Highly Efficient Color‐Stable Deep‐Blue Multilayer PLEDs: Preventing PEDOT:PSS‐Induced Interface Degradation

Highly efficient and stable blue light emission is observed in novel copolymers that are produced from specially designed building blocks. A PEDOT:PSS-induced chemical degradation of the polymer light-emitting diodes (PLEDs) is identified at the interface, and it is found to be accompanied by a shift in the emission color. A method to prevent this highly undesirable interaction is presented. Since the report of the first small-molecule-based thin-film organic light-emitting diode (OLED)1 and the first polymer light-emitting diode (PLED),2 remarkable efforts of the research community and a resilient industrial contribution have led to a number of commercial OLED-based products; they have attracted attention beyond niche markets and have large-area applications in the field of flat-panel displays and lighting devices.3 While fast improvements in the chemical stability of the active materials and the reliability of the devices have led to well-established red and green4 fluorescent and phosphorescent emitters (based on both small molecules and conjugated polymers), there remains a quest for improved stability in materials emitting in the blue spectral range. Among others, poly(para-phenylene) (PPP)- and poly(pyrene)-type5 polymers are rather promising as blue-emitting conjugated polymers for stable PLED applications and therefore of particular interest. In addition to a well-defined chemistry, PPP-type polymers allow for effective emission-color tuning from UV-blue to blue emission by increasing the number of aryl–aryl bridges – stretching from PPP6 with a dominating UV emission, over poly(fluorene)7 (emitting in the UV-blue) and poly(indenofluorene)8 (PIF) (emitting further in the visible) to poly(pentaphenylene),9 yielding a deep blue emission perfectly matching the sensitivity of the human eye for blue. Aside from the chemical stability of the particular emitter material and a high radiative quantum yield of the emissive unit, efficient and bright PLED devices with long lifetimes require a balanced charge carrier injection and an effective and balanced transport of electrons and holes towards the electro-optical active layer. Moreover a pinning of the emission zone to the center of the device, using heterojunctions with an appropriate type-II band level offset to avoid quenching at either of the electrode interfaces, has also been found to be rather favorable.10 Modifications of the poly(3,4-ethylenedioxythiophene):poly(styrenesulphonic acid) (PEDOT:PSS) surface, auxiliary injection layers,11 doped interface regions,12 and multiple tandem structures including interconnecting layers13 have been used to enhance and balance charge injection/transport as well as exciton formation in the devices. Irrespective of the individual preparation techniques, the crucial challenge of solution-processed multilayer PLEDs is how to not redissolve a preceding polymer layer by the solvent used to deposit a subsequent polymer layer. Different approaches have been suggested, ranging from liquid buffer layers between the individual polymer layers14 to in situ converted or crosslinked polymer layers15 to the application of orthogonal solvent systems.16 Alternatively, the stability against the resolubilization of a polymer film can be significantly increased by a thermal bake-out process.17 Independent of the specific device structure, the majority of today's polymer-based light-emitting diodes comprise a PEDOT:PSS layer on top of the transparent bottom electrode due to its reasonable workfunction (4.8 to 5.6 eV), high conductivity range (≈10−5–102 S/cm), and good hole injection ability.18 Since the surface moiety of a PEDOT:PSS layer is sensitive to the preparation/storage conditions, for example, vacuum-annealed samples tend to form PSS-rich surfaces while water vapor supports the decrease of the PSS concentration on top of the layer,19 the overall performance of a PLED can be considerably altered. In particular, enhanced concentrations of the insulating PSS moieties are considered to be responsible for electron blocking at the interface.20 In the case of using desired low-workfunction cathode materials, electrons – representing the charge carrier majority – tend to accumulate at the PSS-rich phase increasing hole injection because of strong electric fields in the anode region, leading to an improved charge carrier balance and higher device efficiencies.21 Even though PEDOT:PSS is widely used for PLEDs due to its beneficial properties, it suffers from many drawbacks such as its acidic character,22 exciton quenching,23 field-induced migration of PEDOT-cations,24 and the possible chemical degradation of the light-emitting polymer.25 These problems can be avoided or at least minimized by integrating an appropriate acid-stable interlayer between the organic–PEDOT:PSS interface. We report here on the electro-optical properties of novel light-emitting copolymers (CPs, designated as CP-AE, CP-ABCD, CP-ACD) comprising different light-emitting and transport-supporting building blocks (Figure 1). The chemical degradation of the emissive units assigned to the acidity of PEDOT:PSS will be addressed, and a route to fabricate highly efficient and color-stable multilayer PLEDs will be presented. Chemical structures of the individual copolymer building blocks: A) 9,9,12,12-tetraorganyl-6,12-dihydroindeno[1,2-b]fluorene; B) (E)-4-organyl-N-phenyl-N-(4-styrylphenyl)aniline; C) 9,10-diorganylphenanthrene; D) N1,N4-bis(4-organylphenyl)-N1,N4-diphenylbenzene-1,4-diamine; and E) 4-organyl-N,N-diphenylaniline; Rn where n is 1–10 represents organyl groups. CP-AE consists of component A and E at a ratio of 50:50; CP-ABCD consists of component A, B, C, and D at an A:B:C:D ratio of 46:2:50:2; and CP-ACD consists of component A, C, and D at an A:C:D ratio of 48:50:2. The molecular design of the copolymers was optimized with particular emphasis on: a) efficient exciton formation, b) light emission in an energy region with reasonable human-eye sensitivity for the blue color, c) enhanced charge carrier transport properties for electrons and holes, d) appropriate energy level positions facilitating both charge injection and blocking in a polymer heterostructure, and e) suppression of the PEDOT:PSS-induced spectral degradation of the emissive units. The alternating copolymer CP-AE is used as the hole-transport and injection layer; it contains component A (indenofluorene) and E (triphenylamine) at an A:E ratio of 50:50. Copolymer CP-ABCD, comprising an indenofluorene unit (A), an (E)-4-organyl-N-phenyl-N-(4-styrylphenyl)aniline emissive unit (B), electron-transport-supporting phenanthrene (C), and hole-transport-supporting amine (D) units at an A:B:C:D ratio of 46:2:50:2, is used as the emissive layer. Additionally the copolymer CP-ACD (A:C:D = 48:2:50) was investigated in place of CP-ABCD for comparison, as the two differ only by the additional emissive unit B. Figure 2a–c show the thin-film absorbance and photoluminescence (PL) spectra of the three utilized CPs as well as the PL film spectra of the CP layers deposited on PEDOT:PSS and the CP films mixed with trifluoroacetic acid (TFA) in order to demonstrate the possibility of chemical degradation of the emissive units of the copolymers. Absorbance and photoluminescence spectra of a) CP-AE, b) CP-ACD, and c) CP-ABCD: thin-film absorbance spectra (filled squares); thin-film PL spectra (open squares); thin-film PL spectra of a PEDOT:PSS/CP double layer (open circles); thin-film PL spectra of a CP–(trifluoroacetic acid) blend (open diamonds). d) Normalized electroluminescence emission spectra of an ITO/PEDOT:PSS/CP-ABCD (60 nm)/Ca (10 nm)/Al (100 nm) device biased at 4 V (first, full squares; third, full circles) or at 14 V (second, open squares; fourth, open circles). As depicted in Figure 2a, CP-AE shows an unstructured absorption spectrum with its maximum at 406 nm. The emission maximum of the CP-AE thin-film PL spectrum was found at 435 nm with the first vibronic peak at 461 nm. Except for a bathochromic shift of ca. 10 meV, which is attributed to the high concentration of copolymerized-amine-based hole transport units,26 both the emission and absorption spectra are prototypical for PIF.8 Figure 2b shows the absorption spectrum of CP-ACD with its maximum located at 404 nm and a slight shoulder at 385 nm. The thin-film emission maxima, found at around 423 and 451 nm, are assigned to the π–π* transition and the first vibronic progression, respectively, of the copolymer. As it becomes evident in Figure 2c, the absorption spectrum of CP-ABCD does not show any significant modifications compared to CP-ACD because the additional emissive component B is only present at a low concentration. Nevertheless, the PL emission spectrum is seriously influenced by the introduced co-emitter B; the maximum of the spectrum is shifted towards longer wavelengths by ca. 20 meV to 457 nm with a shoulder at 480 nm. This significant shift of the PL emission spectrum as well as the strong reduction of the CP-ACD emission peak at 423 nm clearly demonstrates the distinct impact of component B, resulting in a perfect overlay with the sensitivity of the human eye. It is known from literature that polymers containing pyridine units,27 vinylene units28 or distinct end-chain groups29 may be prone to degrade in an acidic environment. Therefore all copolymers have been tested for the possible influence of PEDOT:PSS to the emissive spectrum, and in particular that of the PSS-rich phase at the interface. As depicted in Figure 2c, the thin-film emission spectrum of CP-ABCD on a glass substrate is considerably different from the emission of the copolymer on PEDOT:PSS, where a reduction of the emission peak at 457 nm by more than 60% was found. To clarify the acidic influence on the emission, CP-ABCD was also blended with TFA; when the PL of the thin films was measured, exactly the same reduction in the PL emission at 457 nm was observed. Consistently with the spectra obtained from PEDOT:PSS/copolymer double-layer systems, where essentially no alteration for CP-ACD and CP-AE on PEDOT:PSS was found, the blends of TFA:CP-ACD (Figure 2b and Supporting Information, Figure S1) and TFA:CP-AE (Figure 2a; Figure S1, Supporting Information) did not exhibit any significant changes in the PL spectrum. Comparing the chemical structures of CP-ACD and CP-ABCD, the observed spectral change can be attributed to an electrophilic addition of the acid with the nucleophilic carbon double bond of the vinylene units28 (component B), leading to a strong quenching of the PL at this copolymer site. In addition to PL measurements, the influence of PEDOT:PSS on CP-ABCD was investigated in a PLED device structure (indium tin oxide (ITO)/PEDOT:PSS/CP-ABCD/Ca/Al PLED). As depicted in Figure 2d the electroluminescence (EL) spectrum can be reversibly changed as a function of the bias voltage with respect to the emission peak at 457 nm. At 4 V, the EL spectrum mostly equals the corresponding CP-ABCD single-layer PL spectrum (Figure 2c, open squares) with a maximum at 455 nm and two additional features at 426 and 482 nm. Increasing the bias voltage to 14 V leads to a reduction of the maximum emission peak at 457 nm by more than 50%, comparable to the change found in the PL spectrum from the PEDOT:PSS/CP-ABCD double layer or the TFA:CP-ABCD films (Figure 2c, open circles; open diamonds). Since this behavior is shown to be fully reversible, general device degradation can be excluded, and the change in the EL spectrum can be assigned to a voltage-dependent movement of the recombination zone30 forwards and backwards with respect to the PSS/CP-ABCD interface with degraded vinylene units due to the interaction with PSS. While tentatively discussed in literature,23, 25 these results clearly demonstrates that the acidity of the PEDOT:PSS layer can distinctively alter the device performance/stability due to exciton quenching at the emissive units at the interface, accompanied by changes in the overall spectral emission characteristics and a reduction of the device stability. Particularly when using low-workfunction cathode materials, where the charge carrier recombination zone tends to shift towards the anode, the chemical stability of the emissive units close to the interface with PEDOT:PSS is of crucial importance. In order to impede interactions at the PEDOT:PSS/CP-ABCD interface in PLEDs and to improve hole transport towards the CP-ABCD layer, an additional hole transport layer using CP-AE was integrated. To avoid resolubilization during device preparation, the CP-AE layer was stabilized by thermal curing at 200 °C for 1 h in argon. This bake-out did not result in any significant change of the absorbance or of the PL spectrum (Figure S2, Supporting Information). Atomic force microscopy investigations reveal that CP-AE films cured at 200 °C did not show any noteworthy decrease in film thickness after being washed with pure solvent, whereas films baked at 70 °C can still be completely removed. Compared to devices without CP-AE, ITO/PEDOT:PSS/CP-AE/CP-ABCD/Ca/Al devices did not show any distinct bias-voltage-dependent change in the EL emission spectrum (Figure S3, Supporting Information), but a significant device efficiency enhancement from 1 to 1.9 cd A−1 at 1000 cd m−2 (Table 1). Aside from the enhanced hole transport properties of CP-AE, this significant performance enhancement is a result of an optimal type-II band alignment at the CP-AE/CP-ABCD interface as determined by ultraviolet photoelectron spectroscopy (UPS) (Figure 3a; Figure S4, Supporting Information). Therefore valence-band-region spectra and the secondary electron cut-off was measured after each layer-deposition step. CP-AE's and CP-ABCD's low-binding-energy onsets are 0.6 and 0.9 eV below the Fermi level (EF), respectively; this energy offset corresponds to the hole-injection barrier from PEDOT:PSS. Together with the workfunction determined from the secondary electron cut-off, the corresponding ionization energy can be found at 5.5 and 5.8 eV for CP-AE and CP-ABCD, respectively. Based on the optical bandgaps, the onset of the lowest unoccupied molecular orbital (LUMO) can be assumed to be at 2.6 and 2.8 eV below the vacuum level for CP-AE and CP-ABCD, respectively. Consequently, CP-AE is operative at an additional energy level in this staggered heterojunction, enabling stepwise hole injection from PEDOT:PSS into CP-ABCD. In contrast, electrons are blocked at the CP-AE/CP-ABCD interface, being favorable for device performance. a) Schematic energy level diagram of the PEDOT:PSS/CP-AE/CP-ABCD multilayer structure obtained from layer-by-layer UPS investigation. b) Current density (open squares) and luminance (open circles) as a function of the bias voltage in an ITO/PEDOT:PSS (60 nm)/CP-AE (20 nm)/CP-ABCD (60 nm)/Cs2CO3 (0.15 nm)/Al (100 nm) device. The inset shows the normalized EL spectrum of the device with CIE color space coordinates of x = 0.144 and y = 0.129. (HOMO indicates the highest occupied molecular orbital.) In a final step, further considerable device efficiency enhancement was achieved using a cesium carbonate cathode (Cs2CO3, workfunction of approximately 2.1 eV31) instead of a calcium cathode with its significantly higher workfunction of approximately 2.9 eV.32 Figure 3b shows the current–voltage and luminescence–voltage characteristics as well as the corresponding electroluminescence spectrum of the optimized ITO/PEDOT:PSS/CP-AE/CP-ABCD/Cs2CO3/Al structure. The inset shows the electroluminescence spectrum of the CP-AE/CP-ABCD double-layer device with its maximum at 459 nm and a weak shoulder at around 480 nm. Changing the cathode material from calcium to cesium carbonate nearly doubles the efficiency from 1.9 to 3.7 cd A−1 at 1000 cd m−2. The maximum device efficiency was measured at low current densities to be at 9.7 cd A−1 (Figure S5, Supporting Information). The observed efficiency roll-off in this device structure can be explained by the strongly enhanced electron injection at high bias voltages resulting in a progressive unbalance of electrons and holes. In addition to the high current efficiencies, a high brightness with luminance values of up to 10 600 cd m−2 was found for this multilayer approach. In order to demonstrate the distinct contribution of component B to the device efficiency, the reference copolymer CP-ACD, was examined in an ITO/PEDOT:PSS/CP-AE/CP-ACD/Cs2CO3/Al configuration, leading to a lower overall efficiency of 2.8 cd A−1 at 1000 cd m−2. In conclusion, the fabrication of highly efficient blue-light-emitting multilayer PLEDs was demonstrated to be possible without the need for orthogonally soluble polymers by employing the novel combination of the copolymers CP-AE/CP-ABCD comprising optimized building blocks. We demonstrate that the hole-transporting polymer CP-AE leads to a more balanced charge-carrier injection as well as charge blocking at the polymer interface due to optimal band level alignment. Therefore the recombination zone is shifted away from the PEDOT:PSS interface, which was shown to cause defect-induced exciton quenching and spectral changes, and shifted into the CP-ABCD layer. The introduction of the CP-AE layer not only led to an enhanced spectral stability but also to a clearly increased device efficiency (Table 1). A further efficiency increase was obtained by changing the cathode material from calcium to cesium carbonate. The final ITO/PEDOT:PSS/CP-AE/CP-ABCD/Cs2CO3/Al PLED showed pure blue emission (CIE (Commission Internationale de l'Eclairage) color space coordinates: x = 0.144, y = 0.129), a high luminescence intensity, and notably one of the highest ever reported current efficiencies for blue multilayer PLEDs, 9.7 cd A−1. Polymer Synthesis: All of the copolymers were synthesized using a Suzuki-coupling method as described previously.33 Structures were confirmed by 1H NMR spectroscopy. Purity was confirmed by trace analysis via inductively coupled plasma mass spectrometry (ICP-MS) checking for traces of monomers (Br, B) and catalyst (Pd, P). The following trace contents were found: Br ≤50 ppm, B ≤20 ppm, P ≤200 ppm, and Pd ≤15 ppm. Molecular weights and polydispersities were determined by gel-permeation chromatography calibrated against polystyrene standards. The molecular weights were in the range of 200 000−600 000 g mol−1, with most polymers between 400 000 and 500 000 g mol−1. The polydispersities were in the range of 2.8−4.0. UV-Photon Spectroscopy: Ultraviolet photoemission spectra were recorded using a hemispherical electron spectrometer (Scienta SES 100, resolution 120 meV) at the end station SurICat (beamline PM4) of the synchrotron light source BESSY II, consisting of an interconnected sample preparation and analysis chamber in ultrahigh vacuum (UHV), with a base pressure of 10−8 and 10−10 mbar, respectively. The excitation energy was 35 eV, and for the measurement of the secondary electron cut-off, the sample was biased at –10 V to clear the analyzer workfunction. Optical Spectroscopy: For the basic characterization of the absorption and the PL behavior, a two-beam Perkin−Elmer Lambda 900 spectrometer and a Shimadzu RF-5301 PL spectro-fluorophotometer were used. All the PL spectra were corrected with the detector-specific sensitivity curve. Device Fabrication/Characterization: ITO-covered glass substrates were cleaned in acetone, toluene, and isopropanol and subsequently exposed to oxygen plasma. PEDOT:PSS (H.C. Stark Al 4083) layers were spin-coated under ambient conditions and dried at 120 °C under an inert atmosphere. Each copolymer was dissolved in toluene at a concentration of 3 g L−1 (CP-AE) and 4 g L−1 (CP-ABCD, CP-ACD) and spin-coated on top of the PEDOT:PSS layer, resulting in film thicknesses of about 20 and 60 nm, respectively. The CP-AE interlayer was thermally stabilized at 200 °C for 60 min under argon atmosphere whereas the subsequently coated CP-ABCD was dried at 70 °C for 60 min under high vacuum. All layer thicknesses were measured by atomic force microscopy (Veeco Dimension V and a Nanoscope V Controller). The multilayer cathode was thermally deposited in a vacuum coating unit at base pressures less than 10−6 mbar. The luminescence–voltage measurements were performed using a silicon photodiode and a computer-controlled Keithley 2612 source measurement unit. Spectral characterization was done with a LOT-ORIEL Multispec equipped with a DB 401-UV charge-coupled device (CCD) camera from Andor. Supporting Information is available from the Wiley Online Library or from the author. The authors would like to thank Anna Hayer and Roman Tratting for fruitful discussions and support. For financial support, the Styrian government (Project POLYLED, GZ:A3-22.B-9/2009-9) is acknowledged. As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.