Homepage of Hartmut Yersin

Universität Regensburg


Triplet emitters for OLEDs. Introduction to exciton formation,
charge transfer states, and triplet harvesting

Hartmut Yersin

Universität Regensburg, Institut für Physikalische Chemie, D-93040 Regensburg, Germany

Tel.: ++49 (0)941 943 4464, Fax: ++49 (0)941 943 4488


Homepage: yersin


Organo-transition-metal compounds (triplet emitters) are attractive for optimizations of organic light emitting diodes (OLEDs). This is a consequence of significantly higher efficiencies obtainable with these compounds as compared to organic emitters. In this introduction a basic model is presented, how electron-hole recombination, i.e. the exciton formation process, can be visualized and how the singlet and triplet states of the (doped) luminescent compounds are populated. This takes place by specific singlet and triplet paths which involve dopant-to-matrix charge transfer states. It is also explained, why the excitation energy is harvested in the lowest triplet state of these complexes. In principle, one can obtain a four times higher efficiency than with (small) organic singlet emitter molecules.


Organometallic compounds find an increasing interest due to their large potential for new photophysical and photochemical applications. This is particularly valid for compounds which exhibit high emission quantum yields from the lowest triplet states to the singlet ground states. Their emission colors may lie in the whole visible range from the blue to the red and also in the IR. [1] In this contribution, it is focused on organo-transition-metal triplet emitters due to their successful applications in electro-luminescent devices such as OLEDs. (See for example [2-9].) By use of these compounds, it is possible to obtain, at least in principle, a four times higher electro-luminescence efficiency than with typical singlet emitters. This property is related to the specific mechanisms of exciton formation in the electron-hole recombination zone and to fast and efficient intersystem crossing (ISC) from the excited singlet to the light emitting triplet state. This process of accumulating the excitation energy in the lowest excited triplet state is often called triplet harvesting. These mechanisms are explained in this contribution.

Basic model of exciton formation

It is instructive to discuss, how the process of electron-hole recombination and the formation of a neutral exciton and finally the population of an excited state of the emitter molecule can be visualized. Here, it is focused on processes that occur within the emitter layer of an OLED. This layer of a thickness of about 100 nm consists of an organic matrix which is doped with emitter molecules (dopants). In the subsequent model it is assumed that the recombi­nation of electrons and holes occurs at the dopants. The importance of this process can be deduced from a comparison of photo-luminescence and electro-luminescence properties. [9] These results show that one of the charge carriers, either hole or electron, can indeed be trapped first at the dopant.

Fig. 1 displays a simplified and schematic model to describe the process of exciton formation. [10] The first step is characterized by trapping of a charge carrier. In our model it is assumed that the hole is trapped first at the emitter molecule. This has been proposed, for example, for Ir(F-ppy)3 (= fac-tris[2-(4',5'-difluorophenyl)pyridine]Ir(III)) in a PVK matrix. [11] The described process induces (for a short time interval) the formation of an oxidized Ir(F-ppy)3 complex. Trapping of an electron as first step would result in an equivalent model and might be of relevance for other emitter molecules. The process of charge carrier trapping can induce a reorganization at the emitter molecule. However, this effect is not depicted in the model of Fig. 1. Under an applied external potential, the electron will migrate along the matrix material towards the anode. Usually, this process of electron hopping (more exactly: polaron hopping) requires a thermal activation energy due to inhomogeneities from spatial and energy disorder and due to matrix reorganization effects. The related energy shifts should be less or of the order of the thermal energy kBT with kB being the Boltzmann constant and T the absolute temperature. For clearness, the diagram is simplified and does not show the inhomogeneous distribution of the energy levels of matrix molecules and their energy shifts induced by the external potential.




Fig. 1 Dynamics of exciton formation. It is assumed that at first the hole (+) is trapped on the doped emitter molecule. The exciton formation starts due to Coulomb interaction between the trapped hole and the electron (–) on a matrix molecule. In the be­ginning of the exciton formation the spins of hole and electron are already correlated to one singlet and three triplet substates. This corresponds in a sta­tistical limit to a ratio of 25 % to 75 %. The S-path and T-paths populate the excited states of the emitter molecule. ISC: intersystem crossing, ΔE(e-h): elec­tron (e) – hole (h) binding energy; ΔE(S-T): singlet – triplet splitting.


When the electron is still far from the trapped hole, it will migrate independently from this hole towards the anode. Thus, hole and electron are not bound or correlated. This situation corresponds to the exciton continuum in solid-state semiconductors. However, when the electron migrates into a region given by a specific electron-hole separation R, the positively charged hole will attract the electron. This sepa­ration is reached, when the energy of the Coulomb attraction is of similar size as the larger one of the two values, the mean inhomogeneous distribution of the energy levels of the matrix material or the thermal energy kBT. Due to the Coulomb attraction, an electron (e) - hole (h) binding results. The binding energy ΔE(e-h) is proportional to 1/(ε·R) and thus depends on the separation R and on the dielectric constant e of the matrix material. Induced by this attraction, the exciton is formed. The Coulomb attraction represents a long-range interaction as compared to nearest neighbor separations.

For the further processes, it is very important to take also the spins of both electron and hole into account. The spin of the hole is given by the spin of the residual electron at the emitter mole­cule. In a quantum mechanical treatment, the two spins can be coupled to four new combined states: One singlet state and one triplet state. The triplet consists of three triplet substates. These substates differ from each other mainly by their relative spin orientations. An energy splitting between the resulting singlet and triplet states may be disregarded at large electron-hole separations. Therefore, the corresponding exciton state is shown in Fig. 1 (middle) just by one energy level. In a statistical limit, all four substates of this exciton state will be formed (populated) with an equal probability. Consequently, one obtains a population ratio of one to three for singlet and triplet substates, respectively.

Driven by electron-hole attraction, the electron will further move on matrix molecules towards the trapped hole. At least, when the electron reaches nearest neighbor matrix molecules of the emitter, an overlap of electron and hole wavefunctions has to be taken into account. The resulting (short-range) exchange interaction leads to an energy splitting of singlet (S) and  triplet (T) states  by ΔE(S-T).  This energy is approximately proportional to exp(-a·R), wherein a is a constant which depends on the respective matrix and the emitter material.

In a final step, the electron jumps in a very fast process directly to the emitter molecule and it results an excited emitter. This process may occur as singlet or triplet path (S-path, T-paths) depending on the initial spin orientation of the electron-hole pair. The corresponding time constants are of the order of one picosecond (see next section). The population of Sm and Tm states, as shown in Fig. 1, is only depicted as an example. Subsequently, the system will exhibit the usual behavior of an optically excited emitter molecule with typical relaxation processes to the lowest excited states and typical emission properties. For detailed discussions of photophysical properties of organometallic triplet emitters it is referred to the Refs. [1, 12a, 12b]. The electronic states of the emitter compound are characterized by the small frame in the diagram of Fig. 1.

Charge transfer states and relaxa­tion paths

The final steps of the mechanisms described above can also be discussed in a slightly diffe­rent way, to illustrate the occurrence of specific singlet and triplet paths. The situation of a lacking electron in the ground state of the doped emitter molecule (dopant D) and of addi­tional electron charge density on nearest neighbor matrix molecules M can be characterized by dopant-to-matrix charge transfer (DMCT) states, specifically by singlet (1DMCT) and triplet (3DMCT) states.

It may be visualized that these states belong to a large molecular unit that consists of the doped emitter itself and of the first nearest neighbors of matrix molecules. The energy level diagram of this emitter-matrix-cage unit is schematically displayed in Fig. 2. It corresponds to the large frame depicted in Fig. 1. In particular, the states S and T shown in the frame represent the 1DMCT and 3DMCT states of the large molecular unit. The lower excited states S0, T1, S1 to Sm are largely confined to the doped molecule itself, while the higher lying charge transfer states are spatially more extended to the first matrix coordination sphere. For these latter states, the exchange interaction between the two electrons involved is relatively small. Thus, the energy separation between the 1DMCT state and the 3DMCT state is expec­ted to be much smaller than singlet-triplet separations of the spatially more confined states of the original dopant itself.




Fig. 2 Energy level diagram of an emissive molecule with its first coordination sphere of matrix mole­cules. The states 1DMCT and 3DMCT represent dopant-to-matrix charge transfer states. The lower lying states are largely those of the isolated emitter molecule itself. The relaxations from the 1DMCT and 3DMCT states, respectively, represent the S-path and T-paths of exciton formation as depicted in Fig. 1.>

On the basis of the energy level diagram of Fig. 2, one also obtains an information about the relaxation paths from the excited charge transfer states. In particular, the relaxation from the 1DMCT to lower states will be faster within the system of singlet states than making a spin-flip first. This is due to the fact that spin-orbit coupling in the organic matrix material will be relatively small and thus, intersystem crossing (ISC) is not favored. Consequently, one obtains the fast singlet path that finally populates the S1 state. (Figs. 1 and 2) Subsequently, the population of the S1 state will be followed by an ISC to the T1 state, though usually with a smaller rate. An initial population of the 3DMCT state is similarly followed by a very fast relaxation within the system of triplet states down to the lowest triplet state T1. (Fig. 2) The beginning of this relaxation process corresponds to the triplet paths in the exciton formation model shown in Fig. 1. The relaxation times within the singlet and triplet system, respectively, are of the order of one picosecond or faster, while the ISC processes can be significantly slower.

Triplet harvesting

Spin-orbit coupling induced by the central metal ion of the emitter complex will not strongly alter the mechanism of exciton formation in an organic matrix material, but it will have drastic effects on the efficiency of electro-luminescence in an OLED device. To illustrate this property, we will compare the efficiency which is obtainable with a purely organic molecule to the efficiency achievable with a transition metal complex, if both molecules exhibit equal photo-luminescence quantum yields. If one assumes that the initial process of exciton formation occurs statistically with respect to the spin orientations, one obtains 25 % of excitons  with  singlet  character and  75 %  with triplet character. At least for small molecules, this view is largely accepted. [2-11]

After exciton formation and relaxations according to the specific singlet and triplet paths, as discussed in the preceeding section, the lowest excited singlet and triplet states are popu­lated. This is valid for organic as well as for organo-transition-metal emitter materials. The corresponding processes are schematically displayed in the middle of Fig. 3. The organic molecule can exhibit an efficient emission as S1 → S0 fluorescence, since usually the S1 → T1 intersystem crossing rate is small. On the other hand, since the radiative T1 → S0 rate is also small, the deactivation of the  T1 state  occurs  normally  non-radiatively at am­bient temperature. Therefore, 75% of the excitons, the triplet excitons, are lost. Their energy is transferred into heat. (Fig. 3, left hand side.) The conditions are more favorable for transition metal com­plexes, in which the central metal ion carries significant spin-orbit coupling. (Fig. 3, right hand side.) This is particularly valid for transition metal ions of the third row of the periodic table. For these complexes, ISC to the lowest T1 state is usually very efficient and thus a singlet S1 emission is not observable. Moreover, the radiative T1 → S0 rate can become sufficiently high so that efficient phosphorescence can occur, even at ambient tem­perature. (For a more detailed discussion see Ref. [10].) Consequently, all four possible spin orientations of the excitons can be harvested to populate the lowest T1 state. In conclusion, by this process of triplet harvesting one can in principle obtain a four times larger electro-luminescence efficiency for triplet emitters than for singlet emitters.  



Fig. 3 The diagram explains the effect of triplet harvesting. Due to spin-statistics, electron-hole recombination leads to 25% singlet and 75% triplet state population. This is a consequence of the occurrence of only one singlet state, but of three triplet substates. In an organic molecule, only the singlet emits light (fluorescence), while the triplet state excitation energy is transferred into heat (left hand side). On the other hand, organometallic compounds with transition metal centers do not exhibit fluorescence, but show a fast intersystem crossing (ISC)  to the lowest triplet state. Thus, singlet and triplet excitation energy is harvested in the triplet state and from there can efficiently be converted into light. In principle, a triplet emitter can reach a four times higher light emission efficiency than a singlet emitter.

Exciton formation at matrix molecules versus charge carrier trapping at emitter compounds

Exciton formation and trapping can also occur at matrix sites. Both the lowest singlet and triplet states of matrix molecules will be populated. From there, triplet and singlet exciton diffusion as Frenkel excitons can occur, however, with different transport probabilities. This will alter the spatial distribution of matrix molecules that are excited in singlet and triplet states, respectively, as compared to the situation immediately after excitation. For an efficient OLED, it is mostly requested to harvest the excitation energy completely in an efficiently phosphorescent triplet state of an organometallic dopant. This requires the realization of two distinctly different processes of energy transfer. For example, the singlet excitation energy is transferred by a long-range Förster transfer mechanism from a matrix molecule to the acceptor (dopant). Independently, the triplet excitation energy is also transferred to the acceptor, though by a different process of energy transfer, in particular by the short-range Dexter transfer mechanism. Both transfer mechanisms should be highly efficient and therefore should fulfill the resonance conditions. This can be expressed by non-vanishing spectral overlap integrals of donor emissions and acceptor absorptions. (For background information see, for example [13].) The fulfillment of these two independent conditions for singlet and triplet energy transfer seems to make this concept of material design – apart from specific and selected combinations – less favorable, as compared to systems with charge carrier trapping directly on the triplet emitters, as discussed in Section 2. (Compare also Refs. [10, 14].)font-family: Verdana;


Organo-transition-metal triplet emitters have a great potential to be applied in OLEDs. Thus, an understanding of the processes, which lead to electron-hole recombination (exciton forma­tion) and to the population of the emissive triplet states is presented. In particular, it is shown that the dynamical process of exciton formation and trapping on an emitter molecule involves charge transfer states which result from excitations of the dopant to its nearest neighbor ma­trix environment. Individual singlet and triplet relaxation paths lead to the population of the lowest excited singlet and triplet states of the dopant. In typical organic molecules with weak spin-orbit coupling and highly forbidden triplet-singlet transitions, triplet state population is transferred into heat (at ambient temperature). Only the singlet state can emit radiatively (fluorescence). On the other hand, in organo-transition-metal compounds, fast inter­system crossing induced by spin-orbit coupling (SOC) effectively depopulates the excited singlet into the lowest triplet state. Again due to SOC, the triplet can decay radiatively as phosphorescence even with high emission quantum yield at ambient temperature. In case of validity of spin statistics only 25% of the excitons can be exploited by organic emitters, while for triplet emitters additional 75% of the excitons are harvested. Thus, the efficiency of light emission in an electro-luminescent device with triplet emitters can be by a factor of four higher than with singlet emitters.


Financial support of the Bundesministerium für Bildung und Forschung (BMBF) is acknowledged.


[1] Yersin, H.; Donges, D.: Top. Curr.Chem. 2001, 214, 81
[2] Lin, Y.-Y.; Chan, S.-C.; Chan, M.C.W.; Hou, Y.-J.; Zhu, N.; Che, C.-M.; Liu Y.; Wang Y.: Chem. Eur. J. 2003, 9, 1263
[3] Su, Y.-J.; Huang, H.-L.; Chien-Le, L.; Chien, C.-H.; Tao, Y.-T.; Chou, P.-T.; Datta, S.; Liu, R.-S.: Adv. Mater. 2003, 15, 884
[4] Lamansky, S.; Djurovich, P.I.; Murphy, D.; Abdel-Razzaq, F.; Kwong, R.; Tsyba, I.; Bortz, M.; Mui, B.; Bau, R.; Thompson,M.EInorg. Chem. 2001, 40, 1704
[5] Brooks, J.; Babayan, Y.; Lamansky, S.; Djurovich, P.I.; Tsyba, I.; Bau, R.; Thompson, M.E.: Inorg. Chem. 2002, 41, 3055
[6] Grushin, V.V.; Herron, N.; LeCloux, D.D.; Marshall, W.J.; Petrov, V.A.; Wang, Y.: Chem. Commun. 2001, 1494
[7] Lo, S.-C.; Male, N.A.H.; Markham, J.P.J.; Magennis, S.W.; Burn, P.L.; Salata, O.V.; Samuel, I.D.W.: Adv. Mater. 2002, 14, 975 
[8] Tsuzuki, T.; Shirasawa, N.; Suzuki, T.; Tokito, S.: Adv. Mater. 2003, 15, 1455
[9] Gong, X.; Robinson, M.R.; Ostrowski, J.C.; Moses, D.; Bazan, G.C.; Heeger, A.J.: Adv. Mater. 2002, 14, 581
[10] Yersin, H.: Top. Curr. Chem. 2004, 241, 1
[11] Lamansky, S.; Kwong, R.C.; Nugent, M.; Djurovich, P.I.; Thompson, M.E.: Organic Electronics 2001, 2, 53
[12a] Finkenzeller, W.; Yersin, H.: Chem. Phys. Lett. 2003, 377, 299
[12b] Finkenzeller, W.; Stoessel, P.;Yersin, H.: Chem. Phys. Lett. 2004, 397, 289
[13] Turro, N.J.: Modern Molecular Photo­chemis­try, Benjamin/Cummings Publ, 1978, Menlo Park Calif. USA
[14] Cleave, V.; Yahioglu, G.; LeBarny, P.; Friend, R.H.; Tessler, N.: Adv. Mater. 1999, 11, 285