		81.5(2)		82.2(3)
dihedral angle (deg)	∠CPh–CPh–CIm–NIm	0.9(7)	1.31	1.2(1)	1.84
		2.1(8)		2.1(1)
		1.0(6)		2.2(3)
	∠NIm–CIm–CXy–CXy	86.1(6)	81.4	97.5(1)	96.8
		86.3(8)		107.2(1)
		71.6(9)		85.6(1)

∠CPh–CPh–CIm–NIm: the angle between the phenyl and imidazole planes of the phenanthroline connected to the Ir atom, ∠NIm–CIm–CXy–CXy: the angle between the imidazole plane of the phenanthroline and the xylene plane.

The imidazo-phenanthridine moiety coordinated with the metal center has dihedral angles of 1.31° (av.) for DMP–CH and 1.84° (av.) for DMP–F between the phenyl and imidazole planes of the imidazo-phenanthridine, which means that this imidazo-phenanthridine is arranged in an almost planar structural configuration. Overall, the coordination around the Ir atom is not considerably dependent on the substituent modulations on the cyclometalated C^N ligands. The xylene substituted at the C3 atom site of imidazole has a structure that is nearly perpendicular to the imidazo-phenanthridine plane (DMP–CH: 81.4° (av), DMP–F: 96.8° (av.)). In the packing structure, the π–π interaction between adjacent complexes is usually in the range of 3.34–3.77 Å, which corresponds to the interactions of phenanthridine with xylene (∼3.7 Å) and the interaction of two phenanthridines (∼3.4 Å). The intermolecular interactions of these complexes are shown in Figure b. Compared with DMP–CH, which features a single weak intermolecular interaction (between the phenanthridine plane and xylene ring, 3.72 Å), DMP–F features an additional π–π interaction between phenanthridine planes with a high degree of spatial overlap, which contributes to its photophysical properties.[46,47]

Photophysical Properties

The absorption properties of the imidazo-phenanthridine-based Ir(III) complexes were obtained at 298 K in DCM solution, as shown in Figure a. The Ir(III) complexes depict intense ligand-centered bands (1LC, spin-allowed π–π* transitions) appearing in the region of 230–300 nm (ε > 1.1 × 105 M–1 cm–1), which correspond to the intense absorptions of the free ligands at 225 to 330 nm (ε ≈ 4.5 × 104 M–1 cm–1) (Figure S16). Absorption bands at the near-UV/visible regions (λabs ≈ 350 nm, ε < 2.5 × 104 M–1 cm–1) and the weaker absorption tails detected at wavelengths of λ ≈ 450 nm are assigned to spin-allowed metal-to-ligand charge transfer (1MLCT) and 3MLCT transitions, respectively. When the red-edge of the 3MLCT region is magnified 10 times, a slight difference in the S0–T1 transition between type A (i.e., DMP–H and that with EDG; DMP–CH) and type B (i.e., those with EWGs; DMP–F and DMP–CF) complexes could be observed. The lower energy of the S0–T1 transition of type B complexes, which bear an EWG (−F or −CF3) in the imidazo-phenanthridine moiety, can be explained as a result of the stabilization of the 3MLCT state (relative to type A complexes) by the electron-accepting character of the 7-substituent (−F and −CF3).

Figure 2

(a) UV–Vis absorption spectra of all compounds (ligand: open-circled symbols, Ir(III) complexes: solid line). Emission spectra of Ir(III) complexes in (b) fluid DCM at 298 K and (c) frozen 2-MeTHF at 77 K.

Figure b illustrates the emission spectra of the Ir(III) complexes dissolved in the DCM solvent and measured at 298 K, which show a slightly unstructured peak shape featuring an MLCT characteristic in the emitting state because of the distinct involvement of the metal orbitals in the corresponding transition is related to the metal–ligand vibrational modes.[11,48] These phosphorescence properties are also classified into types A and B depending on the substituent effect: the emission profiles of type B complexes (λem = 494 nm for DMP–F and λem = 480 nm for DMP–CF) are more redshifted by 14–28 nm relative to those of type A complexes (λem = ∼466 nm for DMP–H and DMP–CH). However, the emission energies of rigid films doped with 5 wt. % type A complexes and poly(methyl methacrylate) (PMMA) are almost the same as those in the solution state (Figure S17 and Table ) except for DMP–F. The emission profile of DMP–F, which shows a structureless emission profile in solution, reveals a more structured peak in the rigid environment. Under frozen conditions at 77 K (Figure c), more blue shifting and structured profiles with vibronic progression and separated peaks are observed. The dominant emission peaks are located at the first peaks (i.e., 456 nm for DMP–H and DMP–CH, 467 nm for DMP–F, and 469 nm for DMP–CF), which are assigned to electronic T1 → S0 transitions. The weaker emission maximum at 491–505 nm, which is separated by approximately 1500–1560 cm–1 from the main peaks, is related to the overlapping vibrational satellites of the ground state.

Table 2

Photophysical Properties of the Imidazophenanthridine-Based Ir(III) Complexes

	at 298 Ka						at filmb		at 77 Kc
	Abs (λabs/nm)	Em (λem/nm)	Φem (%)d	τem (μs)	kr (104/s)	knr (105/s)	Em (λem/nm)	Φem (%)d	Em (λem/nm)	τem (μs)
DMP–H	234, 258, 297,	466, 497	7.2	0.23	3.04	40.4	465, 492	36.4	456, 491, 532	3.34
	329, 355, 392, 447
DMP–CH3	237, 258, 296,	466, 497	6.3	0.23	3.04	40.4	464, 492	36.7	456, 491, 532	4.01
	331, 356, 396, 447
DMP–F	235, 258, 303,	494	30.4	1.57	1.85	4.5	477	47.0	467, 502, 544	7.40
	336, 360, 404, 449
DMP–CF3	238, 266, 305,	480, 501	27.6	1.15	2.43	6.2	477, 501	53.4	469, 505, 547	3.63
	340, 364, 399, 452

Measured in degassed DCM (10 μM) at 298 K.

Fabricated on fused silica glass substrates with 5 wt. % Ir(III) complexes and PMMA, which dissolved in toluene.

Measured in a 2-MeTHF glassy matrix at 77 K.

Determined using a Quantaurus-QY measurement system (λex = 350 nm).

The Φem values are higher in the film (Φem = 36.7% for DMP–CH, Φem = 53.4% for DMP–CF) than in the solution state (Φem = 6.3% for DMP–CH and Φem = 27.6% for DMP–CF) because of the suppression of non-radiative processes (Table ). Under deaeration, all Ir(III) complexes (DMP–H, DMP–CH, DMP–F, and DMP–CF) exhibit a common emission decay lifetime of ∼μs in fluid (at 298 K) and frozen solvent media (77 K), indicating the typical triplet emission of the heavy metal complex by singlet-triplet state mixing derived from the spin–orbit coupling.[3] Compared to the triplet lifetimes of Ir(III) complexes at 77 K (τem = 3.34–7.40 μs), the relatively short emission decay lifetimes at 298 K (τem = 0.23–1.57 μs) arise from the thermally activated non-radiative process through the ligand substituents (Table and Figure S18).

The quantities Φem and τem are associated with the two rate constants (radiative (kr) and non-radiative (knr)), which were evaluated by the following relationship: kr = Φem/τem, knr = (1 – Φem)/τem. According to Table , because knr of all Ir complexes is greater than kr, τem is expected to be more dependent on the knr value than kr. The decay time mainly follows the energy-gap law in which the value of knr escalates as the emission energy reduces,[49,50] but this behavior was not applied to the DMP-based Ir complexes used in this study (Figure S19). The difference in τem is speculated to be due to the degree of the thermally activated population of the non-radiative metal-centered state (3d–d) and can be traced to knr of the Ir complexes. The non-radiative rate constants (knr) increased in the order of DMP–F (knr = 4.5 × 105 s–1) < DMP–CF (knr = 6.2 × 105 s–1) ≪ DMP–H, DMP–CH (knr = 40.4 × 105 s–1). The knr values of the type A complexes (DMP–H and DMP–CH) are higher by one order of magnitude compared with those of type B complexes (DMP–F and DMP–CF). This dramatic increase in knr of type A complexes is probably related to the thermal population of 3d–d states, which suppresses the possibility of radiative decay. This hypothesis is supported by the DFT calculation results showing the relatively small energy difference (3MLCT–3d–d states) in the type A complexes (compared to type B complexes) that accelerates non-radiative pathways (vide infra).

The extent of the charge transfer in the excited state for Ir(III) complexes was confirmed in detail by analyzing the emission spectra observed at different solvent polarities. The emissions of the Ir(III) complexes progressively shifted toward lower energies with increasing solvent polarity, as shown in Figure S20; compared with the type A complexes, the type B complexes show greater emission shifting. This result was supported by a Lippert–Mataga model, expressed using eq , which implied that the solvent sensitivity to polarity can be evaluated by the ground and excited state dipole moments (μ and μ) (Figure ).[51−53]where and are absorption and emission wavenumbers and h and c are the Planck’s constant and velocity of light, respectively. ε and n indicate the solvent dielectric constant and the refraction index, which were included in the term of Δf. The Onsager radius (a0) is approximately half the average size of Ir(III) complexes and can be evaluated by crystal structures and ab initio calculations using B3LYP and 6-31G(d,p) functions.

Figure 3

Lippert–Mataga plots of the Ir(III) complexes in various solvents, such as n-hexane (Hx), THF, DCM, and acetonitrile.

The slopes of DMP–H and DMP–CH (type A complexes) in Figure are nearly identical, whereas the slopes of the type B complexes are steeper than those of the type A complexes. Changes in the transition dipole moment (Δμ) of DMP–H, DMP–CH, DMP–F, and DMP–CF were estimated from the slopes of the curves and found to be 9.45, 10.36, 18.07, and 17.67 D, respectively. A charge separation equal to one electron charge placed 1 Å unit apart has a dipole moment of 4.8 Debye.[51] The change in the dipole moment observed in the type A complexes indicates that one electron is transferred over a distance of ∼2.2 Å, which probably corresponds to the distance between the Ir atom and the N atom of the imidazole moiety (distance calculated by DFT: 2.24 Å). Type B complexes have a longer charge separation distance (∼3.7 Å) compared with type A complexes, which means type A and type B complexes have different excited-state dynamics. The type B complexes in this work may be generally expected to have a longer charge separation distance because they bear an EWG (i.e., −F, −CF3).

Electrochemical Properties

The electrochemical redox potentials of the Ir(III) complexes investigated by cyclic voltammetry (CV) are depicted in Figure and listed in Table S7. As previously reported for related derivatives, all the complexes undergo reversibility upon scanning the oxidation waves.

Figure 4

(a) Cyclic voltammograms of the Ir(III) complexes in 1 mM DCM containing 0.1 M TBAP obtained at a scan rate of 0.1 V/s. A Pt disk electrode was used as the working electrode, and a Pt wire and Ag/AgNO3 were used as the counter and reference electrodes, respectively. (b) Correlations between the reduction (top) and oxidation potentials (bottom) versus the Hammett constants (σm) of Ir(III) complexes.

The appearance of one reversible peak at ∼0.8 V is ascribed to the oxidation process of the metal center (Ir3+/Ir4+), in line with the HOMO orbital distribution localized on the Ir center in the DFT calculation (vide infra).[54] The oxidation potential of DMP–H and DMP–CH is quite invariant at 0.56–0.58 V; similarly, that of DMP–F and DMP–CF resides between 0.66 and 0.69 V. Compared with the type A complexes (DMP–H and DMP–CH), the oxidation waves of the type B complexes (DMP–F and DMP–CF) cathodically shifted by 130 mV. This difference could be ascribed to the effect of the substituents of the complexes (e.g., EDG: −CH3, σm = −0.07; EWGs: −F, σm = 0.34; −CF3, σm = 0.43), and the observed change in oxidation potential is linearly correlated with the Hammett parameter (Figure b).[55]

Upon scanning in the negative direction, all complexes exhibited irreversible profiles within the scan limits. The onsets of the reduction peaks of DMP–H, DMP–CH, DMP–F, and DMP–CF were determined to be −2.40, −2.42, −2.20, and −2.14 V, respectively. According to the Hammett parameter, the extent of the observed reduction potential positively shifts in a stepwise manner in the order of DMP–CH, DMP–H, DMP–F, and DMP–CF. Given that the LUMO distribution of Ir(dmp)3 complexes is widely localized over the entire fused dmp ligand, which in contrast to the HOMO orbital distributed on the Ir metal and the phenyl-imidazole moiety (vide infra), it is most likely that the LUMO of the Ir(dmp)3 complex is more sensitive to the electronic properties (EWG or EDG) of the substituent in the chelate than the HOMO level. Indeed, the inductive effect of the substituent on the reduction potential was found to be more noticeable than that on the oxidation potential: the difference in oxidation potential between DMP–CH and DMP–CF is 130 mV, whereas their difference in reduction potential is 280 mV. The significant stabilization of the LUMO energy by substitution of electron-accepting groups (−F and −CF3) resulted in the relatively smaller electrochemical band gaps of the Ir(dmp)3 complex (Eg = 2.86 eV for DMP–F and 2.83 eV for DMP–CF) compared to those of DMP–CH (Eg = 2.98 eV) and DMP–H (Eg = 2.98 eV), which is consistent with the discernible redshifted peaks in the absorption and emission spectra of DMP–F and DMP–CF (Figure and Table ).

DFT Calculations

A theoretical interpretation of molecular orbitals is valuable for understanding the photophysical and electrochemical behaviors of compounds. All compounds involving the ligand and Ir complexes were simulated using the DFT protocol with the B3LYP method, which employed the LANL2DZ basis set for metal and 6-31G for C, H, and N atoms.[56,57] The molecular geometry was optimized with X-ray crystal structures. Figures S21 and S22 compare the relative energies of the molecular orbitals of ligands and Ir complexes, along with the isodensity surfaces of their HOMOs and LUMOs.

Figure compares the DFT calculation results of the free ligands CH3 and CF3 and the Ir complexes DMP–CH and DMP–CF, which shows the most significant difference in electrical properties among our four Ir(III) complexes. The electrical characters of the Ir(III) complexes, including their orbital distribution and energy levels, are closely common to those of the free ligands. Molecular orbital analysis indicates that the HOMOs of the Ir(III) complexes are mainly distributed on the phenyl-imidazole moiety of the phenanthridine ligands and the Ir metal (orbital contributions: 51% Ir(III) atom + 45% phenyl-imidazole moiety), whereas the LUMOs of each Ir(III) complexes are exclusively located throughout the ligand, extending from the phenyl-imidazole moiety to the entire phenanthrolidine segment. Analysis of the percentage of molecular contributions reveals that the HOMO lies primarily at the Ir center (>50%), which could mean that the MLCT is a major transition occurring in the excited state; the results of the time-dependent (TD)-DFT confirm such a transition. The energies of the Ir(III) complexes are sensitive to the phenanthrolidyl ligand pattern. For example, the HOMO of EDG-substituted DMP–CH (−4.56 eV) is more unstable than that of the unsubstituted complex, DMP–H (−4.62 eV), whereas the HOMO of EWG-substituted DMP–CF is remarkably stabilized at −4.96 eV due to the CF3 (σm = 0.43) residue. These results are well-matched with the electrochemical properties presented in Figure and supported by the DFT calculations, as shown in Figure . From the optimized structures, the simulated absorption properties were calculated using the TD-DFT approach. The first vertical triplet and singlet excitations were carefully evaluated and listed in Table S8 and Figure S23. The experimental and calculated forms show good agreement. The initial experimental absorption peaks of the Ir(III) complexes arise from the MLCT transitions, which are well-matched with the theoretical predictions of the large oscillator strength required for the transition from phenyl-imidazole/Ir (HOMO) to the whole imidazo-phenanthridine ligand (LUMO). The peaks are mainly composed of HOMO → LUMO transitions, which could be considered an admixture of intraligand charge transfer (ILCT)- and MLCT-type transitions. The geometries of the triplet state were then optimized using the UB3LYP method. When analyzing the molecular electronic structure in its triplet state, it is necessary to consider two pairs of nearly degenerated molecular orbitals: HOMO–1/HOMO–2 and LUMO+1/LUMO+2 as shown in Figure . This quasidegeneracy of molecules provokes the T1 state to undergo the distortion that, together with correlated changes in bond distances (e.g., Ir–N for DMP–CF: S1 = 2.24 Å for all ligands, T1 = 2.20–2.26 Å for each ligand), results in the higher semioccupied molecular orbital (HSOMO) of the triplet state orbital contribution localized on only one of the three ligands. The calculated SOMO energies (HSOMO and LSOMO) are similar to the HOMO and LUMO energies derived from the measured CV results. The spin densities in the T1 state of DMP–CH and DMP–CF, shown in Figure and Figure S24, possess similar topologies, and the spin density is extended to the phenyl ring to which the substituent is connected, thereby suggesting that the substituents of the imidazo-phenanthridine moiety have a non-negligible effect on the emissive state of the Ir(III) complex. The lowest-lying T1 state is mainly manifested by the transition of LUMO → HOMO, which implies that the emissive state originates from the mixture of 3MLCT and 3ILCT characteristics. Thus, this assignment is well-matched with the above TD-DFT results (Table S8) and the unstructured emission observed at 298 K in DCM solution. In addition, in comparison to DMP–CH and DMP–CF (Figure ), the relatively larger energy gap between HSOMO and the 3d–d state in DMP–CH compared to that of DMP–CF can explain why the knr of DMP–CH is larger than that of DMP–CF by an order of magnitude in photophysical measurements.

Figure 5

Singlet and triplet frontier orbital and energy diagrams of the ligands and Ir(III) complexes.

Device Performance

Because of its attractive optoelectronic properties, DMP–CF was used as an emitter for an OLED device, and the performance of this device was compared with that of a similar device based on DMP–CH. Figure a shows the device configuration [ITO (150 nm, anode)/1,4,5,8,9,11-Hexaazatriphenylene-hexacarbonitrile (HAT-CN) (10 nm)/TAPC (55 nm)/mCBP: Ir(III) complex (DMP–CH or DMP–CF) (30 nm: 18 wt. %)/1,3,5-tri(m-pyridin-3-ylphenyl)benzene (TmPyPB) (30 nm)/Liq (1 nm)/Al (150 nm, cathode)]. HAT-CN was used as the hole injection layer.[58−60] In addition, 4,4′-cyclohexylidene-bis[N,N-bis(4-methylphenyl)benzeneamine] (TAPC), 1,4-bis(triphenylsilyl)benzene, and TmPyPB were used as the hole transport layer, electron-blocking layer (EBL), and electron transport layer materials, respectively. The performance of the fabricated devices was evaluated using plots of current density–voltage–luminance (J–V–L), current efficiency (CE), and power efficiency (PE) characteristics, as shown in Figures b–6d and summarized in Table .

Figure 6

(a) Structure of the device with energy levels of the materials used and ((b–e)) performances of phosphorescent OLEDs: (b) J–V–L properties, (c) CE at different current densities, (d) power efficiency at different current densities, and (e) EQE and normalized electroluminescence spectra (inset) of PhOLEDs based on DMP–CH (blue) and DMP–CF (orange).

Table 3

Device Characteristics of PhOLEDs Doped with Ir(III) Complexesa

	turn-on voltage (V)b	CIE (x, y)c	max EQE (%)	max CE (cd/A)	max PE (lm/W)
DMP–CH3	3.1	(0.15, 0.26)	10.5	18.5	16.2
DMP–CF3	3.7	(0.17, 0.39)	13.5	29.8	20.8

Device conditions: ITO (150 nm)/HAT-CN (10 nm)/TAPC (55 nm)/Ir dopant (DMP–CH or DMP–CF) (18 wt.%):mCBP (30 nm)/TmPyPB (30 nm)/Liq (1 nm)/Al (150 nm).

Turn-on voltage at 1 nit.

Commission Internationale de L’Eclairage (CIE) coordinates at 100 nits.

The device based on DMP–CH showed a maximum current and power efficiency of 18.5 cd/A and 16.2 lm/W, respectively, with an EQE of DMP–CF of 10.5%, whereas the multilayer device with DMP–CF exhibited the maximum CE and PE of 29.8 cd/A and 20.8 lm/W, respectively, with an EQE of 13.5% (Figure b–e and Table ). The higher current and power efficiencies achieved using DMP–CF can be interpreted as a result of the relatively superior emission efficiencies of DMP–CF (in solution and solid state) compared to those of DMP–CH. The EL spectra of both devices exhibited the structured double peak emissions (DMP–CH, λEL = 464 and 494 nm; DMP–CF, λEL = 477 and 502 nm), which are similar to their film PL spectra (Figure S17), suggesting that the electroluminescence profiles are originated from phosphorescence of the Ir dopant (in EML) associated with efficient energy transfer from mCBP to the Ir(III) complex.

Conclusions

In an effort to investigate the excited-state electronic structure of the Ir(dmp)3 dopant, which has been known to be effective for long-lived blue PhOLEDs, imidazophenanthridine-based Ir(III) complexes, namely, DMP–H, DMP–CH, DMP–F, and DMP–CF, were prepared by introducing the EDG (−CH3) or EWGs (−F and −CF3) at the 7-position of the chelating dmp ligand. The photophysical/electrochemical analyses and DFT calculations showed that the electrical modulation significantly changes the excited-state properties of the Ir(dmp)3 complexes, which are associated with a change of the 3MLCT/3ILCT ratio in the T1 excited state. From DFT calculations, it was found that the HOMO of the Ir(dmp)3 complex is predominantly localized on the phenyl-imidazole moiety of the phenanthridine ligands and the Ir(III) metal, whereas the LUMOs are distributed over the entire fused phenanthrolidine ligand, and the LUMO levels being more sensitive to the inductive effect of the 7-substituent than HOMO. As a result of LUMO energy stabilization by the EWG group that suppresses the thermal population from the emissive triplet excited state to the non-radiative 3d–d state, DMP–F and DMP–CF showed relatively higher emission efficiencies (in solution and solid state) than DMP–CH and DMP–H. For the application of PhOLED, DMP–CF-doped multilayer device with CIE chromaticity (0.17, 0.39) exhibited better electroluminescent efficiencies with an EQE of 13.5%, compared to DMP–CH (CIE = 0.15, 0.26; EQE = 10.5%).