Cationic Iridium Complexes with Intramolecular pi-pi Interaction and Enhanced Steric Hindrance for Solid-State Light-Emitting Electrochemical Cells

Cationic Iridium Complexes with Intramolecular

π−π Interaction and

Enhanced Steric Hindrance for Solid-State Light-Emitting

Electrochemical Cells

Hsiao-Fan Chen,

Wen-Yi Hung,*

Shou-Wei Chen,

Ting-Chih Wang,

Shih-Wei Lin,

Shu-Hua Chou,

Chih-Teng Liao,

Hai-Ching Su,*

Hsiao-An Pan,

Pi-Tai Chou,*

Yi-Hung Liu,

and Ken-Tsung Wong*

†_{Department of Chemistry, National Taiwan University, Taipei 10617, Taiwan}

‡_{Institute of Optoelectronic Sciences, National Taiwan Ocean University, Keelung 20224, Taiwan} §_{Institute of Lighting and Energy Photonics, National Chiao Tung University, Tainan 71150, Taiwan}

*

ABSTRACT: Cationic iridium complexes incorporated 4,5-diaza-9,9′-spirobifluorene as N∧N ancillary ligands, in which one (2) or two (3) phenyl groups were introduced onto 4,5-diazafluorene to afford intraligand π−π interactions. The X-ray crystal structures of complexes 2 and 3 show that the pendant phenyl ring forms strong intramolecular face-to-faceπ-stacking with the difluorophenyl ring of the cyclometalated ligand with distances of 3.38 Å for complex 2 and 3.40 and 3.46 Å for complex 3. This π−π stacking interaction minimizes the expansion of the metal−ligand bonds in the excited state, resulting in a longer device lifetime in the light-emitting electrochemical cell (LEC) devices.

■

LECs (light-emitting electrochemical cells) possess several advantages over conventional OLEDs, such as compatibility with air-stable electrodes (e.g., Ag and Au) and single emissive layer which can be easily conducted from a cost-effective solution process.1 The turn-on voltages of LEC devices are generally low and approximately equal to the measured optical band gaps of the emissive materials (close to Eg). The low operating voltages are beneficial for the device power efficiencies. In addition to polymer-based LECs, the use of small-molecule ionic transition metal complexes (iTMCs) in LECs have attracted much attention in recent years.2,3In such devices, no ion-conducting material is needed because these metal complexes are intrinsically ionic. Furthermore, higher electroluminescence efficiencies are expected owing to the phosphorescent nature of the metal complexes.

Although iTMC LECs possess several promising advantages, there are still two important barriers that limit their practical application, that is, poor device stability (expressed as t_1/2, the time for the brightness of the device to decay from the maximum to half of the maximum) and long turn-on time (defined as the time required to reach a brightness of 1 cd cm−2). The intrinsic instability of iTMC is attributed to a water-assisted ligand exchange reaction during the device operation, subsequently forming a new complex for efficient luminance quenching.4 Recently, Bolink et al. demonstrated that the

device lifetime can be significantly improved by increasing the hydrophobicity of the complex by introducing peripheral bulky aromatic groups. The pendant bulky group forms a face-to-face π−π stacking with the metalated ring of the ligand and thus minimizes the expansion of the metal−ligand bonds in the excited state, as well as precludes the ligand exchange by foreign substances such as water. With such a molecular design, the device lifetime can significantly reach to thousands of hours.5

Excited state self-quenching and triplet−triplet annihilation within the solid film are well-known for devastating photo-luminescent efficiency and the device lifetime due to close intermolecular packing. An efficient way to circumvent the close packing motif is to introduce steric bulky ligands. Therefore, the nonradiative pathways can be largely suppressed with increased intermolecular distance.6 We have realized highly efficient LECs based on cationic iridium complexes with enhanced steric hindrance provided by bulky 4,5-diaza-9,9 ′-spirobifluorene (sp) ligand.7The spiro-configured structure of sp ligand has a substantial effect in preventing the intermolecular interaction, as indicated by the closest Ir-to-Ir distance larger than 8.62 Å.8To achieve robust cationic iridium complexes constructed by intramolecularπ−π interaction with enhanced steric hindrance, in this contribution we introduced Received: March 31, 2012

Published: October 26, 2012

one or two pendant phenyl group at 3 and 6 positions of sp to give 3-phenyl-4,5-diaza-9,9′-spirobifluorene (phsp) and 3,6-diphenyl-4,5-diaza-9,9′-spirobifluorene (diphsp), respectively. Furthermore, in order to obtain better intramolecular cofacial π−π interaction, the electron-deficient cyclometalated C∧_N ligand, 1-(2,4-difluorophenyl)pyrazole (dfppz), was selected for π-acceptor. The cationic iridium complexes incorporating phsp ([Ir(dfppz)2(phsp)]+(PF6−), 2) and diphsp ([Ir-(dfppz)₂(diphsp)]+_(PF

6−), 3) ligands were synthesized accord-ing to the reported synthetic route.8Complex without pendant phenyl group ([Ir(dfppz)₂(sp)]+_(PF

6−), 1) was also synthe-sized for comparison (Scheme 1). The distances of the face-to-faceπ−π interactions were measured to be 3.38 Å for complex 2 and 3.40 and 3.46 Å for complex 3. This π−π stacking interaction minimizes the expansion of the metal−ligand bonds in the excited state, resulting in a longer device lifetime in the LEC devices, wherein the device based on 2 showed a 4-fold lifetime enhancement as compared to that of the device based on 1.

■

General Experiments.1_{H and}13_{C NMR spectra of compounds}

were collected on a 400 MHz spectrometer at room temperature. UV−vis absorption spectra were recorded on a spectrophotometer (Hitachi U2800A). PL spectra were measured with a fluorescence spectrophotometer (Hitachi F4500). Photoluminescence quantum yields (PLQYs) for solution and thin-film samples were determined with a calibrated integrating sphere system (Hamamatsu C9920). Phosphorescence lifetime measurements were performed with an Edinburgh FL 900 photon-counting system. Elemental analyses for diphsp were carried out on a Heraeus Vario EL III elemental analyzer at NSC Regional Advanced Instrument Center, National Taiwan University. Mass spectra were recorded on a JEOL JMS-700 mass

spectrometer. Redox potentials of all complexs were determined by cyclic voltammetry (CV) at a scan rate of 100 mV/s in acetonitrile solutions (1.0 mM). A glassy carbon electrode and a platinum wire were used as the working electrode and the counter electrode, respectively. All potentials were recorded versus the Ag/AgCl (saturated) reference electrode and calibrated with the ferrocene/ ferrocenium redox couple. Oxidation CV was performed using 0.1 M tetra-n-butylammonium hexafluorophosphate (TBAPF6) in

acetoni-trile as the supporting electrolyte. For reduction CV, 0.1 M of tetra-n-butylammonium perchlorate (TBAP) in acetonitrile was used as the supporting electrolyte. Cyclometalated Ir(III) dichlorobridged dimer, bis-(μ)-chlorotetrakis(1-(4,6-difluorophenyl)-pyrazolato-C2

,N)-diiridium(III), was synthesized by a literature procedure.9 4,5-Diaza-9,9′-spirobifluorene (sp)7_{and complex 1}2c_{were obtained according to}

our previous publication.

Synthesis of 3-Phenyl-4,5-diaza-9,9′-spirobifluorene (phsp). A mixture of 4,5-diaza-9,9′-spirobifluorene (2.5 g, 7.86 mmol) and diethyl ether (400 mL) was stirred at room temperature, and phenyllithium (9.83 mL, 19.65 mmol, 2 M) was added in one portion. The reaction was stirred for 16 h, quenched with water, and extracted with diethyl ether. The organic layer was then concentrated and dissolved in CH2Cl2(300 mL). Activated MnO2(4.78 g, 55 mmol)

was added to the solution and stirred for 3 h. The mixture wasfiltered off through and successively washed with CH2Cl2 (3 × 100 mL),

CHCl3(3× 100 mL), and EtOAc (3 × 100 mL) to afford ligand phsp

(2.88 g, 93%) as a pale yellow solid. Mp 229°C (DSC).1_{H NMR}

(CDCl3, 400 MHz)δ 8.79 (d, J = 2.8 Hz, 1H), 8.22 (d, J = 7.6 Hz, 2H), 7.88 (d, J = 7.6 Hz, 2H), 7.60 (d, J = 7.6 Hz, 1H), 7.49 (t, J = 7.6 Hz, 2H), 7.45−7.40 (m, 3H), 7.20−7.13 (m, 5H), 6.78 (d, J = 7.6 Hz, 2H).13_{C NMR (CDCl} 3, 100 MHz)δ 159.1, 158.9, 158.3, 150.2, 146.2, 144.2, 142.2, 141.9, 138.9, 132.4, 131.9, 129.1, 128.6, 128.4, 128.2, 127.4, 123.9, 123.6, 120.8, 120.4, 61.4. HRMS (m/z, ESI+) calcd for C29H18N2394.1470, found 394.1473.

Synthesis of 3,6-Diphenyl-4,5-diaza-9,9′-spirobifluorene (diphsp). A mixture of 4,5-diaza-9,9′-spirobifluorene (500 mg, 1.57

mmol) and diethyl ether (80 mL) was stirred at room temperature, and phenyllithium (2 mL, 3.93 mmol, 1.92 M) was added in one portion. After 3 h, the reaction was quenched with water and extracted with diethyl ether. The organic layer was stirred with activated MnO2

(1.36 g) for 6 h. The mixture was dried over MgSO4and thenfiltered

to afford 3-phenyl-4,5-diazaspirobifluorene. The target molecule was obtained by repeating the same procedure. The crude product was purified by column chromatography on silica gel (EtOAc/hexane = 1/ 2) to afford diphsp (650 mg, 88%) as a pale yellow solid. Mp 335 °C (DSC).1_{H NMR (CDCl} 3, 400 MHz)δ 8.22 (d, J = 7.6 Hz, 4H), 7.89 (d, J = 7.6 Hz, 2H), 7.60 (d, J = 7.6 Hz, 2H), 7.50 (t, J = 7.6 Hz, 4H), 7.46−7.42 (m, 4H), 7.22−7.15 (m, 4H), 6.83 (d, J = 7.2 Hz, 2H).13_C NMR (CDCl3, 100 MHz)δ 159.2, 158.4, 146.4, 142.8, 141.9, 139.1, 132.4, 129.1, 128.6, 128.4, 128.2, 127.4, 123.9, 120.8, 120.4, 61.1. HRMS (m/z, ESI+) calcd for C35H22N2 470.1783, found 470.1780;

Anal. Calcd for C35H22N2: C, 89.33; H, 4.71; N, 5.95. Found: C, 89.21;

H, 4.59; N, 5.75.

Synthesis of [Ir(dfppz)2(phsp)]+(PF6−) (2). Bis-(

μ)-chlorotetrakis(1-(4,6-difluorophenyl)-pyrazolato-C2_{,N)diiridium(III)}

(585 mg, 0.5 mmol) and 3-phenyl-4,5-diaza-9,9′-spirobifluorene (433 mg, 1.1 mmol) were dissolved in 1,2-ethanediol (35 mL) under Ar, and the mixture was kept at 150°C for 16 h. The solution was cooled to room temperature, and an aqueous solution of NH4PF6(1.2 g in 10

mL deionic water) was added to yield yellow suspension. The solid was thenfiltered and dried in an oven (80 °C) for 12 h. The crude product was purified by column chromatography on silica gel (CH2Cl2/MeCN = 10/1) to give complex 2 (817 mg, 75%) as an

yellow solid.1H NMR (CDCl3, 400 MHz)δ 8.43 (d, J = 2.8 Hz, 1H), 8.31 (d, J = 2.8 Hz, 1H), 7.88 (dd, J = 8.0, 3.2 Hz, 2H), 7.69 (d, J = 5.2 Hz, 1H), 7.54−7.39 (m, 6H), 7.23−7.24 (m, 4H), 7.12 (t, J = 8.0, 1H), 7.02−6.99 (m, 3H), 6.86−6.78 (m, 5H), 6.63 (td, J = 8.0, 2.4 Hz, 1H), 6.28 (td, J = 8.0, 2.4 Hz, 1H), 5.58 (dd, J = 8.0, 2.4 Hz, 1H), 4.97 (dd, J = 8.0, 2.4 Hz, 1H). MS (m/z, ESI+_{) 551 (50), 945 (100). HRMS}

(m/z, FAB+_{) calcd for C}

47H28F4IrN6945.1935, found: 945.1942.

Synthesis of [Ir(dfppz)2(diphsp)]+(PF6−) (3).

Bis-(μ)-chlorotetrakis(1-(4,6-difluorophenyl)-pyrazolato-C2_{,N)diiridium(III)}

(585 mg, 0.5 mmol) and 3,6-diphenyl-4,5-diaza-9,9′-spirobifluorene (517 mg, 1.1 mmol) were dissolved in 1,2-ethanediol (35 mL) under Ar, and the mixture was kept at 150°C for 16 h. The solution was cooled to room temperature, and an aqueous solution of NH4PF6(1.2

g in 10 mL deionic water) was added to yield a yellow suspension. The solid was thenfiltered and dried in an oven (80 °C) for 12 h. The crude product was purified by column chromatography on silica gel (CH2Cl2/MeCN = 10/1) to give complex 3 (758 mg, 65%) as a

yellow solid.1H NMR (CDCl3, 400 MHz)δ 8.32 (d, J = 2.4 Hz, 2H), 7.90 (d, J = 8.0 Hz, 2H), 7.65 (d, J = 1.6 Hz, 2H), 7.50 (t, J = 7.2 Hz, 2H), 7.42 (d, J = 8.0, 2H), 7.32 (t, J = 7.2 Hz, 2H), 7.20 (d, J = 8.0, 2H), 7.19−7.08 (m, 4H), 7.02 (t, J = 8.0, 4H), 6.88 (t, J = 2.4 Hz, 2H), 6.84 (d, J = 7.2, 4H), 6.11 (td, J = 8.0, 2.4 Hz, 2H), 4.68 (dd, J = 8.0, 2.4 Hz, 2H). MS (m/z, ESI+_{) 551 (100), 577 (95), 1021 (80). HRMS}

(m/z, FAB+_{) calcd for C}

53H32F4IrN61021.2248, found 1021.2236.

Crystal Structure Determinations. Crystallographic data were collected at 295(2) K on a NONIUS Kappa CCD diffractometer using graphite-monochromatized Mo Kα radiation (λ = 0.710 73 Å). Cell parameters were retrieved and refined using HKL Denzo and Scalepack software on all observed reflections.10 _{Data reduction was}

performed with the HKL Denzo and Scalepack software. The structures were solved and refined with SHELX programs.11 _The

hydrogen atoms were included in calculated positions and refined using a riding mode.

Fabrication and Characterization of LEC Devices. ITO-coated glass substrates were cleaned and treated with UV/ozone prior to use. A PEDOT:PSS layer was spin-coated onto the ITO substrate, and then, the structure was baked at 150 °C for 30 min. All solution preparing and spin-coating processes were carried out in glovebox. The emissive layer (∼400 nm) was spin-coated from the CH3CN

solutions of the compound (150 mg/mL). To reduce the turn-on time of the LEC device, the ionic liquid [BMIM+_(PF

6−) ] (20 wt %) was

added to enhance the ionic conductivity of thin films.12 After spin-coating the emissive layer, the samples were baked at 70°C for 10 h in a nitrogen glovebox (oxygen and moisture levels below 1 ppm) and then subjected to thermal evaporation of a 100-nm-thick Al top contact in a vacuum chamber (ca. 10−6 Torr). The electrical and emission characteristics of the LEC devices were measured using a sourcemeter (Keithley 6430) and a picoammeter (Keithley 6487) equipped with a calibration Si-photodiode. All device measurements were performed under a constant bias voltage in a nitrogen glovebox. Computational Methodology. All the calculations were using the density functional theory (DFT) with B3LYP13hybrid functional. Restricted and unrestricted formalisms were adopted in the singlet and triplet geometry optimization, respectively. To investigate the photophysical properties, time-dependent DFT (TDDFT) calcula-tions were performed on the basis of the optimized structures at the ground state. A “double-ζ” quality basis set consisting of Hay and Wadt’s effective core potentials (LANL2DZ)14_{was employed for the}

Ir(III) metal atom, and a 6-31G* basis set15_{was employed for the rest}

of the atoms. The relativistic effective core potential (ECP) replaced the inner core electrons of Ir(III) metal atom, leaving only the outer core valence electrons (5s2_5p6_5d6_{) to be concerned. The electronic}

configurations of 3_{MC (metal-centered triplet) dd states were}

calculated following the literature methodology.16 In brief, the

3_{MLCT state geometry was obtained by performing geometry}

optimization along the triplet state potential energy surface (PES), using the ground state optimized structure as the initial geometry. As Figure 1.Crystal structures of complex 2 (left) and 3 (right). Counteranion PF6−and hydrogen atoms are omitted for clarity. Thermal ellipsoids are

4,5-diazaspirobifluorene, followed by oxidative aromatization with manganese oxide. Complexes 2 and 3 were synthesized according to the literature procedures in 75% and 65% yields, respectively.

Crystal Structures of 2 and 3. Crystals of complexes 2 and 3 suitable for the X-ray diffraction analysis were obtained by slow evaporation of the solvent from the acetone/hexane cosolvent system. The structures of complexes 2 and 3 are depicted in Figure 1, and the crystal data are summarized in Table 1. Complexes 2 and 3 both exhibit a distorted octahedral

geometry around the Ir center as indicated by the small bite angles of C(9)−Ir(1)−N(1) [80.52(18)o_{], C(18)−Ir(1)−N(3)} [79.81(17)o], and N(5)−Ir(1)−N(6) [79.87(14)o] and twisted bond angles of N(1)−Ir(1)−N(3) [175.76(14)o_{], C(9)−} Ir(1)−N(5) [172.79(15)o], and C(18)−Ir(1)−N(6) [176.18(15)o_{] in complex 2, and small bite angles of C(9)−} Ir(1)−N(5) [79.87(19)o], C(18)−Ir(1)−N(4) [80.4(2)o], and N(1)−Ir(1)−N(2) [79.64(14)o_{] and twisted bond angles of} N(4)−Ir(1)−N(5) [172.80(16)o], C(9)−Ir(1)−N(1) [170.05(17)o_{], and C(18)−Ir(1)−N(2) [176.70(18)}o_{] in} complex 3. The Ir−N bond distances between the Ir center

ffect on both sides of Ir−N bonds, as seen with the comparable distance of Ir(1)−N(1) [2.240(4)Å] and Ir(1)−N(2) [2.249(4)Å]. The pendant phenyl rings exhibit very strong intramolecular face-to-face π−π stacking with the cyclometalated difluorophenyl ring of the dfppy ligand as shown in Figure 2. The distances between

two phenyl rings (the pendant phenyl ring and the cyclo-metalated difluorophenyl ring) are about 3.38 Å for complex 2 and 3.40 and 3.46 Å for complex 3, indicating strong aromatic π−π interaction. This interaction diminishes the possibility of foreign substances to react with the metal complex and hence reduces the possible formation of degraded products capable of quenching the luminescence. The closest Ir-to-Ir distances in complex 2 and 3 are as large as 9.47 and 9.97 Å, respectively, indicating sufficient prevention of intermolecular interaction provided by bulky 4,5-diaza-9,9′-spirobifluorene ligand.

Photophysical Properties. Figure 3 shows the UV−vis and PL spectra of 1, 2, and 3 in CH₂Cl₂(10−5M) and in neat films. The photophysical properties of complexes 1, 2, and 3 in solutions (CH₂Cl₂, 10−5M) or in thinfilms are summarized in Table 2. Similar absorption features were observed for all complexes. The intense absorption bands in the UV region of the spectra (240−350 nm) are associated with the ligand-centered (LC) transitions of the ligands. These LC bands are accompanied by weaker and broad bands extending from∼300 nm to the visible region, which are related to the metal-to-ligand-charge-transfer (MLCT) transitions (both allowed and spin-forbidden ones) mediated by the strong spin−orbit coupling of the Ir(III) center. In the PL spectra, all the Table 1. Crystal Data of Complexes 2 and 3

2 3

empirical formula C48H28F10IrN6P C53H32F10IrN6P

fw 1174.85 1208.48

cryst dimensions (mm3_{) 0.25× 0.15 × 0.10 0.25× 0.20 × 0.10}

cryst syst triclinic monoclinic

space group P1̅ P21/c a (Å) 12.4320(2) 16.3211(7) b (Å) 13.5430(2) 16.8506(5) c (Å) 13.9288(3) 19.4315(10) α (deg) 74.532(2) 90 β (deg) 85.7760(10) 114.738(6) γ (deg) 79.5890(10) 90 cell volume (Å)3 _{2222.19(7) 4853.6(4)} Z 2 4 density (calcd) g/cm3 _{1.756 1.654} F(000) 1152 2380 T (K) 295(2) 295(2) wavelength (Å) 0.710 73 0.710 73 no. reflns collected 31 548 21 629

no. indep reflns (Rint) 10 191(0.0415) 9698(0.0331) R(F), wR2 [I > 2σ(I)] 0.0404, 0.1089 0.0350, 0.0973 R(F), wR2 (all data) 0.0536, 0.0626 0.0536, 0.1077

Figure 2.Face-to-faceπ-stacking of the pendant phenyl ring with the cylometalated difluorophenyl ring of complex 2 (left) and 3 (right). Counteranion PF6−and hydrogen atoms are omitted for clarity.

complexes exhibited emission maxima in the range of 500 and 505 nm. The HOMO orbital of a cationic iridium complex was reported to be a mixture of the orbitals of iridium and the phenyl group of the cyclometalated C∧N ligand, whereas the LUMO orbital is generally attributed to the ancillary N∧N ligand.18The slightly red-shifted emissions of 2 and 3 are due to larger conjugation of the N∧N ligands. Intriguingly, 3 emitted with a shoulder at∼485 nm which might be attributed to the component with longer Ir−N distance between the iridium center and the ancillary N∧N ligand due to the introduction of two pendant phenyl groups, causing less electron donating effect of the ancillary ligand toward the HOMO orbital. In thinfilms, the intermolecular interaction is dominant to the emission properties, as observed for red-shifted emission spectra of 1, 2, and 3 as compared with those of the PL in solutions. The slightly lower photoluminescent quantum yield (PLQY) of 3 in solution as compared to those of 1and 2 was possibly attributed to the intraligand interactions which may facilitate the deactivation processes. By calculating the radiative and nonradiative rate constants, the introduction of one phenyl ring increases both constants by a factor of 1.5. The increase of rate constants indicates that the formation of one intramolecular π−π stacking efficiently facilitates the electronic transitions between the excited states and the ground state due to enhanced molecular rigidity. However, doubling the intramolecularπ−π stacking significantly restricts the molecular reorganization in the excited state and thus gives less enhancement for the radiative process. Infilms, it is noted that both 2 and 3 exhibited much lower PLQYs than 1,

indicating that the presence of intramolecular π−π stacking leads to a densely extended π-stacking in the lattice by the formation of solid-state supramolecular-caged structure.5a

Electrochemical Properties. Figure 4 depicts the electro-chemical properties of complexes 1−3 probed by cyclic

voltammetry (CV, see the Experimental Section for details), and the measured redox potentials versus ferrocene/ferroce-nium redox couple are listed in Table 2. Notably, all complexes exhibited nearly identical redox behavior in terms of redox potentials and reversibility. A quasireversible oxidation potential for each complex was observed at around +1.25 V. The highest

Figure 3.Absorption (left axis) and PL (right axis) spectra of (a) 1, (b) 2, and (c) 3 in dichloromethane solutions (10−5M) and in neatfilms.

Table 2. Summary of Physical Properties of Complexes 1−3a

λPL(nm) PL QY

complex solutionb neatfilm solutionc neatfilm τ (μs) kr knr E1/2ox/E1/2red(V)d

1 500 512 0.46 0.20 0.54 8.5× 105 _{1.0× 10}6 _{+1.26/−1.71}

2 504 510 0.47 0.10 0.35 1.3× 106 _{1.5× 10}6 _{+1.25/−1.69}

3 505 510 0.38 0.11 0.38 1.0× 106 _{1.6× 10}6 _{+1.26/−1.68}

a_{At room temperature.} b_{Measured in CH}

2Cl2 (10−5 M). cMeasured in degassed CH2Cl2 solution (10−5 M). dRedox potential vs ferrocene/

ferrocenium redox couple.

Figure 4.Cyclic voltammograms of complexes 1−3 in acetonitrile. 0.1 M Tetrabutylammonium hexafluorophosphate (TBAPF6) and

tetra-butylammonium perchlorate (TBAP) were used as supporting electrolytes for the oxidation and reduction CV, respectively. A glassy carbon electrode was used as the working electrode; scan rate 100 mV/s.

The time-dependent brightnesses of the LEC devices based on complexes 1, 2, and 3 operating at 3.4 V are shown in Figure 5a. The current in the device increased slowly with time after the bias was applied. The times for turning on the device (ton),

of solid-state supramolecular-caged structure, which results in a densely extendedπ-stacking in the lattice and leads to low ionic mobility.5a Such molecular packing motif diminished the currentflux in the device and, as a result, substantially retarded the degradation of the complex and hence increased the device stability as well as EQE. Moreover, the complexes with preferable intramolecular π−π interaction can “lock” the comparable labile ancillary ligand which is prone to dissociation upon electron injecting to the antibonding (π*) orbital. As a consequence, we observed a 4-fold enhancement on the device lifetime (t_1/2) of 2 as compared to that of 1, stemming from less current flux and enhanced molecular stability. Nevertheless, doubling the intramolecular π−π interaction by two pendant groups in 3 did not render a superior device lifetime, as indicated by the shorter t1/2 of 3 as compared to that of 1. Costa et al. have depicted the pendant group effect of LEC performance in the 2,2′-bipyridyl system with DFT calcu-lations.19 The complex with two phenyl groups significantly lowers its metal-centered triplet state (3MC), rendering less luminance yield for radiative decay and more probability on degradation process (vide inf ra). Therefore, although the device based on 3 possesses suppressed currentflux, the unfavorable excited state process (thermal population of 3_{MC state) still} accelerated the degradation of the complex even with two intramolecularπ−π interactions. By inspecting the t_onand t_1/2 for the devices based on 1 and 2, it is suggested that the two device properties are in a trade-off. One cannot extend a device lifetime without sacrificing the response time (i.e., ton) of the device and vice versa. The stability of the ionic complex should be the major obstacle for obtaining high device lifetime with fast response.

Computational Studies. In order to understand the electronic structure of the complexes with the effect of the pendant phenyl group, we performed density functional theory (DFT) calculations at the B3LYP/(6-31G*/LANL2DZ) level on 1, 2, and 3 cations. The features of the frontier HOMO− 1, HOMO, and LUMO orbitals mainly involved in the lower lying electronic transitions are depicted in Figure 6. T1states of 1, 2, and 3 were calculated to lie 2.86, 2.84, and 2.80 eV above their S0states, respectively, which involve different contributions of electronic transitions. The dominant excitation for the T₁states in 1, 2, and 3 are HOMO → LUMO (95%), HOMO → LUMO (23%) with 21% LLCT (fluorene to 4,5-diazafluorene, HOMO− 1 → LUMO), and HOMO − 1 → LUMO (34%; LLCT), respectively. We observed that the percentage of LLCT transition increases along with the number of pendant phenyl ring(s). For example, the T1state of 1 is mainly contributed to Table 3. Characteristics of LEC Devices Based on

Complexes 1−3 3.4 V tona [min] tmaxb [min] t1/2c [min] Lmaxd

[cd/m2_{] [%, cd/A, lm/W]}ηmax _{(x, y)}CIE

1 0.5 54 141 25.4 0.82, 2.13, 1.97 0.23, 0.47 2 106 390 585 5.76 1.13, 3.37, 3.12 0.28, 0.50 3 47 96 102 10.6 1.43, 4.25, 3.92 0.28, 0.54 a

Time required to reach a brightness of 1 cd/m2.bTime required to reach the maximum brightness.c_{The time for the brightness of the}

device to decay from the maximum to half of the maximum under a constant bias voltage.dMaximum brightness achieved at 3.4 voltage.

Figure 5.Dependence of (a) brightness (filled symbols) and current density (empty symbols); (b) EQE (filled symbols) and correspond-ing power efficiency (empty symbols).

the HOMO→ LUMO excitation with 26.8% MLCT (metal d_π to the 4,5-diazafluorene moiety) character, whereas the T1 states of 2 and 3 involve higher percentage of LLCT transitions (21% and 34%, respectively). This result indicates that the dominating electronic transitions in 1−3 are fairly influenced by the presence of intramolecularπ−π stacking, restricting the molecular reorganization process in the excited state for MLCT transitions. Figure 7 illustrates the schematic energy diagram

showing the relative energy of the T1and3MC excited states. It is noted that the T₁and3_{MC excited states for 3 possess almost} the same energies while the3MC excited states are higher by 1.6 and 1.9 kcal mol−1for 1 and 2, respectively. The unexpected small energy difference between T1and3MC excited states for 3 decreases the molecular stability through population of the 3_{MC excited state (Table 2). Moreover, the longer Ir−N bonds} between the iridium and the N∧N ligand in the 3_{MC excited} states are calculated as 2.571, 2.403, and 2.707 Å for 1, 2, and 3, respectively, indicating that 3 has a higher tendency to lose the ancillary ligand in the 3MC excited states, in line with the shorter device lifetime of 3 as compared to that of 2. These Ir− N bonds involve spin density on the nitrogen atoms (Figure S1). The alteration of electron distribution in the3_{MC excited} states makes the bond lengths not simply increase with the increase of steric effect for 1 and 2.

■

In summary, we have successfully synthesized two cationic iridium complexes with one or two pendant phenyl groups on the spiro-configured N∧N ancillary ligand. The photophysical data revealed that the longer Ir−N distance between iridium and the ancillary ligand resulted in a blue-shifted emission spectrum due to less interaction between the ancillary ligand and the HOMO orbital (complex 3). The X-ray crystallography shows substantial intramolecular π−π interaction between pendant phenyl group and the dfppz ligand. The plane to plane distances indicating sufficient π−π interaction are about 3.38 Å for complex 2 and 3.40 and 3.46 Å for complex 3. This π−π interaction minimizes the expansion of the metal−ligand bonds in the excited state and capability of preventing foreign substances to access the metal center. The LEC devices based on 2 and 3 showed much suppressed current densities which resulted from densely extended π-stacking structures. The device lifetime of the devices based on 1, 2, and 3 were 141, 585, 102 min, respectively. The 4-fold enhancement on the device lifetime of 2 as compared to that of 1 is benefited by less current flux and enhanced molecular stability. However, the trade-off of device response and device lifetime limits the practical use of LEC devices. The stability of the ionic complex under the electricfield is considered to be the major obstacle for realizing practical LEC devices. Some possible solutions are foreseen for searching for a new type of stable molecular architecture with ionic character to allow higher durability under the electricfield and developing stable neutral complexes compatible with ionic liquids to give LEC response without using degradable ionic complexes. Work related to improving LEC lifetime without sacrificing device response is under investigation and will be reported in due course.

■

*

Additional tables, figure, and crystallographic data in CIF format. This material is available free of charge via the Internet at http://pubs.acs.org.

■

Corresponding Author

*E-mail: [email protected] (K-T.W.), wenhung@mail. ntou.edu.tw (W.-Y.H.), [email protected] (H.-C.S.), [email protected] (P.-T.C.).

Notes

The authors declare no competingfinancial interest.

■

The work was supported by thefinancial aid of the National Science Council of Taiwan (NSC-98-2119-M-002-007-MY3 and 100-2112-M-019-002-MY3).

■

(1) (a) Pei, Q.; Yu, G.; Zhang, C.; Yang, Y.; Heeger, A. J. Science 1995, 269, 1086. (b) Pei, Q.; Yang, Y.; Yu, G.; Zhang, C.; Heeger, A. J. J. Am. Chem. Soc. 1996, 118, 3922.

(2) (a) Lee, J. K.; Yoo, D. S.; Handy, E. S.; Rubner, M. F. Appl. Phys. Lett. 1996, 69, 1686. (b) He, L.; Duan, L.; Qiao, J.; Wang, R.; Wei, P.; Wang, L. D.; Qiu, Y. Adv. Funct. Mater. 2008, 18, 2123. (c) Colman, E. Z.; Slinker, J. D.; Parker, J. B.; Malliaras, G. G.; Bernhard, S. Chem. Mater. 2008, 20, 388. (d) Su, H.-C.; Chen, H.-F.; Fang, F.-C.; Liu, C.-C.; Wu, C.-C.-C.; Wong, K.-T.; Liu, Y.-H.; Peng, S.-M. J. Am. Chem. Soc. 2008, 130, 3413. (e) Bolink, H. J.; Coronado, E.; Costa, R. D.; Lardiés, Figure 6. Frontier orbitals involved in the lower-lying electronic

transitions for 1, 2, and 3.

Figure 7.Schematic energy diagram showing the relative energy of the

Cornil, J.; Hartmann, D.; Sarfert, W.; Fröhlich, R.; Bizzarri, C.; De Cola, L. Inorg. Chem. 2010, 49, 9891. (q) Chen, H.-F.; Wong, K.-T.; Liu, Y.-H.; Wang, Y.; Cheng, Y.-M.; Chung, M.-W.; Chou, P.-T.; Su, H.-C. J. Mater. Chem. 2011, 21, 768.

(3) (a) Slinker, J. D.; Gorodetsky, A. A.; Lowry, M. S.; Wang, J.; Parker, S.; Rohl, R.; Bernhard, S.; Malliaras, G. G. J. Am. Chem. Soc. 2004, 126, 2763. (b) Hu, T.; He, L.; Duan, L.; Qui, Y. J. Mater. Chem. 2012, 22, 4206.

(4) Kalyuzhny, G.; Buda, M.; McNeill, J.; Barbara, P.; Bard, A. J. J. Am. Chem. Soc. 2003, 125, 6272.

(5) (a) Bolink, H. J.; Coronado, E.; Costa, R. D.; Ortí, E.; Sessolo, M.; Graber, S.; Doyle, K.; Neuburger, M.; Housecroft, C. E.; Constable, E. C. Adv. Mater. 2008, 20, 3910. (b) Graber, S.; Doyle, K.; Neuburger, M.; Housecroft, C. E.; Constable, E. C.; Costa, R. D.; Ortí, E.; Repetto, D.; Bolink, H. J. J. Am. Chem. Soc. 2008, 130, 14944. (c) Lenes, M.; Garcia-Belmonte, G.; Tordera, D.; Pertegás, A.; Bisquert, J.; Bolink, H. J. Adv. Funct. Mater. 2011, 21, 1581. (d) Tordera, D.; Meier, S.; Lenes, M.; Costa, R. D.; Ortí, E.; Sarfert, W.; Bolink, H. J. Adv. Mater. 2012, 24, 897.

(6) (a) Sun, L.; Galan, A.; Ladouceur, S.; Slinker, J. D.; Zysman-Colman, E. J. Mater. Chem. 2011, 21, 18083. (b) He, L.; Duan, L.; Qiao, J.; Zhang, D.; Wang, L.; Qiu, Y. Chem. Commun. 2011, 47, 6467. (c) Rothe, C.; Chiang, C.-J.; Jankus, V.; Abdullah, K.; Zeng, X.; Jitchati, R.; Batsanov, A. S.; Bryce, M. R.; Monkman, A. P. Adv. Funct. Mater. 2009, 19, 2038.

(7) Wong, K.-T.; Chen, R.-T.; Fang, F.-C.; Wu, C.-C.; Lin, Y.-T. Org. Lett. 2005, 7, 1979.

(8) Su, H.-C.; Fang, F.-C.; Hwu, T.-Y.; Hsieh, H.-H.; Chen, H.-F.; Lee, G.-H.; Peng, S.-M.; Wong, K.-T.; Wu, C.-C. Adv. Funct. Mater. 2007, 17, 1019.

(9) (a) Sprouse, S.; King, K. A.; Spellane, P. J.; Watts, R. J. J. Am. Chem. Soc. 1984, 106, 6647. (b) Lohse, O.; Thevenin, P.; Waldvogel, E. Synlett 1999, 1, 45. (c) Nonoyama, M. Bull. Chem. Soc. Jpn. 1974, 47, 767.

(10) Otwinowski, Z.; Minor, W. In Methods in Enzymology: Macromolecular Crystallography, Part A; Carter, C. W., Jr., Sweet, R. M., Eds; Academic Press: San Diego; Vol. 276, 2005.

(11) Sheldrick, G. M. SHELXL-97, Program for the Solution of Crystal Structures; University of Göttingen: Göttingen, Germany, 1997.

(12) Parker, S. T.; Slinker, J. D.; Lowry, M. S.; Cox, M. P.; Bernhard, S.; Malliaras, G. G. Chem. Mater. 2005, 17, 3187.

(13) (a) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785. (b) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.

(14) (a) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270. (b) Wadt, W. R.; Hay, P. J. J. Chem. Phys. 1985, 82, 284. (c) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299.

(15) Hariharan, P. C.; Pople, J. A. Mol. Phys. 1974, 27, 209. (16) Abrahamsson, M.; Lundqvist, M. J.; Wolpher, H.; Johansson, O.; Eriksson, L.; Bergquist, J.; Rasmussen, T.; Becker, H.-C.; Hammarstroem, L.; Norrby, P.-O.; Akermark, B.; Persson, P. Inorg. Chem. 2008, 47, 3540.