Chapter 4: Molecular Design and Synthesis of Metal Complexes as Emitters for TADF‐Type OLEDs – Highly Efficient OLEDs

4
Molecular Design and Synthesis of Metal Complexes as Emitters for TADF‐Type OLEDs

Masahisa Osawa and Mikio Hoshino

RIKEN, 2‐1 Hirosawa, Wako, 351‐0198 Saitama, Japan

4.1 Introduction

Nowadays advances in the organic light‐emitting diode (OLED) technology have promoted the commercialization of new light sources for mercury‐free lighting, flat‐panel displays, and smartphones. Practical applications of OLEDs have been realized not only because of the successful synthesis of new emissive substances and peripheral materials but also because of the significant progress in processing technology. Furthermore, extensive research in device chemistry and physics has accelerated the commercialization of OLEDs.

In 1987, Tang and Van Slyke reported the fabrication of the first efficient OLED [1], which consists of the hole‐transport and emissive layers sandwiched between oppositely charged electrodes. The system, which uses an organic molecule as an emitting dopant, was a fluorescent OLED, and thus, only singlet excitons were responsible for the emission.

It has been well established that both the singlet and triplet excited states are generated by charge recombination in OLEDs. The relative yield of the triplet state is three times higher than that of the singlet state [2, 3]. The triplet state of organic molecules scarcely emits at room temperature because of the very small radiative rate constants. Therefore, the electrochemically formed triplet excited states (75%) in OLEDs, which contain fluorescent organic molecules as emitting dopants, hardly participate in emission at room temperature. Hence, in the case of the fluorescent OLEDs, 75% of the excited states generated by charge recombination are wasted as heat. By taking the light‐extraction efficiency into account, the theoretical upper limit of external quantum efficiency (EQE) (photons per electron) for fluorescent OLEDs is evaluated to be as small as 5%.

In order to achieve large EQE values for OLEDs, it is necessary for the triplet state to participate in the emission processes. Fortunately, there are many metal complexes that show strong phosphorescence emission from the triplet state at room temperature. Heavy metal complexes, such as iridium and platinum complexes, doped in the emissive layer have been found to emit strong phosphorescence, leading to remarkable improvement of the EQE (>20%) [4, 5]. Heavy metals in the complexes are known to increase the rate constants for the intersystem crossing (ISC) process, S1 → T1, and the phosphorescence process, T1 → S0, due to a large spin–orbit interaction induced by the heavy atoms. In some cases, the internal quantum efficiencies of the devices reach 100% owing to the unique ability of the metal complexes to harvest both singlet and triplet excitons. Reverse intersystem crossing (rISC) process from T1 to S1 is frequently observed because of the thermal activation of T1, and thus, relatively large S1–T1 gaps (greater than several thousand cm−1) are essential for strongly phosphorescent materials to avoid rISC.

Up to now, a number of tris‐ or bis(cyclometalated) iridium(III) complexes with octahedral geometries have been successfully synthesized as efficient emitters in phosphorescence‐type organic light‐emitting diodes (PHOLEDs) [59]. The high photoluminescence quantum yields (PLQYs > 80%) and relatively short lifetimes (a few microsecond) are very important to avoid the roll‐off effect, which is frequently observed in the plot of illuminance versus current density. The roll‐off effect sometimes results from triplet–triplet annihilation (TTA). The unique photophysical properties of the metal complexes mentioned above can be elucidated from the nature of the three triplet sublevels with characteristic zero‐field splitting (ZFS) parameters, which are similar to those of [Ru(bpy)3]2+ (bpy = 2,2′‐bipyridine) [1012]. In fact, the highest sublevel of the metal‐to‐ligand charge‐transfer (3MLCT) triplet state of the iridium complex has a very large radiative rate constant (kr) because of the facile mixing with 1MLCT via effective spin–orbit coupling (SOC). According to the studies on comprehensive emission mechanisms of these Ir‐based phosphorescent materials [1316], the strong MLCT character of the Ir complexes in the emitting excited states is found to contribute to the large emission yield and the short lifetime of phosphorescence. Further, an investigation on the symmetry effects in the molecule reveals that the complexes, which have degenerate d orbitals, increase the MLCT character of the emissive excited state. After appropriate optimization of the device that uses the Ir complex as a dopant, high EQEs over 30% in PHOLEDs have been successfully obtained [17, 18].

The most efficient phosphorescent emitters are the complexes containing iridium or platinum, which are very expensive and highly localized on the earth. In addition, Ir(III) or Pt(II) with d6 or d8 configurations have low‐lying d–d* states. These facts are an inevitable impediment for Ir(III) or Pt(II) complexes to achieve low‐priced pure blue phosphorescent OLEDs.

Recently, new materials, which exhibit thermally activated delayed fluorescence (TADF), have been investigated for application as emitting dopants in the OLEDs. Delayed fluorescence (DF) has been originally observed for organic molecules [19]. Although the spectra of DF are the same as those of normal fluorescence (NF), the lifetimes of DF and NF are very different. DF has been divided into two groups depending on the origin of emission: P‐ and E‐types. A representative molecule for the P‐type is pyrene, which exhibits DF from the excited singlet state produced by TTA. An example for an E‐type molecule is eosin. It exhibits DF from the excited singlet state, which is thermally populated from the triplet state of the molecule. Thus, lifetimes of DF for both types are markedly affected by the triplet lifetimes, generally resulting in much longer emission lifetimes than those of NF.

With regard to the E‐type molecules, the S–T energy gaps are small, and consequently, the excited singlet state is thermally accessible from the triplet state by an rISC process. The E‐type DF is synonymous with the TADF.

When (i) the radiative rate constant, kr, of S1 is much larger than the nonradiative rate constant (knr) and (ii) the rate to achieve the equilibration between S1 and T1 is sufficiently faster than the triplet decay rate, fluorescence yields via the TADF process are necessarily very large. Needless to say, the rate for the achievement of the equilibration is governed by the S1–T1 energy gap: The smaller the gap, the faster the rate.

TADF‐type OLEDs with the use of organic emitters were reported by Adachi and coworkers [20]. The organic emitters studied were composed of an electron donor and an acceptor moiety in a molecule that exhibits strong intramolecular charge‐transfer (CT) emission. These emitters inevitably have small S–T energy gaps that are necessary to emit efficient TADF. The S–T energy gaps are governed by the electron exchange terms in the excited state. Since the strong CT interaction in the excited state results in the small electron exchange term, J, the organic emitters reported give intensive CT fluorescence via the TADF process.

In general, luminescent materials that emit phosphorescence at room temperature contain heavy metals to gain the large radiative rate constant at T1 by the effective mixing of T1 with higher Sn (n ≥ 2) states through SOC.

In the past two decades, some complexes containing light metals have been found to exhibit TADF [2125]. These complexes have an S–T energy gap small enough to attain thermal equilibrium between the singlet and the triplet state at room temperature. The use of these metal complexes as guest molecules enables us to construct effective OLEDs that show emission by harvesting both the singlet and triplet excitons of host molecules generated by charge recombination inside the devices. The singlet and the triplet excitons of host molecules undergo energy transfer to the guest molecules, the metal complexes, leading to the formation of both the S1 and T1 states of the guest molecules. The S1 state of the metal complexes is readily converted to the T1 state by ISC, and the T1 state generates the thermally activated S1 state responsible for TADF. Accordingly, the excited states harvested by the metal complexes in the devices are mostly transformed to the triplet state, which emits highly efficient TADF.

The TADF from tetrahedral copper(I) complexes [Cu(NN)2]+ with diimine ligands (NN) in solution was first observed by McMillin and coworkers [26]. The emission occurs from the MLCT excited states. The plots of the emission lifetimes versus temperature, T, measured for [Cu(NN)2]+ have been elucidated on the basis of the Boltzmann distribution between the S1 and T1 states, and thus, the so‐called two‐emitting‐state model of this complex was proposed. The energy gap between the two states was estimated as c. 1800 cm−1. The radiative rate constant, kr, from the upper level, S1, was estimated as ∼107 s−1, and that from the lower T1 level, ∼103 s−1. As predicted by the two‐state model, the S–T energy gap is found to be close to the difference in energy between the emission peaks at room temperature and 77 K. The two‐state model proposed for [Cu(NN)2]+ is consistent with recent results obtained by femtosecond spectroscopic methods and theoretical calculations [2729]. The lifetimes of greenish TADF from [Cu(NN)P2]+ with aryl phosphine ligands (P) in the solid state were measured as a function of temperature to determine the S–T energy gap. It is found that the S–T energy gap (1000 cm−1) [30] between two states is much smaller than that of [Cu(NN)2]+.

The first application of tetrahedral Cu(I) complexes as an emitter in OLEDs was reported by Wang and coworkers [31]. At that time, efficient emission from the copper(I) complexes was not recognized as TADF. However, recent studies on emission from the Cu(I) complexes revealed that many complexes with tetrahedral structures would be TADF emitters. Nowadays, many TADF‐type emitters of Cu(I) complexes doped in devices are found to afford high EQEs in comparison with those of the well‐established Ir(III)‐based materials.

In this chapter, we introduce metal complex emitters exhibiting TADF for OLEDs, particularly d10 metal complexes. In Section 4.2, we discuss the emissive electronic states of tetrahedral copper(I) complexes from the perspective of their molecular orbital (MO) configuration. The correlation between the MO character and the coordinating atoms like P (phosphorus ligands) and N (nitrogen ligands) is briefly described. In Section 4.3, we mention important properties of the mononuclear TADF‐type Cu(I) emitters and summarize the guidelines for the fabrication of efficient OLEDs based on these materials. The conventional OLEDs containing these emitters in the emitting layer exhibit a high efficiency comparable with that of cyclometalated iridium(III)‐based devices, which are the current standards used as a measure of efficiency. In Section 4.4, representative examples of the dinuclear TADF‐type Cu(I) emitters applied in OLEDs are briefly expounded. Other d10 metal complexes (Ag(I) and Au(I)) exhibiting TADF have been presented in Section 4.5. The last section is a short conclusion.

4.2 Cu(I) Complexes for OLEDs

Luminescent Cu(I) complexes have been extensively studied since the 1980s. A vast number of investigations have been carried out for understanding the structural changes occurring in the excited states of [Cu(NN)2]+ due to the pseudo‐Jahn–Teller effect since it has been suggested that the tetrahedral structure of [Cu(NN)2]+ in the excited state transforms into a “flattened” square‐planar‐like form [3237]. The conformational changes occurring in the excited state are an interesting target for fundamental research. However, such conformational changes create problems for Cu(I) complexes with two diimine ligands to achieve high luminescence quantum efficiency. This is because the structural changes occurring in the excited states are commonly accompanied by acceleration in the rate of nonradiative decay processes, and thus, most of the tetrahedral Cu(I) complexes tend to exhibit weak emission. A major challenge in the design and synthesis of efficient emitters based on tetrahedral Cu(I) complexes in the 1990s was to establish a strategy to prevent structural changes in the excited states of the complexes [3842].

4.2.1 Energy Levels of Molecular Orbitals in Tetrahedral Geometries

Electronic transitions in tetrahedral Cu(I) metal complexes have been theoretically elucidated using DFT calculations. The simple orbital diagrams of these complexes are shown in Figure 4.1. The highest occupied molecular orbital (HOMO: dσ*) is explained as a combination of the three degenerate valence orbitals (3dxy, 3dyz, 3dxz) of a central metal with the four orbitals of the coordinating atoms of the ligands located at the vertices of a tetrahedron. The electronic transitions responsible for emission of tetrahedral Cu(I) complexes are principally governed by the nature of the HOMOs and lowest unoccupied molecular orbitals (LUMOs). The former is antibonding in nature because the filled 3d orbitals of Cu(I) interact with the lone pair electrons of the coordinating atoms of the ligand while the latter is the antibonding orbital (π*) of the aromatic group of the ligand. Thus, the first electronic transition of the tetrahedral Cu(I) complexes is roughly assumed to be due to a d → π* transition. This HOMO–LUMO transition is suggested to be MLCT [43]. The changes of constituent orbitals of HOMO depending on the kind of ligands are discussed later.

Figure 4.1Geometries' orbital diagram of tetrahedral Cu(I) complexes.

4.2.2 Ligand Variation

The character of the HOMO (dσ*) in a tetrahedral geometry depends on the nature of the ligands. Here we present two typical examples, [Cu(dmp)2]+ (dmp = 2,9‐dimethyl‐1,10‐phenanthroline) complex 1 and [Cu(dppbz)2]+ (dppbz = 1,2‐bis(diphenylphosphino)benzene complex 2 [44, 45]. Both the ligands, dmp and dppbz, have been frequently used for synthesizing emissive Cu(I) complexes. As shown in Figure 4.2 (A), nonbonding orbitals of the coordinating P atoms in the diphosphine ligand have energies higher than those of the N atoms in the diimine ligand because of the fact that the valence orbitals of phosphorus and nitrogen atoms are 3p and 2p orbitals, respectively.

Figure 4.2 [Cu(dmp)2]+ (1) and [Cu(dppbz)2]+ (2). (a) Chemical structures and orbital diagrams. (b) NTO pairs for the lowest triplet excited state at the So optimized geometry.

The LUMOs of the complexes are the π* orbitals of the ligands, dmp and dppbz, while the HOMOs are principally composed of the fully occupied 3d orbitals of Cu(I) and the nonbonding valence orbitals of the coordinating atoms of the ligands. Since the 3d orbitals of Cu(I) in 1 contribute to the HOMO to a greater extent than in 2, the HOMO–LUMO transition has a large MLCT character in comparison with the latter, and thus, the transition is considered to be (d–π*). Unlike the case of 1, the HOMO of 2 is mainly composed of 3p orbitals of the lone pair electrons on the P atoms, and therefore, the HOMO–LUMO transition is regarded as (σ–π*).

With regard to the HOMOs, the natural transition orbital (NTO) maps obtained from excited‐state calculations clearly demonstrate the difference in orbital compositions between 1 and 2 as shown in Figure 4.2 (B). This figure illustrates the hole (approximately HOMO) distribution for the T1 states of 1 and 2 with geometries optimized at S0. The contributions of Cu(I) and the nitrogen atom in 1 to the hole distribution are 71% and 17%, respectively. In the hole distribution of 2, the contribution of Cu(I) is as small as 30%, while that of the P atoms of the ligands is as large as 42%. As shown in Figure 4.3, these results explain well the absorption spectra of 1 and 2. The absorption spectrum of 1 shows intense ligand‐centered bands responsible for the π–π* transitions of the dmp ligands in the higher‐energy region and the relatively weak MLCT absorption bands with a λmax of 460 nm in the visible region. In contrast, the electronic absorption spectrum of 2 shows only a broad band with λmax = 280 nm in the UV region, which resembles the characteristic spectrum of the dppbz ligand. This is probably because the contribution from MLCT to the electronic transition is decreased and that from intraligand charge transfer (ILCT) is increased by changing from the NN to PP ligands. In agreement with the HOMO–LUMO transition mentioned above, the contribution of MLCT to the excited states is very different between complexes 1 and 2. Despite this difference, both complexes exhibit the pseudo‐Jahn–Teller effect in their excited states. Complex 2 exhibits a weak red emission with λmax = 680 nm in 2‐methyltetrahydrofuran (2‐MeTHF) (M. Osawa, unpublished data). Upon excitation, complex 1 forms the MLCT excited state in which an electron is transferred from Cu(I) to the π* orbital of the ligands. The central Cu(I) atom is formally oxidized to Cu(II) (d9), leading to a change in structure from tetrahedral to a “flattened” square‐planar‐like structure. In fact, the structural change in the excited states in solutions causes a marked redshift of the emission peak and a decrease in the emission yield. Obviously, the HOMO energy levels of the Cu(I) complexes are sensitive to the deformation of the tetrahedral structure. The reduction in symmetry from tetrahedral to square‐planar‐like structure by flattening of the ligands results in the destabilization of the HOMO, and thus, we observe the redshift of the emission maximum.

Figure 4.3Absorption spectra of [Cu(dmp)2]+(1) and [Cu(dppbz)2]+ (2) in 2‐MeTHF.

Source: M. Osawa, unpublished data.

In summary, the character of the HOMO in Cu(I) complexes changes according to the nature of the lone pair of electrons located on the N or P atoms of the ligands, and the energy level is sensitive to the symmetry of the structure.

4.3 Mononuclear Cu(I) Complexes for OLEDs

Figure 4.4 exhibits (I) the diagram for the emission processes of phosphorescent emitters and (II) that of TADF materials in devices. The large SOC interaction gives a key effect for phosphorescent materials to achieve high yields of T1 and a large radiative rate constant (kr) of T1 resulting from the mixing with higher Sn (n ≥ 2) excited states. However, DF materials give a high emission yield without any large SOC interaction caused by heavy atoms. When kr in the S1 state of the TADF materials is much larger than knr, photoluminescence (PL) quantum efficiencies are as high as 100%.

Figure 4.4Electroluminescence processes of (a) phosphorescent and (b) TADF emitters in devices.

According to a simple excited‐state model, one electron is located in the HOMO, while another one in the LUMO. Using the distance (r12) between the two electrons, the electron exchange term J is expressed as

4.1

Here, ΦLUMO and ΦHOMO are the wave functions of the LUMO and HOMO, respectively.

The energy gap, ΔEST, between S1 and T1 is written as

4.2

The J value decreases when the distance between the two electrons (r12) becomes larger. Equations (4.1) and (4.2) indicate that the TADF occurs more efficiently as the J value becomes smaller. This situation is occasionally satisfied in the case of molecules that possess an electron donor (D) and an acceptor (A) unit. By absorbing light, the excited singlet state of D, D*, interacts with the ground state of A, which leads to the formation of an intramolecular exciplex:

4.3

When D and A are almost separated electronically, the HOMO in the molecule, D–A, is confined to D, and the LUMO to A. In this case, the J value decreases with an increase in the distance, rAD, between A and D. Thus, the energy gap between S1 and T1 in the exciplex tends to decrease with an increase in rAD. A strong emission of TADF from D‐A is presumably observable when (i) the radiative rate constant of the exciplex at S1 is much larger than the nonradiative rate constant, (ii) the energy of the triplet exciplex is lower than those of the triplet states of D and A, and (iii) the rate for attainment of the equilibrium between S1 and T1 is faster than that for the decay of the triplet exciplex.

The organic TADF emitter hardly shows any phosphorescence because of the small radiative rate constant of the excited triplet state. In contrast to organic emitters, the Cu(I) complexes afford both phosphorescence and TADF. Tetrahedral Cu(I) complexes with an MLCT character in its excited states are known to have small ΔEST [26, 30], and thus, the S1 and T1 are in thermal equilibrium at a given temperature. The population ratio, [S1]/[T1], is given by the Boltzmann equation:

4.4

Here and are the degeneracies of the S1 and the T1 states, respectively. The spin multiplicity of S1 and T1 gives . Because of the thermal equilibrium between S1 and T1, the fluorescence lifetime from S1 is the same as that of the phosphorescence from T1. It is frequently observed that the emission peak of the Cu(I) complexes is redshifted ongoing from high to low temperatures. The origin of the redshift of the emission peak observed by lowering the temperature is interpreted in terms of the increase in the population of the triplet state, T1: The emission changes from DF to phosphorescence with a decrease in temperature.

The emission mechanism of Cu(I) complexes is usually expressed as follows:

4.5
4.6
4.7
4.8
4.9

Here, K is the equilibrium constant between the S1 and T1 states. S1 gives fluorescence with the radiative constant, krS, and is thermally deactivated with the nonradiative rate constant, knrS. The triplet state, T1, emits phosphorescence with a rate constant, krT, and is deactivated with a nonradiative rate constant, knrT. With the use of the Eqs. (4.5)–(4.9), the rate constant, k, for the decay rate of emission is formulated as

4.10

Here, and are the rate constants for the decay of S1 and T1, respectively.

The equilibrium constant, K, is given by

4.11
4.12

where ΔG, ΔS, and ΔH are the free energy change, the entropy change, and the enthalpy change between the S1 and T1 states, respectively. When the conformational change between S1 and T1 is assumed to be negligibly small, Eq. (4.11) is equivalent to Eq. (4.4), i.e. ΔS = 0 and ΔH = ΔEST. This assumption is found to be satisfied in many cases of TADF observed for the crystals of Cu(I) complexes [16, 4648].

The rate constants, and , are rewritten as

4.13
4.14

Thus, Eq. (4.10) is formulated as

4.15

Equation (4.15) implies that the radiative and nonradiative rate constants, kr and knr, respectively, of the Cu(I) complexes are expressed as

4.16
4.17

On the assumption that the formation yield of the triplet state is close to 1, the emission yield, Φ, is written as

4.18

In cases where and are independent of temperature, the A and ΔH values in Eq. (4.11) are readily obtained by measuring the temperature dependence of k, which is given by Eq. (4.10). The and values in Eq. (4.10) must be the same as those obtained from the Φ values given by Eq. (4.18).

In general, the rate constants, kr(S1) and kr(T1), are independent of temperature, while the nonradiative rate constants, knr(S1) and knr(T1), are represented as a function of temperature:

4.19
4.20

Here, (S1) and (T1) are the temperature‐independent terms, and (S1) and (T1) denote the preexponential factors of the temperature‐dependent terms, respectively. As shown in Eqs. (4.19) and (4.20), S1 and T1 have the temperature‐dependent nonradiative processes with activation energies represented by ΔES and ΔET, respectively. With the use of Eqs. (4.10)–(4.20), the rate constant, k, for the decay of emission is strictly expressed as a function of temperature.

It is noted that the electronic property and spin multiplicity of the excited state are reflected in the magnitude of the radiative rate constant. Thus, the determination of the kr values for the emission process is very important for the elucidation of the electronic nature of the emissive excited state.

Photophysical properties of crystalline Cu(I) complexes, which emit TADF, are very sensitive to the molecular structure. As noted previously, the HOMO energy level is markedly affected by the symmetry of the molecular structure in crystals. The LUMO (mainly π*) level is also changed by intermolecular and/or intramolecular interactions (π–π, CH–π, and hydrogen bond interactions) [49, 50]. The tetrahedral TADF‐type Cu(I) complexes studied hitherto are roughly classified into two groups: (i) the complexes that have emissive NN ligands, which indicates that the LUMO is located on these NN ligands, and (ii) the complexes that have the LUMO on PP ligands as shown in Figure 4.5 and Figure 4.6. The chemical structures of the studied ligands are summarized in Figure 4.7.

Figure 4.5 Category A: The LUMO is on the NN ligand.

Figure 4.6 Category B: The LUMO is on the PP ligand.

Figure 4.7 Chemical structures of studied ligands (A‐2 in Figure 4.5).

The Cu(I) complexes afford little detectable NF because of the very fast ISC process (∼10−11 s) occurring from S1 to T1 [27, 28]. Thus, the triplet yield, ΦST, is estimated to be ∼1. In spite of the fact that the SOC interaction of Cu(I) is weak, the ΦST value of the Cu(I) complexes is outstandingly large.

4.3.1 Bis(diimine) Type

TADF from bisdiimine‐type copper(I) complexes ([Cu(NN)2]+) in solution was first observed in the early 1980s [26]. This type of complexes generally exhibits TADF from the MLCT excited states. Recent experimental results and MO calculations clearly prove the small energy gap between S1 and T1 [42, 51, 52]. Based on extensive research concerning the highly emissive complexes, it has been established that the incorporation of bulky substituents into the diimine ligands on the side of the metal center (A‐1; Figure 4.5) is very effective in preventing the structural changes of the excited states in both the solutions and solid states. For instance, sterically congested alkyl substituents are generally introduced at the 2‐ and 9‐positions of phenanthroline (or the 6,6′‐positions of 2,2′‐bipyridine), resulting in a high ΦPL value. Among these complexes, the homoleptic copper(I) bisdiimine complex [Cu(dtp)2]BF4 3 (dtp = 2,9‐di‐tert‐butyl‐1,10‐phenanthroline) displays the maximum TADF performance: ΦPL = 0.056 and τ = 3.26 µs in CH2Cl2 under a N2 atmosphere [53]. Evidently, introduction of the congested alkyl groups into the ligands improves the photophysical properties of the Cu(I) complexes.

4.3.2 [Cu(NN)(PP)]+ Complexes with phen or bipy Derivatives as Ligands

Heteroleptic complexes composed of the NN (phen‐ or bipy‐based ligand) and PP ligands generally exhibit better TADF performance than bis(diimine) complexes (category A‐2 in Figure 4.5) [54]. These ionic complexes possess counter anions. Exchange of one diimine ligand in [Cu(NN)2]+ for the PP ligand generates σ bonds between the Cu(I) and P atoms. Strong back‐donation of the coordinated P atoms results in electron withdrawing from Cu(I) in the complexes. Hence, the MLCT interaction between Cu(I) and the diimine ligand in [Cu(NN)(PP)]+becomes weak, resulting in the shift of the emission peak toward blue in comparison with the emission peak of [Cu(NN)2]+. The emission yield of [Cu(NN)(PP)]+ is much larger than that of [Cu(NN)2]+, probably due to the energy gap law [54]. However, the emission yields of [Cu(NN)(PP)]+ in solutions are usually not so high (Φ < 0.01), and the pseudo‐Jahn–Teller effect in the excited states is suggested to be responsible for the low emission yields. In 2002, [Cu(dmp)(POP)]BF4 (8) containing the bis[2‐(diphenylphosphino)phenyl]ether (POP) ligand was first reported to display efficient emission with Φ = 0.16 and τ= 16.1 µs in solutions by McMillin [55]. Since this report, Cu(I) complexes with various combinations of diimine and diphosphine ligands have been synthesized to provide strong TADF materials. In contrast to the discussion mentioned above, recent theoretical studies on [Cu(NN)(PP)]+ complexes suggest that the emissive transition from [Cu(NN)(PP)]+ is dominated by MLCT [56, 57].

Photophysical properties are markedly affected by the changes in the environment surrounding the emissive metal complexes. Actually, Cu(I) complexes in the solid state or dissolved in films afford large emission yields owing to the suppression of the structural distortions in the excited state. The emission data obtained with [Cu(NN)(P2 or PP)]+ complexes in the solids and the films are summarized in Table 4.1 (see the chemical structures of the ligands and the complexes in Figure 4.7 and the Appendix 4.A.1) [31, 5865]. Even in rigid environments, substitution of the diimine ligand with bulky alkyl or phenyl groups at positions close to the central Cu(I) is found to exhibit remarkable increase in emission efficiencies of [Cu(NN)(P2 or PP)]+ (426). The substituents of the N–N ligands are responsible for suppression of the structural changes in the MLCT excited states. In fact, as shown in Table 4.1, the emission yields tend to increase when the substituent becomes bulky. Further, the TADF from [Cu(NN)(P2 or PP)]+ with large substituent groups in the N–N ligands are found to be blueshifted more than that with small groups. This finding indicates that the large substituents clearly suppress the structural changes occurring from a tetrahedral to a square‐planar‐like geometry. These facts are explained by assuming that a small geometry distortion of the Cu(I) complexes occurs even in films, leading to the significant effects on TADF.

Table 4.1 Emission properties of [Cu(NN)(P2 or PP)]+ (A‐2 in Figure 4.5).a.

Compound λmax (nm) ΦPL (%) τ (µs) Condition References
[Cu(phen)P2]BF4 4 543 14 8.1
[Cu(dmp)P2]BF4 5 509 32 18.1
[Cu(dbp)P2]BF4 6 504 57 32.9 PMMA (20 wt%) [31]
[Cu(phen)(POP)]BF4 7 555 16 4.6
[Cu(dmp)(POP)]BF4 8 527 49 13.2
[Cu(dbp)(POP)]BF4 9 519 69 20.3
523 71 2.5, 12.2 Neat film [58]
[Cu(bpy)P2]BF4 10 550 0.4 10.0
[Cu(4dmbpy)P2]BF4 11 537 0.2 10.7
[Cu(bpy)(POP)]BF4 12 559 3.3 9.0
[Cu(4dmbpy)(POP)]BF4 13 544 0.3 11.2 PMMA (5 wt%) [59]
[Cu(dmbpy)(POP)]BF4 14 528 14.5 15.7
[Cu(bpy)(xantphos)]BF4 15 559 3.3 9.0
[Cu(tmbpy)(xantphos)]BF4 16 528 14.5 15.7
[Cu(tmbpy)(POP)]BF4 17 555 (555) 55 (74) 13 (11) Powder (after grinding) [60]
[Cu(mbpy)(POP)]PF6 18 550 (567) 10.7 (9.5) 6.0 (2.6)
Powder (PMMA) [61]
[Cu(dmbpy)(POP)]PF6 19 529 (535) 38.4 (43.2) 10.9 (10.5)
[Cu(dmp)(POP)]tfpb 20 517b 88b 26b
[Cu(dmp)(xantphos)]tfpb 21 540b 66b 30.2b Crystalline film [62]
[Cu(dipp)(xantphos)]tfpb 22 513a 95b 38.5b
[Cu(phen)(dppbz)]ClO4 23 553 18.33 2.15, 7.42 Powder [63]
[Cu(bpy)(dppe)]PF6 24 601b <0.5b 0.28, 0.53b Powder
[64]
[Cu(4dmbpy)(dppe)]PF6 25 583 b <0.5b 0.19b Powder
[Cu(dmp)(phanephos)]PF6 26 530 80 14 Powder [65]

aλmax is the peak wavelength of emission spectrum, ΦPL is the photoluminescence quantum yield, and τ is the emission lifetime.

bUnder Ar.

[Cu(dbp)P2]BF4 5 with n‐butyl groups at the 2‐ and 9‐positions of the phenanthroline ligand gives high quantum yield of 57% in PMMA films. Interestingly, [Cu(dbp)(POP)]BF4 (9) shows a Φ value of 69%. Complexes 7 and 9 were the first examples of tetrahedral TADF‐type emitters used for device fabrication [31]. The best efficiency of 11.0 cd A−1 at 1.0 mA cm−2 was obtained with the use of an emitting layer of poly(vinylcarbazole) (PVK) containing complexes 7 or 9 (23 wt%) in 2004. The configuration of the OLED devices is as follows: ITO/PEDOT/PVK: 7 or 9 (23 wt%)/BCP/Alq3/LiF/Al. In 2006, light‐emitting electrochemical cells (LECs: ITO/9/Al) were reported [58]. A current efficiency of 56 cd A−1 at 4 V and an EQE of 16% were achieved by the device with the use of complex 9. In 2012, further optimization, specifically by using the high triplet energy charge transport material as a host and an exciton‐blocking layer, successfully enhanced the performance of the devices based on complex 9. The best efficiency of 49.5 cd A−1 and an EQE of 15% were achieved [66]. This high EQE value suggests that complex 9 is a TADF‐type emitter. The device structure is as follows: ITO/PEDOT: PSS/PYD2: 9 (10 wt%)/DPEPO/LiF/Al.

Luminescent characteristics of structurally similar [Cu(tmbpy)(POP)]BF4 (17) have been investigated in detail [60]. Complex 17 shows efficient yellow luminescence with a λmax of 555 nm and ΦPL = 0.55 in the solid state at 300 K. By cooling down the sample from 300 to 77 K, the emission peak shifts to a longer wavelength by 20 nm (630 cm−1). The S1−T1 gap of less than 720 cm−1 for complex 17 was determined by a curve fitting of the emission lifetimes measured at various temperatures. The small S1−T1 gap and the temperature‐dependent behavior of the emission maximum indicate that complex 17 in the solid state at room temperature gives rise to TADF. The computational results reveal that the origin of TADF is mainly the MLCT excited state: The LUMOs are localized on the π system of the diimine ligand, while the HOMOs are on Cu(I).

Numerous studies on the related heteroleptic Cu(I) complexes have been carried out to elucidate the relationship between the emission properties and molecular structures in both crystals and solutions. It has been reported that complex 4 crystallized from solutions gives a single crystal in which two different conformers are confined. The fact that the single crystal shows no TADF at room temperature is explained by the π‐stacking effects and the intermolecular energy transfer from one conformer to the other [67]. In films, complex 4 is free from the stacking effects and intermolecular energy transfer, resulting in the emission by TADF [31].

As mentioned above, the emission from crystals is quenched by the π‐stacking effects, which effectively take place the nonradiative process in the excited states. Thus, the emission yield in crystals is expected to be increased by removal of the π‐stacking effect. Actually, complex 7 in crystals is free from the π‐stacking effects, and thus, the emission yield is reported to be high [68].

[Cu(NN)(PP)]+, which has the substituent groups on the metal side of the diimine ligands, is unstable in solution and readily yields homoleptic [Cu(NN)2]+ and [Cu(PP)2]+ complexes by a disproportionation reaction [69]. Although isolated heteroleptic complexes [Cu(dmp)(PP)]+ containing two methyl groups in the diimine ligands are very stable in the solid state, the formation of homoleptic complexes via ligand exchange reactions is observed in solutions. The homoleptic/heteroleptic ratio is basically dependent on their relative stabilities. Thus, selection of diphosphine ligands that affect the stability of [Cu(PP)2]+ is very important for the synthesis of heteroleptic Cu(I) complexes with high yields.

Among the PP ligands, the POP ligands are the most popular diphosphine ligands to prepare heteroleptic Cu(I) complexes, because the homoleptic complex [Cu(POP)2]+ is very unstable in solution due to the large bite angle (P–Cu–P = ∼ 115°) [45]. In fact, the isolated homoleptic complex possesses a trigonal geometry with an uncoordinated phosphorus atom more preferentially than a tetrahedral one. As shown in Table 4.1, the emission yields of the [Cu(NN)(POP)]+ complexes are generally larger than those of the heteroleptic complexes with xantphos, dppbz, and dppe ligands.

Recently, [Cu(dmp)(phanephos)]PF6 26 has been found to exhibit efficient TADF [65]. This complex has the rigid diphosphine ligand, phanephos, which has a large bite angle. The P–Cu–P angle of 116° in complex 26 is very similar to that of [Cu(dmp)(POP)]+. Complex 26 displays a strong green emission with a peak maximum at 530 nm. The value of PLQY obtained with powder is as high as 0.80 at 300 K. The energy gap, ΔE(S1–T1), between S1 and T1 was obtained as 1000 cm−1 by measuring the rate constants for the decay of emission at various temperatures, indicating that complex 26 is probably a TADF‐type emission material. These studies indicate that the highly luminescent Cu(I) complexes have the rigid structures to reduce the rates for the nonradiative processes in the excited states.

4.3.3 [Cu(NN)(PP)]+ Complexes with NN Ligands Other Than phen or bipy Derivatives

The energy levels of LUMOs of the heteroleptic ionic Cu(I) complexes with a phen‐ or bipy‐based ligand are very similar to each other because the Cu(I) complex in this category has a LUMO localized on the π* system of the diimine ligands. Replacement of a phen‐ or bipy‐based ligand in [Cu(NN)(PP)]+ complexes with other diimine ligands is a common approach to search for highly luminescent Cu(I) emitters. In this context, such heteroleptic copper complexes, which are used in OLEDs as emitters, have been extensively studied and compiled. The chemical structures of the complexes are shown as category A‐3 in Figure 4.5, and the chemical structures of the NN ligands investigated are represented in Figures 4.8 and 4.9 (see the chemical structures of the complexes in Appendix 4.A.1). The emission and device data of the complexes described in the literature are summarized in Table 4.2. The PP ligands used for the Cu(I) complexes are mostly limited to POP and two P ligands.

Figure 4.8 Chemical structures of NN ligands 1 (A‐3 in Figure 4.5).

Figure 4.9 Chemical structures of NN ligands 2 (A‐3 in Figure 4.5).

Table 4.2 Emission properties and device structure from [Cu(NN)(P2 or POP)]+ (A‐3 in Figure 4.5).a

Compound λPL (nm) ΦPL (%) Device structure λEL (nm) ηext (%) References
[Cu(NN1)P2]BF4 27a 618 8.0
ITO/PEDOT:PSS/PVK: 27a or 27b (10 wt%)/BCP/Alq3/LiF/Al
617 0.08b [70]
[Cu(NN2)P2]BF4 27b 623 10 626 0.08 b
[Cu(NN3)P2]BF4 27c 606 56 ITO/PEDOT:PSS/TCCz: 27c (10 wt%)/TPBI/LiF/Al 606 1.7 b
[Cu(NN1)(POP)]BF4 28a 623 6.0
ITO/PEDOT:PSS/PVK: 28a or 28b (10 wt%)/BCP/Alq3/LiF/Al
625 0.3 b
[Cu(NN2)(POP)]BF4 28b 628 10 629 0.6 b
[Cu(NN3)(POP)]BF4 28c 617 43 ITO/PEDOT:PSS/TCCz: 28c (15 wt%)/TPBI/LiF/Al 618 4.5 b
[Cu(NN4)(POP)]BF4 29 585 ITO/2‐TNATA/NPB/CBP: 29 (6 wt%)/TPBI/LiF/Al 572 [71]
[Cu(NN5)P2]BF4 30 500c 0.1c ITO/PEDOT:PSS/PVK: 30 (3.4 wt%)/BCP/Alq3/LiF/Al 497 [72]
[Cu2(NN6)P4](BF4)2 31 550 15 ITO/PVK: 31 (20 wt%)/F‐TBB/Alq3/LiF/Al 589 [73]
[Cu(NN7)(POP)]BF4 32a 552 ITO/m‐MTDATA/NPB/CBP: 32a (7 wt%)/Bphen/Alq3/LiF/Al 572 [74]
[Cu(NN8)(POP)]BF4 32b 521 ITO/m‐MTDATA/NPB/CBP: 32b (15 wt%)/Bphen/Alq3/LiF/Al 528
[Cu(NN9)P2]BF4 33a 568 ITO/MoO3/NPB/CBP: 33a (15 wt%)/BCP/LiF/Al 570 [75]
[Cu(NN9)(POP)]BF4 33b 568 ITO/MoO3/NPB/CBP: 33b (8 wt%)/BCP/LiF/Al 573
[Cu(NN10)(POP)]BF4 34a 579 12
ITO/PEDOT:PSS/TCCz: 34 (10 wt%)/BCP/Alq3/LiF/Al
547 [76]
Cu(NN11)(POP) 34b 564 16 555
Cu(NN14)(POP) 35 481 35 ITO/TAPC/mCP: 35 (8 wt%)/3TPyMB/TmPyPB/LiF/Al ∼530 6.6 [77]
[Cu(NN15)(POP)]BF4 36a 530 25 2.0 [78]
[Cu(NN16)(POP)]BF4 36b 549 27 ITO/PEDOT:PSS/PYD2: 36 (5 wt%)/DPEPO/LiF/Al 6.1
[Cu(NN17)(POP)]BF4 36c 544 36 7.4
[Cu(NN17)(POP)] BF4 37a 470 8
ITO/m‐MTDATA/NPB/CBP: 37a (23 wt%) or 37b (18 wt%)/Bphen/Alq3/LiF/Al
480 [79]
[Cu(NN18)(POP)] BF4 37b 525 34 532
[Cu(NN19)(POP)]BF4 38 525 c 0.25 c ITO/m‐MTDATA/NPB/CBP: 38 (9 wt%)/Bphen/Alq3/LiF/Al 525 [80]
[Cu(NN20)(POP)]BF4 39a 490d 56d 516 3.18 [81]
[Cu(NN21)(POP)]BF4 39b 465 d 87 d ITO/PEDOT:PSS/26mCPy: 39 (20 wt%)/DPEPO/LiF/Al 504 1.59
[Cu(NN22)(POP)]BF4 39c 492 d 75 d 508 8.47
[Cu(NN23)(POP)]BF4 40 518 d 98 d ITO/PEDOT:PSS/czpzpy: 40 (20 wt%)/DPEPO/TPBI/LiF/Al 516 6.36 [82]
[Cu(NN24)(POP)]BF4 41a 532 16 537 3.5 [83]
[Cu(NN25)(POP)]BF4 41b 537 14 ITO/PEDOT:PSS/PYD2: 41 (5 wt%)/DPEPO/LiF/Al 546 4.6
[Cu(NN26)(POP)]BF4 41c 516 48 526 6.7
[Cu(NN27)(POP)]BF4 41d 517 37 516 8.7

aSee the Appendix 4.A.1 and 4.A.2 for abbreviations and molecular structures of materials for OLEDs; λPL is the peak wavelength of photoluminescence spectra of films, ΦPL is the photoluminescence quantum yield in films, and ηext is the external quantum efficiency.

bMeasured at 10.0 mA cm−2.

cIn CH2Cl2 solutions.

dIn the solid state.

By extending the π system of diimine ligands, the LUMO levels in NN1–NN3 (Figure 4.8) become lower in energy, and the distortions in the excited states are reduced [70]. Complexes 27c and 28c with the most bulky and rigid NN3 ligands are found to exhibit high PLQYs in films, and their emission maxima are located at long wavelengths around 620 nm. The device performance of 28c possessing a POP ligand is much better than that of 27c with two P ligands. After optimization of the device structures and the dopant concentration, the device with the structure ITO/PEDOT/TCCz: 28c/TPBI/LiF/Al gives a current efficiency up to 6.4 cd A−1 and an EQE of 4.5%.

NN4 in the complex, 29, and NN5 in the complex, 30, are both dppz derivatives. The device containing 29 fabricated by the vacuum vapor deposition technique gives a turn‐on voltage as low as 4 V, a maximum current efficiency of 11.3 cd A−1, and a peak brightness of 2322 cd m−2 [71]. The oxadiazole unit attached to the NN5 ligand is expected to act not only as an electron‐transporting role but also as a fence for prevention of π stacking between the molecules in the deposited films. The devices made of 30 as an emitter afford the brightness of 47 cd m−2 at 50 A cm−2 [72].

Metal complexes having the heterocyclic aromatic ligand, dipyrido[3,2‐a:2′,3′‐c]phenazine (dppz), have been known to show interesting emission properties [84, 85]. Re(I) complexes with dppz were first used as emitters in OLEDs [86].

Diimines consisting of a pyridine and a five‐membered heterocyclic compound, e.g. imidazole, benzimidazole, pyrazole, triazole, and tetrazole (NN6–NN27 in Figure 4.9), are commonly used as ligands for the preparation of luminescent metal complexes. The nitrogen atom of these heterocyclic rings is a stronger electron donor than that of pyridine, resulting in shorter Cu–N bond lengths in the Cu(I) complexes. The stability of the complexes increases by using this type of ligand.

In 2005, the dinuclear Cu(I) complex [Cu2(NN6)2P4](BF4)2 31, possessing two 2‐(2′‐pyridyl)benzimidazole ligands, was reported for the first time [73]. With an increase in temperature from 77 K to ambient temperature, complex 31 shows a blueshift of the emission maximum from 563 to 550 nm in a PMMA film. This result suggests that the emission is ascribed to TADF. The MO of the related mononuclear complex indicate that the electronic nature of the HOMO is principally dominated by 3d orbitals of the Cu(I) ion. However, the contribution from the POP ligand cannot be ignored. The LUMO is mainly distributed on the 2‐(2′‐pyridyl)benzimidazole unit. Accordingly, the lowest electronic transition is considered to be an MLCT mixed with ligand‐to‐ligand' charge transfer (LL'CT) (CT from P atoms of the POP ligand to π* orbitals of the diimine ligand) [87]. The luminescence efficiency of the prototype device [ITO/PVK: 31(20 wt%)/F‐TBB/Alq3/LiF/Al] is low, and yellow‐orange electroluminescence (EL) is observed. Most complexes in this category have counterions because the two bidentate ligands are neutral. These complexes are unstable toward sublimation and poorly soluble and/or unstable in nonpolar solvents, and hence, these are not amenable to vacuum deposition or solution processing methods for the preparation of OLEDs. However, the 2‐(2′‐pyridyl)benzimidazole ligand is attractive because of thermal stability. Thermal analyses demonstrate that the 2‐(2′‐pyridyl)benzimidazole ligand and NN8 are decomposed at 671 and 588 K, respectively [74]. On the other hand, there seems no obvious sign of decomposition with regard to NN7 at these temperatures. No thermal decomposition of both NN7 and NN8 takes place below 873 K. The Cu(I) complexes, 32a and 32b, containing the NN7 and the NN8 ligands are found to be thermally stable below 627 and 595 K, respectively. Therefore, the two complexes are stable enough for us to construct OLED devices by the vacuum deposition method at 573 K. Furthermore, the oxadiazolyl and carbazolyl arms incorporated into NN7 and NN8, respectively, are useful for the electron‐ and hole‐transport processes in the devices.

The OLEDs utilizing 32a and 32b as dopants in the CBP emissive layer were fabricated by the vacuum deposition method. The 7 wt% 32a‐ and 15 wt% 32b‐doped devices afford the peak EL efficiency of 2.8 cd A−1 at 1.1 mA cm−2 and 2.2 cd A−1 at 1.4 mA cm−2, respectively. The maximum brightness of the device with 32a was 8669 cd m−2 at 14 V. This value is much higher than previously reported values of the devices made from other Cu(I) complexes.

Recently, the relationship between the structural rigidity of the Cu(I) complexes and the performance of devices has been examined with the use of similar complexes, 33a and 33b [75]. Both complexes having the NN9 ligand are stable enough to be sublimated during the course of the EL device fabrication. The device fabricated using 33b, which contains a rigid POP ligand, showed a better performance in comparison with that using 33a, which is composed of two P ligands having a rigid structure less than POP. The rigidity of the copper emitters seems to be one of the important factors that dominate the device performance. In fact, the device doped with 8 wt% 33b shows a strong yellow EL with a maximum brightness of 4758 cd m−2 at 12.3 V.

Luminescence properties and the device performance are expected to differ between the neutral and ionic Cu(I) complexes. 2‐(2′‐Pyridyl)benzimidazole ligands are readily transformed from a charge neutral to an anionic ligand by losing a proton in the presence of bases. Thus, it is possible to make both the neutral and ionic Cu(I) complexes with the use of these ligands. Neutral mononuclear Cu(I) complexes (34b) and ionic complexes with counterions (34a) were synthesized using the NN10 and NN11 ligands in order to compare their luminescence properties and device performance [76]. The neutral complex, 34b, shows a blueshifted emission with longer lifetimes in comparison with the ionic complex. This result is explained by assuming that although both 34b and 34a have radiative transition based on MLCT plus ligand‐centered π–π* transition (LC), the former possesses LC character much larger than the latter. By doping 34 in TCCz, OLEDs are fabricated with the device structure of ITO/PEDOT: PSS/TCCz: 34(10 wt%)/BCP/Alq3/LiF/Al. The device with the charge‐neutral ligand 34b exhibits a higher current efficiency than that with 34a.

A series of neutral Cu(I) complexes with 5‐(2‐pyridyl)tetrazolate (NN13) and various phosphine ligands is found to show better luminescence properties than the ionic Cu(I) complexes containing the neutral NN12 ligand [88]. Although PLQYs of the ionic complexes are as small as 4–46%, those of the neutral complexes are 89% in the solid state. MO calculations indicate that the neutral complexes emit from the excited state, (ML + IL)CT. With an increase in the temperature from 77 K to ambient temperature, the neutral complex shows a blueshift of the emission maximum in the solid state, which implies that this emission would be TADF.

The charge‐neutralized Cu(I) complex affords a high PLQY and undergoes facile sublimation under vacuum. The charge‐neutral complex 35 having the anionic ligand NN14 is readily doped in OLEDs by the vacuum deposition method [77], and the peak EL efficiency of the device is obtained as 6.6%. It is suggested that 35 emits phosphorescence because of its strong MLCT character in the excited state, a substantially high triplet yield, ΦISC, and large radiative rate constant, kr [88].

An approach to enhance the emission efficiencies of cationic complexes is to suppress the C–H vibrations [78]. According to the energy gap law, high‐frequency vibrational modes such as C–H effectively induce nonradiative processes [89, 90]. By replacing these modes with those of lower frequency, the PLQY is expected to increase. The effects of suppression of the excited‐state distortion and the C–H vibrational quenching on the PL quantum efficiency have been studied with the use of the complexes, 36a–c. The ligands of these complexes are systematically changed from N15 to N17 to examine the effects of the C–H vibrational modes on the PLQY. The films doped with the complexes give the PLQY as 0.25 for 36a, 0.27 for 36b, and 0.36 for 36c, respectively. From the Stokes shifts observed for these complexes, it was assumed that the large PLQY obtained with 36c could be due to the suppression of C–H vibrational quenching. The increase in PLQYs improves the performance of OLEDs: The maximum EQE is 2.0% for 36a, 6.1% for 36b, and 7.4% for 36c [78].

Blue emitters made of Cu(I) complexes are an attractive target to produce full‐color displays. A good strategy to create a blue emitter is to synthesize the Cu(I) complex with a large energy separation between the HOMO and LUMO. It is known that (i) the HOMOs of the Cu(I) complexes are mainly confined to the d orbitals of the Cu(I) ion and (ii) the LUMOs are generally distributed over the π* orbitals of the NN ligand. Thus, a high energy of the LUMOs of the N–N ligand is necessarily required to synthesize the blue emitter. In particular, electron donors of N and/or S introduced into the NN ligands at the α‐ and α′‐positions of a CN bond raise the energy level of LUMOs as is seen in NN17 and NN18. Actually, emission maxima are observed at 470 nm for 37a having NN17 and at 525 nm for 37b having NN18 [79]. These peaks are certainly shifted to high energy in comparison with those of complexes 3335, which have low‐energy LUMOs. The OLEDs utilizing 37a and 37b as dopants were fabricated with a general structure of ITO/m‐MTDATA/NPB/CBP:37a (23 wt%) or 37b (18 wt%)/Bphen/Alq3/LiF/Al. The EL emission peaks are located at 480 nm for the 37a‐based device and at 532 nm for the 37b‐based device. Although the 37a‐based device affords blue emission color sufficient for practical use, the peak EL efficiency is as low as 1.47 cd A−1. In order to improve the EL efficiency, a structurally similar ligand, NN19, which has the carbazolylbutyl unit as an additional electron‐donating group, was prepared. However, contrary to expectations, complex 38 with NN19 gave green emission: The peak wavelength of the emission spectrum is located at 525 nm. The OLEDs utilizing 38 are fabricated by the vacuum deposition method. The 9 wt% 38‐doped device gives a peak EL efficiency of 1.71 cd A−1 and a maximum brightness of 1500 cd m−2 [80].

As mentioned above, Cu(I) complexes, which emit blue light, have sufficient energy separation between the HOMO and LUMO. Electron‐rich pyridylpyrazole ligands NN20–NN22, which are considered to have high‐energy LUMOs, were prepared for the synthesis of Cu(I) complexes as blue emitters [81]. Powdered complexes 39ac with NN20–NN22 afford emission maxima at 490, 465, and 492 nm with high PLQYs of 56%, 87%, and 45%, respectively. Detailed investigation of the emission lifetimes at various temperatures reveals that these complexes are typical TADF materials with small energy gaps between S1 and T1 (∼1400 cm−1). The MO calculations carried out for these complexes support an observation that the energy gaps between S1 and T1 are small enough to show TADF [91]. The solution‐processed OLEDs using 39c exhibit an emission maximum at 508 nm, an EQE of 8.47%, a peak current efficiency of 23.68 cd A−1, and a maximum brightness of 2033 cd m−2.

A new solution‐processed method to make OLEDs containing complex 40 is reported [82]. The emissive layers of 40 are readily prepared by spin‐coating of the CH2Cl2 solution‐dissolved [Cu(CH3CN)4(POP)]BF4 and excess NN23. In this process, NN23 having the carbazole moiety acts as both a ligand and the host material, and thus, no purification of 40 is necessary. The OLED fabricated by this process affords an emission maximum at 514 nm, an EQE of 6.36%, a peak current efficiency of 17.53 cd A−1, and a maximum brightness of 3251 cd m−2 at 14.3 V. This emission efficiency is comparable with that of the device made by conventional solution processes using the isolated complex 40. From the examination of the emission lifetimes at various temperatures, complex 40 was assumed to emit TADF: The energy gap between S1 and T1 was estimated as 1049 cm−1.

Complexes 41ad with triazolylpyridine ligands N24–N27 have been used as emitters in OLEDs [83]. The methyl group introduced into the pyridine moiety at a position in the metal side improves the photophysical properties of the Cu(I) complexes: The structural distortion of the complexes is suppressed by the methyl group in the excited state as seen in 41a–d. However, the carbazole unit introduced into the triazole ring exhibits no effect on the photophysical properties. The OLED using 41d gives the highest current efficiency of 26.2 cd A−1 and EQE of 8.7%.

In summary, owing to the rapid development of TADF‐type heteroleptic copper(I) emitters, [Cu(NN)(PP)]+, prototype OLEDs have provided new light sources for practical applications. Interestingly, the HOMO and LUMO of [Cu(NN)(PP)]+ are localized separately: The LUMO distribution is largely confined to the π* of the NN ligand, whereas the HOMO distribution is essentially confined to the Cu(I) atom and partly to the two P atoms of PP ligand. Thus, localization of the HOMO and LUMO in the Cu(I) complexes is one of the useful guiding principles for the synthesis of the Cu(I) complexes that show TADF.

4.3.4 Tetrahedral Cu(I) Complexes with the LUMO on the PP Ligand

As described in the previous sections, the key units in the [Cu(NN)(PP)]+ complexes responsible for TADF are the aromatic groups of the NN ligand. The LUMO is principally localized on the NN ligand and the HOMO on the Cu(I) atom. The emission occurs from the MLCT (d,π*) excited states. This section describes mononuclear Cu(I) complexes with the LUMO localized on diphosphine ligands as given in categories B‐1 and B‐2 in Figure 4.6. The structures of the complexes are shown in Figure 4.10, and the PL and the corresponding device data are summarized in Table 4.3.

Figure 4.10 Chemical structures of studied Cu(I) complexes (B‐1 and B‐2 in Figure 4.6).

Table 4.3 Emission properties and device structures from [Cu(NN)n(PP)2−n]+ or 0 (B‐1 in Figure 4.6).a.

Compound λPL ΦPL Device structure λEL ηext References
(nm) (%) (nm) (%)
[Cu(Ph2Bpz2)(dppbz)] 42 545 50  552 11.9
[Cu(Ph2Bpz2)(dppbz‐F)] 43 534 63 ITO/TAPC/mCP: 4244 (10 wt%)/ 3TPYMB/LiF/Al  545 16.0 [92]
[Cu(Ph2Bpz2)(dppbz‐CF3)] 44 523 68  528 17.7
Cu(H2Bpz2)(POP) 45 436b 45 [25]
Cu(Bpz4)(POP) 46 447 b 90
Cu(Ph2Bpz2)(POP) 47 464 b 90
[Cu(dppbz)2]BF4 48 497 b 56 ITO/PVK: 48 or 49 (12.5 wt%)/Al ∼590
[44] (M. Osawa, unpublished data)
[Cu(dppbz)(POP]BF4 49 494c  2 ∼620
Cu(dppbz)(PS) 50 545 23 ITO/PEDOT:PSS/ PVK/mCP: 50 (10 wt%), TAPC(30 wt%)/ 3TPYMB/LiF/Al   550  7.8 [93]

aSee the Appendix 4.B for abbreviations and molecular structures of materials for OLEDs; λPL is the peak wavelength of photoluminescence spectra of films, ΦPL is the photoluminescence quantum yield in films, and ηext is the external quantum efficiency.

bIn the solid state.

cIn CH2Cl2 solutions.

Charge‐neutral Cu(I) complexes are very attractive as promising emitters for vacuum‐deposited OLEDs because they can be readily sublimed under vacuum. From this viewpoint, Cu(I) complexes 4244 have been prepared [92]. These tetrahedral Cu(I) complexes were composed of dppbz derivatives and the anionic bidentate ligand diphenyl‐bis(pyrazol‐1‐yl)borate (pz2Bph2) as shown in Figure 4.10a.

Dppbz and POP are common diphosphine ligands used to prepare transition metal complexes. It contains two types of aromatic groups: a bridging o‐phenylene group and auxiliary phenyl groups. The pz2Bph2 ligand of complexes 4244 possesses a relatively high‐energy π* orbital in comparison with that of dppbz derivatives, and thus, the LUMOs of these complexes are the π* orbitals of dppbz derivatives. As indicated in Figure 4.2, there is a strong relationship between the photophysical properties of dppbz and that of the Cu(I) complex because the transition responsible for TADF is regarded as a σ → π* [MLCT + ILCT].

Thermogravimetric analyses of complexes 42–44 under vacuum indicated that, with an increase in the fluorine contents of the ligands, the Cu(I) complexes tend to become more sublimable. Complexes 42–44 in vacuum‐deposited amorphous films have been found to show a strong green emission. Figure 4.11 displays the emission spectra of 42–44 in films at 293 and 77 K. Most probably, the excited‐state structures of 42–44 are immobilized in the amorphous films due to the rigidity of the 1,3‐bis(carbazol‐9‐yl)benzene (mCP) host. The bright green emission is considered to arise from an excited state with a tetrahedral structure. Although the half‐widths and λmax of emission spectra at 77 K are very similar to those at 293 K, the emission edges of 42–44 on the short wavelength side have been redshifted. Furthermore, the lifetimes of emission at 77 K are one or two orders of magnitude longer than those at 293 K. These observations indicate that the emission from 42 to 44 at room temperature would be a TADF. An NTO analysis demonstrated that the origin of green luminescence from 42 to 44 is mainly due to a σ → π* transition (MLCT + ILCT).

Figure 4.11 Corrected emission spectra of complexes 4244 in films at 300 (thick line) and 77 K (thin line).

Bottom‐emitting devices with a conventional three‐layer structure of ITO (110 nm)/TAPC (30 nm)/mCP + 10% 4244 (25 nm)/3TPYMB (50 nm)/LiF (0.5 nm)/Al (100 nm) were fabricated. Properties of the devices are shown in Figure 4.12. All the three devices emit bright green light with the emission peak wavelength at 552 nm for 42, 545 nm for 43, and 528 nm for 44. The maximum current efficiencies are determined as 34.6 for 42, 46.7 for 43, and 54.1 cd A−1 for 44. These values are obtained with the current density of 0.02 mA cm−2 for 42, 0.20 mA cm−2 for 43, and 0.02 mA cm−2 for 44. The maximum EQEs are evaluated as 11.9%, 16.0%, and 17.7% for devices containing 42, 43, and 44, respectively. These high EQEs are close to those of the PHOLEDs based on rare metal complexes.

Figure 4.12 Properties of OLEDs containing complexes 4244. (a) Electroluminescence spectra. (b) The dependence of the EQE on the current density. (c) I–V characteristics. (d) The dependence of the current efficiency on the current density.

There are some fabricated examples of TADF‐type OLEDs containing Cu(I) complexes as emitters [21, 22]. However, each of the devices including ours is still prototype. As shown in Figure 4.12, roll‐off is serious for our devices. We have not measured the lifetimes of the devices. Such efforts are certainly future works for us after improvement of the devices' performance, e.g. the optimization of peripheral materials.

The complexes in the excited state show an MLCT character that causes flattening motions of the complex in the excited state, resulting in an increase in the rate of nonradiative decay. Thus, the solutions of 42–44 exhibited weak luminescence with Φ < 0.02 at 293 K. However, as mentioned above, complexes 42–44 in rigid films afford strong emissions at room temperature. The EL performance of the devices is strongly related to the PL properties of Cu(I) complexes in the amorphous films. Thus, the detailed studies on the PL properties of the complexes in amorphous films are inevitable prior to manufacturing the devices. The copper(I) complexes presented here possess simple structures and are attractive as emitters for OLEDs because of the fact that these complexes are readily sublimed in vacuo for facile doping in the emissive layer of the devices.

Strong blue TADF from structural analogues of the Cu(I) complexes 4547 was reported [25]. The chemical structures of 4547 are shown in Figure 4.10b. It is expected that these Cu(I) complexes would be applicable to OLEDs. Heteroleptic or heteroleptic bis(diphosphine) Cu(I) complexes, [Cu(dppbz)2]BF4 (48) and [Cu(dppbz)(POP)]BF4 (49), represented in Figure 4.10c were applied to OLEDs as emitters [44]. Although complex 48 showed a strong emission with the PLQY of 56% in the solid state, the corresponding device performances are moderate at 490 cd m−2 for 48 and 330 cd m−2 for 49. In both cases, the EL spectra are redshifted and significantly broader in comparison with the PL spectra observed in the solid states. This result suggests that the EL performance of the devices is not directly related to the PL properties in the solid states.

Figure 4.10d shows a charge‐neutral tetrahedral Cu(I) complex, 50, containing dppbz and a bidentate anionic ligand, 2‐diphenylphosphinobenzenethiolate (PS) [93]. The luminescence from an amorphous film of mCP doped with 10% complex 50 was investigated. The emission peak wavelength observed at 293 K (λmax = 545 nm) is redshifted to λmax = 561 nm at 77 K, and the lifetime at 293 K (τ = 2.1 µs) becomes c. 350 times shorter than that at 77 K (τ = 780 µs). These observations suggest that luminescence from 50 in the amorphous film at 293 K is ascribed to TADF. In fact, the energy gap between S1 and T1 in 50 was estimated to be around 600 cm−1 from the temperature dependence of the decay rate constants of emission. NTO analyses of the MO along with the hole and electron maps reveal that the major transitions responsible for TADF are concerned with two types of LL'CT: One is the CT transition of an electron from sulfur to an empty antibonding π orbital on the phenylene and phenyl rings in the dppbz ligand (S → π*), and another is that from the π orbital on the phenylene ring in the PS ligand to an empty antibonding π orbital on the phenylene and phenyl rings in the PP ligand (π → π*). The thiolate ligand (PS) with electron‐donating character reduces the contribution of the metal orbitals to the HOMOs of the complexes, resulting in the decrease in the MLCT character of the excited states. It is interesting that complex 50 exhibits TADF from an LL′CT transition, but not from the MLCT or σ → π* transitions. The quantum efficiency of complex 50 in films is 23% at most.

Complex 50 is thermally stable with a decomposition temperature higher than 300 °C. However, vacuum deposition of 50 was unsuccessful because of the extremely low vapor pressure of the complex. Thus, a prototype OLED containing 50 was prepared via a wet process. This device containing 50 exhibited green luminescence with a current efficiency of 21.3 cd A−1 and a maximum EQE of 7.8%. Since ΦPL of complex 50 in films is 23%, the EQE value is reasonable as shown in Table 4.3.

4.3.5 Charge‐Neutral Three‐Coordinate Cu(I) Complexes

In this section, we describe the preparation and the photophysical properties of charge‐neutral three‐coordinate Cu(I) complexes containing phosphine ligands. The structure of the complex is shown as category B‐3 in Figure 4.6.

Monodentate halide anions such as Cl, Br, and I are very popular ligands to prepare luminescent multinuclear Cu(I) complexes, particularly dinuclear and tetranuclear complexes possessing diamond {Cu2X2}n units (X = Cl, Br and I, n = 1 or 2). These units are naturally occurring core structures. Hitherto, over a thousand complexes composed of {Cu2X2}n cores and various organic ligands have been reported [50, 94, 95]. In 2007, dinuclear Cu(I) complexes, [(dppbz)CuX]2, were found to exhibit efficient green emission with ΦPL = 0.6–0.9 in the solid state (see next Section 4.4 in detail) [96]. This finding encouraged us to prepare a mononuclear three‐coordinate Cu(I) complex with the use of dppbz derivatives having congested structures to avoid the formation of dinuclear Cu(I) complexes.

The three‐coordinate complexes having molecular weights lower than those of the dinuclear complexes are expected to sublime readily in vacuum and are suitable for the preparation of OLEDs by the vacuum deposition method. We initially prepared simple three‐coordinate Cu(I) complexes [(LMe)CuX] [X = Cl (51), Br (52), or I(53)] by using a chelating diphosphine ligand [LMe = 1,2‐bis(o‐ditolylphosphino)benzene] [97]. Further, diphosphine ligands LEt and LiPr, possessing the ethyl and isopropyl substituents, respectively, were synthesized [98]. The luminescence properties of diphosphine ligands, LMe, LEt, and LiPr, were found to be almost identical. The structures of Cu(I) complexes [(LEt)CuBr] 54 and [(LiPr)CuBr] 55 are also shown in Figure 4.13. Listed in Table 4.4 are the emission peak wavelengths, λmax, the quantum yields, Φ, and the lifetime, τ, observed for the Cu(I) complexes in films at 293 K.

Figure 4.13(a) Molecular structures of 5155. (b) ORTEP view of 52. Thermal ellipsoids are drawn at the 50% probability level, and H atoms have been omitted for clarity. (c) Core structure of 52.

Table 4.4 Emission properties of charge‐neutral three‐coordinate Cu(I) complexes (B‐2 in Figure 4.6).a.

Compound T = 293 K T = 77 K References
λmax (nm) Φ (%) τ (µs) λmax (nm) Φ (%) τ (ms)
(LMe)CuCl 51 517 67 6.1, 4.6 518 84 2.8, 0.67
(LMe)CuBr 52 513 71 5.5, 3.9 506 95 1.3, 0.16
(LMe)CuI 53 504 57 3.2, 1.4 487 83 0.51, 0.15 [97, 98]
(LEt)CuBr 54 510 66 8.8, 3.4
(LiPr)CuBr 55 508 64 8.4, 2.6
(LMe)CuSPh 56 488b 95 b 6.6 b 481 95 1.1

[99]
(LiPr)CuSPh 57 500 b 95 b 5.0 b 504 95 1.9
P2Cucbz 58 461c 24 c 11.7 c
P2CuNPh2 59 521 c 23 c 3.17 c [100]
(POP)CuNPh2 60 563 c 18 c 1.70 c
[(NHC1)Cu(phen)]OTf 61 1.5 0.23, 1.1
(NHC1)Cu(NN12) 62 630 35 24.7
(NHC1)Cu(NN28) 63 560d 17 d 10 d 570 c 58 c
(NHC1)Cu(NN29) 64 594 d 3.2 d 1.3 d 574 c 68 c [101104]
(NHC1)Cu(NN30) 65 592 d 1.4 d 0.5 d 560 c 77 c
(NHC1)Cu(NN31) 66 590 d 2.4 d 1.1 c 555 c 61 c
(NHC1)Cu(NN32) 67 475 b 76 b 11 b 490 b 91 b 34 µs b
(NHC2)Cu(NN32) 68 575 b 73 b 34 b 585 b 80 b 21 µs b

aλPL is the peak wavelength of photoluminescence spectra of films, ΦPL is the photoluminescence quantum yield in films, and τ is the decay time of emission.

bIn the solid state.

cIn methylcyclohexane solutions.

dIn cyclohexane solutions.

Single‐crystal X‐ray diffraction studies reveal that complexes 51–53 and 55 possess monomeric three‐coordinate structures. The molecular structure of 52 is shown in Figure 4.3b as an example. The coordination geometries of the copper centers in 51–55 are trigonal planar. The sums of the angles around the Cu(I) center are 359.66° for 51, 359.37° for 52, 359.43° for 53, and 360.00° for 55. The o‐methyl groups of LMe are essential for the formation of the three‐coordinate complexes. In fact, dppbz, which lacks methyl groups, only forms the halogen‐bridged binuclear copper complex [(dppbz)Cu(μ‐X)]2. This is probably because the Cu2X2 diamond core in [(LMe)Cu(μ‐X)]2 would be highly unstable due to the steric hindrance of the o‐methyl groups located on the side of the metal centers, and thus, unusual monomeric three‐coordinate structures are produced. The introduction of a bulky substituent on the ligand appears necessary to produce the three‐coordinate copper complexes.

The PL properties of 51–55 are presented in Table 4.4. Complexes 51–55 emit intense green phosphorescence in degassed CH2Cl2 solutions at 293 K. The luminescence quantum yields (Φ) of 0.43, 0.47, 0.60, 0.43, and 0.50 were obtained for 51, 52, 53, 54, and 55, respectively. The phosphorescent lifetimes (τ) of the three complexes were measured by laser excitation at 355 nm: τ = 4.9, 5.4, 6.5, 3.8, and 8.3 µs for 51, 52, 53, 54, and 55, respectively. Emission maxima (λmax) of 51–53 are in the order 51 < 52 < 53, suggesting that λmax is affected by the ligand field strength (I < Br < Cl). Presumably the electronic nature of the triplet excited state of 5153 is influenced to some extent by X → π*(LMe) CT transitions.

NTO analysis reveals that these transitions can be described as single hole–electron pairs and reproduces over 95% of the change in electron density upon excitation. The hole (approximately HOMO) and electron (approximately LUMO) distributions of 52 in the excited triplet state with the T1 optimized geometry are shown in Figure 4.14 (a). The hole distribution of 52 is essentially confined to the halogen atom but also extended slightly to the σ orbitals between the Cu(I) and P atoms. It is known that Cu d orbitals mix well with p orbitals of the Br atoms as in the case of the {Cu2X2} core [105, 106]. Thus, the hole distribution spreads over the Br, Cu, and P atoms. On the other hand, the electron distribution is largely confined to the o‐phenylene group in the LMe ligand. Excited‐state calculations carried out for 51–53 indicate that the emission results from the transition (σ + X) ← π*.

Figure 4.14(a) NTO pairs for T1 of (LMe)CuBr 52 at the T1 optimized geometry. (b,c) The optimized core structure of models in S0 (left) and T1 (right) and molecular orbital diagrams (below). (b) [Cu(PMe3)3]+ and (c) (LMe)CuBr 52.

Omary observed a Jahn–Teller distortion from Y‐ to T‐shaped geometry in the triplet excited state of trigonal planar Au(I) complexes containing three monophosphine ligands ([Au(PR3)3]+) [107, 108]. Our calculations suggest that the Cu(I) complex, [Cu(PMe3)3]+, with a structure analogous to that of [Au(PMe3)3]+, undergoes similar distortion in the excited state. The optimized core structures of [Cu(PMe3)3]+ and complex 52 in the ground state and triplet excited state are shown in Figure 4.14b,c. This type of distortion seems to be inevitable for [Cu(PMe3)3]+ with an equilateral triangle structure because the two orbitals, 3dxy and 3dx2−y2, are degenerate in the ground state. Since, upon excitation, an electron is transferred from the HOMO to the LUMO (4pz plus 4 s), the degeneracy of the two orbitals is resolved in the excited state, leading to the T shape of the complex. On the other hand, complex 52 has an isosceles triangle structure, suggesting that the two 3d orbitals are not fully degenerate in the ground state. Besides, the P–Cu–P angle is tightly fixed to c. 90° by the rigid LMe ligand. The structural features of 51–53 largely prevent the distortion. The small distortion of the excited states of 51–53 is assumed to reduce the rate of nonradiative decay, leading to a high Φ.

Complex 52 emits intense green phosphorescence in various matrices, dichloromethane, tetrahydrofuran, ethanol, toluene, films, and crystals at 293 K as shown in Figure 4.15. Spectral shifts hardly depend on the nature of matrices, suggesting that there is little structural change between the ground and excited states. This observation is consistent with the NTO analysis.

Figure 4.15 Corrected emission spectra of (LMe)CuBr 52 in various matrices.

Figure 4.16 shows the emission lifetime, τ, measured for complex 52 in crystals between 77 and 300 K. The plot of τ versus T is well explained by the model of two excited states, S1 and T1. The S1–T1 energy gap, ΔE(S1–T1), of complex 52 is obtained as 810 cm−1 by the curve fitting of the plot of τ versus T with the use of Eq. (4.21) [98]:

4.21

Figure 4.16 Emission decay time of (LMe)CuBr 52 in crystals versus temperature. The parameters described in the inset were determined from a fit of Eq. 4.21 for (LMe)CuBr 52.

Complexes 51–55 afford small S1–T1 energy gaps, 600–830 cm−1, in the solid state, indicating that the emission at room temperature is probably due to TADF.

The spectrum of phosphorescence from T1 is commonly located at wavelengths longer than that of fluorescence from S1. The TADF mechanism claims that the emission spectrum shows a redshift upon cooling [46]. Since the thermal population of the lower excited state, T1, becomes dominant at low temperatures, phosphorescence from the T1 state precedes TADF from the S1 state, resulting in the redshift of emission from TADF materials. However, complexes 5253 show a small blueshift in the emission spectra when the temperature is decreased from 293 to 77 K, as shown in Table 4.4. Such a phenomenon has been occasionally found for the materials with TADF [9799]. The blueshift of the emission observed by lowering the temperature is presumably related to the energy relaxation processes in the excited states. However, further studies are necessary for full understanding of the blueshift of emission observed for TADF systems at low temperatures.

Conventional bottom‐emitting devices containing 5155 were fabricated using the vacuum deposition method. The devices had a three‐layer structure of indium tin oxide ITO/TAPC/mCP: 10 wt% complex 5155/3TPYMB/LiF/Al. All of the devices containing complexes 5155 emit bright green light with a maximum emission peak at 513–529 nm. The emission spectra of the OLEDs are in good agreement with the PL spectra of the complexes in the films. The maximum current efficiencies of OLEDs using complexes 5155 were 55.6–69.5 cd A−1 at a current density of 0.01 mA cm−2. At the maximum current efficiencies, the maximum EQEs were obtained as 18.6–2.5%. The excellent performance of the devices can be ascribed to the TADF‐type emission in these systems. The Cu(I) complexes 5155 emit from a thermally populated excited singlet state at an ambient temperature because of a small singlet–triplet (S1–T1) energy gap. Figure 4.17 shows (i) the EL spectra of the devices, (ii) I–V characteristics, and (iii) the dependence of the EQE on the current density. Table 4.5 lists Von, λmax, EQEmax, ηc,max, and CIE.

Figure 4.17 Properties of OLEDs containing 5155. (a) EL spectra. (b) I–V characteristics. (c) The dependence of the EQE on the current density.

Table 4.5 OLEDs' performance characteristics.

Complex Von (V)a EL emission, λmax (nm) EQEmax (%)b ηc,max (cd A−1)c CIE coord. x/y
51 3.3 527 21.1 67.7 0.30/0.55
52 3.0 517 21.3 65.3 0.29/0.54
53 3.1 513 21.2 62.4 0.26/0.51
54 3.3 529 22.5 69.4 0.32/0.54
55 3.3 515 18.6 55.6 0.26/0.51

aThe voltage required to reach a brightness of 1 cd m−2.

bMaximum external quantum efficiency.

cMaximum current efficiency.

There are some fabricated examples of TADF‐type OLEDs containing Cu(I) complexes as emitters [21, 22]. Each of the devices including ours, however, is still prototype. We believe that further studies are necessary to manufacture practical devices with high luminance and long operating life after improvement of the devices' performance, e.g. the optimization of peripheral materials.

The three‐coordinate Cu(I) complexes, Cu(LMe)(SPh) 56 and Cu(LiPr)(SPh) 57, which afford TADF from the LL'CT excited state, are prepared by replacing halides with arylthiolate anions [99]. The chemical structures of 56 and 57 are shown in Figure 4.18a. The Cu(I) complexes produce intense blue‐green emission with the quantum yields as high as approximately 1.0 at both 293 and 77 K in the solid state. Small S1–T1 gaps (ΔE(S1–T1) = 690 cm−1 for 56 and 630 cm−1 for 57) indicate that the emission from 56 and 57 in the solid state at room temperature is presumably ascribed to TADF.

Figure 4.18 Chemical structures of three‐coordinate Cu(I) complexes 5668.

MO calculations reveal that the major transitions (∼ 95%) that contribute to the emission from 56 and 57 are attributed to the two types of LL'CT. One is the CT from the sulfur atom to an empty antibonding π orbital on the phenylene and tolyl rings of the diphosphine ligand (S → π*), and another is from the π orbital on the aryl ring of the thiolate ligand to an empty antibonding π orbital on the phenylene and tolyl rings of the diphosphine ligand (π → π*). Conversely, MLCT contributions are very small: ∼2–3% for 56 and 57. Although the transitions responsible for the emission in complexes 56 and 57 are almost the same as that of complex 50, the PL efficiencies are very different: PLQY of complex 50 is as low as 0.23. Further studies are necessary to elucidate this difference. Complexes 56 and 57 were thermally unstable, and thus, we could not make OLEDs with them using the vacuum deposition method.

LL'CT emission from other three‐coordinate Cu(I) complexes, 58–60, composed of P(P) and amide ligands has been reported [100]. The chemical structures of 5860 are shown in Figure 4.18b, and their photophysical data are summarized in Table 4.4. The calculated HOMO is localized over the entire amide ligand, while the LUMO is localized on the π* orbital of the phosphine ligands. The emission, therefore, is assumed to occur from the LL'CT excited states. On the other hand, the MLCT character is very small. It seems possible to tune the emission color by a careful choice of amide and phosphine ligands as is exemplified by complexes 58, 59, and 60, which exhibit blue, green, and yellow emission, respectively. Relatively short lifetimes imply that the emission could be ascribed to TADF.

N‐Heterocyclic carbene (NHC) Cu(I) complexes with three‐coordinate structures (61–68) have been reported [101104]. The chemical structures of 61–68 are shown in Figure 4.18c and the Appendix 4.A.1., and their photophysical data are summarized in Table 4.4. The carbene complexes possessing a LUMO confined in the NN ligands emit both phosphorescence and TADF. According to the MO calculations, the structural changes due to the rotational motion around the CNHC–Cu bond are closely related to the luminescent characteristics. It was found that the change in the torsion angle between the two ligands induces the change in the energy gap between S1 and T1. A large torsion angle results in a large energy gap (ΔE(S1–T1)), leading to an increase in the intensity of phosphorescence and a decrease in the intensity of TADF. Actually, complex 68 with a large torsion angle exhibits pure phosphorescence (τ = 18 µs) at ambient temperature, whereas complex 67 with a small torsion angle displays emission (τ = 11 µs; phosphorescence 38% and TADF 62%). The energy gaps of 67 and 68 have been estimated as 740 and >3000 cm−1, respectively. This system gives a good guideline for designing TADF‐type materials [104].

The synthesis and photophysical properties of emissive charge‐neutral three‐coordinate Cu(I) complexes were described in this section. These results suggest that three‐coordinate copper(I) complexes are promising EL materials in terms of emission efficiency and thermal stability.

4.4 Dinuclear Cu(I) Complexes for OLEDs

Since the early 1970s, multinuclear, in particular di‐, tri‐, and tetranuclear, Cu(I) complexes displaying intriguing emission characteristics have been extensively studied [50, 94, 96, 109, 110]. The {Cu2(μ‐X)2}n complexes (where X = Cl, Br, and I, n = 1 or 2) have a common core structure, Cu2(μ‐X)2, frequently found in such multinuclear Cu(I) complexes, which exhibit luminescence thermochromism and luminescence rigidchromism caused by the structural changes in the excited state. Such sensitivity to external stimuli is an unfavorable character for optoelectronic materials. Thus, prevention of structural changes in the excited state might be necessary to synthesize strongly luminescent multinuclear Cu(I) complexes. In this section, dinuclear Cu(I) complexes that are used in OLEDs have been briefly surveyed. The chemical structures of the complexes examined are illustrated in Figure 4.19.

Figure 4.19 Chemical structures of dinuclear Cu(I) complexes 6978.

4.4.1 Dinuclear Cu(I) Complexes Possessing {Cu2(μ‐X)2} Cores

In 2007, the preparation of dinuclear copper(I) complexes [(dppbz)Cu(μ‐X)]2 (X=Cl (69), Br (70), and I (71)) and their application to the vapor‐deposited OLEDs were reported for the first time [96]. Complexes exhibited an efficient green emission with ΦPL = 0.6–0.9 in the solid state, and the origin of emission was regarded as (M + X)LCT from DFT calculations. By measuring the temperature‐dependent emission spectra and lifetimes at various temperatures, the energy gap between the 1(M + L)CT and 3(M + L)CT states in complex 71 was found to be as small as ∼2.0 kcal mol−1. Thus, the efficient green emission is probably ascribed to TADF. Since the EL peak (565 nm) is redshifted in comparison with the PL peak (502 nm) observed in the solid state, the EL is assumed to occur from the “flattened” square‐planar‐like structure in the excited state. Probably, the viscosity of the amorphous host is lower than that of the solid state, allowing the structural change from tetrahedral to a flattened one at room temperature. The configuration of OLED devices was as follows: ITO/PF01/CBP: 71 (10 wt%)/Bphen/KF/Al. The current efficiency of 10.4 cd A−1, power efficiency with 4.2 lm W−1 at 93 cd m−2, and maximum EQE with 4.8% were obtained with the device. In 2012, further optimization was carried out by using the high triplet energy charge transport material as a host with an exciton‐blocking layer. The best efficiency of 30.6 cd A−1 and an EQE of 9.0% were achieved [66]. These results suggest that the dinuclear Cu(I) complexes are promising emitters of OLED.

In 2011, a simple method to fabricate devices via coevaporation of (CuI)2 and pyridine‐based ligands was reported and is shown in Figure 4.19 [111]. The distinctive feature of this process is that the excess organic ligand for Cu(I) also serves as a host matrix. The chemical structure of the Cu(I) complex in a codeposited CuI ligand film was characterized by X‐ray absorption fine structure (XAFS) including X‐ray absorption near edge structure (XANES) and extended EXAFS. By using the codeposited film as an emissive layer, the best OLED (ITO/MoO3/CBP/CuI: CPPyC 73 (4 wt%)/TPBi/LiF/Al) gives a maximum EQE as high as ∼15.7% at a luminance of 100 cd m−2 [112].

The application of halogen‐bridged dinuclear Cu(I) complexes possessing an additional bridging ligand, diphenylphosphinopyridine, in OLEDs was reported in 2012 [113]. The core structure of the complex is shown in Figure 4.19. The geometry of the {Cu2(μ‐X)2} unit in this core is not planar but butterfly‐shaped due to the bridging ligand [114]. The combination of the {Cu2(μ‐X)2} core with various mono‐ and bidentate N‐heteroaromatic ligands and an ancillary ligand, PPh3, is a well‐known method to synthesize efficiently luminescent Cu(I) complexes [95, 110]. This synthetic strategy was also applied to prepare luminescent mononuclear Cu(I) complexes [115]. Nonetheless, the core structure is well designed to have a separated HOMO and LUMO. The LUMO is localized on the pyridine moiety, whereas the HOMO is distributed on the {Cu2(μ‐X)2} unit. The separation of the HOMO and LUMO is an important factor for the complexes to afford highly efficient TADF (see Section 4.3) and enables us to change the emission color by the introduction of substituents into the pyridine ring (R group in the core structure in Figure 4.19). A series of Cu(I) complexes having this core structure with various substituents on the pyridine ring and other series of complexes possessing the bridging ligand bearing five‐membered heterocyclic moieties (see Section 4.3.3) have been reported [116121]. This strategy mentioned above has been successfully applied to the synthesis of Cu(I) complexes exhibiting emission in a wide range of color from 458 to 713 nm.

A double‐bridged dinuclear Cu(I) complex, 74, was recently demonstrated to give an internal quantum efficiency close to unity in a solution‐processed film. The ΔE(S1–T1) of 74 is estimated as 726 cm−1 by measuring the emission peaks at room temperature and at 77 K. The chemical structure of 74 was identified by performing X‐ray absorption spectroscopy. Furthermore, it is reported that an optimized device (ITO/PEDOT:PSS/PLEXCORE UT‐314/PYD2:74(30 wt%)/3TPYMB/LiF/Al) gives a maximum EQE of 23% (73 cd A−1) [122].

4.4.2 Other Dinuclear Cu(I) Complexes

TADF‐type OLEDs containing the neutral dinuclear Cu(I) complex 75 having the two amido‐bridged core were reported in 2010 and are shown in Figure 4.19 [123]. The rigid structure of complex 75 with two bis(diisobutylphenylphosphino)amido (PNP) ligands certainly enhances the PLQY up to 68% in solutions. The small S1–T1 energy gap of 740 cm−1 determined by fitting of the plot of τ (emission lifetimes) versus T (80–295 K) indicates that the efficient emission from 75 is TADF. A calculation of the HOMO and LUMO of 75 reveals that the HOMO is mainly distributed on the N atoms of the amido groups and slightly on the Cu atoms. The LUMO distribution is largely confined to the phenylene groups on the PNP ligands. Thus, the HOMO–LUMO transition is almost pure CT (MLCT + ILCT) in nature, leading to the small energy gap between S1 and T1 of complex 75. Vapor‐deposited OLEDs (ITO/CFx/TAPC/CBP: TAPC(25 wt%)/CBP: TAPC(20 wt%): 75(0.2 wt%)/CBP/BAlq‐13/LiF/Al) doped with the complex in the emissive layer gave a maximum EQE of 16.1%. The roll‐off observed for the EL devices with increasing current density may be explained by the quenching of the excited state by holes and/or ionization of the excited state under the influence of the electric field on the basis of the examination of PL intensity from the hole‐only device (ITO/CFx/TAPC/TAPC: 75(8 wt%)/TAPC/Al).

Dinuclear Cu(I) complexes (76–78) exhibiting relatively short‐lived emissions from both S1 and T1 in the thermal equilibrium at an ambient temperature have been reported [124]. These complexes are similar to complex 67 (see Section 4.3.4) displaying mixed emission (τ = 11 µs; phosphorescence 38% and TADF 62%). This emission mechanism can be regarded as a four‐excited‐state model in which S1 and three triplet substates are in the thermal equilibrium. Complex 76 in powder form shows bright blue emission at 300 K with an emission peak wavelength at 485 nm, a decay time of 8.3 µs, and a quantum yield of 0.92. The emission consists of 20% phosphorescence and 80% TADF. Obviously a large SOC is necessary for the triplet state to emit phosphorescence. The large splittings of triplet substates of 76 are determined as 15 and 7 cm−1 for ΔE(III–I) and ΔE(II–I), respectively. On the other hand, the energy gap between S1 and T1 is 930 cm−1. A balance of ΔE(ZFS) and ΔE(S1–T1) is very important to design this type of complexes.

4.5 Another Group of Metal Complexes Exhibiting TADF

In Sections 4.3.4 and 4.3.5, Cu(I) complexes exhibiting efficient TADF with dppbz derivative ligands have been described. The complexes are divided into two groups: One is the complexes emitting from an ILCT excited state, and another is the complexes from an LL'CT excited state. Both transitions do not require the d orbitals of the copper(I) atom with regard to emissive transitions, indicating that the central metal need not be limited to the copper(I) atom.

Figure 4.20 shows the molecular structures of silver(I) or gold(I) complexes (79–86) with dppbz derivative ligands [93, 125129] (M. Osawa, unpublished data). The photophysical properties are summarized in Table 4.6. All these complexes exhibit blue‐orange luminescence with λmax of 447–610 nm and ΦPL = 0.12–0.98 in the solid state at 300 K. At 77 K, the emission peaks of these complexes shift to the longer wavelength by 6–25 nm compared with those at 300 K. These results suggest that the emission from 79 to 85 in the solid state at room temperature is ascribed to TADF. According to the DFT calculations, the emissive excited states of complexes 79, 80, and 82–86 have a poor MLCT character (<15%). With regard to complexes 79 and 80, the P atoms are found to contribute largely to the HOMO. The HOMO of three‐coordinate Ag(I) complexes 83 and 84 essentially distributes on both the halogen and P atoms. Complexes 85 and 86 have the HOMOs mainly distributed on the S atom and the phenylene moiety. Thus, TADF is originating from the ILCT excited state for 79 and 80, the (X + I)LCT excited state for 82–84, and the LL'CT excited state for 85 and 86, respectively. It is likely that the TADF of 79–84 is principally dependent on the electronic nature of the dppbz derivatives coordinated to Cu(I).

Figure 4.20 Chemical structures of silver(I) and gold(I) complexes 7986.

Table 4.6 Emission properties of Ag(I) and Au(I) complexes 7985.

T = 300 K T = 77 K
Compound λmax (nm) Φ (%) τ (µs) λmax (nm) Φ (%) τ (µs) ΔE(S1–T1) (cm−1) References
[Ag(dppbz)2]BF479 447 45 1.0, 4.9 453 85 1200 [126, 127]
[Au(dppbz)2]NO380 486 95 3.7 492 95 18.6 630 [125, 128]
[(dppbz)CuCl]2 81 480 93 15 96 1100 980 [129]
[(LMe)AgBr]2 82 487 56 2.6,9.2 500 90 1200 (M. Osawa, unpublished data)
(LEt)AgBr 83 463 70 2.8, 12 482 98 2, 1000
(LiPr)AgBr 84 463 98 5.7, 19 479 98 300, 1300
[Ag(dppbz)(PS)] 85 505 32 0.6, 2.2 530 83 570 [93]
[Au(dppbz)(PS)] 86 610 12 8.4, 2.6 630 18 52

The temperature dependence of the emission decay time of the gold(I) complex 80 is shown in Figure 4.21. The emission is concluded to be TADF on the basis of the small energy gap ΔE(S1–T1) = 630 cm−1 and natural decay times of S1 (70 ns) and T1 (19 µs). It is well known that, because of the heavy atom effects, the Au(I) complexes often emit phosphorescence from the aromatic ligands at room temperature. Au(I) brings forth an increase in the rate constant of spin‐forbidden processes, S → T and T → S, due to SOC induced by the heavy gold atoms [130132]. The TADF emission also suffers similar heavy atom effects. Both the ISC and rISC rates would be increased by the effects of Au(I), and thus, 80 exhibits intense blue TADF with a λmax of 486 nm and ΦPL = 0.95 in the solid state at 300 K.

Figure 4.21 Emission decay time of 80 in crystals versus temperature. The parameters described in the inset were determined from a fit of Eq. (4.21) for 80.

4.6 Conclusion

Over the past three decades, the tetrahedral d10 copper(I) complexes having diimine and/or diphosphine (or two monophosphine) ligands have been found to exhibit TADF. These Cu(I) complexes have been extensively studied as luminescent materials for OLEDs over the last 10 years due to the low cost and stable supply of copper metal.

In this chapter, studies on TADF‐type Cu(I) complexes have been summarized and described. TADF is principally observed from the Cu(I) complexes having a small energy gap between T1 and S1. The nature of the excited states responsible for TADF is classified into three groups: (i) the MLCT and (M + X)LCT excited states, (ii) the ILCT excited state, and (iii) the LL'CT excited state.

A drawback of the TADF‐type Cu(I) complexes is the structural changes occurring in the excited states due to the pseudo‐Jahn–Teller effect. These structural changes bring forth low emission quantum yields for these complexes. This problem is resolved by constructing a “rigid structure” for both mono‐ and dinuclear Cu(I) complexes, which suppresses the structural changes in the excited states. With regard to the [Cu(NN)(PP)]+ complexes, an effective method is to introduce bulky alkyl or phenyl groups into the (NN) ligands at positions facing toward the central metal. Similarly, the bridging ligands of the dinuclear copper(I) complexes with {Cu2(μ‐X)2} units (X = Cl, Br, and I) are found to increase the quantum efficiencies of TADF in amorphous films.

There are still problems for the application of the TADF‐type Cu(I) complexes to an OLED display. For example, an improvement in the thermal and redox stability of the Cu(I) emitter is a key requirement to ensure a long lifetime for the OLEDs fabricated by the vacuum deposition method. For thermal stability, the phosphine ligands seem to be more suitable than imine ligands. However, the metal‐centered oxidation potential of phosphine‐based copper(I) complexes is often close to that of the phosphine ligands, and thus, the complexes are readily decomposed by oxidation. At present, emissive Cu(I) complexes with both high redox stability and thermal stability have not yet been thoroughly researched, but this is an unavoidable challenge for the manufacturing of OLEDs for practical use.

Acknowledgments

The Integrated Collaborative Research Program between RIKEN and Canon Inc. is acknowledged for the funding of our research. M.O. is grateful to the Ministry of Education, Culture, Sports, Science and Technology of the Japanese Government (Grant‐in‐Aid for Scientific Research, No.24410080) for financial support of this research. We are grateful to Dr. S. Okada, Dr. A. Tsuboyama, Dr. K. Suzuki, Mr. M. Yashima, Dr. I. Kawata, Mr. S. Igawa, Dr. M. Hashimoto, and Mr. R. Ishii (Canon Inc.) for the fruitful collaboration.

Appendix

4.A.1 Schematic Structures of 1–86

4.A.2 Abbreviations and Molecular Structures of Materials for OLEDs

ITO: indium tin oxide
PEDOT:PSS: poly(3,4‐ethylene‐dioxythiophene)/poly(styrene sulfonate)
PVK: poly(vinylcarbazole)
Alq3 or AlQ: aluminum 8‐hydroxyquinolinate
BCP: 2,9‐dimethyl‐4,7‐diphenyl‐1,10‐phenanthroline
PYD2 or 26mCPy: 2,6‐bis(N‐carbazolyl)pyridine
DPEPO: bis[2‐(diphenylphosphino)phenyl] ether oxide
TCCz: N‐(4‐(carbazol‐9‐yl)phenyl)‐3,6‐bis(carbazol‐9‐yl) carbazole
TPBI or TBi: 1,3,5‐tris(N‐phenylbenzimidazol‐2‐yl)benzene
2‐TNATA: 4,4′,4″‐tris[(2‐naphthyl)phenylamino]triphenylamine
CBP: 4,4′‐N,N′‐dicarbazole‐biphenyl
F‐TBB: 1,3,5‐tris(4′‐fluorobiphenyl‐4‐yl)benzene
NPB: N,N′‐di(1‐naphthyl)‐N,N′‐diphenyl‐(1,1′‐biphenyl)‐4,4′‐diamine
m‐MTDATA: 4,4′,4″‐tris[(3‐methylphenyl)phenylamino]triphenylamine
Bphen: 4,7‐biphenyl‐1,10‐phenanthroline
MoO3: molybdenum trioxide
TAPC: di‐[4‐(N,N‐ditolyl‐amino)‐phenyl]cyclohexane
mCP: 1,3‐bis(carbazol‐9‐yl)benzene
3TPYMB: tris(2,4,6‐trimethyl‐3‐(pyridine‐3‐yl)phenyl)borane
TmPyPB: 1,3,5‐tri[(3‐pyridyl)phen‐3‐yl]benzene
Czpzpy: 2‐(9H‐carbazolyl)‐6‐(1H‐pyrazolyl)pyridine
CDBP: 4,4′‐bis(9‐carbazolyl)‐2,2′‐dimethyl‐biphenyl
PF01: 4,4′‐bis[phenyl(9,9′‐dimethylfluorenyl)amino]biphenyl
CPPyC: 3‐(carbazol‐9‐yl)‐5‐((3‐carbazol‐9‐yl)phenyl)pyridine
CFx: plasma polymerization of CHF3 at low frequencies
BAlq‐13: bis(2‐methyl‐quinolin‐8‐olato)(2,6‐diphenylphenolato)aluminum(III)

References

  1. 1 Tang, C.W. and VanSlyke, S.A. (1987). Appl. Phys. Lett. 51: 913–915.
  2. 2 Baldo, M.A., Lamansky, S., Burrows, P.E., Thompson, M.E., and Forrest, S.R. (1999). Appl. Phys. Lett. 75: 4–6.
  3. 3 Baldo, M.A., Thompson, M.E., and Forrest, S.R. (2000). Nature 403: 750–753.
  4. 4 Baldo, M.A., O'Brien, D.F., You, Y., Shoustikov, A., Sibley, S., Thompson, M.E., and Forrest, S.R. (1998). Nature 395: 151–154.
  5. 5 Adachi, C., Baldo, M.A., Thompson, M.E., and Forrest, S.R. (2001). J. Appl. Phys. 90: 5048–5051.
  6. 6 Tanaka, D., Sasabe, H., Li, Y.‐J., Su, S.‐J., Takeda, T., and Kido, J. (2007). Jpn. J. Appl. Phys., Part 2 46: L10–L12.
  7. 7 Tanaka, D., Agata, Y., Takeda, T., Watanabe, S., and Kido, J. (2007). Jpn. J. Appl. Phys., Part 2 46: L117–L119.
  8. 8 Kondakova, M.E., Pawlik, T.D., Young, R.H., Giesen, D.J., Kondakov, D.Y., Brown, C.T., Deaton, J.C., Lenhard, J.R., and Klubek, K.P. (2008). J. Appl. Phys. 104: 094501.
  9. 9 Chopra, N., Lee, J., Zheng, Y., Eom, S.‐H., Xue, J., and So, F. (2008). Appl. Phys. Lett. 93: 143307.
  10. 10 Harrigan, R.W., Hager, G.D., and Crosby, G.A. (1973). Chem. Phys. Lett. 21: 487–490.
  11. 11 Nozaki, K., Takamori, K., Nakatsugawa, Y., and Ohno, T. (2006). Inorg. Chem. 45: 6161–6178.
  12. 12 Yersin, H., Humbs, W., and Strasser, J. (1997). Coord. Chem. Rev. 159: 325–358.
  13. 13 Lacky, D.E., Pankuch, B.J., and Crosby, G.A. (1980). J. Phys. Chem. 84: 2068–2074.
  14. 14 Marchetti, A.P., Deaton, J.C., and Young, R.H. (2006). J. Phys. Chem. 110: 9828–9838.
  15. 15 Hofbeck, T. and Yersin, H. (2010). Inorg. Chem. 49: 9290–9299.
  16. 16 Yersin, H., Rausch, A.F., and Czerwieniec, R. (2012). Organometallic Emitters for OLEDs: Triplet Harvesting, Singlet Harvesting, Case Structures, and Trendsin Physics of Organic Semiconductors, 2e (ed. W. Brutting and C. Adachi), 371–424. Wiley‐VCH.
  17. 17 Kim, K.‐H., Moon, C.‐K., Lee, J.‐H., Kim, S.‐Y., and Kim, J.‐J. (2014). Adv. Mater. 26: 3844–3847.
  18. 18 Lee, C.W. and Lee, J.Y. (2013). Adv. Mater. 25: 5450–5454.
  19. 19 Parker, C.A. (1964). Advanced in Photochemistry, vol. 2 (ed. W.A. Noyes Jr., G.S. Hammond and J.N. Pitts Jr.), 305–383. New York: Wiley.
  20. 20 Endo, A., Ogasawara, M., Takahashi, A., Yokoyama, D., Kato, Y., and Adachi, C. (2009). Adv. Mater. 21: 4802–4806.
  21. 21 Tao, Y., Yuan, K., Chen, T., Xu, P., Li, H., Chen, R., Zheng, C., Zhang, L., and Huang, W. (2014). Adv. Mater. 26: 7931–7958.
  22. 22 Dumur, F. (2015). Org. Electron. 21: 27–39.
  23. 23 Czerwieniec, R., Leitl, M.J., Homeier, H.H.H., and Yersin, H. (2016). Coord. Chem. Rev. 325: 2–28.
  24. 24 Leitl, M.J., Schinabeck, A., Yersin, H., Zink, D.M., Baumann, T., and Volz, D. (2016). Top. Curr. Chem. 374: 25.
  25. 25 Czerwieniec, R., Yu, J.‐B., and Yersin, H. (2011). Inorg. Chem. 50: 8293–8301.
  26. 26 Kirchhoff, J.R., Gamache, R.E. Jr., Blaskie, M.W., Del Paggio, A.A., Lengel, R.K., and McMillin, D.R. (1983). Inorg. Chem. 22: 2380–2384.
  27. 27 Iwamura, M., Takeuchi, S., and Tahara, T. (2015). Acc. Chem. Res. 48: 782–791.
  28. 28 Iwamura, M., Takeuchi, S., and Tahara, T. (2007). J. Am. Chem. Soc. 129: 5248–5256.
  29. 29 Iwamura, M., Watanabe, H., Ishii, K., Takeuchi, S., and Tahara, T. (2011). J. Am. Chem. Soc. 133: 7728–7736.
  30. 30 Blasse, G. and McMillin, D.R. (1980). Chem. Phys. Lett. 70: 1–3.
  31. 31 Zhang, Q., Zhou, Q., Cheng, Y., Wang, L., Ma, D., Jing, X., and Wang, F. (2004). Adv. Mater. 16: 432–436.
  32. 32 McMilllin, D.R. and McNett, K.M. (1998). Chem. Rev. 98: 1201–1219.
  33. 33 Everly, R.M., Ziessel, R., Suffert, J., and McMillin, D.R. (1991). Inorg. Chem. 30: 559–561.
  34. 34 Sakaki, S., Mizutani, H., and Kase, Y. (1992). Inorg. Chem. 31: 4575–4581.
  35. 35 Shinozaki, K. and Kaizu, Y. (1994). Bull. Chem. Soc. Jpn. 67: 2435–2439.
  36. 36 Eggleston, M.K., Fanwick, P.E., Pallenberg, A.J., and McMillin, D.R. (1997). Inorg. Chem. 36: 4007–4010.
  37. 37 Eggleston, M.K., McMillin, D.R., Koenig, K.S., and Pallenberg, A.J. (1997). Inorg. Chem. 36: 172–176.
  38. 38 Miller, M.T., Gantzel, P.K., and Karpishin, T.B. (1998). Inorg. Chem. 37: 2285–2290.
  39. 39 Miller, M.T., Gantzel, P.K., and Karpishin, T.B. (1999). J. Am. Chem. Soc. 121: 4292–4293.
  40. 40 Miller, M.T., Gantzel, P.K., and Karpishin, T.B. (1999). Inorg. Chem. 38: 3414–3422.
  41. 41 Miller, M.T. and Karpishin, T.B. (1999). Inorg. Chem. 38: 5246–5249.
  42. 42 Felder, D., Nierengarten, J.F., Barigelletti, F., Ventura, B., and Armaroli, N. (2001). J. Am. Chem. Soc. 123: 6291–6299.
  43. 43 Kutal, C. (1990). Coord. Chem. Rev. 99: 213–252.
  44. 44 Moudam, O., Kaeser, A., Delavaux‐Nicot, B., Duhayon, C., Holler, M., Accorsi, G., Armaroli, N., Seguy, I., Navarro, J., Destruel, P., and Nierengarten, J.‐F. (2007). Chem. Commun. 3077–3079.
  45. 45 Kaeser, A., Moudam, O., Accorsi, G., Seguy, I., Navarro, J., Belbakra, A., Duhayon, C., Armaroli, N., Delavaux‐Nicot, B., and Nierengarten, J.‐F. (2014). Eur. J. Inorg. Chem. 2014: 1345–1355.
  46. 46 Yersin, H., Rausch, A.F., Czerwieniec, R., Hofbeck, T., and Fischer, T. (2011). Coord. Chem. Rev. 255: 2622–2652.
  47. 47 Leitl, M.J., Kuechle, F.‐R., Mayer, H.A., Wesemann, L., and Yersin, H. (2013). J. Phys. Chem. A 117: 11823–11836.
  48. 48 Czerwieniec, R. and Yersin, H. (2015). Inorg. Chem. 54: 4322–4327.
  49. 49 Zhang, X., Chi, Z., Zhang, Y., Liu, S., and Xu, J. (2013). J. Mater. Chem. C 1: 3376–3390.
  50. 50 Cariati, E., Lucenti, E., Botta, C., Giovanella, U., Marinotto, D., and Righetto, S. (2016). Coord. Chem. Rev. 306: 566–614.
  51. 51 Siddique, Z.A., Yamamoto, Y., Ohno, T., and Nozaki, K. (2003). Inorg. Chem. 42: 6366–6378.
  52. 52 Asano, M.S., Tomiduka, K., Sekizawa, K., Yamashita, K.‐I., and Sugiura, K.‐I. (2010). Chem. Lett. 39: 376–378.
  53. 53 Green, O., Gandhi, B.A., and Burstyn, J.N. (2009). Inorg. Chem. 48: 5704–5714.
  54. 54 Armaroli, N., Accorsi, G., Cardinali, F., and Listorti, A. (2007). Top. Curr. Chem. 280: 69–115.
  55. 55 Cuttell, D.G., Kuang, S.‐M., Fanwick, P.E., McMillin, D.R., and Walton, R.A. (2002). J. Am. Chem. Soc. 124: 6–7.
  56. 56 Yang, L., Feng, J.‐K., Ren, A.‐M., Zhang, M., Ma, Y.‐G., and Liu, X.‐D. (2005). Eur. J. Inorg. Chem. 2005 (10): 1867–1879.
  57. 57 Zou, L.‐Y., Cheng, Y.‐X., Li, Y., Li, H., Zhang, H.‐X., and Ren, A.‐M. (2014). Dalton Trans. 43: 11252–11259.
  58. 58 Zhang, Q., Zhou, Q., Cheng, Y., Wang, L., Ma, D., Jing, X., and Wang, F. (2006). Adv. Funct. Mater. 16: 1203–1208.
  59. 59 Andres‐Tome, I., Fyson, J., Baiao Dias, F., Monkman, A.P., Iacobellis, G., and Coppo, P. (2012). Dalton Trans. 41: 8669–8674.
  60. 60 Linfoot, C.L., Leitl, M.J., Richardson, P., Rausch, A.F., Chepelin, O., White, F.J., Yersin, H., and Robertson, N. (2014). Inorg. Chem. 53: 10854–10861.
  61. 61 Keller, S., Constable, E.C., Housecroft, C.E., Neuburger, M., Prescimone, A., Longo, G., Pertegas, A., Sessolo, M., and Bolink, H.J. (2014). Dalton Trans. 43: 16593–16596.
  62. 62 Smith, C.S., Branham, C.W., Marquardt, B.J., and Mann, K.R. (2010). J. Am. Chem. Soc. 132: 14079–14085.
  63. 63 Li, X.‐L., Ai, Y.‐B., Yang, B., Chen, J., Tan, M., Xin, X.‐L., and Shi, Y.‐H. (2012). Polyhedron 35: 47–54.
  64. 64 Nishikawa, M., Wakita, Y., Nishi, T., Miura, T., and Tsubomura, T. (2015). Dalton Trans. 44: 9170–9181.
  65. 65 Czerwieniec, R., Kowalski, K., and Yersin, H. (2013). Dalton Trans. 42: 9826–9830.
  66. 66 Zhang, Q., Komino, T., Huang, S., Matsunami, S., Goushi, K., and Adachi, C. (2012). Adv. Funct. Mater. 22: 2327–2336.
  67. 67 Coppens, P., Sokolow, J., Trzop, E., Makal, A., and Chen, Y. (2013). J. Phys. Chem. Lett. 4: 579–582.
  68. 68 Zhang, L., Li, B., and Su, Z. (2009). Langmuir 25: 2068–2074.
  69. 69 Kaeser, A., Mohankumar, M., Mohanraj, J., Monti, F., Holler, M., Cid, J.‐J., Moudam, O., Nierengarten, I., Karmazin‐Brelot, L., Duhayon, C., Delavaux‐Nicot, B., Armaroli, N., and Nierengarten, J.‐F. (2013). Inorg. Chem. 52: 12140–12151.
  70. 70 Zhang, Q., Ding, J., Cheng, Y., Wang, L., Xie, Z., Jing, X., and Wang, F. (2007). Adv. Funct. Mater. 17: 2983–2990.
  71. 71 Che, G., Su, Z., Li, W., Chu, B., Li, M., Hu, Z., and Zhang, Z. (2006). Appl. Phys. Lett. 89: 103511–103513.
  72. 72 Walsh, P.J., Lundin, N.J., Gordon, K.C., Kim, J.‐Y., and Lee, C.‐H. (2009). Opt. Mater. 31: 1525–1531.
  73. 73 Jia, W.L., McCormick, T., Tao, Y., Lu, J.‐P., and Wang, S. (2005). Inorg. Chem. 44: 5706–5712.
  74. 74 Si, Z., Li, J., Li, B., Liu, S., and Li, W. (2008). J. Lumin. 129: 181–186.
  75. 75 Yu, T., Liu, P., Chai, H., Kang, J., Zhao, Y., Zhang, H., and Fan, D. (2014). J. Fluoresc. 24: 933–943.
  76. 76 Min, J., Zhang, Q., Sun, W., Cheng, Y., and Wang, L. (2011). Dalton Trans. 40: 686–693.
  77. 77 Hsu, C.‐W., Lin, C.‐C., Chung, M.‐W., Chi, Y., Lee, G.‐H., Chou, P.‐T., Chang, C.‐H., and Chen, P.‐Y. (2011). J. Am. Chem. Soc. 133: 12085–12099.
  78. 78 Wada, A., Zhang, Q., Yasuda, T., Takasu, I., Enomoto, S., and Adachi, C. (2012). Chem. Commun. 48: 5340–5342.
  79. 79 Zhang, L., Li, B., and Su, Z. (2009). J. Phys. Chem. C 113: 13968–13973.
  80. 80 Zhang, D. (2010). J. Lumin. 130: 1419–1424.
  81. 81 Chen, X.‐L., Yu, R., Zhang, Q.‐K., Zhou, L.‐J., Wu, X.‐Y., Zhang, Q., and Lu, C.‐Z. (2013). Chem. Mater. 25: 3910–3920.
  82. 82 Chen, X.‐L., Lin, C.‐S., Wu, X.‐Y., Yu, R., Teng, T., Zhang, Q.‐K., Zhang, Q., Yang, W.‐B., and Lu, C.‐Z. (2015). J. Mater. Chem. C 3: 1187–1195.
  83. 83 Zhang, Q., Chen, X.‐L., Chen, J., Wu, X.‐Y., Yu, R., and Lu, C.‐Z. (2015). Dalton Trans. 44: 10022–10029.
  84. 84 Lundin, N.J., Walsh, P.J., Howell, S.L., McGarvey, J.J., Blackman, A.G., and Gordon, K.C. (2005). Inorg. Chem. 44: 3551–3560.
  85. 85 Lundin, N.J., Walsh, P.J., Howell, S.L., Blackman, A.G., and Gordon, K.C. (2008). Chem. Eur. J. 14: 11573–11583.
  86. 86 David, G., Walsh, P.J., and Gordon, K.C. (2004). Chem. Phys. Lett. 383: 292–296.
  87. 87 McCormick, T., Jia, W.‐L., and Wang, S. (2006). Inorg. Chem. 45: 147–155.
  88. 88 Li, E.Y.‐T., Jiang, T.‐Y., Chi, Y., and Chou, P.‐T. (2014). Phys. Chem. Chem. Phys. 16: 26184–26192.
  89. 89 Stein, G. and Wurzberg, E. (1975). J. Chem. Phys. 62: 208–213.
  90. 90 Haas, Y., Arein, G., and Wurzberg, E. (1974). J. Chem. Phys. 60: 258–263.
  91. 91 Li, Q., Xu, S.‐X., Wang, J.‐L., Xia, H.‐Y., Zhao, F., and Wang, Y.‐B. (2014). Int. J. Quantum Chem. 114: 1685–1691.
  92. 92 Igawa, S., Hashimoto, M., Kawata, I., Yashima, M., Hoshino, M., and Osawa, M. (2013). J. Mater. Chem. C 1: 542–551.
  93. 93 Osawa, M., Kawata, I., Ishii, R., Igawa, S., Hashimoto, M., and Hoshino, M. (2013). J. Mater. Chem. C 1: 4375–4383.
  94. 94 Kyle, K.R., Palke, W.E., and Ford, P.C. (1990). Coord. Chem. Rev. 97: 35–46.
  95. 95 Tsuge, K., Chishina, Y., Hashiguchi, H., Sasaki, Y., Kato, M., Ishizaka, S., and Kitamura, N. (2016). Coord. Chem. Rev. 306: 566–614.
  96. 96 Tsuboyama, A., Kuge, K., Furugori, M., Okada, S., Hoshino, M., and Ueno, K. (2007). Inorg. Chem. 46: 1992–2001.
  97. 97 Hashimoto, M., Igawa, S., Yashima, M., Kawata, I., Hoshino, M., and Osawa, M. (2011). J. Am. Chem. Soc. 133: 10348–10351.
  98. 98 Osawa, M., Hoshino, M., Hashimoto, M., Kawata, I., Igawa, S., and Yashima, M. (2015). Dalton Trans. 44: 8369–8378.
  99. 99 Osawa, M. (2014). Chem. Commun. 50: 1801–1803.
  100. 100 Lotito, K.J. and Peters, J.C. (2010). Chem. Commun. 46: 3690–3692.
  101. 101 Krylova, V.A., Djurovich, P.I., Whited, M.T., and Thompson, M.E. (2010). Chem. Commun. 46: 6696–6698.
  102. 102 Krylova, V.A., Djurovich, P.I., Aronson, J.W., Haiges, R., Whited, M.T., and Thompson, M.E. (2012). Organometallics 31: 7983–7993.
  103. 103 Krylova, V.A., Djurovich, P.I., Conley, B.L., Haiges, R., Whited, M.T., Williams, T.J., and Thompson, M.E. (2014). Chem. Commun. 50: 7176–7179.
  104. 104 Leitl, M.J., Krylova, V.A., Djurovich, P.I., Thompson, M.E., and Yersin, H. (2014). J. Am. Chem. Soc. 136: 16032–16038.
  105. 105 Araki, H., Tsuge, K., Sasaki, Y., Ishizaka, S., and Kitamura, N. (2005). Inorg. Chem. 44: 9667–9675.
  106. 106 Araki, H., Tsuge, K., Sasaki, Y., Ishizaka, S., and Kitamura, N. (2007). Inorg. Chem. 46: 10032–10034.
  107. 107 Barakat, K.A., Cundari, T.R., and Omary, M.A. (2003). J. Am. Chem. Soc. 125: 14228–14229.
  108. 108 Sinha, P., Wilson, A.K., and Omary, M.A. (2005). J. Am. Chem. Soc. 127: 12488–12489.
  109. 109 Tsuge, K. (2013). Chem. Lett. 42: 204–208.
  110. 110 Liu, Z., Djurovich, P.I., Whited, M.T., and Thompson, M.E. (2012). Inorg. Chem. 51: 230–236.
  111. 111 Liu, Z., Qayyum, M.F., Wu, C., Whited, M.T., Djurovich, P.I., Hodgson, K.O., Hedman, B., Solomon, E.I., and Thompson, M.E. (2011). J. Am. Chem. Soc. 133: 3700–3703.
  112. 112 Liu, Z., Qiu, J., Wei, F., Wang, J., Liu, X., Helander, M.G., Rodney, S., Wang, Z., Bian, Z., Lu, Z., Thompson, M.E., and Huang, C. (2014). Chem. Mater. 26: 2368–2373.
  113. 113 Volz, D., Baumann, T., Fluegge, H., Mydlak, M., Grab, T., Baechle, M., Barner‐Kowollik, C., and Brase, S. (2012). J. Mater. Chem. 22: 20786–20790.
  114. 114 Volz, D., Nieger, M., Friedrichs, J., Baumann, T., and Brase, S. (2013). Langmuir 29: 3034–3044.
  115. 115 Ohara, H., Kobayashi, A., and Kato, M. (2014). Dalton Trans. 43: 17317–17323.
  116. 116 Volz, D., Zink, D.M., Bocksrocker, T., Friedrichs, J., Nieger, M., Baumann, T., Lemmer, U., and Brase, S. (2013). Chem. Mater. 25: 3414–3426.
  117. 117 Zink, D.M., Volz, D., Baumann, T., Mydlak, M., Fluegge, H., Friedrichs, J., Nieger, M., and Brase, S. (2013). Chem. Mater. 25: 4471–4486.
  118. 118 Zink, D.M., Baechle, M., Baumann, T., Nieger, M., Kuehn, M., Wang, C., Klopper, W., Monkowius, U., Hofbeck, T., Yersin, H., and Braese, S. (2013). Inorg. Chem. 52: 2292–2305.
  119. 119 Zink, D.M., Baumann, T., Friedrichs, J., Nieger, M., and Braese, S. (2013). Inorg. Chem. 52: 13509–13520.
  120. 120 Volz, D., Wallesch, M., Grage, S.L., Goettlicher, J., Steininger, R., Batchelor, D., Vitova, T., Ulrich, A.S., Heske, C., Weinhardt, L., Baumann, T., and Braese, S. (2014). Inorg. Chem. 53: 7837–7847.
  121. 121 Wallesch, M., Volz, D., Zink, D.M., Schepers, U., Nieger, M., Baumann, T., and Braese, S. (2014). Chem. Eur. J. 20: 6578–6590.
  122. 122 Volz, D., Chen, Y., Wallesch, M., Liu, R., Flechon, C., Zink, D.M., Friedrichs, J., Fluegge, H., Steininger, R., Goettlicher, J., Heske, C., Weinhardt, L., Braese, S., So, F., and Baumann, T. (2015). Adv. Mater. 27: 2538–2543.
  123. 123 Deaton, J.C., Switalski, S.C., Kondakov, D.Y., Young, R.H., Pawlik, T.D., Giesen, D.J., Harkins, S.B., Miller, A.J.M., Mickenberg, S.F., and Peters, J.C. (2010). J. Am. Chem. Soc. 132: 9499–9508.
  124. 124 Hofbeck, T., Monkowius, U., and Yersin, H. (2015). J. Am. Chem. Soc. 137: 399–404.
  125. 125 Osawa, M., Kawata, I., Igawa, S., Hoshino, M., Fukunaga, T., and Hashizume, D. (2010). Chem. Eur. J. 16: 12114–12126.
  126. 126 Osawa, M. and Hoshino, M. (2008). Chem. Commun. 6384–6386.
  127. 127 Igawa, S., Hashimoto, M., Kawata, I., Hoshino, M., and Osawa, M. (2012). Inorg. Chem. 51: 5805–5813.
  128. 128 Osawa, M., Kawata, I., Igawa, S., Tsuboyama, A., Hashizume, D., and Hoshino, M. (2009). Eur. J. Inorg. Chem. 2009: 3708–3711.
  129. 129 Yersin, H., Leitl, J.M., and Czerwieniec, R. (2014). Proceedings Volume 9183, Organic Light Emitting Materials and Devices XVIII, 91830N/1‐91830N/11
  130. 130 Osawa, M., Hoshino, M., Akita, M., and Wada, T. (2005). Inorg. Chem. 44: 1157–1159.
  131. 131 Osawa, M., Hoshino, M., and Hashizume, D. (2007). Chem. Phys. Lett. 436: 89–93.
  132. 132 Osawa, M., Hoshino, M., and Hashizume, D. (2008). Dalton Trans. 2248–2252.