Copper(I) Complexes for Thermally Activated Delayed Fluorescence: From Photophysical to Device Properties

  • Markus J. Leitl
  • Daniel M. Zink
  • Alexander Schinabeck
  • Thomas Baumann
  • Daniel Volz
  • Hartmut Yersin
Part of the following topical collections:
  1. Photoluminescent Materials and Electroluminescent Devices


Molecules that exhibit thermally activated delayed fluorescence (TADF) represent a very promising emitter class for application in electroluminescent devices since all electrically generated excitons can be transferred into light according to the singlet harvesting mechanism. Cu(I) compounds are an important class of TADF emitters. In this contribution, we want to give a deeper insight into the photophysical properties of this material class and demonstrate how the emission properties depend on molecular and host rigidity. Moreover, we show that with molecular optimization a significant improvement of selected emission properties can be achieved. From the discussed materials, we select one specific dinuclear complex, for which the two Cu(I) centers are four-fold bridged to fabricate an organic light emitting diode (OLED). This device shows the highest efficiency (of 23 % external quantum efficiency) reported so far for OLEDs based on Cu(I) emitters.


Thermally activated delayed fluorescence TADF Phosphorescence Fluorescence OLED Emitter Triplet harvesting Singlet harvesting Emission properties Electroluminescence Cu(I) Copper 

1 Introduction

Organic light emitting diodes (OLEDs) are capable of converting electrical energy into light. This property may be used to substitute established lighting technologies, such as light bulbs, energy-saving lamps, or fluorescent tubes. The possibility of making thin, flexible, and large-sized OLEDs inspired scientists and engineers to create entirely new applications: semi-transparent or bendable displays and light emitting labels for packaging already exist as prototypes. However, to turn these prototypes into products, new materials are required. Especially regarding light emitting materials, there is much room for improvement in terms of efficiency, realization of deep-blue emitting devices, sustainability, and cost efficiency.

In this chapter, we introduce organo-metallic emitter materials that potentially satisfy all these requirements. These materials are based on Cu(I) complexes that exhibit certain key properties allowing to harvest all excitons, singlets and triplets, that are generated in the emission layer for the generation of light. Hereby, the underlying photophysical process is the singlet harvesting effect [1, 2, 3] that is based on the molecular mechanism of thermally activated delayed fluorescence (TADF) [4].

1.1 Exciton Harvesting Mechanisms and Historical Developments

The properties of electroluminescent devices depend essentially on the chemical and physical nature of the emitter material. In this section, we briefly introduce three emission mechanisms that are currently being used in OLEDs: fluorescence, phosphorescence, and thermally activated delayed fluorescence (TADF) [1, 2].

1.1.1 Fluorescent Emitters

The initial launch of OLEDs was realized by Van Slyke and Tang in 1987. Their OLED prototype had a quantum efficiency of 1 %, a brightness of 1,000 cd m−2, and a turn-on voltage of less than 10 V [5, 6]. In this device, tris(8-hydroxyquinolinato)aluminum or short Alq3, a fluorescent metal complex, was used as emitter material. In the following years, many other fluorescent molecules have been applied as emitters in OLEDs [7, 8, 9, 10]. Some examples for such purely fluorescent materials are given in (Fig. 1 top). Soon after the realization of OLEDs with small molecules1, Burroughes et al. published work on the first OLED with a polymeric emitter in 1990, using poly-phenylene-vinylene (PPV) [11]; other conjugated polymers have also been successfully applied as emitters (Fig. 1 bottom) [12, 13, 14, 15, 16, 17].
Fig. 1

Examples of small molecule (top) and polymer (bottom) fluorescent OLED materials [5, 6, 7, 10, 11, 14, 15, 16, 17]

However, it was soon realized that only a limited efficiency is achievable with these conventional fluorescent materials because of spin statistics: There are two different types of excitons formed during OLED operation, singlet and triplet excitons, which are generally formed in a 1:3 ratio [18, 19, 20]. As demonstrated in Fig. 2, fluorescent materials can only use singlet excitons for the generation of light. Hereby, the emission originates from the spin-allowed S1 → S0 transition. Triplet excitons cannot be harvested, as the T1 → S0 transition is strongly spin-forbidden in this material class. Consequently, 75 % of all excitons are lost for the generation of light and transferred into heat [19]. Nevertheless, fluorescent materials are still used today in electroluminescent devices, especially, to achieve stable deep blue-emitting OLEDs due to the lack of satisfying alternatives [21].
Fig. 2

Exciton harvesting mechanism for conventional fluorescent materials. Only singlet excitons (25 %) can be used for the generation of light, as all triplet excitons (75 %) are lost and transferred into heat

1.1.2 Phosphorescent Emitters

To overcome the issue of losing 75 % of the excitons, organo-metallic compounds containing heavy metals such as iridium and platinum were investigated as emitter materials. These so-called phosphorescent materials or triplet emitters exhibit strong spin–orbit coupling (SOC) with respect to the lowest excited states. This changes the emission and exciton harvesting mechanism: First, SOC leads to a fast relaxation from the populated lowest excited singlet to the triplet state by means of fast intersystem crossing (ISC). The ISC time for these emitters is less than about 100 fs [22, 23]. Second, the T1 → S0 transition that is normally spin-forbidden for purely organic molecules becomes by orders of magnitude more allowed [1] and thus, can efficiently generate photons. Hence, both singlet and triplet excitons can be used for the emission of light (compare Fig. 3). Since all excitons are harvested in the triplet state, this mechanism is called the triplet harvesting effect [1, 2, 19, 24]. As a consequence, phosphorescent emitters are able to reach an internal quantum efficiency of up to 100 % [25].
Fig. 3

Schematic diagram to illustrate the triplet harvesting mechanism. All excitons, singlets and triplets, can be used for the generation of light. Prior to emission, all excitations are harvested in the first excited triplet state T1

The use of phosphorescent organo-metallic emitters represented the birth of modern OLED emitters in 1998 [24, 26, 27]. The first working examples used the platinum complex PtOEP [24] (Fig. 4). Nowadays, the most efficient phosphorescent materials are based on Ir(III) and also Pt(II) complexes. Several selected examples are displayed in Fig. 4, such as FIrpic (sky-blue light emitter), Ir(ppy)3 (green) [1, 21, 23, 25, 26, 28, 29], Ir(dm-2-piq)2(acac) (red) [1], I(piq)3 (red) [30, 31] Pt(O^N^C^N) (green) [32], and PtON7-dtb (deep blue) [33, 34] and related complexes [35, 36].
Fig. 4

Examples of phosphorescent platinum and iridium OLED materials [1, 24, 27, 30, 31, 32, 33, 37, 38, 39, 40]

1.1.3 TADF Emitters

The most recent developments in this field represent emitters that show thermally activated delayed fluorescence (TADF). Unlike conventional fluorescent or phosphorescent materials, TADF-materials are designed to exhibit a very small energy splitting ΔE(S1 − T1) between the first excited singlet and triplet state. Because of this, up-intersystem crossing (up-ISC) or reverse ISC (RISC) from T1 to S1 is possible. The reverse ISC processes are thermally activated at ambient temperatures, which means that ΔE(S1 − T1) should not be much larger than around 1,000 cm−1 (≈120 meV) so that the thermal energy k B T of ≈ 210 cm−1 at ambient temperature is still sufficient to distinctly thermally populate the S1 from the T1 state. Since the spin-allowed S1 → S0 transition from the thermally activated singlet state to the singlet ground state possesses much larger oscillator strength compared to the spin-forbidden T1 → S0 transition, emission occurs as delayed fluorescence from the singlet state S1 and shows a significantly shorter (radiative) emission decay time than the triplet state. Unlike phosphorescent materials, which harvest both exciton types, singlets and triplets, in the triplet state, TADF emitters harvest the excitation in the singlet state and hence, are often referred to as singlet harvesting materials [1, 2, 3].

Two important material classes were identified that exhibit TADF. One class is based on Cu(I) emitters. The corresponding mechanism is displayed in Fig. 5, right. For these compounds, the processes of ISC are rather fast (order of 10 ps for down-ISC [41, 42]) due to SOC induced by the Cu(I) center(s). Hence, significant prompt S1 → S0 florescence is not observed. Properties of this class of emitters are in the focus of this article (see below).
Fig. 5

Graphical illustration of the singlet harvesting mechanism based on the molecular TADF effect [1, 2, 3]. All excitons, singlets and triplets, can be used for the generation of light. The excitations are harvested in the first excited singlet state S1 prior to emission. The mechanism is displayed for purely organic (left) as well as metal-organic (right) emitter materials. Note the differences with respect to the occurrence of the prompt fluorescence. Both material classes usually exhibit phosphorescence at low temperatures

A second class of TADF materials is based on purely organic molecules. The corresponding molecular effect was already discovered in 1961 by Parker and Hatchard [4]. However, it took decades until it was realized that TADF emitters based on organic molecules can be highly interesting for application in electroluminescent devices. Mainly, this can be attributed to the fact that despite the TADF mechanism, emission decay times can be as long as several milliseconds and thus, these emitters caused problems regarding OLED efficiency at higher current densities due to roll-off effects [43]. These long decay times are dictated by the long ISC times for organic molecules [44]. However, it was recently demonstrated that these materials can be strongly optimized in this regard. Accordingly, the latest generation of organic TADF materials can exhibit suitably short decay times [45, 46, 47, 48, 49]. In addition, it was shown that with these materials highly efficient OLEDs can be realized [45, 46, 47, 48]. In Fig. 6, selected examples are displayed for illustration.
Fig. 6

Examples for purely organic TADF materials [45, 49]

2 Luminescent Cu(I) Emitters: A Brief Overview

More than 30 years ago, luminescent copper(I) complexes started to gain the interest of the scientific community. Beginning in the late 1970s and 1980s, McMillin investigated mononuclear phenanthroline complexes in various studies and, in the course of his results, proposed that the emission stems from two different excited electronic states, which is nowadays confirmed and known as TADF [50]. A large variety of luminescent copper(I) compounds has been investigated since that time and their first successful application as emitting materials in OLEDs was reported in 1999 [51, 52]. Subsequently, the number of scientific investigations using this new kind of TADF emitters is increasing rapidly and a short overview of several important structure classes together with examples with respect to OLED applications is given in this section (compare also Sect. 4). A comprehensive review concerning an overview of Cu(I) compounds can be found in the literature [53].

2.1 Mononuclear Compounds

Mononuclear complexes can be grouped into three main categories: cationic tetrahedrally coordinated (Fig. 7), neutral tetrahedrally coordinated (Fig. 8), as well as trigonally coordinated compounds (Fig. 9). With respect to an application as OLED emitters, in particular, phenanthroline, bipyridyl-, pyrazolyl- as well as tetrazolyl-based [Cu(N^N)(P^P)] complexes have been reported with external quantum efficiencies in electroluminescent devices of around 15 % [54, 55, 56, 57, 58, 59, 60].
Fig. 7

Examples of mononuclear cationic copper(I) complexes [54, 55, 56, 61, 62, 63, 64]

Fig. 8

Examples of mononuclear neutral copper(I) complexes [3, 61, 65]

Fig. 9

Examples of trigonally coordinated Cu(I) complexes [68, 69, 70, 73]

Neutral complexes with tetrahedral coordination are an important class of mononuclear compounds (Fig. 8). A steadily increasing number of copper(I) complexes, which feature an anionic ligand in order to compensate the positive charge of the copper ion instead of a counterion, bear witness to the great potential of this class of materials. Neutral complexes offer the advantage that they are lacking the counterion, which might have unexpected effects at high electric fields in the operating OLED device. A large number of anionic ligands such as borates, tetrazolates, or thiolates have been used so far to synthesize emitter complexes, along with blue-light emitters [3, 65], with outstanding photoluminescence quantum yields (PLQY) [3, 61] as well as high device efficiencies (of almost 18 %) of external quantum efficiencies (EQE) [66, 67, 68].

Finally, the class of mononuclear complexes consisting of the trigonally three-coordinated copper(I) ions should be mentioned. Here, a monocoordinating anionic ligand can be used to compensate the positive charge of the copper ion instead of a bidentate anionic ligand (Fig. 9). Halide as well as thiolate complexes have been reported to show high PLQY values of 71 % together with device EQEs of over 20 % [68]. Alternatively, the charge of the Cu(I) ion can be compensated by using a monodentate neutral carbene and a bidentate negatively charged boron ligand [69, 70, 71, 72].

2.2 Dinuclear Compounds

Dinuclear halide- or pseudo halide-bridged Cu(I) complexes form a further large group of frequently brightly emitting materials. Various compounds featuring monodentate or chelating phosphines as well as nitrogen-containing ligands are known and have been tested in OLEDs. In 2007, the first devices using iodo-bridged complexes ([Cu(μ-I)(1,2-bis[diphenylphosphino]benzene)]2) (Fig. 10) were reported [74, 75]. In most cases, such compounds were synthesized classically by solution-based reactions. However, dinuclear CuI-based complexes featuring the general formula L2Cu2I2 (L = pyridine derivative) can also be synthesized via the evaporation of CuI together with the organic ligand. This new strategy paves the way for OLED fabrication using the well-established sublimation technique, which is often not successful for compounds with high molecular weight or insufficient thermal stability. This approach has been proven to enable device construction with EQEs of up to 15.7 % [76, 77].
Fig. 10

Examples of dinuclear halide-bridged complexes with P and N ligands [74, 78, 79, 80, 81]

Further, it was shown that Cu(I) complexes with chelating ligands exhibiting N and P donors can be synthesized [78]. With this strategy, it was possible to develop blue-emitting compounds (Fig. 10 bottom), which is particularly interesting considering the lack of efficient alternatives for blue fluorescent emitters for OLEDs.

A different class of dinuclear complexes exhibits a butterfly-shaped copper-halide core. The two copper atoms are bridged by one P^N ligand. A fourth coordination site of each copper ion is represented by phosphine ligands (Fig. 11) [82, 83, 84]. A related structure has been presented in Refs. [75, 85]. By introducing additional substitutions, such as solubility enhancing or crosslinking precursor groups, the complex properties can be strongly modified [86, 87, 88, 89, 90, 91, 92, 93]. For example, the emission of these compounds can be tuned from deep blue to red by modifications of the bidentate P^N ligand [84]. The complexes show extraordinarily high photoluminescence quantum efficiencies and also very high EQE values up to 23 % when applied in OLEDs [94, 95]. Photophysical properties of these types of compounds are discussed in Sect. 3.2, while an OLED stack with this record-efficiency will be presented in Sect. 4.
Fig. 11

Examples of NHetPHOS-type Cu(I)-emitters [82, 83, 84, 88, 94]

3 Photophysical Properties—Case Studies

The vast majority of Cu(I)-compounds that have been used in electroluminescent devices are either mononuclear or dinuclear copper complexes with bridging halides and chelating P ligands [56, 60, 66, 68, 94, 95, 96, 97, 98]. For this reason, we will focus on examples of these two families in the following sections by presenting case studies. First, we will investigate the impact of specific molecular features on the excited-state properties of mononuclear Cu(I) complexes. Second, we will investigate examples of dinuclear, NHetPHOS-type emitters (compare compounds 5 and 6, below) [93], which have been used in OLED devices recently.

3.1 Mononuclear Compounds: Rigidity and Emission Properties

A structure motif often found for Cu(I) compounds is represented by complexes with pseudo-tetrahedral coordination of two bidentate ligands [1, 2, 3, 62, 63, 64, 65, 66, 96, 97, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110]. Frequently, these complexes suffer from nonradiative deactivations since strong geometry changes can occur upon excitation. In this section, we focus on these mechanisms and show that reducing the extent of such distortions can significantly enhance emission quantum yields. Two different strategies are promising in order to control and limit excited-state distortions. The first approach is based on an optimization of the matrix material (environment) to limit excited-state distortions of the embedded or doped emitter material. The second approach is based on modifying the chemical structure of the compounds so that geometry distortions upon excitation are reduced by sterical interactions of the ligands. In this latter strategy, excited-state distortions are already hindered at a molecular level.

In this case study, a series of the three compounds [Cu(dmbpy)(pop)]BF4 (1), [Cu(tmbyp)(pop)]BF4 (2), and [Cu(dmp)(phanehphos)]PF6 (3) is investigated [62, 102]. The corresponding structures are displayed in Fig. 12. Hereby, dmbpy = 4,4′-dimethyl-2,2′-bipyridine, tmbpy = 4,4′,6,6′-tetramethyl-2,2′-bipyridine, pop = bis[2-(diphenylphosphino)-phenyl]ether, dmp = 2,9-dimethyl-1,10-phenanthroline, and phanephos = 4,12-bis(diphenylphosphino)-[2.2]paracyclophane.
Fig. 12

Structures of the investigated compounds 1 [62, 63], 2 [62, 63], and 3 [102]. The intrinsic rigidity of the complexes due to sterical interactions of the two ligands increases from left to right [62, 102]

3.1.1 Emission and Rigidity Effects

In this section, we will discuss and compare properties of the three compounds displayed in Fig. 12. At first, we focus on photophysical properties of [Cu(dmbpy)(pop)]BF4 (1) in different environments. In Fig. 13, emission data are displayed for the compound dissolved in ethanol (EtOH) and compared to data of the powder material. In addition, absorption spectra of [Cu(dmbpy)(pop)]BF4 as well as of the free ligands pop and bpy recorded in the same solvent are shown2.
Fig. 13

Absorption spectra of the compound [Cu(pop)(dmbyp)]BF4 (1) and of the ligands pop and bpy recorded in EtOH. Emission spectra are displayed for the compound as powder and dissolved in deoxygenated EtOH solution. The samples were excited at λ exc = 350 nm. All spectra were recorded at ambient temperature [62]

For complex 1, intense absorption bands are observed in the wavelength range between 230 and 330 nm. These bands are also present in the spectra of the free ligands, which indicates that they result from ligand-centered transitions. In contrast, complex 1 exhibits an absorption band between 330 and 450 nm, which does not occur in the spectra of the free ligands. Consequently, this band is assigned to result from metal-to-ligand charge transfer (MLCT) transitions.

This assignment is supported by density functional theory (DFT) calculations, which are frequently and successfully applied for investigating electronic states of transition metal complexes [62, 65, 67, 68, 69, 72, 81, 84, 87, 102, 111, 112, 113] and, in particular [114, 115].  The calculations for 1 and 2 show that the highest occupied molecular orbital (HOMO) is largely located at the copper center and is mainly of 3d character, whereas the lowest unoccupied molecular orbital (LUMO) is distributed on the bipyridine moiety of the complex (Fig. 14). Furthermore, time-dependent density functional theory (TDDFT) calculations reveal that the first excited singlet S1 and triplet state T1 are determined by transitions between these two frontier orbitals, which clearly underlines the 1,3MLCT character of these states.
Fig. 14

Highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) of [Cu(pop)(dmbpy)]+ (1) displayed for the optimized triplet state (T1) geometry. Calculations were performed on the B3LYP/def2-SVP level of theory. Iso-contour values were set to 0.05. Hydrogen atoms were omitted for clarity. Note the pronounced spatial shift from HOMO to LUMO. For completeness it is noted that the HOMO and LUMO contour plots are largely similar for compound 2

At ambient temperature, the powder of compound 1 exhibits relatively weak orange luminescence (λ max = 575 nm) under illumination with UV light. The corresponding emission spectrum as displayed in Fig. 13 is broad and featureless, which is in agreement with the charge transfer character of the emitting state(s). The emission quantum yield is moderate, amounting to Φ PL = 9 %. When complex 1 is dissolved in EtOH, the emission is red-shifted to λ max = 655 nm and the quantum yield decreases significantly to Φ PL <1 %.

This behavior is a consequence of the pronounced MLCT character of the emitting states. Since on excitation a significant amount of charge is transferred from the Cu(I) center to the bipyridine ligand, the copper center is partially oxidized towards Cu(II). Furthermore, Cu(II) prefers a planar coordination compared to the tetrahedral configuration of Cu(I). Thus, the charge transfer in the Cu(I) complexes is generally connected with a substantial molecular reorganization [41, 103, 116]. For the investigated compound, this reorganization is represented by a flattening distortion from the pseudo-tetrahedral ground state to a more planar excited-state configuration. This can also be seen from the DFT calculations [62]. Such distortions are driven by a lowering of the excited-state energies which represents a red shift of the emission energies. Furthermore, strong distortions from the ground to the excited-state geometry lead to a pronounced increase of non-radiative deactivations to the ground state and thus, to a lowering of the emission quantum yield. This behavior can be explained in a model in which the potential energy surfaces of the excited and ground states can come close or even intersect (compare Fig. 15). This leads to an increased overlap of the vibrational wavefunctions of the excited and the ground state and therefore, enlarges the corresponding Franck–Condon factors that govern the non-radiative processes [44]. The tendency of excited-state distortions to occur is especially pronounced in non-rigid environments, such as fluid solutions. In contrast, in more rigid environments such as powders, these distortions are partly suppressed. Consequently, it is expected that the quantum yield strongly depends on the rigidity of the environment. Indeed, in a rigid environment the quantum yield is substantially higher than in a non-rigid, fluid environment (compare also Table 1 and Refs. [3, 62, 102]).
Fig. 15

Simplified illustration of the dependence of the emission energy on the degree of distortion occurring in the excited state. In the case of Cu(I) compounds with two bidentate ligands, the distortion coordinate mainly represents the dihedral angle between the two ligands [62]. Solid lines represent radiative transitions, dotted lines refer to resonant ISC and/or internal conversion processes

As discussed above, a rigidity increase of the environment provides an effective strategy to reduce excited-state distortions and, as a consequence, non-radiative deactivations. This process enhances emission quantum yields. However, frequently, molecular distortions that occur as a consequence of excitation can be more efficiently reduced by introducing sterically demanding groups at the ligands, i.e., within the molecules themselves. An impressive demonstration for this approach is illustrated by the series of the compounds 1, 2, and 3.

Compound 1 can be modified by adding two methyl groups at the 6,6′ positions of the bipyridine ligand, giving compound 2 (Fig. 12). This modification should limit distortions that can occur upon excitation due to the sterical interaction of the methyl groups of the modified bipyridine and the phenyl groups of the pop ligand. Consequently, the non-radiative deactivation to the ground state should be reduced and the emission quantum yield should increase. Furthermore, as excited-state distortions result in a red-shifted emission, a less pronounced distortion is expected to induce a smaller red shift. Indeed, comparing the two compounds shows that both effects, the increase of the emission quantum yield and the blue shift of the emission, are observed experimentally. In the powder phase, the emission maximum of compound 2 lies at λ max = 555 nm (18,020 cm−1) compared to that of compound 1 with λ max = 575 nm (17,390 cm−1). More importantly, the powder of compound 2 exhibits an emission quantum yield of Φ PL = 55 % compared to Φ PL  = 9 % found for the powder of compound 1. These trends are also observed for the compounds dissolved in EtOH solution, giving quantum yields of Φ PL = 6 % for 2 and Φ PL  < 1 % for 1.

For completeness, it is noted that also the electron-donating character of the methyl groups might have an influence on the emission energy. This would result in a shift of the LUMO to higher energy and consequently also to a blue shift of the emission. However, in the absorption spectra of both compounds (not displayed) only a slight blue shift of the MLCT absorption band from compound 1 to 2 is observed. In emission, a shift of similar energy would be expected. However, the observed shift of the emission energy is significantly larger and therefore, is mainly rationalized by the rigidity effect and not by an electron-donating effect of the methyl groups.

Compound 3 represents an example in which possible distortions on excitation are even more hindered due to the interaction of the bulky phanephos and the dimethylphenantroline ligand. Consequently, the emission quantum yield is even higher than for compound 2, amounting to 80 % in the powder phase. Importantly, also in fluid dichloromethane solution the quantum yield is very high, amounting to 40 % compared to 9 % for 2 and to <1 % for 1. This strongly indicates that geometry distortions upon excitation can effectively be reduced at a molecular level. An overview of the emission parameters of the three compounds is given in Table 1.
Table 1

Emission data of the compounds Cu(pop)(dmbyp)BF4 (1), Cu(pop)(tmbyp)BF4 (2), and Cu(dmp)(phanephos)PF6 (3) measured in solution and powder, respectively












Temperature (K):









 λ max (nm)







 τ (µs)







 Φ PL (%)







 k r (s−1)


2.4 × 104


4.0 × 104

0.5 × 104

 k nr (s−1)


38 × 104


6.0 × 104

0.3 × 104



 λ max (nm)







 τ (µs)







 Φ PL (%)







 k r (s−1)


5.0 × 104

0.5 × 104

5.7 × 104

0.3 × 104

 k nr (s−1)


4.1 × 104

0.6 × 104

1.4 × 104

0.1 × 104

For compounds 1 and 2 the solution data were recorded for the complexes dissolved in ethanol (EtOH), while compound 3 was dissolved in dichloromethane (DCM). λ max represents the wavelength at the emission peak, τ is the emission decay time, and Φ PL is the photoluminescence quantum yield. The radiative rate k r and nonradiative rate k nr were calculated according to k r = Φ PL τ −1 and k nr = (1 − Φ PL ) τ −1

aStrongly deviating from a monoexponential decay behavior

Interestingly, the occurrence of geometry distortions even in the powder material has another important consequence. Typically, powder samples are not well-suited to investigate molecular emission properties, as these can be significantly masked by intermolecular effects, for example, by energy transfer to quenching impurities. However, the geometry distortions of the Cu(I) compounds are sufficient (even in the powder phase) to lower the excited-state energy to such an extent that the condition for energy transfer to adjacent non-excited molecules is not fulfilled. Therefore, the excitation may be regarded as trapped at the initially excited emitter molecule (self-trapped) [1, 2, 3, 117]. Consequently, even powder samples can be used well to study emission properties of such Cu(I) complexes. In the following, this self-trapping effect is the basis for the studies of temperature-dependent emission properties of compounds 1, 2, and 3 using powder samples.

3.1.2 Low-Temperature Phosphorescence—High-Temperature TADF

For a deeper understanding of the photophysical properties of the complexes, we investigated powder samples in the temperature range between 77 and 300 K. As the photophysical behavior of compound 1 is mainly governed by non-radiative processes, we want to focus the discussion on compounds 2 and 3.

At T = 77 K, the emission decay time of compound 2 amounts to τ(77 K) = 87 µs and is assigned to be a phosphorescence originating from the lowest excited triplet state T1. With increasing temperature to T = 300 K, the decay time decreases by a factor of about 8 to τ(300 K) = 11 µs, while the emission quantum yield stays almost constant over the whole temperature range [Φ PL(77 K) = 47 % and Φ PL(300 K) = 55 %]. This allows us to determine the radiative rates k r according to k r = Φ PL τ −1, giving k r(77 K) = 0.5 × 104 s−1 and k r(300 K) = 5.0 × 104 s−1. Thus, the radiative rate increases by a factor of ten with temperature increase. This behavior is paralleled by a blue shift of the emission from λ max(77 K) = 575 nm to λ max(300 K) = 555 nm, corresponding to an energy difference of 630 cm−1 (compare Fig. 16).
Fig. 16

Emission spectra of compounds 1 [62], 2 [62], and 3 [102] as powders at different temperatures. The blue shift of the emission maximum with increasing temperature is clearly visible

Fig. 17

Emission decay time versus temperature for powders of compounds 2 [62] and 3 [102]. The gray solid line represents a fit according to Eq. (1) and the values given in the figure represent the data as obtained by the fitting procedure. The decay time at a given temperature is monoexponential in the entire temperature range

Both effects, the increase of the radiative rate and the blue shift of the emission upon heating can be rationalized by a simple TADF model (Fig. 18, see below). At low temperature, only emission from the lowest excited triplet state T1 is observed. With increasing temperature, a thermal population of the energetically higher-lying first excited singlet state S1 occurs. As the spin-allowed S1 → S0 transition carries significantly larger allowedness than the spin-forbidden T1 → S0 transition, population of the S1 state results in a drastic reduction of the decay time. Also, the blue-shifted emission at higher temperatures can be explained by this model, as the energy of the S1 state is higher than that of the T1 state. Such a mechanism corresponds to a thermally activated delayed fluorescence (TADF). The energy separation ΔE(S1 − T1) between the first excited singlet and triplet state can roughly be estimated from the shift of the emission spectra with temperature increase3 from T = 77 to 300 K amounting to ΔE(S1 − T1) = 630 cm−1 for compound 2.
Fig. 18

Energy level diagram for compounds 2 and 3. The energy splittings ΔE(S1 − T1) and the phosphorescence decay time τ(T1) were obtained with the data from Fig. 17 by applying Eq. (1). The TADF decay time at ambient temperature was determined according to \(\tau \left( {{\text{T}}_{ 1} } \right)^{ - 1} +\, \tau \left( {\text{TADF}} \right)^{ - 1} = \, \tau \left( {300{\text{ K}}} \right)^{ - 1}\). Hereby, τ(300 K) = τ(phos. + TADF) and τ(T1) = τ(phos.). Note that for compounds 2 and 3, τ (phos) >> τ(TADF) and therefore, τ(phos. + TADF) ≈ τ(TADF). A prompt fluorescence was not observed

A more accurate approach for the determination of the emission parameters is represented by an investigation of the temperature dependence of the emission decay time. The obtained data can be fitted by a modified Boltzmann relation according to the following equation: [1, 2, 118].
$$\tau \left( T \right) = \frac{{3 + \exp \left[ { - \frac{{\Delta E({\text{S}}_{ 1} - {\text{T}}_{ 1} )}}{{k_{\text{B}} T}}} \right]}}{{3 \tau ({\text{T}}_{ 1} )^{ - 1} + \tau ({\text{S}}_{ 1} )^{ - 1} \exp \left[ { - \frac{{\Delta E({\text{S}}_{ 1} - {\text{T}}_{ 1} )}}{{k_{\text{B}} T}}} \right]}}$$

In this equation, ΔE(S1 − T1) represents the energy separation between the first excited singlet S1 and triplet T1 state, τ(S1) and τ(T1) the intrinsic emission decay times of the individual states, and k B the Boltzmann constant. If Eq. (1) is used to fit the data points displayed in Fig. 17, these molecular parameters can be determined. From the fitting procedure, a value of ΔE(S1 − T1) = 720 cm−1 is obtained, which is in good agreement with the value resulting from the spectral shift (630 cm−1). For the decay time of the first excited singlet state, a value of τ(S1) = 160 ns was found. Such a value is in agreement with the singlet nature of this state, however, being connected with a rather low oscillator strength of the S1 → S0 transition. An emission originating as a prompt fluorescence was not found. This is due to the competing and significantly faster ISC process from the S1 to the T1 state, which has been reported for other Cu(I) complexes to be on the order of 10 ps [116, 119, 120]. It is noted that Eq. (1) can only be applied if the states participating in the emission process are in a thermal equilibrium [121, 122]. In the temperature range between 77 and 300 K this condition is fulfilled, which is indicated by the strictly mono-exponential decay behavior found for the entire temperature range.

Similar investigations have been performed for compound 3 [102]. It was found from the fitting procedure that the singlet–triplet energy splitting amounts to ΔE(S1 − T1) = 1,000 cm−1, which is in good agreement with the spectral shift occurring on heating from T = 77 to 300 K of 1,070 cm−1. Furthermore, the fitting procedure gives values for the intrinsic decay times of the S1 and T1 states amounting to τ(S1) = 40 ns and τ(T1) = 240 µs, respectively. Interestingly, an alternative approach to determine τ(S1) from absorption spectra based on the Strickler–Berg relationship gives a value of τ(S1) ≈ 80 ns. Taking into account the fundamentally different nature of these two independent methods and the connected errors, both values are in good agreement [102].

It is noted that the determination of the singlet–triplet energy splitting via the indirect method based on the measurement and fitting of the temperature dependence of the emission decay is not suitable for compound 1, as in this case the excited state deactivation is mainly governed by non-radiative processes, which are strongly temperature-dependent. However, the energy separation between the S1 and T1 state can be estimated from the spectral shift to ΔE(S1 − T1) = 580 cm−1 for the powder material (compare Fig. 16). This value is similar to the one found for compound 2. As complex 1 and 2 exhibit similar structures differing only by two methyl groups, similar shifts are expected to occur. Although these groups have a strong impact on the emission quantum yields, they do not seem to strongly alter the electronic structures of the compounds.
Table 2

Properties of the excited S1 (1MLCT) and T1 (3MLCT) states for powders of compounds 1, 2, and 3














 ΔE(S1 − T1) (cm−1)






 ΔE(S1 − T1) (cm−1)




 τ(S1) (ns)




 τ(T1) (µs)




The energy splitting ΔE(S1 − T1) was determined by the spectral shifts occurring between 77 and 300 K and by an indirect method, based on the measurement and fitting of the temperature-dependent emission decay time. In addition, the intrinsic decay times τ(S1) and τ(T1) are given

For all investigated compounds, the energy splitting ΔE(S1 − T1) between the first excited singlet and triplet state is very small compared, for example, to conventional purely organic molecules, for which the singlet–triplet energy splitting is frequently of the order of many 103 cm−1 [44]. Also, for transition metal complexes, e.g., Pt(II) complexes, larger values have been reported [22, 123]. The rather small singlet–triplet energy splitting found for the investigated Cu(I) complexes is rationalized by the pronounced charge transfer character of the T1 (3MLCT) and S1 (1MLCT) states. For the investigated compounds, the HOMO is mainly located on the copper center, whereas the LUMO is mainly distributed over one of the ligands. Consequently, the spatial overlap of these frontier orbitals is small. This results in a small exchange interaction of the two involved electrons and thus, in a small ΔE(S1 − T1) singlet–triplet splitting. Interestingly, the small spatial overlap of HOMO and LUMO results also in a smaller oscillator strength of the S1 → S0 transition. Thus, it is expected that with decreasing ΔE(S1 − T1), the singlet decay time τ(S1) increases. Indeed, this trend is displayed when comparing compounds 2 (τ(S1) = 160 ns, ΔE(S1 − T1) = 720 cm−1) and 3 (τ(S1) = 40 ns, ΔE(S1 − T1) = 1,000 cm−1). (see also [3, 65]).

The results are summarized in Table 2 for the compounds investigated in this section. Furthermore, in Fig. 16, for compounds 2 and 3 energy level diagrams visualize the singlet-triplet splittings and the decay times of the phosphorescence τ(phos), the thermally activated delayed fluorescence τ(TADF), and give the decay times (at ambient temperature) for the combined emission τ(phos + TADF). Compare also Ref. [69].

3.2 Dinuclear Compounds: Stability and Photophysical Properties

Cu(I) complexes with two metal centers have also gained high attention recently [111]. Frequently, these complexes exhibit a structure in which the two Cu(I) centers are bridged by two halides [74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84]. However, these bridges are often unstable, especially in fluid environments [78, 124, 125]. On the other hand, bridging the two copper centers with an additional bidentate ligand can significantly enhance the stability. A relatively stable structure type has been proposed in Refs. [82, 83, 84]. Furthermore, in this complex class, the HOMO is often not just localized on one copper center but is delocalized over the entire copper halide core. On excitation, an oxidation of the entire core and not only of one copper center is occurring, which helps to limit excited state distortions [94]. Two representatives of this class and a typical Cu(I) dimer without an additional bridge are discussed in this section, in particular with regard to their photophysical properties. Moreover, it will be demonstrated in a case study, presented in Sect. 4, that one of the compounds represents an excellent emitter for highly efficient OLEDs.

3.2.1 Photophysical Introduction

The three compounds [Cu(µ-I)(PNMe2)]2 (4), Cu2I2(MePyrPHOS)(PPh3)2 (5) and Cu2I2(MePyrPHOS)(dpph) (6) (MePyrPHOS = 2-Diphenylphosphino-4-methyl-pyridin, dpph = 1,6-Bis diphenylphosphino hexan) represent dinuclear Cu(I) complexes in which the two copper centers are bridged by two iodine ions. In compounds 5 and 6, the copper centers are further bridged by a P^N ligand and in the case of compound 6 by an additional alkyl bridge representing a fourth bridge (Fig. 19). Accordingly, within the series of the three complexes, the stability is enhanced from compound 4 to 6.
Fig. 19

Increasing stabilization by introduction of bridges. Structures of compounds 4 [78], 5 [82, 83, 91], and 6 [94]

At ambient temperature, the powder of compound 4 exhibits a bright blue emission with a maximum at λ max = 464 nm, an emission quantum yield of Φ PL = 65 %, and an emission decay time of τ = 4.6 µs. For compound 5, a bright green emission is seen with a maximum at λ max = 511 nm, an emission quantum yield being close to Φ PL = 100 %, and a decay time of τ = 5.0 µs is found, while compound 6 exhibits an emission maximum at λ max = 519 nm and a quantum yield of Φ PL = 88 %. However, the emission decay time is significantly longer amounting to τ = 24 µs (Table 3).
Table 3

Emission data of powder of [Cu(µ-I)(PNMe2)]2 (4), Cu2I2(MePyrPHOS)(PPh3)2 (5) and Cu2I2(MePyrPHOS)(dpph) (6)












This work

This work

Temperature (K):









 λ max (nm)







 τ (µs)







 Φ PL (%)







 k r (s−1)

14 × 104

0.4 × 104

19 × 104

5 × 104

3.7 × 104

0.7 × 104

 k nr (s−1)

7.6 × 104


0.6 × 104


0.5 × 104

0.2 × 104

λ max is the wavelength maximum of the emission, τ the emission decay time, and Φ PL is the photoluminescence quantum yield. The radiative rates k r and nonradiative rates k nr were calculated according to k r = Φ PL τ −1 and k nr = (1 − Φ PL ) τ −1, respectively

In the following we want to focus on the brightly emitting compounds 5 and 6. A detailed discussion of the properties of compound 4 can be found in Ref [78].

To gain a preliminary insight into the electronic structures of the compounds, DFT and TDDFT calculations were performed on the optimized T1 state geometry. The calculations reveal that for compound 5, the HOMO is located at both copper centers to similar amounts (Cu(1) 27 %, Cu(2) 14 %). Furthermore, a significant contribution to the HOMO is also located at the bridging iodines (I(1) 18 %, I(2) 16 %). In contrast, the LUMO is localized at the pyridine moiety of the bridging P^N ligand (Fig. 20). The first excited singlet S1 and triplet T1 states result from transitions between these frontier orbitals. Thus, we can assign these states as being of (metal + halide)-to-ligand charge-transfer [(M + X)LCT] character (compare also Ref. [84]).
Fig. 20

Highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) of Cu2I2(MePyrPHOS)(PPh3)2 (5) and Cu2I2(MePyrPHOS)(dpph) (6) displayed for the optimized triplet state (T1) geometry. Calculations were performed on the B3LYP/def2-SVP level of theory. Iso-contour values were set to 0.05. Hydrogen atoms were omitted for clarity. Note that in the case of compound 6 essentially only one Cu(I) center is involved in the HOMO, whereas for compound 5 both copper centers contribute with largely similar amounts

For compound 6, the situation is slightly different. Although, the iodine contribution to the HOMO is similar [I(1) 22 %, I(2) 16 %], the contribution of the copper centers is different compared to 5, as a significant amount to this molecular orbital is contributed by only one copper center [Cu(1) 39 %, Cu(2) 3 %]. Again, the LUMO is located at the pyridine moiety of the P^N ligand. Also here, the S1 and T1 states result from HOMO–LUMO transitions, which again allows assigning these states as being of (M + X)LCT character.

TDDFT calculations based on the T1 state geometry give singlet–triplet splittings for compound 5 of ΔE(S1 − T1) = 510 cm−1 and for compound 6 of 640 cm−1. These values may be regarded as an orientation for the amount of splitting, while the experimentally determined values are presented in the next section.

3.2.2 Thermally Activated Delayed Fluorescence

Frequently, emission spectra measured at different temperatures might already give a first indication of an occurrence of TADF due to a blue shift of the spectra with temperature increase. Such spectra are displayed in Fig. 21. Indeed, for compounds 5 and 6, the blue shift is obvious. On the other hand, a determination of ΔE(S1 − T1) based only on this spectral shift might be misleading as the shift can easily be masked by temperature-induced broadening effects, in particular, for small singlet–triplet gaps. Therefore, a better access to this gap (and other important photophysical parameters) is gained from the dependence of the emission decay time on temperature (compare Sect. 3.1.2). The results for powders of compounds 5 and 6 are displayed in Fig. 22. For compound 6, it is found that the energy splitting amounts to ΔE(S1 − T1) = 830 cm−1 (which is somewhat larger than predicted by TDDFT calculations), whereas for compound 5 it is with ΔE(S1 − T1) = 270 cm−1, distinctly smaller than predicted by the calculations. These deviations might be related to short-comings of the TDDFT approach with regard to the description of charge-transfer-state energies and/or to the fact that the calculations were performed for the gas phase but not for the relatively rigid crystalline environment, for which a different molecular geometry might exist compared to the one assumed for the model calculations.
Fig. 21

Emission spectra of compounds [Cu(µ-I)(PNMe2)]2 (4) [78], Cu2I2(MePyrPHOS)(PPh3)2 (5) [82, 83, 91], and Cu2I2(MePyrPHOS)(dpph) (6) [94] as powders at different temperatures. The blue shift of the emission maximum with increasing temperature is clearly visible

Fig. 22

Emission decay time versus temperature for powders of compounds 5 and 6. The gray solid line represents a fit according to Eq. (1). Fit data are also displayed. Note the different time scales for the decay time axes

Fig. 23

Energy level diagrams for compounds 5 and 6. The energy splittings ΔE(S1 − T1) and the phosphorescence decay time τ(T1) were obtained with the data from Fig. 22 by applying Eq. (1). The TADF decay time at ambient temperature was determined by applying \(\tau \left( {{\text{T}}_{ 1} } \right)^{ - 1} + \, \tau \left( {\text{TADF}} \right)^{ - 1} = \, \tau \left( { 3 0 0 {\text{ K}}} \right)^{ - 1}\). Hereby, τ(300 K) = τ(phos. + TADF) and τ(T1) = τ(phos.). Note that in contrast to the situation shown in Fig. 18, phosphorescence and TADF decay times, especially for compound 5, are not very different. Consequently, the phosphorescence channel also contributes to the emission at ambient temperature [69, 126]. A prompt fluorescence was not observed

Fig. 24

Chemical structures of selected Cu(I) emitters for highly efficient OLEDs [66, 68, 97, 111, 137]

Table 4

Properties of the excited S1 (1(M + X)LCT) and T1 (3(M + X)LCT) states

Compound (powder):









This work

This work



 ΔE(S1 − T1) (cm−1)






 ΔE(S1 − T1) (cm−1)




 τ(S1) (ns)




The energy splitting ΔE(S1 − T1) was determined from the spectral shift and from the temperature dependence of the emission decay time, respectively.  In addition, the intrinsic decay times τ(S1) and τ(T1) of the S1 and T1 state, respectively, are given

The singlet–triplet energy splitting ΔE(S1 − T1) for compound 5, amounting to ΔE(S1 − T1) = 270 cm−1, is remarkably small and to the best of our knowledge represents the smallest value that has been reported so far. Consequently, the thermal population of the singlet state S1 from the T1 state reservoir should be highly effective at ambient temperature and result in a very pronounced shortening of the TADF decay time.

The resulting energy level diagrams and decay times for compounds 5 and 6 are summarized in Fig. 23.

However, at ambient temperature, the emission decay time for a TADF system is not solely determined by the singlet–triplet gap but also by the intrinsic decay τ(S1) of the first excited singlet state S1 [compare Eq. (1)]. For compounds 5 and 6, these decay times amount to τ(S1) = 570 ns (5) and τ(S1) = 190 ns (6), respectively (Table 4). In particular, the τ(S1) value for compound 5 is unusually long for a spin-allowed transition. This value is of similar size as that found for spin-forbidden T1 → S0 transitions for Ir(III) complexes. For example, Ir(ppy)3 (ppy = phenylpyridine) exhibits an ambient-temperature triplet decay time of τ(T1) = 1.4 µs [1, 28]. Moreover, the “spin-forbidden” transition from the triplet sublevel III to the S0 ground state exhibits an even shorter decay time of τ(III) = 200 ns. Accordingly, it is by a factor of almost three shorter than that of the spin-allowed S1 ↔ S0 transition of compound 5.

The occurrence of a long S1 decay time for compounds with a very small ΔE(S1 − T1) value is, however, not unexpected. As already discussed in the previous section, a small spatial overlap of HOMO and LUMO does not only give a small singlet–triplet splitting, but also a low oscillator strength for the S1 ↔ S0 transition. Consequently, the costs for a small singlet–triplet gap are paid by a long singlet decay time—as displayed by our experimental results.

Furthermore, such long intrinsic singlet decay times explain also why no prompt fluorescence is observed. Compared to singlet decay times of several 100 ns, the S1 → T1 ISC time of 10 ps represents a much faster and therefore dominating decay route [116, 119, 120]. Consequently, at low temperature, the S1 state is dominantly depopulated via ISC before prompt fluorescence can occur. However, at higher temperatures up-ISC processes are very fast and therefore, a favorite situation is given for an effective thermal activation (up-ISC) of the S1 state from the long-lived T1 reservoir.

4 Electroluminescence with Cu(I) Compounds

An OLED device consists of a number of layers, such as anode, hole-injection layer, hole transporting layer, emission layer, electron transporting layer, electron injection layer, and cathode [21, 127, 128]. An example is discussed below. To obtain an optimized electron and hole transport and charge carrier recombination, an adjustment of the redox potentials of the respective layers and of the emitter is required. Here, we want to restrict ourselves only to considerations with respect to the emission layer that consists of the host material and the Cu(I) complex emitter.

4.1 Introductory Remarks

In most cases, the HOMOs of Cu(I)-compounds are either located on a copper-centered orbital or—in the case of dinuclear complexes—on a Cu2X2-localized orbital. For this reason, most Cu(I)-emitters have HOMO energies that lie in the range between −5.0 and −5.4 eV [88, 94]. The LUMO energies are somewhat more spread among the different Cu(I)-compounds, because these values are dictated by the respective ligands [84]. The determination is often difficult, because many Cu(I)-emitters cannot be measured with standard methods of, for example, cyclic voltammetry [129], and due to the fact that indirect determination of the LUMO energy based on a known HOMO energy and the optical transition energy may be faulty, especially, in the case of Cu(I)-emitters that frequently exhibit very broad emission spectra and large Stokes shifts. This latter issue has been discussed above with respect to the “self-trapping” behavior [1, 2, 3, 117]. As a rule of thumb, the LUMO energy of many Cu(I)-emitters lies between −3.0 and −2.2 eV, at least for dinuclear compounds [88].

The “bandgap” ΔE(S1 − S0) of Cu(I) complexes is often large. For instance, even for green-emitting Cu(I) emitters, the singlet energies are relatively high as displayed by the very broad TADF emission spectra. These energies can be estimated from the high-energy flanks of the emission spectra and it is found that, for green emitters, they lie near 2.8 eV (compare the spectra reproduced in Fig. 21 and in Refs. [3, 58, 69, 78, 102, 126]). The information given above is important, for example, for the selection of an adequate host material. In particular, its triplet energy has to be higher than the singlet energy of the TADF emitter. This is (coarsely) fulfilled, for example, for the host material PYD2 (E(T1) = 2.93 eV). The chemical structure of this molecule is displayed in Fig. 25 (see below) [58]. Furthermore, it is required that the HOMO and LUMO energies of the host fit to the corresponding values of the Cu(I) based emitters. This is also fulfilled for the PYD2 host material, exhibiting HOMO and LUMO energies at −5.9 and −2.2 eV, respectively [130].
Fig. 25

Energy level diagram for the device ITO (130 nm) // PEDOT:PSS (30 nm) // UT-314 (45 nm) // compound 6/PYD2 (30/70 [wt%]) (27 nm) // 3TYPMB (15 nm) // LiF(2 nm) // Al (100 nm) and molecular structures of the materials used [94]

Frequently, Cu(I) emitters are applied in rather high doping concentrations of much more than 10 % in the host materials. This is advantageous for the construction of efficient OLED devices due to the absence of concentration quenching, as has been proposed in Refs. [1, 3, 65, 69, 78, 102, 126, 131, 132]. On the other hand, the emitters can represent effective traps for holes in the emission layer or even significantly contribute to the charge transport [133]. This requires that the emitters are very stable against redox reactions in order to achieve a high operational stability in OLED devices. The fact that many publications showed problems concerning measurements of  reversible oxidation or reduction potentials [109, 129, 134, 135, 136] suggests that this is an open issue concerning the development of devices with long-term stability.

4.2 Literature OLED Examples

Indeed, a number of valuable investigations using Cu(I) emitters in OLEDs have been carried out. However, it is not in the scope of this contribution to present those details. Here, we will only refer to examples reported in these studies. Early stage-investigations were carried out with tetranuclear compounds [51, 52]. Subsequently, compounds as displayed in Fig. 24, for example, were applied. In particular, compound (LMe)Cu(Br) could be vacuum sublimed and reached, as already shown in 2011, a very high efficiency of 65.3 cd/A (EQE 21.3 %) [98].

4.3 OLED Case Study—Solution-Processed Device Achieving nearly 100 % Internal Quantum Efficiency

In Sect. 3.2, we presented photophysical properties of dinuclear compounds. According to these studies, compound 5 should be best-suited for an OLED application. However, compound 6 (Fig. 19) shows several advantages [94]. Presumably, due to its four-fold bridged Cu(I) centers, it exhibits a high thermal stability (no decomposition up to T = 290 °C in nitrogen atmosphere derived from thermo-gravimetric analysis (TGA)). Moreover, when doping compound 6 into the PYD2 host an emission quantum yield of 92 % is found, being slightly higher than the value registered for the powder material with Φ PL = 88 % (Table 3). Probably, this is induced by the specific thin-film morphology, an effect that has recently been found also for complexes with similar structures [88, 90]. Moreover, good film formation properties and little crystallization tendency are observed even when preparing neat thin films of compound 6.

The OLED device was fabricated after extensive optimization leading to the stack architecture, as displayed in Fig. 25 with layer thicknesses given in the figure caption. The hole injection layer (PEDOT:PSS (AI4803)) was spin-coated onto the indium tin oxide (ITO) substrate and baked at 140 °C for 30 min in air. The hole transport layer PLEXCORE [138, 139] was spin-coated onto PEDOT:PSS and annealed at 180 °C for 30 min under nitrogen for crosslinking. Then, the emitting layer was spin-coated from toluene solution on top of the hole transporting layer and annealed to dry. All other layers were thermally evaporated under vacuum [94].

Note that the emission layer exhibits a relatively high doping concentration of the emitter without showing concentration quenching, as was already proposed earlier [1, 3, 102, 132]. Both, 3TPYMB and PLEXCORE UT-314 were chosen due to their high triplet energies in order to prevent quenching by energy transfer [58].

Figure 26 displays the current efficiency versus electro-luminance of the optimized device. The turn-on voltage was 2.6 V and the maximum brightness 10,000 cd/A was achieved at 10 V. The current efficiency amounts to 71 cd/A at 100 cd/m2, while the peak current efficiency reaches even 73 cd/A (at ≈40 cd/m2) with 63 lm/W, corresponding to an external quantum efficiency of 23 %. This is the highest value reported for OLEDs so far for solution- and vacuum-processed emission layers based on Cu(I) emitters. It is remarked that no outcoupling-enhancing strategies were applied. This value is comparable to the state-of-the-art efficiency of vacuum sublimed devices based on Ir(III) emitters.
Fig. 26

Left current efficiency vs. luminance (EL). Right electroluminescence spectrum as obtained with the device according to Fig. 25

For completeness, it is remarked that the device shows a pronounced roll-off of the current efficiency with increasing luminance. This behavior can be attributed to a loss of charge balance as most of the hole transport does not occur via the host material but via the emitter itself. At higher current densities, the relatively long emission decay of the emitter amounting to 24 µs could become problematic and lead to saturation effects [43].

5 Conclusion

The development and understanding of new luminescent materials was stimulated by their potential application in electroluminescent devices. Since more than 15 years research in this respect was focused on phosphorescent Ir(III) and Pt(II) compounds [35]. Since about 5 years, Cu(I) complexes and metal-free TADF emitters became interesting, as these materials might replace high-cost rare metal compounds that are currently most-often applied in commercial OLEDs.

Both phosphorescent and TADF emitters can harvest all excitons that are generated after the electron–hole recombination process, i.e., according to the triplet harvesting [1, 2, 19, 24, 140] and the singlet harvesting mechanism [1, 2, 3], respectively. Obviously, the required progress in material development can only proceed if a deep scientific research concerning structure–property relations is carried out. Thus, in this contribution, we presented—after an introduction to the mentioned mechanisms—a detailed insight into the photophysics of Cu(I) compounds and how their properties can deliberately be modified. Accordingly, we present an emitter compound representing a multi-bridged dinuclear Cu(I) complex (compound 6) that opens the door to very efficient OLEDs. With this material, we obtained an external quantum efficiency of 23 %, which represents the highest value reported so far for Cu(I)-based OLEDs. However, although these investigations are very promising, further developments have to be carried out to optimize the device performance, especially with respect to more sophisticated architectures on the device-level and a shortening of the TADF decay time (at high emission quantum yield) on the material level. A new promising approach based on a combined TADF and triplet-state emission at high spin–orbit coupling may open the required progress [69, 126, 141].


  1. 1.

    ‘Small molecule' is a widely-used, yet ambiguous term. It is most often used to distinguish isolated molecules with a molecular weight of less than 1,000 Da from polymers.

  2. 2.

    Bipyridine (bpy) was used instead of dmbpy. However, the absorption spectrum of the methylated ligand is not expected to deviate significantly from that of bpy.

  3. 3.

    It is remarked that an estimation of ΔE (S1 − T1) from the spectra is only possible if both states, S1 and T1, result from transitions between the same molecular orbitals. For the investigated compounds, this condition is fulfilled.



The authors thank the German Ministry for Education and Research (BMBF) for funding in the scope of the cyCESH project (FKN 13N12668). The authors (T.B., D.V., D.M.Z.) gratefully acknowledge the collaboration with the groups of Prof. Franky So (NCSU), Prof. Christopher Barner-Kowollik (KIT), Prof Clemens Heske (KIT, UNLV), Prof. Uli Lemmer (KIT), and Prof. Stefan Bräse (KIT), as well as the scientific division of CYNORA.


  1. 1.
    Yersin H, Rausch AF, Czerwieniec R, Hofbeck T, Fischer T (2011) Coord Chem Rev 255:2622CrossRefGoogle Scholar
  2. 2.
    Yersin H, Rausch AF, Czerwieniec R (2012) In: Brütting W, Adachi C (eds) Physics of organic semiconductors. Wiley VCH, Weinheim, p 371Google Scholar
  3. 3.
    Czerwieniec R, Yu J, Yersin H (2011) Inorg Chem 50:8293CrossRefGoogle Scholar
  4. 4.
    Parker CA, Hatchard CG (1961) Trans Faraday Soc 57:1894CrossRefGoogle Scholar
  5. 5.
    Tang CW, VanSlyke SA (1987) Appl Phys Lett 51:913CrossRefGoogle Scholar
  6. 6.
    Tang CW, VanSlyke SA, Chen CH (1989) J Appl Phys 65:3610CrossRefGoogle Scholar
  7. 7.
    Shinar J (ed) (2003) Organic light-emitting devices: a survey. Springer, New YorkGoogle Scholar
  8. 8.
    Müllen K, Scherf U (eds) (2006) Organic light emitting devices: synthesis, properties and applications. Wiley VCH, WeinheimGoogle Scholar
  9. 9.
    He SJ, Wang ZB, Wang DK, Jiang N, Lu ZH (2013) Appl Phys Lett 103:083301CrossRefGoogle Scholar
  10. 10.
    Aziz H, Popovic ZD (2004) Chem Mater 16:4522CrossRefGoogle Scholar
  11. 11.
    Burroughes JH, Bradley DDC, Brown AR, Marks RN, Mackay K, Friend RH, Burns PL, Holmes AB (1990) Nature 347:539CrossRefGoogle Scholar
  12. 12.
    Bocksrocker T, Hoffmann J, Eschenbaum C, Pargner A, Preinfalk J, Maier-Flaig F, Lemmer U (2013) Org Electron 14:396CrossRefGoogle Scholar
  13. 13.
    Bocksrocker T, Eschenbaum C, Preinfalk JB, Hoffmann J, Asche-Tauscher J, Maier-Flaig F, Lemmer U (2012) Renew energy environ opt photonics congr [OSA technical digest (online)]. Optical Society of America, Eindhoven, p LM3A4Google Scholar
  14. 14.
    Braun D, Heeger AJ, Kroemer H (1991) J Electron Mater 20:945CrossRefGoogle Scholar
  15. 15.
    Braun D, Heeger AJ (1991) Appl Phys Lett 58:1982CrossRefGoogle Scholar
  16. 16.
    Spreitzer H, Becker H, Kluge E, Kreuder W, Schenk H, Demandt R, Schoo H (1998) Adv Mater 10:1340CrossRefGoogle Scholar
  17. 17.
    Gustafsson G, Cao Y, Treacy GM, Klavetter F, Colaneri N, Heeger AJ (1992) Nature 357:477CrossRefGoogle Scholar
  18. 18.
    Helfrich W, Schneider WG (1966) J Chem Phys 44:2902CrossRefGoogle Scholar
  19. 19.
    Yersin H (2004) Top Curr Chem 241:1CrossRefGoogle Scholar
  20. 20.
    Reufer M, Walter MJ, Lagoudakis PG, Hummel AB, Kolb JS, Roskos HG, Scherf U, Lupton JM (2005) Nat Mater 4:340CrossRefGoogle Scholar
  21. 21.
    Reineke S, Thomschke M, Lüssem B, Leo K (2013) Rev Mod Phys 85:1245CrossRefGoogle Scholar
  22. 22.
    Yersin H, Donges D (2001) Top Curr Chem 214:81CrossRefGoogle Scholar
  23. 23.
    Hedley GJ, Ruseckas A, Samuel IDW (2008) J Phys Chem A 113:2CrossRefGoogle Scholar
  24. 24.
    Baldo MA, O’Brien DF, You Y, Shoustikov A, Sibley S, Thompson ME, Forrest SR (1998) Nature 395:151CrossRefGoogle Scholar
  25. 25.
    Adachi C, Baldo MA, Thompson ME, Forrest SR (2001) J Appl Phys 90:5048CrossRefGoogle Scholar
  26. 26.
    Baldo MA, Lamansky S, Burrows PE, Thompson ME, Forrest SR (1999) Appl Phys Lett 75:4CrossRefGoogle Scholar
  27. 27.
    Baldo MA, Thompson ME, Forrest SR (2000) Nature 403:750CrossRefGoogle Scholar
  28. 28.
    Hofbeck T, Yersin H (2010) Inorg Chem 49:9290CrossRefGoogle Scholar
  29. 29.
    Yersin H, Finkenzeller WJ (2008) In: Yersin H (ed) Highly efficient OLEDs with phosphorescent materials. Wiley-VCH, Weinheim, p 1Google Scholar
  30. 30.
    Tsuzuki T, Tokito S (2007) Adv Mater 19:276CrossRefGoogle Scholar
  31. 31.
    Tsuboyama A, Iwawaki H, Furugori M, Mukaide T, Kamatani J, Igawa S, Moriyama T, Miura S, Takiguchi T, Okada S, Hoshino M, Ueno K (2003) J Am Chem Soc 125:12971CrossRefGoogle Scholar
  32. 32.
    Cheng G, Kui SCF, Ang WH, Ko M-Y, Chow PK, Kwong CL, Kwok CC, Ma C, Guan X, Low KH, Su SJ, Che CM (2014) Chem Sci 5:4819CrossRefGoogle Scholar
  33. 33.
    Fleetham T, Li G, Wen L, Li J (2014) Adv Mater 26:7116CrossRefGoogle Scholar
  34. 34.
    Li G, Fleetham T, Turner E, Hang X-C, Li J (2015) Adv Opt Mater 3:390CrossRefGoogle Scholar
  35. 35.
    Yersin H (ed) (2008) Highly efficient OLEDs with phosphorescent materials. Wiley-VCH, WeinheimGoogle Scholar
  36. 36.
    Deaton JC, Castellano FN (2016) In: Zysman-Colman E (ed) Iridium(III) in optoelectronic and photonics applications. Wiley VCH, Weinheim, Germany (in press)Google Scholar
  37. 37.
    Giebink NC, Forrest SR (2008) Phys Rev B 77:235215CrossRefGoogle Scholar
  38. 38.
    Che CM, Kwok CC, Lai SW, Rausch AF, Finkenzeller WJ, Zhu N, Yersin H (2010) Chem Eur J 16:233CrossRefGoogle Scholar
  39. 39.
    Adachi C, Kwong RC, Djurovich P, Adamovich V, Baldo MA, Thompson ME, Forrest SR (2001) Appl Phys Lett 79:2082CrossRefGoogle Scholar
  40. 40.
    Rossi E, Colombo A, Dragonetti C, Roberto D, Ugo R, Valore A, Falciola L, Brulatti P, Cocchi M, Williams JAG (2012) J Mater Chem 22:10650CrossRefGoogle Scholar
  41. 41.
    Iwamura M, Takeuchi S, Tahara T (2015) Acc Chem Res 48:782CrossRefGoogle Scholar
  42. 42.
    Iwamura M, Takeuchi S, Tahara T (2007) J Am Chem Soc 129:5248CrossRefGoogle Scholar
  43. 43.
    Murawski C, Leo K, Gather MC (2013) Adv Mater 25:6801CrossRefGoogle Scholar
  44. 44.
    Turro NJ (1978) Modern molecular chemistry. Benjamin/Cummings Publishing Company Inc., Menlo ParkGoogle Scholar
  45. 45.
    Uoyama H, Goushi K, Shizu K, Nomura H, Adachi C (2012) Nature 492:234CrossRefGoogle Scholar
  46. 46.
    Zhang QS, Li B, Huang SP, Nomura H, Tanaka H, Adachi C (2014) Nature Photonics 8:326CrossRefGoogle Scholar
  47. 47.
    Goushi K, Yoshida K, Sato K, Adachi C (2012) Nature Photonics 6:253CrossRefGoogle Scholar
  48. 48.
    Nakanotani H, Masui K, Nishide J, Shibata T, Adachi C (2013) Sci Rep 3:2127CrossRefGoogle Scholar
  49. 49.
    Yersin H, Mataranga-Popa NL, Czerwieniec R, Bolz A (2015) WO 2015/121239 A1Google Scholar
  50. 50.
    Blasse G, McMillin DR (1980) Chem Phys Lett 70:1CrossRefGoogle Scholar
  51. 51.
    Ma Y, Che CM, Chao HY, Zhou X, Chan WH, Shen J (1999) Adv Mater 11:852CrossRefGoogle Scholar
  52. 52.
    Ma YG, Chan WH, Zhou XM, Che CM (1999) New J Chem 23:263CrossRefGoogle Scholar
  53. 53.
    Dumur F (2015) Org Electron 21:27CrossRefGoogle Scholar
  54. 54.
    Zhang Q, Zhou Q, Cheng Y, Wang L, Ma D, Jing X, Wang F (2004) Adv Mater 16:432CrossRefGoogle Scholar
  55. 55.
    Keller S, Constable EC, Housecroft CE, Neuburger M, Prescimone A, Longo G, Pertegas A, Sessolo M, Bolink HJ (2014) Dalton Trans 43:16593CrossRefGoogle Scholar
  56. 56.
    Chen XL, Yu RM, Zhang QK, Zhou LJ, Wu CY, Zhang Q, Lu CZ (2013) Chem Mater 25:3910CrossRefGoogle Scholar
  57. 57.
    Che G, Su Z, Li W, Chu B, Li M, Hu Z, Zhang Z (2006) Appl Phys Lett 89:103511CrossRefGoogle Scholar
  58. 58.
    Zhang Q, Komino T, Huang S, Matsunami S, Goushi K, Adachi C (2012) Adv Funct Mater 22:2327CrossRefGoogle Scholar
  59. 59.
    Chen X-L, Yu R, Zhang Q-K, Zhou L-J, Wu X-Y, Zhang Q, Lu C-Z (2013) Chem Mater 25:3910CrossRefGoogle Scholar
  60. 60.
    Wada A, Zhang Q, Yasuda T, Takasu I, Enomoto S, Adachi C (2012) Chem Commun 48:5340CrossRefGoogle Scholar
  61. 61.
    Bergmann L, Friedrichs J, Mydlak M, Baumann T, Nieger M, Bräse S (2013) Chem Commun 49:6501CrossRefGoogle Scholar
  62. 62.
    Linfoot CL, Leitl MJ, Richardson P, Rausch AF, Chepelin O, White FJ, Yersin H, Robertson N (2014) Inorg Chem 53:10854CrossRefGoogle Scholar
  63. 63.
    Linfoot CL, Richardson P, Hewat TE, Moudam O, Forde MM, Collins A, White F, Robertson N (2010) Dalton Trans 39:8945CrossRefGoogle Scholar
  64. 64.
    Cuttell DG, Kuang SM, Fanwick PE, McMillin DR, Walton RA (2002) J Am Chem Soc 124:6CrossRefGoogle Scholar
  65. 65.
    Czerwieniec R, Yersin H (2015) Inorg Chem 54:4322CrossRefGoogle Scholar
  66. 66.
    Igawa S, Hashimoto M, Kawata I, Yashima M, Hoshino M, Osawa M (2013) J Mater Chem C 1:542CrossRefGoogle Scholar
  67. 67.
    Osawa M, Kawata I, Ishii R, Igawa S, Hashimoto M, Hoshino M (2013) J Mater Chem C 1:4375CrossRefGoogle Scholar
  68. 68.
    Osawa M, Hoshino M, Hashimoto M, Kawata I, Igawa S, Yashima M (2015) Dalton Trans 44:8369CrossRefGoogle Scholar
  69. 69.
    Leitl MJ, Krylova VA, Djurovich PI, Thompson ME, Yersin H (2014) J Am Chem Soc 136:16032CrossRefGoogle Scholar
  70. 70.
    Krylova VA, Djurovich PI, Conley BL, Haiges R, Whited MT, Williams TJ, Thompson ME (2014) Chem Commun 50:7176CrossRefGoogle Scholar
  71. 71.
    Krylova VA, Djurovich PI, Aronson JW, Haiges R, Whited MT, Thompson ME (2012) Organometallics 31:7983CrossRefGoogle Scholar
  72. 72.
    Krylova VA, Djurovich PI, Whited MT, Thompson ME (2010) Chem Commun 46:6696CrossRefGoogle Scholar
  73. 73.
    Osawa M (2014) Chem Commun 50:1801CrossRefGoogle Scholar
  74. 74.
    Tsuboyama A, Kuge K, Furugori M, Okada S, Hoshino M, Ueno K (2007) Inorg Chem 46:1992CrossRefGoogle Scholar
  75. 75.
    Araki H, Tsuge K, Sasaki Y, Ishizaka S, Kitamura N (2007) Inorg Chem 46:10032CrossRefGoogle Scholar
  76. 76.
    Liu Z, Qayyum MF, Wu C, Whited MT, Djurovich PI, Hodgson KO, Hedman B, Solomon EI, Thompson ME (2011) J Am Chem Soc 133:3700CrossRefGoogle Scholar
  77. 77.
    Liu ZW, Qiu J, Wei F, Wang JQ, Liu XC, Helander MG, Rodney S, Wang ZB, Bian ZQ, Lu ZH, Thompson ME, Huang CH (2014) Chem Mater 26:2368CrossRefGoogle Scholar
  78. 78.
    Leitl MJ, Küchle FR, Mayer HA, Wesemann L, Yersin H (2013) J Phys Chem A 117:11823CrossRefGoogle Scholar
  79. 79.
    Tsuge K, Chishina Y, Hashiguchi H, Sasaki Y, Kato M, Ishizaka S, Kitamura N (2016) Coord Chem Rev 306:636CrossRefGoogle Scholar
  80. 80.
    Tsuge K (2013) Chem Lett 42:204CrossRefGoogle Scholar
  81. 81.
    Yersin H, Leitl MJ, Czerwieniec R (2014) Proc SPIE 9183:91830NCrossRefGoogle Scholar
  82. 82.
    Yersin H, Monkowius U, Fischer T, Hofbeck T (2009) DE 10 2009 030 475 A1Google Scholar
  83. 83.
    Yersin H, Monkowius U, Fischer T, Hofbeck T, Baumann T, Grab T (2015) EP 2 408 787 B1Google Scholar
  84. 84.
    Zink DM, Bächle M, Baumann T, Nieger M, Kuhn M, Wang C, Klopper W, Monkowius U, Hofbeck T, Yersin H, Bräse S (2013) Inorg Chem 52:2292CrossRefGoogle Scholar
  85. 85.
    Musina EI, Shamsieva AV, Strelnik ID, Gerasimova TP, Krivolapov DB, Kolesnikov IE, Grachova EV, Tunik SP, Bannwarth C, Grimme S, Katsyuba SA, Karasik AA, Sinyashin OG (2016) Dalton Trans 45:2250CrossRefGoogle Scholar
  86. 86.
    Volz D, Wallesch M, Fléchon C, Danz M, Verma A, Navarro JM, Zink DM, Bräse S, Baumann T (2015) Green Chem 17:1988CrossRefGoogle Scholar
  87. 87.
    Volz D, Hirschbiel AF, Zink DM, Friedrichs J, Nieger M, Baumann T, Bräse S, Barner-Kowollik C (2014) J Mater Chem C 2:1457CrossRefGoogle Scholar
  88. 88.
    Volz D, Zink DM, Bocksrocker T, Friedrichs J, Nieger M, Baumann T, Lemmer U, Bräse S (2013) Chem Mater 25:3414CrossRefGoogle Scholar
  89. 89.
    Volz D, Baumann T, Flügge H, Mydlak M, Grab T, Bächle M, Barner-Kowollik C, Bräse S (2012) J Mater Chem 22:20786CrossRefGoogle Scholar
  90. 90.
    Volz D, Nieger M, Friedrichs J, Baumann T, Bräse S (2013) Langmuir 29:3034CrossRefGoogle Scholar
  91. 91.
    Zink DM, Volz D, Baumann T, Mydlak M, Flügge H, Friedrichs J, Nieger M, Bräse S (2013) Chem Mater 25:4471CrossRefGoogle Scholar
  92. 92.
    Zink DM, Baumann T, Friedrichs J, Nieger M, Bräse S (2013) Inorg Chem 52:13509CrossRefGoogle Scholar
  93. 93.
    Wallesch M, Volz D, Zink DM, Schepers U, Nieger M, Baumann T, Bräse S (2014) Chem Eur J 20:6578CrossRefGoogle Scholar
  94. 94.
    Volz D, Chen Y, Wallesch M, Liu R, Fléchon C, Zink DM, Friedrichs J, Flügge H, Steininger R, Göttlicher J, Heske C, Weinhardt L, Bräse S, So F, Baumann T (2015) Adv Mater 27:2538CrossRefGoogle Scholar
  95. 95.
    Wallesch M, Volz D, Fléchon C, Zink DM, Bräse S, Baumann T (2014) Proc SPIE 9183:918309CrossRefGoogle Scholar
  96. 96.
    Ni T, Liu X, Zhang T, Bao H, Zhan G, Jiang N, Wang J, Liu Z, Bian Z, Lu Z, Huang C (2015) J Mater Chem C 3:5835CrossRefGoogle Scholar
  97. 97.
    Cheng G, So GKM, To WP, Chen Y, Kwok CC, Ma CS, Guan XG, Chang X, Kwok WM, Che CM (2015) Chem Sci 6:4623CrossRefGoogle Scholar
  98. 98.
    Hashimoto M, Igawa S, Yashima M, Kawata I, Hoshino M, Osawa M (2011) J Am Chem Soc 133:10348CrossRefGoogle Scholar
  99. 99.
    Zhang Q, Ding J, Cheng Y, Wang L, Xie Z, Jing X, Wang F (2007) Adv Funct Mater 17:2983CrossRefGoogle Scholar
  100. 100.
    Moudam O, Kaeser A, Delavaux-Nicot B, Duhayon C, Holler M, Accorsi G, Armaroli N, Seguy I, Navarro J, Destruel P, Nierengarten JF (2007) Chem Commun: 3077Google Scholar
  101. 101.
    Qin L, Zhang Q, Sun W, Wang J, Lu C, Cheng Y, Wang L (2009) Dalton Trans: 9388Google Scholar
  102. 102.
    Czerwieniec R, Kowalski K, Yersin H (2013) Dalton Trans 42:9826CrossRefGoogle Scholar
  103. 103.
    Vorontsov II, Graber T, Kovalevsky AY, Novozhilova IV, Gembicky M, Chen YS, Coppens P (2009) J Am Chem Soc 131:6566CrossRefGoogle Scholar
  104. 104.
    Liu XF, Li RF, Ma LF, Feng X, Ding YQ (2016) New J Chem 40:619CrossRefGoogle Scholar
  105. 105.
    Chen XL, Lin CS, Wu XY, Yu R, Teng T, Zhang QK, Zhang Q, Yang WB, Lu CZ (2015) J Mater Chem C 3:1187CrossRefGoogle Scholar
  106. 106.
    Capano G, Rothlisberger U, Tavernelli I, Penfold TJ (2015) J Phys Chem A 119:7026CrossRefGoogle Scholar
  107. 107.
    Hsu CW, Lin CC, Chung MW, Chi Y, Lee GH, Chou PT, Chang CH, Chen PY (2011) J Am Chem Soc 133:12085CrossRefGoogle Scholar
  108. 108.
    Kaeser A, Mohankumar M, Mohanraj J, Monti F, Holler M, Cid JJ, Moudam O, Nierengarten I, Karmazin-Brelot L, Duhayon C, Delavaux-Nicot B, Armaroli N, Nierengarten JF (2013) Inorg Chem 52:12140CrossRefGoogle Scholar
  109. 109.
    Femoni C, Muzzioli S, Palazzi A, Stagni S, Zacchini S, Monti F, Accorsi G, Bolognesi M, Armaroli N, Massi M, Valenti G, Marcaccio M (2013) Dalton Trans 42:997CrossRefGoogle Scholar
  110. 110.
    Armaroli N, Accorsi G, Holler M, Moudam O, Nierengarten JF, Zhou Z, Wegh RT, Welter R (2006) Adv Mater 18:1313CrossRefGoogle Scholar
  111. 111.
    Deaton JC, Switalski SC, Kondakov DY, Young RH, Pawlik TD, Giesen DJ, Harkins SB, Miller AJM, Mickenberg SF, Peters JC (2010) J Am Chem Soc 132:9499CrossRefGoogle Scholar
  112. 112.
    Ohara H, Kobayashi A, Kato M (2014) Dalton Trans 43:17317CrossRefGoogle Scholar
  113. 113.
    Cid JJ, Mohanraj J, Mohankumar M, Holler M, Monti F, Accorsi G, Karmazin-Brelot L, Nierengarten I, Malicka JM, Cocchi M, Delavaux-Nicot B, Armaroli N, Nierengarten JF (2014) Polyhedron 82:158CrossRefGoogle Scholar
  114. 114.
    Jesser A, Rohrmüller M, Schmidt WG, Herres-Pawlis S (2014) J Comput Chem 35:1CrossRefGoogle Scholar
  115. 115.
    Niehaus TA, Hofbeck T, Yersin H (2015) RSC advancesGoogle Scholar
  116. 116.
    Hua L, Iwamura M, Takeuchi S, Tahara T (2015) Phys Chem Chem Phys 17:2067CrossRefGoogle Scholar
  117. 117.
    Gliemann G, Yersin H (1985) Clusters. Springer, Berlin, p 87CrossRefGoogle Scholar
  118. 118.
    Azumi T, O’Donnell CM, McGlynn SP (1966) J Chem Phys 45:2735CrossRefGoogle Scholar
  119. 119.
    Tschierlei S, Karnahl M, Rockstroh N, Junge H, Beller M, Lochbrunner S (2014) Chem Phys Chem 15:3709CrossRefGoogle Scholar
  120. 120.
    Iwamura M, Watanabe H, Ishii K, Takeuchi S, Tahara T (2011) J Am Chem Soc 133:7728CrossRefGoogle Scholar
  121. 121.
    Yersin H, Strasser J (2000) Coord Chem Rev 208:331CrossRefGoogle Scholar
  122. 122.
    Yersin H, Humbs W, Strasser J (1997) Top Curr Chem 191:153CrossRefGoogle Scholar
  123. 123.
    Yersin H, Donges D, Nagle JK, Sitters R, Glasbeek M (2000) Inorg Chem 39:770CrossRefGoogle Scholar
  124. 124.
    Arnby CH, Jagner S, Dance I (2004) CrystEngComm 6:257CrossRefGoogle Scholar
  125. 125.
    Araki H, Tsuge K, Sasaki Y, Ishizaka S, Kitamura N (2005) Inorg Chem 44:9667CrossRefGoogle Scholar
  126. 126.
    Hofbeck T, Monkowius U, Yersin H (2015) J Am Chem Soc 137:399CrossRefGoogle Scholar
  127. 127.
    Xiao L, Chen Z, Qu B, Luo J, Kong S, Gong Q, Kido J (2011) Adv Mater 23:926CrossRefGoogle Scholar
  128. 128.
    Brütting W, Frischeisen J, Schmidt TD, Scholz BJ, Mayr C (2013) Phys Status Solidi A 210:44CrossRefGoogle Scholar
  129. 129.
    Leiva AM, Rivera L, Loeb B (1991) Polyhedron 10:347CrossRefGoogle Scholar
  130. 130.
    Williams EL, Haavisto K, Li J, Jabbour GE (2007) Adv Mater 19:197CrossRefGoogle Scholar
  131. 131.
    Gneuß T, Leitl MJ, Finger LH, Rau N, Yersin H, Sundermeyer J (2015) Dalton Trans 44:8506CrossRefGoogle Scholar
  132. 132.
    Yersin H, Czerwieniec R, Monkowius U (2011) DE 10 2011 000282 A1Google Scholar
  133. 133.
    Wei F, Zhang T, Liu X, Li X, Jiang N, Liu Z, Bian Z, Zhao Y, Lu Z, Huang C (2014) Organ Electron 15:3292CrossRefGoogle Scholar
  134. 134.
    Franco E, López-Torres E, Mendiola M, Sevilla M (2000) Polyhedron 19:441CrossRefGoogle Scholar
  135. 135.
    Manbeck GF, Brennessel WW, Eisenberg R (2011) Inorg Chem 50:3431CrossRefGoogle Scholar
  136. 136.
    Yang L, Powell DR, Houser RP (2007) Dalton Trans: 955Google Scholar
  137. 137.
    Yersin H, Czerwieniec R, Monkowius U (2010) DE10 2010 031831Google Scholar
  138. 138.
    Liaptsis G, Hertel D, Meerholz K (2013) Angew Chem Int Ed 125:9742CrossRefGoogle Scholar
  139. 139.
    Xiang C, Chopra N, Wang J, Brown C, Ho S, Mathai M, So F (2014) Organ Electron 15:1702CrossRefGoogle Scholar
  140. 140.
    Yersin H (2004) Proc SPIE 5214:124CrossRefGoogle Scholar
  141. 141.
    Czerwieniec R, Hofbeck T, Leitl MJ, Monkowius U, Yersin H (2013) DE 10 2013 106 426 A1Google Scholar

Copyright information

© Springer International Publishing Switzerland 2016

Authors and Affiliations

  • Markus J. Leitl
    • 1
  • Daniel M. Zink
    • 2
  • Alexander Schinabeck
    • 1
  • Thomas Baumann
    • 2
  • Daniel Volz
    • 2
  • Hartmut Yersin
    • 1
  1. 1.Institut für Physikalische Chemie, Universität RegensburgRegensburgGermany
  2. 2.Cynora GmbHBruchsalGermany

Personalised recommendations