Sebastian B. Meierab, Wiebke Sarfert*b, José M. Junquera-Hernándezc, Manuel Delgadoc, Daniel Torderac, Enrique Ortí*c, Henk J. Bolinkc, Florian Kessler‡d, Rosario Scopellitid, Michael Grätzeld, M. Khaja Nazeeruddind and Etienne Baranoff*de
aDepartment of Materials Science VI: Materials for Electronics and Energy Technology, Friedrich-Alexander University of Erlangen-Nuremberg, Martensstr. 7, DE-91058 Erlangen, Germany
bSiemens AG, Corporate Technology, CT RTC MAT MPV-DE, Günther-Scharowsky Strasse 1, DE-91058 Erlangen, Germany. E-mail: wiebke.sarfert@siemens.com
cInstituto de Ciencia Molecular, Universidad de Valencia, ES-46980 Paterna, Spain. E-mail: enrique.orti@uv.es
dLaboratory of Photonics and Interfaces, Institute of Chemical Sciences and Engineering, School of Basic Sciences, Swiss Federal Institute of Technology (EPFL), CH-1015 Lausanne, Switzerland
eSchool of Chemistry, University of Birmingham, Edgbaston, Birmingham B15 2TT, UK. E-mail: e.baranoff@bham.ac.uk
First published on 23rd October 2012
We report here a new cationic bis-cyclometallated iridium(III) complex, 1, with deep-blue emission at 440 nm and its use in Light-emitting Electrochemical Cells (LECs). The design is based on the 2′,6′-difluoro-2,3′-bipyridine skeleton as the cyclometallating ligand and a bis-imidazolium carbene-type ancillary ligand. Furthermore, bulky tert-butyl substituents are used to limit the intermolecular interactions. LECs have been driven both at constant voltage (6 V) and constant current (2.5 mA cm−2). The performances are significantly improved with the latter method, resulting overall in one of the best reported greenish-blue LECs having fast response (17 s), light intensity over 100 cd m−2 and a reasonable efficiency of almost 5 cd A−1.
Consequently, charged blue-emitting iridium complexes are highly searched for, which has resulted in various reports about new blue-emitting complexes based on different substitution patterns, main ligand skeleton, and ancillary ligands.11,12,15–27 Only a few blue emitters based on emitting charged iridium(III) complexes have been reported in LECs, and all these devices showed poor performances (see details in the ESI†). A particular issue is the color point of the emitted light. The electroluminescence spectrum very often shifts significantly to the red with increasing operation time and true blue emission is only observed during the first stages of operation. While the exact reasons for this red shift in electroluminescence (EL) are still unclear, it is paralleled by the red shift of photoluminescence (PL) observed when increasing the concentration of iTMCs in thin films.28 The large spectral red-shift in EL compared to PL (84 nm) observed for the blue-emitting iridium(III) complex [Ir(ppy-F2)2Me4phen][PF6] was attributed to the high concentration of the iTMC in the thin film resulting in a change of the emissive excited state. Supported by quantum-chemical calculations, the origin of the concentration-dependent emission was ascribed to the population of three closely contiguous low-energy triplet states lying within an energy difference of only 0.1 eV.28 Moreover, Margapoti et al. found evidence that molecular rearrangement into aggregates in LECs based on the iTMC [Ir(4,5-diphenyl-2-methylthiazolo)2(5-methyl-1,10-phenanthroline)][PF6] favors the formation of excimeric states induced by the local electric field in a LEC.10 These states are shown to be irreversibly formed and to be the preferential emitting states even in devices left unbiased for a long period of time. Accordingly, a change of the molecular environment in the stressed devices over time could reasonably explain the color shift. Phosphorescent materials with emission independent of the surrounding medium, in particular in neat solid, are expected to solve this issue. The introduction of bulky substituents such as dendrimers on the periphery of the emitting center is a classical approach in neutral phosphorescent emitters to limit the intermolecular interactions impacting the output color.29–32 However, in the case of LECs, the iTMCs act both as charge transporting materials and light emitters. Addition of too large electro-inactive substituents would strongly isolate the iTMCs and decrease the charge transport performances in the device with an overall negative effect on the efficiency. A good balance should therefore be obtained to limit the red shift of emission due to intermolecular interactions on the one hand, while keeping close contact to allow for efficient charge transport through iTMCs on the other hand. In this respect, small and bulky substituents appear as a promising compromise.
Here we report the synthesis, characterization and electroluminescent properties of a new cationic bis-cyclometallated iridium(III) complex, 1 in Scheme 1, with deep-blue emission centered at 440 nm. The complex has been designed by using a 2′,6′-difluoro-2,3′-bipyridine skeleton as the main ligand and a bis-imidazolium carbene-type ancillary ligand. Furthermore, we used bulky tert-butyl substituents to limit the intermolecular interactions. LECs have been studied both at constant voltage (6 V) and constant current (2.5 mA cm−2). The performances are significantly improved with the latter method, resulting overall in one of the best reported greenish-blue LECs with fast response (17 s), light intensity over 100 cd m−2 and a reasonable efficiency of almost 5 cd A−1. In addition, although a red shift of the color point was observed under stress, the blue component of the electroluminescence spectrum remained stable over time.
Scheme 1 Chemical structure of complex 1 and previously reported similar complexes with a carbene-based ancillary ligand. |
Fig. 1 ORTEP drawing of the X-ray crystal structure of 1 (only the Λ-isomer is shown; the PF6 anion and solvent molecules have been omitted for clarity). Selected bond lengths (Å) and angles (°) of the Λ-isomer: Ir(1)–N(1) 2.057(6), Ir(1)–N(4) 2.074(6), Ir(1)–C(10) 2.047(8), Ir(1)–C(11) 2.059(7), Ir(1)–C(29) 2.111(8), and Ir(1)–C(35) 2.113(7); C(29)–Ir(1)–C(35) 84.1(3), C(10)–Ir(1)–C(29) 173.8(3), C(11)–Ir(1)–C(35) 173.8(3), N(1)–Ir(1)–N(4) 171.0(2), C(10)–Ir(1)–C(11) 83.1(3), C(10)–Ir(1)–C(35) 100.5(3), and C(11)–Ir(1)–C(29) 92.8(3). |
Fig. 2 Absorption (left, in acetonitrile) and emission (right: bold black line, in acetonitrile at room temperature; thin black line, in 2-methyltetrahydrofuran at 77 K; blue line, 5 wt% of 1 in PMMA; green line, mixture 1/[BMIM][PF6] 4:1) spectra of complex 1. |
The emission spectrum of 1 in acetonitrile (excitation at 370 nm) is shown in Fig. 2. The highest energy emission band occurs at 440 nm with a photoluminescence quantum yield of 0.13 and an observed excited-state lifetime of 1.16 μs. The radiative constant (kr = 1.1 × 105 s−1) is similar to that reported for complex B (same ancillary ligand, kr = 0.96 × 105 s−1 in CH2Cl2),23 while significantly higher than for A (similar main ligand, kr = 0.23 × 105 s−1).17 On the other hand, the nonradiative constant (knr = 7.5 × 105 s−1) is higher than for A (knr = 0.94 × 105 s−1) and B (kr = 3.8 × 105 s−1 in CH2Cl2) but half of the value reported for C (kr = 14.2 × 105 s−1).17,23 The differences with B could be attributed to a solvent effect as the reported data were measured in CH2Cl2 while we use acetonitrile, a more polar solvent. When cooled to 77 K in 2-methyltetrahydrofuran, the emission spectrum is only slightly blue shifted (λmax = 436 nm) and is highly structured with two vibronic progressions, about 1000 and 1500 cm−1. This points to an emitting triplet state with mainly 3LC character, as supported by theoretical calculations (see below).
Solid-state photophysical data were obtained for thin solid films consisting of 5% by weight of 1 in poly(methyl-methacrylate) (PMMA) and of 1 and the ionic liquid (IL) 1-butyl-3-methyl-imidazolium hexafluorophosphate [BMIM][PF6] in a molar ratio of 4 to 1 using an excitation wavelength of 280 nm. The PL emission spectra of both thin solid films (Fig. 2) show a structured band with maxima at 440 and 466 nm and a lower intensity shoulder at around 495 nm. The higher-energy peak is more intense for the film of 1 in PMMA than for the 1:IL film, whereas the shoulder is more pronounced for the complex mixed with the IL. The photoluminescence spectra in the solid state are very similar to those recorded in solution. This effect is due to the bulky tert-butyl substituents attached to the emitting ligand that prevent intermolecular stacking, as seen above in the X-ray crystal structure. Photoluminescence quantum yields of 0.59 and 0.17 are measured for 1 in a 5% complex in a PMMA film and in a more concentrated film similar to that used in LEC devices, respectively. The significant decrease in photoluminescence quantum yield when increasing the concentration of the complex in a solid film indicates that significant exciton hopping occurs in a more concentrated film. Hence, the tert-butyl substituents prevent emission shifts yet they do not shield the molecules that much to prevent exciton hopping and concentration quenching.
The geometry of the complexes in their ground electronic state (S0) was fully optimized without imposing any symmetry restriction. For complex 1, calculations correctly reproduce the near-octahedral coordination of the Ir metal observed in the X-ray structure and predict geometric parameters in good accord with the experimental data (see Table S2 in ESI†). For instance, the values computed for the Ir(1)–C(10) and Ir(1)–C(11) distances (2.050 and 2.058 Å, respectively) and for the bite angle of the bis-carbene ancillary ligand (84.5°) agree well with the X-ray values (2.047(8) Å, 2.059(7) Å and 84.1(3)°, respectively). The largest difference between computed and X-ray data corresponds to the coordinate Ir–N and Ir–C(carbene) bonds, whose distances are theoretically overestimated by 0.03 and 0.05 Å, respectively, as is frequently observed when using the B3LYP functional.36 To maximize the interaction with the iridium core, the bis-carbene ligand loses its planarity and the imidazolium rings are twisted by an average angle of 52.0°.
Fig. 3 compares the energy and electron density contours calculated for the highest-occupied and lowest-unoccupied molecular orbitals (HOMO and LUMO, respectively) of 1 with those obtained for complexes A, B and C. The electron density distribution of the HOMO shows important differences for the four complexes. The HOMO of 1 calculated at −5.98 eV is composed of a mixture of Ir dπ orbitals (t2g) and π orbitals of the ancillary ligand, with a small contribution from the π orbitals of the main ligands. This is not what is usually observed for commonly designed Ir-iTMCs incorporating two cyclometallating C⁁N ligands, the most popular being 2-phenylpyridines (ppy), and a bidentate diimine N⁁N ancillary ligand, for which the HOMO is a mixture of Ir dπ orbitals and phenyl π orbitals of the ppy ligands.9,22,37 The different topology calculated for the HOMO of 1 is due to the high energy of the HOMO of the bis-imidazolium carbene ligand. On passing to complex A, the bis-carbene:C^C:ligand is replaced by a pyridine–carbene N⁁C:ligand and the HOMO retrieves the participation of the main C⁁N ligands (Fig. 3). As a consequence, the HOMO is stabilized by 0.26 eV in perfect agreement with the increase of the oxidation potential from 1 (1.35 V) to A (1.61 V). The effect of the main ligand is reflected on going to complex B, for which the cyclometallating 2,3-bipyridine ligands are replaced by the less electron-withdrawing ppy ligands and the HOMO increases in energy by 0.32 eV. This destabilization determines the lower oxidation potential recorded for B (0.99 V) and is slightly smaller than that obtained on passing from A to C (0.47 eV) due to the tert-butyl donor groups present in 1. It is to be mentioned that whereas the HOMO-1 of 1 shows an electron density distribution similar to the HOMO, the HOMO-2, HOMO-3 and HOMO-4, which are calculated within a range of 0.55 eV below the HOMO, are mainly composed of different mixtures of Ir dπ orbitals and π orbitals of the main ligands (Fig. S2†).
Fig. 3 Schematic diagram showing the electron density contours (0.03 e bohr−3) and the energy values calculated for the HOMO and LUMO of complexes 1, A, B and C. The tert-butyl groups in complex 1 and the methyl groups in complexes 1 and B have been omitted to facilitate the viewing of the orbitals. |
In contrast to the HOMO, the LUMO shows a similar electron density distribution for the four complexes (see Fig. 3); it resides on the main C⁁N ligands, showing no contribution from the ancillary ligand. Nonetheless, the change of the ancillary ligand, from a bis-carbene (1 and B) to a pyridine–carbene (A and C), stabilizes the LUMO in good agreement with the anodic shift recorded for the first reduction potential (1: −2.33 V vs.A: −2.21 V and B: −2.32 V vs.C: −2.26 V). The slightly larger stabilization predicted in passing from 1 to A (0.19 eV compared with 0.08 eV from B to C) agrees well with the experiment and is attributed to the presence of the tert-butyl donor groups in 1 as proposed above from the electrochemical properties. The LUMO + 1 is calculated slightly above the LUMO (0.05 eV for 1) and shows a similar electron density distribution. Indeed, no molecular orbital shows a relevant contribution from the ancillary ligand in a range of up to 1 eV above the LUMO for complex 1 (Fig. S2†).
The evolution of the HOMO–LUMO energy gap along the series B (4.13 eV), C (4.16 eV), 1 (4.41 eV) and A (4.48 eV) follows the trend observed for the electrochemical gap (ΔE1/2redox = E1/2ox − E1/2red, B: 3.31 V; C: 3.45 V; 1: 3.68 V; A: 3.81 V) and justifies, in a first approach, the blue shift that the replacement of the main C⁁N ligand (2-phenylpyridine by 2,3′-bipyridine) causes in the emission spectra in passing from B (452 nm, CH2Cl2)23 to 1 (440 nm, CH3CN) or from C (450 nm, CH3CN) to A (435 nm, CH3CN).17
The atomic orbital composition calculated for the frontier orbitals does not support the fact that emission in complex 1 takes place from the main ligand (3LC character), as the experiment suggests, since the HOMO implies the Ir dπ orbitals (t2g) and the π orbitals of the ancillary ligand whereas the LUMO resides on the main C⁁N ligands. As the simple description of the HOMO and LUMO is not always representative of the emitting triplet state, the lowest-energy triplet state of complex 1 was calculated using the spin unrestricted UB3LYP approach. After full-geometry relaxation, the calculation converges to a triplet state (T1) for which the two unpaired electrons mostly reside on one of the C⁁N ligands with a small contribution (0.09 e) from the metal (Fig. 4). This is also the case for complexes A (0.10 e), B (0.15 e) and C (0.15 e). Calculations therefore suggest that excitation to the emitting triplet state of complex 1, and also of complexes A, B and C, mainly implies an electron promotion within the cyclometallated ligands and has a 3LC electronic nature in good agreement with experimental evidence. The large LC character of T1 explains the structural aspect of the emission band observed experimentally and the absence of any rigidochromic shift upon decreasing the temperature (Fig. 2). The T1 state is computed to lie 2.97 (1), 2.91 (A), 2.81 (B) and 2.84 eV (C) above the electronic ground state S0 (adiabatic energy differences). These energy differences also support the blue shift observed for the emission of complexes 1 and A, bearing 2′,6′-difluoro-2,3′-bipyridine C⁁N ligands, compared with B and C.
Fig. 4 UB3LYP spin-density contours (0.005 e bohr−3) computed for the optimized T1 state of complexes 1, A, B and C. |
To further investigate the nature of the excited states, the low-lying triplet states of complex 1 were calculated using the time-dependent DFT (TD-DFT) at the optimized geometry of S0. TD-DFT calculations predict two low-energy triplets, T1 and T2, lying at very close energies (3.06 and 3.11 eV above S0, respectively) and showing a complex nature resulting from a mixture of MLCT, LLCT and LC excitations (see Table S3†). The geometry of these two states was optimized at the TD-DFT level as a further refinement. After full-geometry relaxation, both states evolve to 3LC states, each one localized on one C⁁N ligand (see Fig. S3†). TD-DFT optimizations therefore confirm that the two lowest-energy triplet states of complex 1, that lie almost degenerate at 2.70 (T1) and 2.75 eV (T2) above S0 (adiabatic energy differences), correspond to π–π* 3LC states involving the C⁁N ligands. It is interesting to note that at the optimized geometries of T1 and T2 both the HOMO and the LUMO are localized on one of the cyclometallated ligands. This is illustrated in Fig. 5 for T1. Therefore, excitation to T1 and T2 slightly modifies the coordination sphere of the iridium core and mainly affects the geometry of the C⁁N ligand involved in the excitation (Table S2 and Fig. S4†).
Fig. 5 Schematic diagram showing the electron density contours (0.03 e bohr−3) calculated for the HOMO and LUMO of complex 1 at the TD-DFT optimized geometry of T1. The tert-butyl and the methyl groups have been omitted to facilitate the viewing of the orbitals. |
The first 40 singlet states of complex 1 were also calculated using TD-DFT. The calculations provide a picture of the absorption spectrum that reproduces the main bands observed in the experimental spectrum. An interesting feature is the lack of any calculated singlet state with energy below 3.76 eV (330 nm). This points to triplet states as responsible for the very weak absorption bands detected experimentally at 402 and 428 nm, which would explain their low intensity (ε ≈ 160 and 60 M−1 cm−1, respectively). Triplets T1 and T2 are well separated in energy from the rest of excited triplet states (Table S3†) and are the main candidates to account for those bands.
Fig. 6a represents the time-dependent electroluminescence spectra of 1 in the aforementioned LEC device configuration at a fixed voltage of 6 V. The initial spectrum (black line) was recorded as soon as emission could be observed (after two minutes). The device exhibits greenish EL with a peak centered at around 500 nm, coincident with the PL shoulder emission determined in acetonitrile solution and in thin film (Fig. 2). Moreover, the EL spectrum exhibits a minor tail extending into the red part of the visible spectrum. The CIE 1931 coordinates are (0.27, 0.43). As the device operates, the low energy tail of the electroluminescence spectrum becomes more and more pronounced and color shifts towards the red, which is more clearly visualized in the CIE 1931 diagram in Fig. 6b. After ten minutes of operation the CIE 1931 coordinates have changed to (0.36, 0.45) and the EL maximum at that operation time is at 532 nm. Nevertheless, the blue part of the spectrum remains interestingly fairly stable, and the red-shift observed is best described as a broadening of the spectrum.
Fig. 6 (a) Normalized time-dependent electroluminescence spectra of 1:IL = 3:1 at a fixed voltage of 6 V. (b) Corresponding time-dependent CIE coordinates (CIE 1931). |
The onset of the gradual red-shift of the EL spectrum during operation matches very well with the point in time after maximum luminance was achieved. The temporal dependence of the luminance and current density of the LEC device is illustrated in Fig. 7a. Upon applying a fixed voltage of 6 V to the device, the typical behavior of light-emitting electrochemical cells, a gradual increase in both luminance and current with proceeding time, is observed. The slow rise reflects the operational mechanism of LEC devices. In response to the applied electrical field, the small PF6− counter-anions of 1 migrate to the respective electrode interface (anode) leaving an excess of positively charged complex molecules at the counter-electrode (cathode), which on the one hand lowers the barrier for charge injection at both electrodes,39–41 and on the other hand results in an electrochemical doping process of the active layer.42–44
Fig. 7 (a) Luminance and current density and (b) current and luminous efficiency over time of the LEC operated at a fixed voltage of 6 V. |
The brightness reaches a maximum (Lmax) of 15 cd m−2 after 3.6 minutes (Fig. 7a), defined to be the turn-on time (ton) of the device, and maximum efficiency values of 1.76 cd A−1 and 0.92 lm W−1, respectively, were attained (Fig. 7b). After having reached its maximum, the light intensity dropped rapidly resulting in a low lifetime (t1/2, defined to be the time for the luminance to decay from the maximum to half of the maximum) of around three minutes, which is partly attributed to the irreversible oxidation behavior of 1 (see the Electrochemistry section) resulting in degradation of the complex. The current, however, showed a further increase with time after crossing the point of maximum luminance. This opposing behavior of luminance and current is due to the continuous growth of the doped regions beyond what is necessary for an efficient operation of LEC devices and results in device degradation.44
To further study the material, the LEC was also tested under constant current conditions (galvanostatic mode), an operational protocol previously reported to be advantageous for the performance of the devices such as accelerated turn-on time.45 A constant current density of 2.5 mA cm−2 was applied to the LEC and the temporal evolution of the luminance, voltage and efficiency is illustrated in Fig. 8. Compared to the constant voltage mode, a nearly eight-fold increase in maximum brightness (113 cd m−2) accompanied by a more than two-fold gain in current efficiency (4.71 cd A−1 and 1.95 lm W−1, respectively) was achieved. The device response was also very fast, with a turn-on time of around 17 s. Another striking feature is the bluish-green appearance of the device at the moment of maximum brightness demonstrated by the CIE 1931 coordinates of (0.24, 0.40).
Fig. 8 (a) Luminance and voltage and (b) current and luminous efficiency over time of the LEC operated at a fixed current density of 2.5 mA cm−2. |
A summary of the electrical characteristics of the LEC incorporating 1 is given in Table 1. We now compare complex 1 to the previously published blue-emitting Ir-iTMCs in LEC environments in combination with an ionic liquid.11,12,15,19,23 We stress that in the present case no optimization has been carried out concerning the device configuration and active layer composition.
In 2008 He et al. published a greenish-blue LEC having CIE coordinates of (0.20, 0.28) and moderate brightness (39 cd m−2) and current efficiency (0.65 cd A−1).11 However, the very slow device response resulted in maximum brightness achieved only after 7 hours. Ionic liquid was added to reduce the turn-on time to 1 hour at the expense of the performance. The light intensity and efficiency decreased to 23 cd m−2 and 0.51 cd A−1, respectively, and the color point significantly red-shifted (CIE: 0.33 and 0.45), which was attributed to the electrical excitation. Under a high electrical field in LEC devices, the molecular orbitals of the ionic complex might be polarized and, thus, the energy of the emitting state would decrease, resulting in the red-shift of the EL spectrum.46 When using complex 1 in a LEC configuration with IL similar to that used by He et al., and by applying a similar driving voltage (6 V), we achieve a more bluish LEC (CIE: 0.28, 0.43) with a much faster turn-on time of only 3.6 minutes and a threefold increase in efficiency to 1.76 cd A−1 at a comparable light intensity.
Yang et al. presented blue-green emitting LECs with light intensities and response times similar to our devices under constant voltage conditions.23 Although devices using complex B showed deeper blue electroluminescence (CIE: 0.20 and 0.34) than our device with 1, efficiencies were lower by a factor of two.
In 2009 He et al.12 demonstrated for the first time reasonable current efficiencies as high as 8.4 cd A−1 for blue-emitting LECs, and even high efficiencies of nearly 20 cd A−1 have recently been demonstrated by the same group.15 Although, these high values have been achieved at very low light levels of only 39 and 15 cd m−2. Moreover, the turn-on time of those LECs was very long (20 to 30 minutes). In contrast, we present LEC devices with fast response (17 s) having a moderate light intensity of over 100 cd m−2 and a reasonable efficiency of almost 5 cd A−1 with similar CIE coordinates as shown in the aforementioned papers.
Concerning the lifetime of our devices, we want to stress that stability in general seems to be a limiting factor in blue-emitting LEC devices to date. To the best of our knowledge, there are only a few publications that explicitly display stability values of blue LECs ranging from minutes to few hours.16,19,23,26
All the calculations were performed in the presence of the solvent (acetonitrile). Solvent effects were considered within the self-consistent reaction field (SCRF) theory using the SMD keyword that performs a polarized continuum model (PCM)56–58 calculation using the solvation model of Truhlar et al.59 The SMD solvation model is based on the polarized continuous quantum chemical charge density of the solute (the “D” in the name stands for “density”).
Time-dependent DFT (TD-DFT)60–62 calculations of the lowest-lying triplets and singlets were performed in the presence of the solvent at the minimum-energy geometry optimized for the ground state (S0). Up to 40 singlet and 40 triplet states were calculated. The geometries of the two lowest-energy triplets (T1 and T2) of complex 1 were fully optimized at the TD-DFT level.
LEC devices were prepared as follows. Three percent by weight of complex 1 and [BMIM][PF6] in a molar ratio of 3:1 were dissolved in 0.5 mL of MEK. The solution was filtered using a 0.1 μm PTFE-filter and spin-coated on top of a ITO/PEDOT:PSS substrate under ambient conditions giving rise to a transparent thin film. The pre-coated 100 nm PEDOT:PSS layer was deposited to planarize the ITO anode, increasing the yield and reproducibility of working devices. After coating, the substrate was transferred to a N2 filled glovebox (O2 and H2O < 1 ppm) and annealed on a hotplate at 100 °C for 1 hour resulting in a final active layer thickness of 100 nm. Finally, a 150 nm thick aluminum (Al) cathode was thermally evaporated on top under high vacuum (<1 × 10−6 mbar) using a shadow mask, defining 8 pixels per substrate, each having an active lighting area of 4 mm2. The device was encapsulated by a glass capping layer to protect the organic layers against penetrating oxygen and water.
The thicknesses of the films were determined using a KLA Tencor P·15 profilometer. The electro-optical characterization was carried out in a Botest OLED Lifetime Test System (OLT) under a constant voltage of 6 V and a constant current density of 2.5 mA cm−2, respectively, at a constant temperature of 20 °C. The photo-diodes were calibrated using a Photo Research PR-650 SpectraScan® Colorimeter and simultaneously the electroluminescent spectra of the LEC devices were detected in the visible range between 380 and 780 nm at a constant voltage of 6 V.
Footnotes |
† Electronic supplementary information (ESI) available. CCDC 892956. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c2tc00251e |
‡ Present address: Siemens AG, Corporate Technology, CT RTC MAT MPV-DE, Günther-Scharowsky Strasse 1, DE-91058 Erlangen, Germany. |
This journal is © The Royal Society of Chemistry 2013 |