Bulky, dendronized iridium complexes and their photoluminescence

The emission efficiencies of bulky phosphorescent emitters are reduced by elongated metal–ligand bonds, which provides insights for designing better emitters.


Introduction
Even though OLED-based products, e.g. flat-panel displays and solid-state lighting are coming into people's lives slowly, 1-3 the prices are still quite high due to the complicated fabrication of multilayer devices with costly ultrahigh vacuum (UHV) deposition and inefficient use of materials. 4 Solution-processed methods, e.g. spin-coating and inkjet printing, on the other hand, could dramatically reduce the cost for the fabrication of devices, and therefore attract great attention among the communities of flexible electronics and are believed to play vital roles in the manufacture of inexpensive and next-generation OLEDs. 5 Therefore, developing high-performance and solution-processable emitters has been and is still a hot topic. 4,6,7 Dendrimers and polymers are more popular candidates for designing solution-processed materials than small molecules which are usually prone to form crystalline rather than the preferred amorphous films for OLEDs. 4 Dendrimers are monodisperse and highly-branched macromolecules with a tailor-made core, a shell and a surface structure. 8 They feature several desired attributes for the design of solution-processable emitters in comparison to polymers, i.e. absolute reproducibility, high photoluminescence quantum yield (PLQY) by encapsulation of the emitter in the core, and layer-by-layer substitution to create multi-functional materials with potential for greatly reducing the complexity of the device. 6,[9][10][11][12] It is essential to select an appropriate core, i.e. an emitter for the development of efficient dendrimer-based light-emitting materials. Phosphorescent emitters (PEs) such as iridium and platinum complexes are much more efficient than conventional fluorescent ones (e.g. pyrene derivatives) because PEs can generate both singlet and triplet excitons in OLEDs (corresponding to 100% internal quantum yield (IQE) and 20% external quantum yield (EQE) in theory) but the other kind can only harvest singlet ones (not more than 25% IQE and 5% EQE). [13][14][15][16] The neat films of small-molecular PEs are liable to experience severe self-quenching. Therefore, a matrix is usually employed to accommodate the emitters (also called dopants) to ensure high PLQYs. The matrix composed of host materials, e.g. N,N 0dicarbazolyl-3,5-benzene (mCP), serves as a medium for charge transport and energy transfer to the dopants. 17 However, this method still encounters uneven dispersions of the dopants in the host with UHV depositions. Dendrimers on the other hand effectively avoid this problem. 6,[9][10][11][12] Within a rigid and shapepersistent dendritic architecture, a PE is well encapsulated in the core by the bulky dendrons, and the surface is functionalized with host moieties to exert charge transport and surface-to-core energy transfer. Besides, one can precisely manipulate the ratio and distances between the hosts and the PE by molecular design. 10,11 On account of these merits, dendrimer-based PEs have gained significant breakthroughs within the last two decades. 6,18-23 Both highlyefficient green and red dendrimer PEs with comparable OLED performances to small molecules have been reported. [21][22][23] As to dendrimer-based blue PEs, even though a few examples are available, they all suffer from either very low device efficiencies or poor colour purities. [24][25][26][27][28] Therefore, a high-performance, dendrimer-based and pure blue PE is still missing.
PEs based on shape-persistent polyphenylene dendrimers (PPDs) have been studied as well. 29,30 For example, Qin, et al. prepared several PPDs as efficient green PEs. 29 In addition, our group has demonstrated that core-surface-substituted, firstgeneration PPDs have a good architecture for the design of better dendrimer-based emitters by judicious selection of peripheral moieties to establish efficient surface-to-core energy transfer and intermolecular charge transport. 31,32 Extending the conjugation of a PE with the dominant 3 p-p* ( 3 LC) character of the emissive excited states has been found by us to give a strongly red-shifted emission. The comparison of the fac-(dfpypy) 3 Ir-based dendrimer (D3) with fac-(dfpypy) 3 Ir (9) (red shift: 50 nm) is shown (Fig. 1). 33 In addition, elongated conjugation within the ligands tends to strengthen the 3 LC characteristics of its emissive excited states. 34 We thought to make use of these findings to design new dendrimer-based blue PEs. Thus, utilizing a PE (with ultraviolet (UV) emission) as the core of the dendrimer could possibly push the emission to a pure blue region. Schildknecht et al. reported that fac-(dpbic) 3 Ir (Fig. 1) as a near UV emitter (l max : 400 nm) furnished a PLQY of 0.19. 35 We therefore envisaged fac-(dpbic) 3 Ir as the core to develop new dendrimer-based blue PEs. As depicted in Fig. 1, a TIPSE-substituted fac-(dpbic) 3 Ir (2) and two first-generation, fac-(dpbic) 3 Ir-based PPDs, i.e. D1 and D2 which contain peripheral carbazoles to facilitate charge transport and energy transfer are synthesized and characterized. Their photophysical properties and the OLED performances of compound 2 and D2 are discussed as well.
The 1 H NMR spectra (Fig. S1, ESI †) characterized Ir-complex 2 as a facial isomer with a total of eleven proton signals in the aromatic region (eleven aromatic protons in one ligand). Single crystals of compound 2 were obtained by slow addition of methanol to a dichloromethane solution. As depicted by the crystal structure in Fig. 2, the molecule adopts a quasi-octahedral geometry. Its three Ir-C carbene and Ir-C phenyl bonds are slightly longer than those of fac-(pmb) 3 Ir 34 (Table S2, ESI †) due to the bulky TIPSE moieties in Ir-complex 2. 41 The non-coordinated benzene rings are highly twisted from the benzimidazole-based carbene plane due to the strong steric hindrance between the benzene ring and a carbene moiety from a nearby ligand. In contrast to fac-(dfpypy) 3 Ir 42 and fac-(pmb) 3 Ir ( Fig. 1), 34 intermolecular pÁ Á Áp close interactions are not observed for compound 2 due to the protection by the bulky TIPSE segments. This suggests Ir-complex 2 as a promising candidate for application in non-doped solution-processed OLEDs. [24][25][26]31,32,43,44 The reduced intermolecular interactions of Ir-complex 2 are also in accordance with its longer intermolecular IrÁ Á ÁIr distance (B12.90 Å) than those in fac-(dfpypy) 3 Ir (B9.10 Å) 42 and fac-(pmb) 3 Ir (B9.38 Å). 34 The dendrimers D1 and D2 were characterized by 1 H and 13 C NMR, MALDI-TOF mass and high-resolution mass spectroscopy. The MALDI-TOF mass spectra of each dendrimer showed a single peak of the molecular ion. In addition, the highresolution MALDI-TOF mass spectra of the dendrimers show isotope patterns of the molecular ion in good agreement with the calculated results ( Fig. S2 and S3, ESI †).

Photophysical characterization
Regarding the absorption of compound 2, a major band with several shoulders in the UV region is due to the transitions of the TIPSE-functionalized, 1,3-diphenylbenzimidazole-based carbene ligand, which is consistent with the absorption of TIPSEsubstituted, 1,3-diphenylbenzimidazolium chloride (3) (Fig. 1,  3 and Table 1). The small band above 350 nm is attributed to a mixture of metal-to-ligand charge transfer (MLCT) and 3 LC, based on a comparison with the transitions of fac-(pmb) 3 Ir. 34 Ir-complex 2 exhibits strong and pure-blue emission (l max : 440 nm and 469 nm) in solution under argon protection because oxygen easily quenches the triplet states. 45,46 The peak emission of compound 2 undergoes a bathochromic shift of 40 nm compared with that of (dpbic) 3 Ir. 35 This red shift is due to the dominant 3 LC nature of the emissive excited states of compound 2 together with its elongated conjugation. 33 In addition, the emission of Ir-complex 2 at 77 K displays intense and nearly identical spectra to those measured at room temperature (Fig. 3b). This supports the major 3 LC nature of the transition because a dominant 3 MLCT-type of emission usually undergoes a hypsochromic shift in the solid matrix in contrast to that in the solution state. 47,48 The PLQY was measured to be 0.19 with an integration sphere. The photoluminescence lifetime extracted from the corresponding decay  curve of Ir-complex 2 (Fig. S7, ESI †) is around 0.49 ms which suggests the phosphorescent nature of Ir-complex 2. 49 The phosphorescence lifetime is shorter than that of most known triplet emitters, 49 and its decay pathways are currently under study. To the best of our knowledge, luminophore 2 is the first reported pure-blue Ir complex with alkynyl moieties in the ligand. Moreover, the thin film of compound 3 shows a strong and featureless deep blue emission, which is similar to those of other reported highly emissive benzimidazole-based organic salts; 50,51 for instance, Boydston et al. synthesized a series of benzobis(imidazolium) salts with high PLQYs. 50 As to the absorption of dendrimers D1 and D2 (Fig. 4), the bands between 250 and 350 nm are due to ligand-and polyphenylene-centered transitions; 52 for example, the peak at 297 nm of D2 is characteristic of a carbazole absorption. 53 The weak shoulders above 350 nm are attributed to a mixture of MLCT and 3 LC. The dendrimers in solutions are not emissive at room temperature even under argon protection but exhibit strong sky-blue emission at 77 K with nearly identical peak positions (l max : B462 nm and 488 nm) and a slight bathochromic shift of D2 compared with D1. The quenching of the dendrimer emission at ambient temperatures could be due to the small energy barrier between the nonradiative excited state (NR) and the emissive excited state (T 1 ), similar to the situation prevailing in other reported heavy-metal complexes, e.g. fac-Ir(ppz) 3 (Fig. 1). 34,54 Alternatively, the decay of the emissive excite state could be depleted by the vibrations or rotations of the polyphenylene dendrons. The latter assumption is supported by the detected weak emission of the dendrimers in thin films (Fig. 4). We claim that this is not the major origin of the phosphorescence quenching of the dendrimers at room temperature. We rather postulate that it must be ascribed to the rapid transition between T 1 and NR states at ambient temperature.

Exploration of the emission efficiencies through theoretical calculations
Many transition-metal complexes are strongly emissive at 77 K but their emission is severely quenched at room temperature, which in many cases is ascribed to the thermal population of the nonradiative triplet metal-centered charge transfer state ( 3 MC). [54][55][56][57][58][59] For example, Sajoto et al. demonstrated that the emission efficiencies of many Ir-complexes, such as fac-(ppz) 3 Ir and fac-(pmb) 3 Ir (Fig. 1), were primarily determined by the energy gap between the T 1 and NR states, i.e. 3 MC. 54 Recently, Zhou et al. concluded that the thermal population of 3 MC was the major nonradiative decay pathway for N-heterocyclic carbene-chelated Ir complexes. 60 Thereby, the energy level of 3 MC is an important factor for evaluating the emission efficiencies of many Ir complexes. 3 MC originates from transitions between the non-degenerate d orbitals of the metal atom in heavy metal complexes as predicted by crystal field theory (five d orbitals are split into three occupied and two unoccupied ones, called t 2g and e g respectively) (Fig. 5). The energy gap between t 2g and e g is determined by the arrangement and type of the ligands, 61 and it is known that strengthening the metal-ligand bond destabilizes the 3 MC state. 34 Sajoto et al. reported that fac-(pmb) 3 Ir (Fig. 1) showed a higher energy level of 3 MC than many carbon-and-nitrogenchelated Ir-complexes (Ir(C^N) 3 ); this is argued from calculations  and is consistent with the shorter Ir-C carbene bond length of fac-(pmb) 3 Ir than the Ir-N bond length of Ir(C^N) 3 , e.g. fac-(ppz) 3 Ir. 54 Therefore, comparing the bond lengths of Ir-C carbene among our targeted fac-(dpbic) 3 Ir-containing molecules is justified to assess their relative 3 MC levels and thereby to explain their different emission efficiencies at room temperature. Based on the crystal structures and structural optimizations of several Ir-complexes as shown in Fig. 6(a), it is suggested that the average bond length of Ir-C carbene increases with more bulky ligands and the longest one is from the bulkiest dendrimer D1. Therefore, the energy level of 3 MC should decrease in the order: fac-(dpbic) 3 Ir, Ir-complex 2 and D1. To explain the emission efficiency of these molecules, the energy barriers for the transition of T 1 -3 MC were calculated through a constrained potential energy surface (PES) scan along the longest Ir-C bond length (ESI † for the detailed method of calculation). 57 The energy difference between the highest point of the PES and the T 1 state is the barrier height ( Fig. 6(b)). The calculations reveal that, for fac-(dpbic) 3 Ir and Ir-complex 2, the energy barriers are about 0.80 eV, however, for dendrimer D1, the computed barrier is only 0.25 eV. Owing to the very complex structure of D2, we did not attempt a calculation of it; however, it can be inferred that D2 should possess an even lower energy of 3 MC than D1 as indicated in the trend (Fig. 6a) because of the bigger size of D2 than D1; due to their very similar peak emission at 77 K (E T : 2.68 eV for both) (Fig. 4 and Table 1), we could envisage that the energy barrier (T 1 -3 MC) for D2 is even smaller than that of D1 (Fig. 5). Therefore, we conclude that the lack of emission from D1 and D2 under ambient conditions is due to the easy access to the non-radiative 3 MC state from the emissive state T 1 (Fig. 5).

Electrochemistry
From the CV measurements, it is deduced that the HOMO energy (À5.23 eV) of Ir-complex 2 is increased and its LUMO energy (À2.11 eV) is decreased compared with the parent fac-(dpbic) 3 Ir due to the extended conjugation. 17 This qualifies it for OLED applications because many available charge-transporting materials could correspond to the HOMO and LUMO energies of compound 2 for efficient charge injection and transport in OLEDs, such as PEDOT:PSS (HOMO: À5.10 eV, for hole injection and transport) 31 and 4-(triphenylsilyl)phenyldiphenylphosphine oxide (TSPO1) (LUMO: À2.52 eV and E T : 3.36 eV for electron transport and exciton blocking). 13 As to dendrimer D2, the HOMO energy was calculated to be À5.53 eV from the oxidation onset of the CV curve of D2. This is consistent with that of the peripheral carbazoles, 53 probably owing to the large number of carbazole groups per Ir complex segment within one molecule (12 : 1).

OLED performance
The Ir-complex 2 and dendrimer D2 were tested in OLEDs. The device structure for Ir-complex 2 is ITO/PEDOT:PSS/Ir-complex 2/TPCz/TmPyPB/LiF/Al (TPCz: 3,6-bis(diphenylphosphoryl)-9-(4 0 -(diphenylphosphoryl)phenyl)carbazole; TmPyPB: 1,3,5-tri(mpyrid-3-yl-phenyl)benzene). 64 In this device, Ir-complex 2 was deposited by spin coating and the layers of TPCZ and TmPyPB were generated by vacuum evaporation (ESI †). Interestingly, Ir-complex 2 in this non-doped OLED exhibits a pure blue emission with peak emission at 442 and 469 nm, a maximum current   6 The relationship between the bulkiness of the molecules (quantified with the van der Waals volume 62 of the groups in the molecule) and the Ir-C carbene bond length (a) (the dark and red curves represent the calculated average and longest bond length; the calculations were performed using the Gaussian software, 63 DFT, the B3LYP hybrid functional and the LanL2DZ basis set; the pink and blue curves represent the experimental longest and average bond length, from the crystal structures); (b) computed PES scan plots for fac-(dpbic) 3 Ir (black line), Ir-complex 2 (red line) and dendrimer D1 (blue line). efficiency of 1.04 cd A À1 and a CIE xy (0.18, 0.18) (Table 1). Surprisingly, the emission of the Ir-complex 2 in the device is nearly the same as that measured in dilute solution and the thin film (Fig. 3). The lack of an aggregation-induced red-shifted emission is, again, due to the very bulky TIPSE moieties. As far as we know, this is the first reported non-doped small-molecule Ir-complex with pure blue emission in solution-processed OLEDs. This finding provides a new design concept for high-performance, pure-ordeep-blue emitters in solution-processed devices. For dendrimer D2, all charge transporting and emitting layers were spin coated with selected solvents, except that CuSCN was deposited by inkjet printing (ESI †). The devices have a structure of ITO/PEDOT:PSS/ CuSCN/D2/Ca/Al. As shown in Table 1 and Fig. S9 (ESI †), the device displays green emission (l max : B520 nm with a shoulder around 600 nm), consistent with the photoluminescence of the film. The overall poor device performances are due to the very easy access to the nonradiative T 1 -3 MC transition for D2 at room temperature (Fig. 5). The OLED performance of fac-(dpbic) 3 Ir was reported in 2005 (Table 2). 35 Due to the difficulty in finding an appropriate host material for fac-(dpbic) 3 Ir, utilizing essentially insulating poly(methyl methacrylate) (PMMA) as the matrix rendered a moderate efficiency and the emission was kept deep blue desirably.

Conclusions
TIPSE-functionalized Ir-complex 2 is a blue phosphorescent emitter under ambient conditions, whereas dendrimers are barely emissive at room temperature. The failure of the dendrimers to emit at room temperature is related to the much reduced energy level of 3 MC states. This, in turn, is due to the observed notable lengthening of the Ir-C carbene bonds induced by the bulky polyphenylene dendrons. This explanation is further supported by our PES calculations, which reveal that the barrier height for the transition from the emissive T 1 state to the non-emissive 3 MC state is only 0.25 eV for the dendrimer. Therefore, we conclude that, even though the introduction of bulky moieties into a PE can effectively inhibit excimer and triplet-triplet annihilation, 65 the bulky groups are detrimental to the PLQY of a PE if the steric hindrance is too high. Therefore, the impact of the bulkiness of the molecules on the emission of metal-organic complex-based phosphorescent emitters is significant for the design of highly emissive dendrimer-based PEs by proposing: (i) to adjust the steric hindrance by the dendrons and (ii) to utilize more rigid phosphorescent emitters as the core of a dendrimer. Examples would be square-planar platinum complexes which minimize the impact of the steric hindrance. 66,67 In addition, Ir-complex 2, as the first-reported pure-blue small-molecular PE in solution-processed and non-doped OLEDs, stimulates the design of new small-molecular phosphorescent materials.

Conflicts of interest
There are no conflicts of interest to declare.