Are copper(I) complexes tough enough to be processed from solution?
Organic light-emitting diodes (OLEDs) suggest fascinating possibilities: they promise to solve existing problems such as large-area lighting and display manufacturing in an elegant, efficient way. In so doing, they could enable innovative products such as flexible, transparent devices and smart packaging. However, OLEDs also present two key challenges: how to substitute commonly used compounds containing very rare metals (iridium and platinum) with readily available materials; and how to simplify a laborious vacuum process using wet-processing techniques such as coating and printing.
The use of luminescent copper (Cu) complexes1,2 could address both aspects, but there is a catch. Not every copper complex is suitable for solution processing. Owing to their electronic properties, only complexes with copper having the oxidation number +1 are suitable as emitters in luminescent devices. Such complexes contain one or more positively charged copper(I) ions that are connected to several ligands (binding molecules). Depending on the charge of the ligands, the complexes may contain counterions to obtain an electronically neutral compound. Some of these compounds are not air stable and are vulnerable to oxygen and water: see Figure 1(a). Dealing with this aspect is tedious during synthesis, but it has already been resolved. Both water and air must be carefully avoided in any event to produce OLED devices. In fact, some of the reported best copper(I) OLEDs contain substances very sensitive to air.3
Similarly, some copper(I) complexes show intrinsic weaknesses. They dissociate in solution, potentially forming other compounds and can therefore not be used in OLED devices.4 The key to properly controlling the structure and properties of copper(I) emitter complexes is to understand the coordination chemistry (i.e., interaction of ligands with metal centers) behind the degradation and proceed accordingly.
For many complexes, the cleavage (splitting) of one copper-ligand bond per copper atom is easy to achieve, especially in coordinating solvents. Consequently, several different, though related, chemical species are often observed in solution.5–7 This does not necessarily mean that the complexes degrade in an irreversible way. Chemical equilibria between the species often allow for self-regeneration upon film processing, provided certain design rules are satisfied.
The first rule is to avoid thermodynamic traps (i.e., undesired by-products). The different species found in solution often have very similar energies. This causes the lattice energy (a measure of bond strength) to strongly favor the formation of thermodynamic products that precipitate irreversibly out of solution. The thermodynamically preferred situation is thus ruled by solubility. When dealing with copper(I) complexes, species and intermediates that are insoluble have a high crystallization tendency in solution. For example, dissolving the complex shown in Figure 1(b) gives rise to an equilibrium between the ‘homoleptic’ compounds with only copper(I) bromide (CuBr) and one type of ligand and the desired, ‘heteroleptic’ complex with two different types of ligands. On standing, the insoluble complex [CuBr(PPh3)]4—tetrakis-[bromotriphenylphosphinecopper(I)]—precipitates, while the more soluble species [CuBr(dimethylpyridine)]nremain in solution (‘n’ denotes a repeating unit that ranges from 1 to infinity).
The formation of thermodynamic products is also associated with degradation of the (kinetic) product, which is of interest to the materials scientist. Consequently, a second rule is to strengthen the kinetic favoritism of the emitter. Such kinetic control promotes formation of the desired molecule during development of the amorphous thin film. Because the time scale from the evaporation of the solvent to the formation of the solid film is always short during wet processing, and by definition kinetic products form fast, thermodynamic products can often be inhibited.
Several strategies have been applied by us and others to exploit this kinetic control (see Figure 2). One popular approach is to use chelate (i.e., multidentate or multiply binding) ligands to reduce the number of ligands. When charged ligands are employed, heteroleptic complexes are strongly favored and homoleptic species are destabilized.8 Bridging chelate ligands often lead to strongly favored motifs. An example is the PyrPHOS (pyridyl diphenylphosphine) system, a NˆP ligand (the ‘hat’ indicates the bridge) that leads to corresponding complexes of the type Cu2I2(NˆP)(P)2, which includes two equivalents of copper halide, two equivalents of phosphine ligands, and one equivalent of bridging PyrPHOS. The synthetic procedure is straightforward and can easily be scaled up. The properties of this system are remarkable. We showed how emission can be tuned from 481 to 713nm by exchanging the N-heterocycles1 (ring structures with different types of atoms) and how these specially modified dinuclear complexes are able to automatically crosslink the emission layer for higher stability.9 In addition, the solubility can be fine-tuned in a straightforward way.10
Fabricating cost-efficient OLEDs depends specifically on the materials (their abundance) used and the processes needed for production. Brightly luminescent copper(I) complexes are both suitable for solution processing and, for a metal, rather abundant, which make them ideal candidates. By following the basic design rules described above, our group was able to establish several complex families with high stability that meet these demands. Using these dinuclear copper(I) halide complexes, it is possible to cover the entire color spectrum, tune the materials' polarity and solubility from cyclohexane to ethanol, and apply a new crosslinking technique for multilayer architectures. Copper(I) halide complexes have been successfully employed in early-stage OLED devices (see Figure 3). Our next step will be to optimize our OLED stack to maximize both efficiency and device lifetime.
We acknowledge the Deutsche Forschungsgemeinschaft (project B2 of SFB/TRR 88) and the German Federal Ministry of Education and Research through the funding program cyCESH.
Karlsruhe Institute of Technology