An isomeric strategy for enhancing phosphorescence efficiency of iridium(III) complexes with N-heterocyclic naphthyridine ligands: A theoretical study

Organic Electronics 22 (2015) 180–190

Contents lists available at ScienceDirect

Organic Electronics journal homepage: www.elsevier.com/locate/orgel

An isomeric strategy for enhancing phosphorescence efficiency of iridium(III) complexes with N-heterocyclic naphthyridine ligands: A theoretical study Ming-Shuo Ma a,b, Lu-Yi Zou a, Yan Li a,c, Ai-Min Ren a,⇑, Ji-Kang Feng a a b c

State Key Laboratory of Theoretical and Computational Chemistry, Institute of Theoretical Chemistry, Jilin University, Changchun 130023, People’s Republic of China Center of Analysis and Measurement, Jilin Institute of Chemical Technology, Jilin City 132022, People’s Republic of China School of Chemical and Pharmaceutical Engineering, Jilin Institute of Chemical Technology, Jilin 132022, People’s Republic of China

a r t i c l e

i n f o

Article history: Received 21 January 2015 Received in revised form 14 March 2015 Accepted 27 March 2015 Available online 28 March 2015 Keywords: Naphthyridine Spin–orbit coupling Phosphorescence quantum efficiency

a b s t r a c t The electronic structures and photophysical properties of six isomeric Ir(III) complexes with different N-heterocyclic naphthyridine ligands were investigated by density functional theory (DFT) and time dependent DFT (TD-DFT) approach. The radiative transition rates (kr) were determined through calculated the spin–orbital coupling (SOC) matrix elements hT m jHSOC jSn i and the energy levels (ESn and ETm). The non-radiative transition rates (knr) were estimated through analysis of the structural distortions, the d-orbital splittings and the energy differences between the S0 and T1 states DE(T1  S0).   As the results, the ESn, the ETm and the energy splittings DES1 T m and DET m T m1 can be regulated by the position of two nitrogen atoms in naphthyridine ring for studied complexes. Moreover, Ir(III) complex inclusive of quinoxaline heterocyclic ring presents large kr and knr, so its phosphorescence quantum efficiency is difficult up to be 100%. While two Ir(III) complexes bound to quinazoline heterocyclic ring show weakly emissive because of large knr. Notably, the presence of the cinnoline heterocyclic ring in the Ir(III) complex makes singlet–triplet intersystem (ISC) rate and kr fast but knr slow, then leads to its high phosphorescence quantum efficiency. Ó 2015 Elsevier B.V. All rights reserved.

1. Introduction Phosphorescent transition-metal complexes have come into the focus of research, due to their potential application as highly efficient electroluminescent emitters in organic light emitting diodes (OLEDs) [1]. The strong spin–orbital coupling of heavy-atom effect, increases the probability of efficient intersystem crossing (ISC) from the excited singlet state to the light emitting excited triplet state [2], and breaks the principle of triplet spin-forbidden, makes the complexes harvest both singlet and triplet excitons, so the internal quantum efficiency even can reach as much as 100% [3]. Thompson and Förrest synthesized a series of homoleptic or heteroleptic iridium(III) complexes with strong-field C^N or N^N cyclometalating ligand [4]. Ir(III) complexes have been recognized as the excellent phosphors because of their highly phosphorescence quantum yields, good photo- and thermal stabilities, facile color tuning through ligand structure control, relatively short phosphorescence lifetime, and high color purity [5]. Moreover, ⇑ Corresponding author. E-mail address: [email protected] (A.-M. Ren). http://dx.doi.org/10.1016/j.orgel.2015.03.037 1566-1199/Ó 2015 Elsevier B.V. All rights reserved.

the third-row transition complexes with quasi-octahedral geometries, have a large Ddd⁄ (the splitting between the highest occupied and the lowest unoccupied d orbitals) and a small Dddocc (the energy gap between the two highest occupied d-orbitals). A large Ddd⁄ may result in an thermal unaccessible metal-centered (MC) dd excited states and reduce non-radiative quenching [6]. A small Dddocc means a strong spin–orbital coupling (SOC) and a higher radiative transition rate kr [7]. Iridium complexes are the most successful family of phosphors for OLEDs in the visible range especially for green and blue emitters, however, in contrast to the well-developed short-wavelength phosphorescent materials, long-wavelength orange to red-emitting iridium complexes are prone to low quantum yields due to the smaller energy gap [8]. Therefore, how to overcome its intrinsic defects via modifying cyclometalating ligand is still a cutting-edge research topic for high efficient red-emitting complexes. The most commonly adopted strategy is inserted an additional sp2-hybridized N in the pyridyl chelating ring to form naphthyridine-containing heterocycles such as quinoxaline [9], quinazoline [9b,10], and cinnoline [11] derivatives etc. This tuning strategy, which involving direct nitrogenfor-carbon substitution at the p-framework, should be as good as

M.-S. Ma et al. / Organic Electronics 22 (2015) 180–190

or even better than the traditional method of using the electrondonating or electron-withdrawing substituents. It exhibits shorter excited-state lifetimes, small efficiency roll-off and more thermally stabilities in theses dinitrogen-containing heterocycle complexes compared to those normal nitrogen-containing counterpares [8b,9a–12]. However, the phosphorescence quantum efficiency of these complexes are far from satisfied and need to be further explored. Unfortunately, to our knowledge, there has been no systematic comparison and analysis of the effect of the non-chelating sp2-hybridized N on the phosphorescence quantum efficiency. Recently, one iridium complex bearing two 2-(4-fluorophenyl)-3methyl-quinoxaline (fpq) ligands and an ancillary ligand: triazolylpyridine (trz) was reported and proved to be strong phosphorescent material with short-living phosphorescent decay. Complex (fpq)2Ir(trz) (hereafter noted by complex 1) shows a superior operating lifetime in device compared to [Ir(piq)3] (a widely used kind of red emission material in OLED devices) [13]. Based on the structure of complex 1, we altered the relative position of the two nitrogen atoms in the naphthyridine ring and proposed several additional modified structures [trans-2-fp-4-mqz]2Ir(trz) (2), [cis2-fp-4-mqz]2Ir(trz) (3), [4-fp-2-mqz]2Ir(trz) (4), [1-fp-4-mp l]2Ir(trz) (5), [3-fp-4-mcn]2Ir(trz) (6), in which fp = fluorophenyl, mpl = methylphthalazine, mcn = methylcinnoline, respectively (in Scheme 1). In this work, we are focused on the influence of the replacement position of the non-chelating nitrogen atom in the naphthyridine-containing heterocyclic ring on the phosphorescent properties of the isomeric Ir(III) complexes. Toward this goal, we implement a comprehensive quantum chemistry calculation on complexes 1–6 to explore the effect of the introducing positions of the non-chelating nitrogen atom (para-, meta- and ortho-positions) on the photophysical property of these complexes. The electronic structures, the frontier molecular orbitals (FMOs) and UVabsorption spectra are investigated, moreover, we explored mechanism of the phosphorescent emission for the series of Ir(III) complexes from microscopic theoretical perspectives, such as the relationship between the spin–orbital coupling (SOC) matrix elements hT m jHSOC jSn i, the energy splittings of the Sn and Tm states, and the radiative transition rates (kr), meanwhile we analyzed the factors affecting the non-radiative transition rates for the

181

complexes. Finally, we predicted their phosphorescence quantum efficiency (/p). 2. Computational details The ground-state geometries were optimized by density functional theory (DFT) [14] with the Becke’s three-parameter hybrid method combined with the Lee–Yang–Parr correlation functional (denoted as B3LYP) [15]. There were no symmetry constraints on these complexes. Vibrational frequencies were performed at the same theoretical level to confirm that each configuration is stable structure. The lowest lying triplet excited state geometries were optimized by the unrestricted B3LYP (UB3LYP) [16], and the calculated spin contamination are rather small (the expectation values of spin operator hS2i were all below 2.03 for triplet excited state), which can be neglected. Recent calculations with TDDFT method for transition-metal complexes have been supported by experimental spectra and the reliability of TDDFT method was valid [17]. Thus, at the respective optimized geometries of the ground and excited states, the molecular orbital compositions, absorption and emission spectra of the complexes in dichloromethane (CH2Cl2) solvent were calculated by time-dependent DFT (TDDFT) [18] method associated with the polarized continuum model (PCM) [19]. In the calculation, the choice of appropriate functionals and basis sets is crucial step for rational and precise prediction. So besides B3LYP, several other frequently used functionals including the PBE0, X3LYP, O3LYP, B3P86, B3PW91, CAM-BLYP and LC-BLYP [20] were picked out to evaluate the transition properties of complex 1. With regard to basis sets, the SDD ECP basis set for iridium is adequate to describe the ground state geometries of the Ir(III) complexes, which has been discussed by Li et al. [21]. As to the nonmetal atoms (C, N, H and F atoms), we employed a couple of basis sets to obtain more accurate excitation energies and electronic transition properties closer to the experimental values. Finally, the combination of the SDD of ECP basis set for iridium together with the D95V basis set for rest of atoms performs better performance over the other collocations. The comparative details

Scheme 1. Chemical structures sketch of Ir(III) complexes 1–6.

182

M.-S. Ma et al. / Organic Electronics 22 (2015) 180–190

Fig. 1. Calculated absorption wavelength (kabs, nm) for complex 1 by TD-DFT with different functionals (a) and basis sets (b).

were shown in Fig. 1. All the calculations were performed with Gaussian 09 software package [22]. 3. Results and discussion 3.1. Geometries in the ground state S0 and triplet excited state T1 The optimized geometry results show that all the complexes are distorted octahedron and the optimized ground-state geometrical structure of complex 1 is illustrated as a representation in Fig. 2. The three ligands around the metal center are almost orthogonal to each other, because the dihedral angles N1–Ca1–Ir–Cb1 and N5–N6–Ir–Cb1 are approximately to 90°. The trigeminal structures separate the donor (trz) from acceptor (fpq) moieties, limit orbital overlap and nonradiative decay [23]. In complexes 5 and 6, the IrN1(Ir–N3) bond is apparently shorter than that in others and this can be attributed to the interaction enhancement of bond N@N than bond N@C. According to this effect, it possibly increases the radiative transition rate due to the fpq ligands in 5 and 6 strongly coordinate with the iridium atom. In addition, introducing the electron-donating substituent –CH3 into 4-position in the complex 6 will be feasible to concentrate electrons on the cinndinyl ring compared with introducing it into 3-position in the complex 5, thus, 6 has a more tight and compact structure, which is benefit decreasing the non-radiative transition rate [19]. In complexes 2 and 3, the Ir-ligand bonds lengthened, indicating that the interactions between the iridium atom and the ligands weaken, which might decrease the spin–orbital coupling effects, and then facilitates the non-radiative transition.

From Fig. 2, the ancillary ligand triazolylpyridine (trz) remains planarity, and these rigid skeletal structure may effectively prevent the non-radiative decay. While the two cyclometalating ligands (fpq) are not in a plane, the fluorophenyl substituent bends slightly out from the naphthyridine moieties plane and the torsion angle sizes follow the order of complex 4 > 6 > 1 > 5 > 2 > 3. These torsions mainly depend on the interaction between the iridium atom and the cyclometalating ligands. For example, the fluorophenyl fragment in complex 4 possesses a much larger bending degree to the quinazoline plane than that in complexes 2 and 3, which can be ascribe to the cooperative effect of p-electron deloclization and the electron-donating –CH3 substituents. In general, a large torsion degree means a strong interaction between the metal and the ligands, which might increase the spin–orbital coupling (SOC) effects, and facilitate the radiative transition of triplet excited state towards to singlet ground state. It is known that the structural differences of the T1 excited state from S0 are related to the non-radiative transition rate [6], that is, larger structural distortion accelerates the non-radiative transition. For this reason, we separately optimized the S0 and T1 state geometries and have obtained the geometric parameters at the S0 and T1 states (listed in Table 1). To visualize the change, the deviations of bond lengths and dihedral angles between the two states were depicted in Fig. 3. The changes of bond lengths mainly occurred among bond lengths 1, 2, 3, 4, 7 as well as 10 and 11 and the bonds directly connecting to the metal atom evidently changed. For complexes 1 and 4, the bond lengths of Ir–Cb1 and Ir–N3 reduce apparently, but Ir– Ca1 and Ir–N1 are slightly longer on going from the S0 to the T1

Fig. 2. Optimized ground and chemical structure of complex 1 at the B3LYP level.

183

M.-S. Ma et al. / Organic Electronics 22 (2015) 180–190 Table 1 Main parameters of optimized geometries at the B3LYP level for 1–6. 1 S0 Bond length/Å Ir–N1 Ir–N3 Ir–N5 Ir–N6 Ir–Ca1 Ir–Cb1 Bond angle/° Ca1–Ir–N1 Cb1–Ir–N3 N1–Ir–N6

a

2.107 2.101 2.258 2.144 2.011 2.020

2 T1

Expt [13]

2.132 2.037 2.272 2.139 2.013 2.016

2.075 2.059 2.213 2.147 1.992 1.980

3

S0 2.119 2.113 2.250 2.141 2.021 2.031

T1 2.042 2.144 2.255 2.126 2.017 2.042

4

S0 2.119 2.113 2.256 2.142 2.022 2.031

T1 2.130 2.117 2.260 2.135 2.016 2.011

S0 2.114 2.114 2.240 2.132 2.012 2.023

5 T1 2.133 2.072 2.247 2.137 2.017 2.013

S0 2.047 2.053 2.201 2.115 2.020 2.032

6 T1 1.988 2.085 2.211 2.092 2.028 2.046

S0 2.040 2.047 2.207 2.118 2.021 2.047

T1 2.048 2.068 2.227 2.095 1.999 2.028

79.2 79.2 81.4

79.0 80.6 82.1

78.9 79.6 80.9

79.8 79.8 82.9

81.2 79.5 85.3

80.4 80.6 82.8

80.5 80.6 82.6

79.4 79.4 82.1

79.3 80.9 82.2

79.3 79.0 87.8

80.9 78.3 92.6

80.0 79.7 87.6

80.0 79.9 86.8

Dihedral angle/° 88.1 h1a h2 81.4 h3 79.6

88.4 82.3 80.1

87.6 78.2 72.5

87.1 82.4 81.1

87.7 82.2 83.9

88.2 81.9 80.9

86.4 80.7 82.4

89.9 81.4 80.4

89.9 81.6 80.1

90.0 87.5 87.1

89.3 87.5 87.9

89.8 87.5 87.0

89.6 86.7 86.5

h1 = h(N1–Ca1–Ir–Cb1), h2 = h(N5–N6–Ir–Cb1), h3 = h(N5–N6-Ir–Ca1).

Fig. 3. Deviations of the bond lengths and dihedral angles between T1 excited state and the S0 state.

state. In result, the interaction between the fpq2 ligand and metal atom enhance in the T1 excited state. Another variation occurs in complex 2, the distance between the fpq1 ligand and metal atom is shortened from the S0 to T1 state. In addition, in complexes 1

and 4, there are nearly no changes in the bonds 13–18. It is worth paying attention that in complex 6, the deviation values narrowly fluctuate around zero, and that is, its bond lengths have the smallest change among the isomeric iridium complexes.

184

M.-S. Ma et al. / Organic Electronics 22 (2015) 180–190

The variations of dihedral angles are also different. Take 2 and 3 for example, their maximum deviations are 2.8° and 1.7°, respectively, which are more pronounced than that of 4 (0.03°) and 5 (0.01°). It indicates that complexes 2 and 3 undergo significant structural distortions, while the geometric relaxations of complexes 4 and 5 change slightly. Considering from the two aspects, 2 and 3 undergo maximum structural distortion, next 1 and 5, 4 and 6. These results presage that the non-radiative transition rates of 4 and 6 may be slower than that of 2 and 3.

3.2. The frontier molecular orbital properties At the optimized ground-state structures, the frontier orbital energy levels, energy gaps and electronic density contour diagrams of the highest occupied molecular orbitals (HOMOs) and the lowest unoccupied molecular orbitals (LUMOs) for these complexes were represented in Table 2 and Fig. 4, and the d-orbital compositions of iridium atom at the S0 and T1 states were listed in Tables S2–S7, respectively.

Table 2 Molecular orbital compositions in the ground state for complexes 1–6. Orbital

Energy [eV]

Contribution [%] Ir

a

Assign

fpq1

fpq2

trz

Nan1

Fph1a

Nan2

Fph2

1 L+2 L+1 L H H1 H2 H3 H4

1.98 2.71 2.85 5.83 5.99 6.48 6.70 6.74

3.3 4.6 4.7 17.2 35.0 25.9 10.4 25.2

1.1 10.6 67.2 3.7 5.1 13.3 5.1 12.4

1.1 2.8 12.1 3.6 14.5 11.3 5.8 7.4

1.6 68.9 9.6 2.4 6.9 18.0 10.2 10.4

1.9 10.6 4.0 3.7 23.2 21.6 10.8 12.3

91.0 2.4 2.5 69.3 15.3 9.9 57.9 32.3

p⁄(trz) p⁄(Nan1 + Nan2 + Fph2 p⁄(Nan1 + Fph1) d(Ir) + p(trz) d(Ir) + p(Fph1 + Fph2 + trz) d(Ir) + p(Nan1 + Fph1 + Nan2 + Fph2 + trz) d(Ir) + p(Nan2 + Fph2 + trz) d(Ir) + p(Nan1 + Nan2 + Fph2 + trz)

2 L+2 L+1 L H H1 H2 H3 H4

1.94 2.48 2.55 5.73 5.91 6.32 6.55 6.73

3.4 3.2 3.2 27.2 30.1 28.5 41.5 12.3

1.1 4.1 80.9 3.3 4.2 10.9 14.6 3.5

1.0 1.3 9.3 5.9 12.1 10.4 12.3 6.4

1.8 82.6 3.4 3.6 4.2 16.2 11.5 3.8

1.8 7.6 1.6 9.0 18.1 22.2 17.2 8.9

90.9 2.0 1.5 50.9 31.3 11.9 2.8 65.3

p⁄(trz) p⁄(Nan2) p⁄(Nan1 + Fph1) d(Ir) + p(Fph2 + trz) d(Ir) + p(Fph1 + Fph2 + trz) d(Ir) + p(Nan1 + Fph1 + Nan2 + Fph2 + trz) d(Ir) + p(Nan1 + Fph1 + Nan2 + Fph2) d(Ir) + p(Fph2 + trz)

3 L+2 L+1 L H H1 H2 H3 H4

1.97 2.39 2.45 5.73 5.91 6.42 6.53 6.64

3.4 2.6 2.4 28.2 25.7 5.0 53.2 7.0

2.7 1.0 89.4 3.0 2.2 9.0 7.4 22.1

1.4 1.5 4.0 8.2 7.9 18.6 2.7 34.0

1.9 90.7 1.1 3.7 5.4 24.7 6.9 4.6

2.0 2.7 2.1 11.5 18.5 40.3 7.6 24.7

88.6 1.5 1.1 45.5 40.3 2.5 22.2 7.7

p⁄(trz) p⁄(Nan2) p⁄(Nan1) d(Ir) + p(Fph1 + Fph2 + trz) d(Ir) + p(Fph2 + trz) p(Nan1 + Fph1 + Nan2 + Fph2) d(Ir) + p(trz) p(Nan1 + Fph1 + Fph2)

4 L+2 L+1 L H H1 H2 H3 H4

1.97 2.62 2.77 5.81 6.02 6.44 6.66 6.71

3.0 3.3 4.0 15.2 41.1 36.1 2.8 26.9

2.1 4.0 70.8 2.5 4.0 11.3 1.1 4.9

0.8 1.6 18.2 2.2 14.2 6.5 2.9 8.8

1.2 62.4 3.1 1.4 5.4 17.1 7.8 18.9

1.8 27.4 1.9 2.2 25.3 17.2 7.4 18.1

91.0 1.2 2.0 76.5 10.0 11.8 77.9 22.4

p⁄(trz) p⁄p(Nan2 + Fph2) p⁄(Nan1 + Fph1) d(Ir) + p(trz) d(Ir) + p(Fph1 + Fph2 + trz) d(Ir) + p(Nan1 + Nan2 + Fph2 + trz) p(trz) d(Ir) + p(Fph1 + Nan2 + Fph2 + trz)

5 L+2 L+1 L H H1 H2 H3 H4

1.81 2.41 2.48 5.60 5.77 6.17 6.35 6.64

1.6 3.3 4.0 13.5 44.9 40.4 51.8 1.0

94.1 1.3 77.1 2.3 3.9 9.4 13.1 0.4

1.2 1.9 14.2 1.2 15.6 8.2 10.3 1.3

1.2 78.6 1.4 1.9 5.9 14.1 6.7 2.5

0.5 13.7 1.9 1.5 23.7 11.7 9.2 3.3

1.3 1.2 1.4 79.5 6.0 16.3 9.0 91.4

p⁄(Nan1) p⁄(Nan2 + Fph2) p⁄(Nan1 + Fph1) d(Ir) + p(+trz) d(Ir) + p(Fph1 + Fph2) d(Ir) + p(Nan1 + Fph1 + Nan2 + Fph2 + trz) d(Ir) + p(Nan1 + Fph1 + Fph2 + trz) p(trz)

6 L+2 L+1 L H H1 H2 H3 H4

1.92 2.59 2.62 5.66 5.73 6.31 6.44 6.53

2.3 4.1 3.5 10.6 41.7 12.9 33.2 11.1

71.8 2.1 91.5 2.1 4.0 10.7 6.2 1.6

22.6 1.2 0.7 0.9 15.9 18.4 8.2 1.5

0.7 91.0 2.1 1.6 7.1 19.0 10.7 2.4

1.2 0.7 1.0 1.1 26.9 30.8 6.0 1.7

1.5 1.0 1.2 83.7 4.4 8.2 35.6 81.7

p⁄(Nan1 + Fph1) p⁄(Nan2) p⁄(Nan1) d(Ir) + p(trz) d(Ir) + p(Fph1 + Fph2) d(Ir) + p(Nan1 + Fph1 + Nan2 + Fph2 + trz) d(Ir) + p(Fph1 + Nan2 + trz) d(Ir) + p(trz)

Here, Nan represents naphthyridine moiety, Fph represents fluorophenyl moiety.

M.-S. Ma et al. / Organic Electronics 22 (2015) 180–190

185

Fig. 4. Frontier molecular orbital energy levels and contour plots of HOMO and LUMO for 1–6.

In according to the difference in components of the HOMOs, the six complexes can be divided into two groups: 1, 4, 5 and 6 belong to group I, 2 and 3 belong to group II. In group I, the HOMOs of the complexes are mainly localized on the 5d (Ir) and the p-orbital of the trz ligand, whereas their LUMOs are delocalized on the p⁄-orbitals of fpq1 ligand. In group II, the HOMOs of the complexes are more delocalized compared to that in group I, there are additional contributions from the fpq ligands. From the frontier molecular orbital compositions at the S0 state (Table 2) and the T1 state (Tables S2–S7), we found that these complexes have large percentage of metallic orbital composition in the HOMOs. The detailed data are: 1 (S0: 17.2%, T1: 13.1%), 2 (S0: 27.2%, T1: 28.2%), 3 (S0: 28.2%, T1: 32.0%), 4 (S0: 15.2%, T1: 13.4%), 5 (S0: 13.5%, T1: 17.8%) and 6 (S0: 10.6%, T1: 14.0%), respectively. In complexes 2, 3, 5 and 6, the percentage of metal atom participating in HOMOs in the T1 state is greater than that in the S0 state, but the cases in complexes 1 and 4 are just reversed. In addition, for 1–6, the dorbital space spreading compositions in the HOMOs are different. In the S0 state, they are: dxz, dx2  y2 and dxy for 1, dxz, dyz, dx2  y2, and dxy for 2, dxz, dyz and dxy for 3, and dx2  y2 and dxy for 4, 5, and 6, respectively. In the T1 state, the d-orbital compositions of 1 change to dx2  y2 and dxy, and that of 2 change to dxz, dx2  y2 and dxy, but that of other complexes unchange. According to the d-orbital SOC matrix element values in Table S1, the dxy orbital with dx2  y2 orbital coupling will generate a larger SOC, facilitate faster radiative transition. Therefore, these

complexes can have stronger spin–orbital coupling except for complex 3. In the view of the chemical structure, the location of two nitrogen atoms in the naphthyridine moieties would greatly affect the compositions and the energy levels of the frontier molecular orbitals, the HOMO energy levels follow the order of 5 and 6 (N2 ortho to N1) > 2, 3 and 4 (N2 meta to N1) > 1 (N2 para to N1). The LUMO of 6 has predominant contribution from naphthyridine (Nan) p⁄ orbital character, however, the LUMO of 5 possess naphthyridine (Nan) and fluorophenyl (Fph) p⁄ orbital characters, suggesting the electronic cloud distribution of LUMOs localized on the naphthyridine moiety rather than the fluorophenyl moiety when the electron-donating substituent –CH3 is separated by the non-chelating nitrogen atom. 3.3. Absorption spectra in CH2Cl2 solvent Simulated absorption spectra of these complexes in CH2Cl2 solvent were shown in Fig. 5. The detailed information, such as excitation energies, oscillator strengths (f), dominant configurations, transition nature, and the experimental values were listed in Table 3. The calculated absorption spectra of complex 1 can well reproduce the experimental absorption profile in band position, intensity and separation [13]. Their lowest singlet–singlet excitation energies increase in the order of 1 (2.32 eV) < 6 (2.40 eV) < 4 (2.44 eV) < 5 (2.49 eV)2 < (2.53 eV) < 3 (2.67 eV), which is consistent with the

186

M.-S. Ma et al. / Organic Electronics 22 (2015) 180–190

Fig. 5. Simulated absorption spectra in CH2Cl2 for 1–6.

law of the energy gaps except for complex 6. The reason may be that the lowest singlet excitated state in 6 is inclusive of not only the electron promotion from HOMO to LUMO [24], but also other

low-lying one-electron excitation configurations, which will be mentioned in the next section. For these complexes, there are two dominant regions in the absorption spectra, the calculated lowest-lying absorption bands of 1–5 at 534, 490, 465, 508, 497 nm, are assigned to the S0 ? S1 electronic transition (at least 89%), mainly arising from the HOMO ? LUMO configurations. For complex 1, the calculated singlet ? triplet transition is 593 nm with featureless band shapes which is very close to the experimental value 590 nm. For 1, 4 and 5, the HOMO ? LUMO transitions show mixed MLCT [d(Ir) ? p⁄(fpq1 + fpq2)]/LLCT[p(trz) ? p⁄(fpq1 + fpq2)] transition. For 2 and 3, the transition characters are assigned as possessing MLCT [d(Ir) ? p⁄(fpq1 + fpq2)]/LLCT[p(trz) ? p⁄(fpq1 + fpq2)]/ ILCT[p(fpq1 + fpq2) ? p⁄(fpq1)] transitions. While in 2 and 3, the occurrence of the intraligand transition motions (ILCT) in the same ligand will weaken the metal–ligand bond, deduced SOC effect, result in low phosphorescence quantum efficiency. Therefore, it is sensible to avoid the hazardous transition for deed red phosphorescence complexes [25]. Therefore, in complexes 2 and 3, the presence of electronic intraligand transitions might be responsible for a small kr. For complex 6, the absorption sprectrum at 516 nm is composed of HOMO ? LUMO (67%) and HOMO-3 ? LUMO (30%) transitions, these multiple transitions may involve more metalto-ligand charge transfer, increase the spin–orbital coupling effect.

Table 3 Calculated absorption spectra in CH2Cl2 at TD-B3LYP level for 1–6.

1

2

3

4

5

6

States

k (nm)/E [eV]

Oscillator

Main configurations

Assign

Expt [13]

S1 S4 S9

534/2.32 463/2.68 376/3.30

0.0086 0.0299 0.1628

372/3.34

0.0578

S11

363/3.41

0.0553

S12

360/3.45

0.1026

S15

343/3.61

0.0457

MLCT/LLCT MLCT/LLCT MLCT/LLCT/ILCT MLCT/LLCT/ILCT MLCT/LLCT MLCT/LLCT MLCT/ILCT MLCT/LLCT/ILCT MLCT/ILCT MLCT/LLCT MLCT/LLCT

539 449 372

S10

H ? L (89%) H  1 ? L + 1 (89%) H  3 ? L (53%) H  4 ? L (31%) H  4 ? L + 1 (32%) H  3 ? L + 1 (31%) H  1 ? L + 2 (43%) H  4 ? L + 1 (21%) H  1 ? L + 2 (40%) H  4 ? L + 1 (20%) H  5 ? L + 1 (34%)

S1 S4 S9

490/2.53 434/2.86 368/3.37

0.0152 0.0336 0.0667

S14

342/3.62

0.1140

H ? L (93%) H  1 ? L + 1 (93%) H  1 ? L + 2 (47%) H ? L + 3 (37%) H  4 ? L (54%) H  1 ? L + 3 (28%)

MLCT/LLCT MLCT/LLCT/ILCT MLCT/LLCT/ILCT MLCT/LLCT MLCT/LLCT MLCT/LLCT

S1 S4 S8

465/2.67 419/2.96 370/3.35

0.0100 0.0035 0.0655

S14

348/3.56

0.1014

H ? L (95%) H  1 ? L + 1 (95%) H  1 ? L + 2 (51%) H ? L + 4 (31%) H  2 ? L + 1 (55%) H  1 ? L + 4 (20%)

MLCT/LLCT/ILCT MLCT/LLCT/ILCT MLCT/LLCT MLCT/LLCT LLCT/ILCT MLCT/LLCT

S1 S4 S9

508/2.44 444/2.79 364/3.38

0.0187 0.0790 0.1941

S13

350/3.54

0.0839

H ? L (88%) H  1 ? L + 1 (89%) H  4 ? L (39%), H  3 ? L (27%) H  3 ? L + 1 (39%) H  4 ? L + 1 (32%)

MLCT/LLCT MLCT/LLCT MLCT/LLCT LLCT LLCT MLCT/LLCT

S1 S4 S11

497/2.49 449/2.76 370/3.35

0.0105 0.0716 0.0733

S15 S16

345/3.60 337/3.68

0.1004 0.1007

H ? L(80%) H  1 ? L + 1 (75%) H  3 ? L + 1 (36%), H  1 ? L + 2 (22%) H  4 ? L (87%) H  4 ? L + 1 (84%)

MLCT/LLCT MLCT/LLCT MLCT/LLCT MLCT/LLCT LLCT LLCT

S1

516/2.40

0.0033

S2

510/2.43

0.0143

S4

486/2.55

0.0296

S14 S29

375/3.37 316/3.92

0.3248 0.1727

H ? L (67%) H  1 ? L (30%) H ? L + 1 (49%) H  1 ? L + 1 (47%) H  1 ? L + 1 (49%), H ? L + 1 (48%) H  4 ? L (36%) H  2 ? L + 2 (63%)

MLCT/LLCT MLCT/LLCT MLCT/LLCT MLCT/LLCT MLCT/LLCT MLCT/LLCT LLCT LLCT

361 342

187

M.-S. Ma et al. / Organic Electronics 22 (2015) 180–190

The absorption spectra ranging from 300 to 450 nm mainly origin from the characteristic absorption of ligand-structure-skeleton (overlapping of cyclometalating ligands and ancillary ligand, in Fig. S1), which can be assigned as the interligand charge transfer (LLCT) or the intraligand charge transfer (ILCT). It is deserved concern that ligand 2 (L2) and ligand 3 (L3) are identical in structure before coordination with iridium atom, so the L2 plot overlaps with the L3 plot in Fig. S1. After forming complexes 2 and 3, their photochemical properties change a lot. One of the factors is the symmetry of cyclometalating ligand (inflected by substituent – CH3), the other is the two nitrogen atoms (N1 and N3) in cyclometalating ligands connected with the metal centers residing at the trans location, and Ca1 and Cb1 at the cis location. 3.4. The phosphorescence quantum efficiency in CH2Cl2 The phosphorescence quantum efficiency /p depends on three processes: (1) the intersystem crossing (ISC) between the singlet and triplet excited states; (2) radiative transition from the triplet excited state to the ground state; (3) non-radiative transition from the excited state to the ground state [6]. In general, the /p is formulated as [25a,26]:

kr /p ¼ kr þ knr

ð1Þ

where kr and knr are the radiative and non-radiative rate constants, respectively. To obtain an efficient phosphorescent material, a large kr and a small knr are expected. The knr is expressed as:

knr ¼

a bDET m S

e

ð2Þ 0

a, b are constants, and b is related to the structural distortion between the Tm and S0 states, a small energy gap and a large structural distortion (respond to b) between the Tm and S0 states will cause a fast non-radiative decay. The kr is usually expressed as

k r ðT m Þ ¼

HSOC ¼

g2 E3Tm X hT m jHSOC jSn i 1:5

n

ESn  ETm

!2 

fs ESn

X 1 X Ze2 1 l s ¼ fc li si 2 c2 r 3 i i 4 p e 2m 0 e i i

ð3Þ

ð4Þ

where g represents the refractive index of the medium (denoted as CH2Cl2 solution in this system and g = 1.424), ESn and ETm are the emission energy of the nth singlet and mth triplet excited states, respectively, hT m jHSOC jSn i is the SOC matrix element between the Tm and Sn states, HSOC is the SOC Hamiltonian, ^l and ^s are angular momentum operators for orbital and spin, respectively, fs is the oscillator strength, and fc presents the one-electron spin–orbital coupling constant. In addition, the excited-state wavefunction (1,3W) obtained by TD-DFT can be expressed as a linear combination of one-electron excitation configurations (1,3wi) from an occupied molecular orbital to an unoccupied molecular orbital (eg. from HOMO to LUMO): 1;3

W¼

X ai 1;3 wi

ð5Þ

i

X cki vk



3

 

ð6Þ

k

Here ai is the coefficient of the configuration 1,3wi to the excitedstate wavefunction 1,3W, and cki is the coefficient of AO vk to MO wi.

1 

^ SOC  Wn Wm  H

*

 ¼

* ¼

X

3

i

X

   ^ X1 1 ai 3 wi H aj wj SOC 

+

j

3

! !+   X X    ^ X1 cki 3 ðdki p Þ H aj clj 1 dlj p SOC 

ai

i

j

k

l

ð7Þ

According to Eq. (7), the spin–orbital coupling SOC matrix elements are determined by the coefficients (ai, aj) of the MLCT oneelectron excitation configurations and the coefficients (cki, clj) of the occupied d orbitals contributing to the transitions. Because the SOC matrix elements can be negative or positive, which depend on the signs of the coefficients of the configuration 1,3wi to the excited-state wavefunction 1,3W (ai) obtained by TD-DFT, a large SOC matrix elements do not mean a fast radiative decay rate since the sum of all the SOC matrix elements may lead to a small radiative decay rate (kr). Moreover, the spin–orbital integrals  E D   ^ 1 3 can be approximately expressed as follows w H SOC  w i

j

[18,19,24e,25,26]:

D

3

  E   D E  ^ 1   3 wi H dki p ^l^s1 dlj p SOC  wj ¼ fc The matrix elements

ð8Þ

  E   3 dki p ^l^s3 dlj p , including the d orbital

D

coefficients are listed in Table S1 (Supporting information) In this paper, theoretical value of fc is 4430 cm1 for the Ir(III) 5d electron. 3.4.1. d-Orbital splittings Yersin and our previous research [6,24d,e] have stated that the d-orbital splittings are best criteria for evaluating the phosphorescence efficiency of the transition metal complexes. There are two effects can be considerated: On the one hand, a large Ddd⁄ can efficiently prevent the non-radiative quenching [6,26b]. On the other hand, a smaller Dddocc is conductive to a larger HSOC matrix element and a faster radiative decay rate constant. Table 4 presented the calculated values of Dddocc and Ddd⁄ based on both the S0 and T1 optimized geometries, together with the energy gaps of the two states. In addition the frontier orbital energies for all the complexes were listed in Tables S2–S7 in the Supporting information. Complexes 1–6 all have small Dddocc and large Ddd⁄ at the optimized S0 geometries. However, at T1 geometries, the Ddd⁄ of all complexes are reduced and the Dddocc of 2, 3, 5 and 6 increased while that of 1 and 4 decrease, the most change occurs in 5, which is from 0.17 eV (Dddocc)/6.46 eV (Ddd⁄) to 0.34 eV (Dddocc)/5.26 (Ddd⁄), respectively. This indicates that large geometric relaxations take place at T1 state, especially for 5. In respect to prediction of the phosphorescence efficiency, the Dddocc and Ddd⁄ gained on the basis of d-orbital splitting at the excited-state T1 geometry are more advisable than that at the S0 geometry [20a]. According to this view, although complex 1 has the smaller Dddocc (0.08 eV), its Ddd⁄ is 3.75 eV, which is almost half of that of the others. It can be deduced that 1 has larger radiative decay, and non-radiative

Table 4 The values of Dddocc [eV] and Ddd⁄ [eV], ET 1 !S0 energy gap [kcal mol1]. Complex

The molecular orbitals (wi) can be expanded using natural atomic orbitals (vk):

wi ¼

The SOC matrix elements hT m jHSOC jSn i can be written as follows [16]

1 2 3 4 5 6

S0

T1

ET 1 !S0

Dddocc

Ddd⁄

Dddocc

Ddd⁄

0.16 0.18 0.18 0.21 0.17 0.07

5.57 5.84 5.47 5.85 6.46 6.43

0.08 0.25 0.24 0.17 0.34 0.17

3.75 5.43 5.42 5.46 5.26 5.60

46.1 50.5 54.0 49.1 47.7 51.4

188

M.-S. Ma et al. / Organic Electronics 22 (2015) 180–190

decay rate at the same time. Remarkably, whether on the basis of T1 or S0 state, 6 has larger Ddd⁄ (5.60 or 6.43 eV), and the smaller Dddocc (0.17 or 0.07 eV). Thus, we can infer that 6 will be able to efficiently prevent non-radiative quenching from metal centered dd excited state and has fast radiative decay rate. From analysis of the calculated Dddocc and Ddd⁄ data, a little change in the structure of cyclometalating ligands will cause significantly effect on d-orbital splittings and phosphorescence efficiency. Take 5 and 6 (N2 ortho to N1) for instance, introducing ligand 5 (L5) results in a small kr, while introducing ligand 6 (L6) brings out a large kr, these can be attributed that L6 possesses a maximum absorption wavelength range (154–412 nm) (see Fig. S1). In the absorption spectra of L5 and L6, the higher energy absorption peaks, ranging from 220 to 300 nm, are assigned to spin allowed 1LC character, while the absorption spectrum of L6 extends up to 412 nm with featureless band shape is assigned to 3 LC character [25a,27], the position of this band overlaps with the 1MLCT absorption bands of complex 6 (Fig. 5). In contrast, there is no obviously overlap between the absorption spectrum of L5 and complex 5. It is generally accepted that the phosphorescence in iridium complexes is attributed to a mixture of MLCT and LC characters, and the increase in MLCT character will increase the radiative rate because of spin–orbit coupling. Consequently, after coordinating with metal, complex 6 will have larger kr than complex 5. In addition, the energy gap DE(T1  S0) is another factor effecting the knr and kr, the knr increases as the DE(T1  S0) decreases. The DE(T1  S0) of the complexes follow the order of 1 (46.1 kcal/mol) < 5 (47.7 kcal/mol) < 4 (49.1 kcal/mol) < 2 (50.5 kc al/mol) < 3 (51.4 kcal/mol) < 6 (54.0 kcal/mol) (listed in Table 4). Therefore, according to the analysis of d-orbital splittings together with the energy gaps DE(T1  S0) of the complexes, it can be concluded that introducing L1 might lead to enhance both kr and knr, introducing L5 might result in a small kr and a large knr, and introducing L6 might bring out a large kr and small knr. Thus, 1 will not have large phosphorescent quantum efficiency because of its large knr, and 6 will be expected to have the higher phosphorescence efficiency. 3.4.2. The non-radiative rate constant knr From Eq. (2), the knr is inversely proportional to the energy gaps DE(T1  S0) and the structure distortions between the S0 and T1 states. Having considered all factors above, a large knr could be resulted in following situations: (1) The structure distortions between the S0 and T1 states of 4 and 6 are small, while that of 2 and 3 are large, thus, 2 and 3 will show faster knr. (2) In 2 and 3, the presence of intraligand charge transfer will induce a large knr. (3) The knr of 1 will be considerable large because of its smallest Ddd⁄ on the basis of T1 state. (4) The d-orbital splitting of 5 is largest on going from S0 state to T1 state, which will lead to a Jahn–Teller distortion and even to fast knr. (5) 6 possesses both largest Ddd⁄ and energy gap between the S0 state and T1 states, which will be able to efficiently prevent non-radiative transition from metal centered dd excited state. 3.4.3. The radiative rate constant kr From Eq. (3), the radiative rate constant kr is proportional to the square of the spin–orbital coupling SOC, the phosphorescence emission energy (ETm) and oscillator (fs) but inversely proportional to the square of the energy difference (DEST) between the mth triplet (ETm) and the nth singlet (ESn) states. So as to comprehend the

impact of the four factors on the kr, we calculated the energy levels of the Sn and Tm states by TD-DFT based on the T1 states and the SOC matrix elements hT m jHSOC jSn i by Eqs. (3), (7) and (8) based on the T1 optimized geometries, and the results were shown in Table 5 and Tables S8 and S9 (Supporting information). According to the direct SOC assumption, two rule-of-thumbs can be drawn: (1) A small energy difference between the singlet and triplet excited states will fast the corresponding intersystem crossing (ISC) rate and reverse intersystem crossing (RISC) rate [23]. (2) The 1MLCT and 3LC one-electron excitation configurations contributing to the Sn and Tm excited states, must involve the same unoccupied p⁄ orbital but different occupied d orbital [5,14]. Take 1 for example (see Table S8), the SOC between the S3 and T1 excited states is invalid, the reason is that there is no same unoccupied p⁄ orbital between them, the T1 excited state is derived from four transitions: the H  2 ? L (ai = 0.39), H  1 ? L (ai = 0.52), H ? L (ai = 0.19), and H  3 ? L (ai = 0.12), but the S3 is derived mainly from the H  1 ? L + 1 (ai = 0.22) and H ? L + 1 (ai = 0.66). While the SOC between the S1/S2 and T1 states are much large, which are 167.15 cm1 and 109.10 cm1, respectively. The energy levels (ESn and ETm) and energy splittings (DESn–Tm) can be regulated by introducing different cyclometalating ligands. In 2, 3 and 5, there are only one triplet state (T1) below the S1 state in energy level, and the energy splittings DES1–T1 are 0.48 eV, 0.42 eV and 0.60 eV, respectively. In 1 and 4, there are two triplet states (T1 and T2) below the S1 state, in 6, there are three triplet states (T1–T3) below the S1 state, and the energy levels of the S1 state (2.10 eV) and T3 state (2.09 eV) are almost identical, such small singlet–triplet splitting would offer potential ISC S1 ? T3 and RISC T3 ? S1 processes. Accordingly, in Table 5, the calculated results of average kr (Tm ? S0): 1 kr (T1 ? S0) = 1.37  104 s1), kr (T2 ? S0) = 2.71  105 s1); 2 (kr (T1 ? S0) = 1.66  104 s1); 3 (kr (T1 ? S0) = 0.36  104 s1); 4 (kr (T1 ? S0) = 3.81  104 s1), (kr (T2 ? S0) = 5.55  104 s1); 5 (kr (T1 ? S0) = 0.67  104 s1); 6 (kr (T1 ? S0) = 2.87  104 s1), (kr (T2 ? S0) = 2.98  104 s1), (kr (T3 ? S0) = 1.25  105 s1), respectively. Recently, Adachi’s group reported a series of organic compounds with a small DES1–T1 6 0.1 eV and a reasonable radiative decay rate to overcome competitive non-radiative decay pathways, exhibited efficient thermally activated delay fluorescence (TADF) [23a–c], in addition, Ma’s groups reported another series of ‘‘hot exciton’’ compounds through reduce the DESn–Tm to avoid emission from long-time T1 state [23d–f,28]. Chi et al. have found that for Os(II) and Ag(I) complexes, ultrafast intersystem crossing (1011– 1012 s1) can happen between higher electronic excitations, while the ISC rate between S1 and T1 state is anomalously slow [29]. Accordingly, to our titled Ir(III) complexes, in 1 (DES1–T2 = 0.1 eV) 4 (DES1–T2 = 0.13 eV) and 6 (DES1–T3 = 0.01 eV), the ISC/RISC rate between S1 and Tm state may be comparable internal-conversion (IC) rate from the Tm to the T1 state, as commonly observed in heavy transition metal complexes. Therefore, it is feasible that in complexes 1, 4 and 6, a fraction of singlet excitons may be converted into high-lying triplet excitons through the S1 ? Tm channel before them relax to the S0 state, finally lead to high phosphorescence quantum efficiency. For complex 1, there is another interesting phenomenon when compare Stokes shift between experimental and theoretical approaches. We calculated the Stokes shift between 3MLCT absorption and phosphorescent emission, the results calculated by TDDFT exhibit the values are 34 nm (derived from T2 excited state) and 199 nm (derived from T1 excited state), respectively and reference 13 exhibits a small Stokes shift at the value of 27 nm, therefore, the experimental emission peak of 605 nm largely show the high-lying triplet emission feature. It is unexpected, there is no longer-lived emission was observed in the experiment, a reasonable explanation is the radiative rate has increased by as large as

189

M.-S. Ma et al. / Organic Electronics 22 (2015) 180–190 Table 5 Calculated triplets emission wavelengths, related transition properties, radiative rate constants in CH2Cl2 solution by DFT/TDDFT calculations for 1–6.

1

kcalcd emm [nm]

Main configuration

794(T1)

H1?L H2?L H?L H2?L

629(T2)

Assignment

kcalcd [104 s1] r

kexptl emm [nm]

Uexptl p

sexptl [ls] p

53% 30% 74% 16%

MLCT/LLCT MLCT/LLCT/ILCT MLCT/LLCT MLCT/LLCT/ILCT

1.37(T1)

605

0.42

2.6

27.06(T2)

2

706(T1)

H?L H2?L

58% 30%

MLCT/LLCT/ILCT MLCT/LLCT/ILCT

1.66(T1)

3

655(T1)

H?L H2?L

44% 33%

MLCT/LLCT/ILCT MLCT/LLCT/ILCT

0.36(T1)

4

702(T1)

H?L H2?L H?L H2?L

53% 32% 70% 19%

MLCT/LLCT MLCT/LLCT/ILCT MLCT/LLCT MLCT/LLCT/ILCT

3.81(T1)

592(T2)

5.55(T2)

5

767(T1)

H?L H2?L

61% 21%

MLCT/LLCT MLCT/LLCT/ILCT

0.67(T1)

6

648(T1)

H?L H1?L H1?L+1 H?L+1 H?L+1 H?L

45% 21% 38% 22% 61% 8%

MLCT/LLCT MLCT/LLCT/ILCT MLCT/LLCT/ILCT MLCT/LLCT MLCT/LLCT MLCT/LLCT

2.87(T1)

640(T2) 590(T3)

20-fold from T1 to T2 excited state, leading to undetected longer emission in the experiment. According to the above mentioned, it can be concluded that the emission of 2, 3 and 5 are originated from the T1 excited state, while the emission of 1, 4 and 6 would show more high-lying triplet emission feature arising from the ultrafast ISC rate between singlet and triplet states. Based on Eq. (1), high phosphorescence quantum efficiency /p can be attained either by reducing knr or increasing kr. All factors above prove that 1 possesses both a fast kr and a large knr, leading to a value of 0.42 lower than it was expected, 2, 3 and 5 possess small kr but large knr, thus resulting in their low /p; with regard to 6, its non-radiative constant knr will be much smaller than 1, together with its large kr, so 6 will be an excellent luminescent materials with high phosphorescence quantum efficiency. 4. Conclusions A family of iridium complexes with isomeric naphthyridine heterocyclic ligands were investigated to explore the effect of the position of two nitrogen atoms in the naphthyridine moieties on the phosphorescent properties. The electronic structures, frontier molecular orbitals (FMOs), absorption spectra and phosphorescence quantum efficiency were investigated through theoretical calculations. Our results reveal that subtle structural adjustment would greatly affect the composition and the energy levels of the FMOs. Their HOMO energy levels reduce graduately wen nonchelating atom N2 is far from chelating N1, namely, the HOMO levels of 5 and 6 (non-chelating N2 ortho to chelating N1) are the highest, 2, 3 and 4 (non-chelating N2 meta to chelating N1) are the second and 1 (non-chelating N2 para to chelating N1) is comparatively lowest. The electronic cloud distribution of LUMOs localized on the naphthyridine moiety rather than the fluorophenyl moiety when the electron-donating substituent –CH3 is separated to the non-chelating nitrogen atom, suggesting that the electron-donating group –CH3 and electron-withdrawing group fluorophenyl will cooperatively affect energy compositions of LUMOs. We approximately measured the radiative transition rate, 1 kr (T1 ? S0) = 1.37  104 s1), kr (T2 ? S0) = 2.71  105 s1); 2 (kr (T1 ? S0) = 1.66  104 s1); 3 (kr (T1 ? S0) = 0.36  104 s1); 4 (kr (T1 ? S0) = 3.81  104 s1), (kr (T2 ? S0) = 5.55  104 s1); 5 (kr (T1 ? S0) = 0.67  104 s1); 6 (kr (T1 ? S0) = 2.87  104 s1), (kr (T2 ? S0) = 2.98  104 s1), (kr (T3 ? S0) = 1.25  105 s1),

2.98(T2) 12.50(T3)

respectively. Alternatively, we predict the emission of 2, 3 and 5 are originated from the T1 excited state, while the emission of 1, 4 and 6 would show more high-lying triplet emission feature arising from the ultrafast ISC rate between singlet and triplet states. Finally, we predicted their phosphorescence quantum efficiency (/p) order, the /p of 1 will be moderate as result of both large kr and knr, 2, 3 are emitting weakly with larger knr, while for 4, another quinazoline heterocycle counterparts, its knr is small due to the absence of intraligand and distinctly structure distortions. It is should be point out that 6, in which the C^N = N ligand strengthens the interaction between metal and ligands, and the methyl substituent on the opposite side to non-chelating nitrogen atom minimize the feasibility of non-radiatve transition, so complex 6 has large kr and small knr, while for 5, the –CH3 substituent is next to the non-chelating nitrogen atom, increase the knr. In summary, in the naphthyridine heterocyclic ring, the non-chelating N next to the chelating N will increase the radiative transition rate, on the other hand, separating the two electron-donating groups (the non-chelating N and the methyl substituent) is benefit to decrease the non-radiatve transition rate. Acknowledgments This work is supported by the Natural Science Foundation of China (Nos. 21473071, 21173099 and 20973078), the Major State Basis Research Development Program (Grant 2013CB 834801), special funding to basic scientific research projects for Central Colleges. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.orgel.2015.03. 037. References [1] (a) S.R. Förrest, M.A. Baldo, D.F. O’Brien, Y. You, A. Shoustikov, S. Sibley, M.E. Thompson, Nature 395 (1998) 151–154; (b) P.T. Chou, Y. Chi, Eur. J. Inorg. Chem. 17 (2006) 3319–3332; (c) Z.M. Hudson, C. Sun, M.G. Helander, H. Amarne, Z.H. Lu, S. Wang, Adv. Funct. Mater. 20 (2010) 3426–3439; (d) C.H. Fan, P. Sun, T.H. Su, C.H. Cheng, Adv. Funct. Mater. 23 (2011) 2981– 2985;

190

[2] [3]

[4]

[5]

[6] [7]

[8]

[9]

[10] [11]

[12] [13]

[14] [15] [16]

[17]

M.-S. Ma et al. / Organic Electronics 22 (2015) 180–190 (e) V.K. Au, K.M. Wong, D.P. Tsang, M.Y. Chan, N. Zhu, V.W. Yam, J. Am. Chem. Soc. 132 (2010) 14273–14278. H. Yersin, Top. Curr. Chem. 241 (2004) 1–26. (a) C. Adachi, M.A. Baldo, S.R. Förrest, M.E. Thompson, Appl. Phys. Lett. 77 (2000) 904–906; (b) Y. Kawamura, K. Goushi, J. Brooks, J.J. Brown, H. Sasabe, C. Adachi, Appl. Phys. Lett. 86 (2005) 071104–071107; (c) B. Tong, Q.B. Mei, S.J. Wang, Y. Fang, Y.Z. Meng, B.J. Wang, Mater. Chem. 18 (2008) 1636–1639. (a) S. Lamansky, P. Djurovich, D. Murphy, F.A. Razzaq, H.E. Lee, C. Adachi, P.E. Burrows, S.R. Förrest, M.E. Thompson, J. Am. Chem. Soc. 123 (2001) 4304– 4312; (b) A.B. Tamayo, B.D. Alleyne, P.I. Djurovich, S. Lamansky, I. Tsyba, N.N. Ho, R. Bau, M.E. Thompson, J. Am. Chem. Soc. 125 (2003) 7377–7387; (c) S. Lamansky, P. Djurovich, D. Murphy, F. AbdelRazzaq, R. Kwong, L. Tsyba, M. Bortz, R. Bau, M.E. Thompson, Inorg. Chem. 40 (2001) 1704–1711. (a) E. Orselli, G.S. Kottas, A.E. Konradsson, P. Coppo, R. Frohlich, L. De Cola, A. van Kijken, M. Buchel, H. Borner, Inorg. Chem. 46 (1093) (2007) 11082–11093; (b) S.J. Lee, K.M. Park, K. Yang, Y. Kang, Inorg. Chem. 48 (2009) 1030–1037; (c) S.L. Lai, S.L. Tao, M.Y. Chan, M.F. Lo, T.W. Ng, S.T. Lee, W.M. Zhao, C.S. Lee, J. Mater. Chem. 21 (2011) 4983–4988; (d) Y. Liu, K. Ye, Y. Fan, W.F. Song, Y. Wang, Z. Hou, Chem. Commun. 25 (2009) 3699–3701. G.S. Tong, C.M. Che, Chem. Eur. J. 15 (2009) 7225–7237. (a) A.F. Rausch, H.H.H. Homeier, P.I. Djurovich, M.E. Thompson, H. Yersin, Proc. SPIE-Int. Soc. Opt. Eng. 6655 (2007) 66550F; (b) H. Yersin, Highly Efficient OLEDs with Phosphorescent Materials, WileyVCH, Weinheim, Germany, 2008. (a) C. Adachi, M.A. Baldo, M.E. Thompsom, S.R. Förrest, J. Appl. Phys. 90 (2001) 5048–5051; (b) Y. Chi, P.T. Chou, Chem. Soc. Rev. 39 (2010) 638–655. (a) F.M. Hwang, H.Y. Chen, C.S. Liu, Y. Chi, C.F. Shu, F.L. Wu, P.T. Chou, S.M. Peng, G.H. Lee, Inorg. Chem. 44 (2005) 1344–1353; (b) B.S. Du, C.H. Lin, Y. Chi, J.Y. Hung, M.W. Chun, T.Y. Lin, G.H. Lee, K.T. Wong, P.T. Chou, W.Y. Hung, H.C. Chiu, Inorg. Chem. 49 (2010) 8713–8723; (c) R. Tao, J. Qiao, G.L. Zhang, L. Duan, L.D. Wang, Y. Qiu, J. Phys. Chem. C 116 (2012) 11658–11664; (d) B. Wiegmann, P.C. Jones, G. Wagenblast, C. Lennartz, I. Münster, S. Metz, W. Kowalsky, H.H. Johannes, Organometallics 31 (2012) 5223–5226; (e) J.P. Duan, P.P. Sun, C.H. Cheng, Adv. Mater. 15 (2003) 224–228. Y.H. Song, S.J. Yeh, C.T. Chen, Y. Chi, C.S. Liu, J.K. Yu, Y.H. Hu, P.T. Chou, S.M. Peng, G.H. Lee, Adv. Funct. Mater. 14 (2004) 1221–1226. (a) B.X. Mi, P.F. Wang, Z.Q. Gao, C.S. Lee, S.T. Lee, H.L. Hong, X.M. Chen, M.S. Wong, P.F. Xia, K.W. Cheah, C.H. Chen, W. Huang, Adv. Mater. 21 (2009) 339– 343; (b) Z.Q. Gao, B.X. Mi, H.L. Tam, K.W. Cheah, H.S. Chen, M.S. Wong, S.T. Lee, C.S. Lee, Adv. Mater. 20 (2008) 774–778. M.S. Lowry, S. Bernhard, Chem. Eur. J. 12 (2006) 7970–7977. D. Schneidenbach, S. Ammermann, M. Debeaux, A. Freund, M. Zöllner, C. Daniliuc, P.G. Jones, W. Kowalsky, H.H. Johannes, Inorg. Chem. 49 (2010) 397– 406. E. Runge, E.K.U. Gross, Phys. Rev. Lett. 52 (1984) 997–1000. (a) C. Lee, W. Yang, T.R.G. Parr, Phys. Rev. B. 37 (1988) 785–789; (b) A.D. Becke, J. Chem. Phys. 98 (1993) 5648–5652. (a) E.U. Gross, J.F. Dobson, M. Petersilka, Top. Curr. Chem. 181 (1996) 81; (b) I. Avilov, P. Minoofar, J. Cornil, L. De Cola, J. Am. Chem. Soc. 129 (2007) 8247–8258; (c) J. Autschbach, T. Ziegler, S.J.A. Gisbergen, E.J. Baerends, J. Chem. Phys. 116 (2002) 6930–6940. (a) C.L. Li, Y.J. Su, Y.T. Tao, P.T. Chou, C.H. Chien, C.C. Cheng, R.S. Liu, Adv. Funct. Mater. 15 (2005) 387–395; (b) R. Zhu, J. Lin, G.A. Wen, S.J. Liu, J.H. Wan, J.C. Feng, Q.L. Fan, G.Y. Zhong, W. Wei, W. Huang, Chem. Lett. 34 (2005) 1668–1669; (c) K.C. Robson, B. Sporinova, B.D. Koivisto, E. Schott, D.G. Brown, C.P. Berlinguette, Inorg. Chem. 50 (2011) 6019–6028;

[18] [19] [20]

[21] [22]

[23]

[24]

[25]

[26]

[27]

[28] [29]

(d) Y. Zhang, J.A. Garg, C. Michelin, T. Fox, O. Blacque, K. Venkatesan, Inorg. Chem. 50 (2011) 1220–1228; (e) D.A. Lutterman, A.A. Rachford, J.J. Rack, C.J. Turro, Phys. Chem. Lett. 1 (2010) 3371–3375; (f) K. Li, X.G. Guan, C.W. Ma, W. Lu, Y. Chen, C.M. Che, Chem. Commun. 47 (2011) 9075–9077; (g) L. Wang, Y. Wu, G.G. Shan, Y. Geng, J.Z. Zhang, D.M. Wang, G.C. Yang, Z.M. Su, J. Mater. Chem. C 2 (2014) 2859–2868. M.E. Casida, J. Mol. Struct. Theochem. 914 (2009) 3–18. (a) V. Barone, M. Cossi, J. Chem. Phys. 107 (1997) 3210–3221; (b) M. Cossi, G. Scalmani, N. Regar, V. Barone, J. Chem. Phys. 117 (2002) 43–54. (a) P. Fuentealba, H. Preuss, H. Stoll, L.V. Szentpály, Chem. Phys. Lett. 89 (1982) 418–422; (b) F.D. Angelis, F. Santoro, M.K. Nazeruddin, V.J. Barone, Phys. Chem. B 112 (2008) 13181–13183. X. Li, B. Minaev, H. Ågren, H. Tian, J. Phys. Chem. 115 (2011) 20724–20731. M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G.A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H.P. Hratchian, A.F. Izmaylov, J. Bloino, G. Zheng, J.L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M.T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J.A. Montgomery Jr., J.E. Peralta, F. Ogliaro, M. Bearpark, J.J. Heyd, E. Brothers, K.N. Kudin, V.N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J.C. Burant, S.S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J.M. Millam, M. Klene, J.E. Knox, J.B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W. Ochterski, R.L. Martin, K. Morokuma, V.G. Zakrzewski, G.A. Voth, P. Salvador, J.J. Dannenberg, S. Dapprich, A.D. Daniels, O. Farkas, J.B. Foresman, J.V. Ortiz, J. Cioslowski, D.J. Fox, Gaussian 09, Revision A.02, Gaussian Inc., Wallingford CT, 2009. (a) H. Uoyamal, K. Goushi, K. Shizu, H. Nomura, C. Adachi, Nature 492 (2012) 234–238; (b) Q.S. Zhang, B. Li, S.P. Huang, H. Nomura, H. Tanaka, C. Adachi, Nat. Photonics 8 (2014) 326–332; (c) H. Nakanotani, T. Higuchi, T. Furukawa, K. Morimoto, M. Numata, H. Tanaka, Y. Sagara, T. Yasuda, C. Adachi, Nat. Commun. 5 (2014) 4016–4023; (d) Y.Y. Pan, W.J. Li, S.T. Zhang, L. Yao, C. Gu, H. Xu, B. Yang, Y.G. Ma, Adv. Opt. Mater. 2 (2014) 1–6; (e) L. Yao, S.T. Zhang, R. Wang, W.J. Li, F.Z. Shen, B. Yang, Y.G. Ma, Angew. Chem. Int. Ed. 53 (2014) 2119–2123; (f) W.J. Li, Y.Y. Pan, R. Xiao, Q.M. Peng, S.T. Zhang, D.G. Ma, F. Li, F.Z. Shen, Y.H. Wang, B. Yang, Y.G. Ma, Adv. Funct. Mater. 24 (2014) 1609–1614. (a) P.J. Hay, J. Phys. Chem. A 106 (2002) 1634–1641; (b) A. Curioni, W. Andreoni, R. Treusch, F.J. Himpsel, E. Haskal, P. Seidler, C. Heske, S. Kakar, T. van Buuren, L. Terminello, J. Appl. Phys. Lett. 72 (1998) 1575–1577; (c) S.Y. Hong, D.Y. Kim, C.Y. Kim, R. Hoffmann, Macromolecules 34 (2001) 6474–6481; (d) L.Y. Zou, M.S. Ma, Z.L. Zhang, H. Li, Y.X. Cheng, A.M. Ren, Org. Electron. 13 (2012) 2627–2638; (e) Y. Li, L.Y. Zou, A.M. Ren, M.S. Ma, J.X. Fan, Chem. Eur. J. 20 (2014) 4671– 4680. (a) Y. You, S.Y. Park, Dalton Trans. 8 (2009) 1267–1282; (b) Y.S. Yeh, Y.M. Cheng, P.T. Chou, G.H. Lee, C.H. Yang, Y. Chi, C.F. Shu, C.H. Wang, ChemPhysChem 7 (2006) 2294–2297. (a) K.J. Nozaki, Chin. Chem. Soc. 53 (2006) 101–112; (b) J.A.G. Williams, Top. Curr. Chem. 281 (2007) 205–268; (c) Z.A. Siddique, Y. Yamanoto, T. Ohno, K. Nozaki, Inorg. Chem. 42 (2003) 6366–6378. (a) K. Dedeian, J. Shi, E. Forsythe, D.C. Morton, P.Y. Zavalij, Inorg. Chem. 46 (2007) 1603–1611; (b) C.L. Lee, R.R. Das, J. Kim, J. Chem. Mater. 16 (2004) 4642–4646. N.J. Turro, Modern Molecular Photochemistry, University Science Books, Sausalito, 1991. p. 148. C.C. Hsu, C.C. Lin, P.T. Chou, C.H. Lai, C.W. Hsu, C.H. Lin, Y. Chi, J. Am. Chem. Soc. 134 (2012) 7715–7724.