Diimide nanoclusters play hole trapping and electron injection roles in organic light-emitting devices

Abstract

We report thermally stable diimide nanoclusters that could potentially replace the conventional thick electron transport layer (ETL) in organic light-emitting devices (OLEDs). Bis-[1,10]phenanthrolin-5-yl-bicyclo[2.2.2]oct-7-ene-2,3,5,6-tetracarboxylic diimide (Bphen-BCDI) was synthesized from the corresponding dianhydride and amine moieties, and its purified product exhibited a high glass transition temperature (232 °C) and a wide band gap (3.8 eV). The Bphen-BCDI subnanolayers deposited on substrates were found to form organic nanoclusters, not a conventional layer. The OLED made with a subnanolayer of Bphen-BCDI nanoclusters, instead of a conventional ETL, showed greatly improved efficiency (about 2-fold) compared with an OLED without the diimide nanoclusters. The role of the BPhen-BCDI nanoclusters was assigned to hole trapping and electron injection in the present OLED structure.

---

Introduction

Organic light-emitting devices (OLEDs) have been extensively studied for the last two decades, and small size OLED displays have been commercialized for use in mobile phones, MP3 players, car audio/controllers etc.^1–4 However, in terms of large size OLED displays for digital television and/or computer monitors, progress in commercialization remains slow because of a variety of technical issues, including difficulty in controlling uniformity over large areas and a lack of effective heat extraction from the active area (millions of pixels).^1,2Heat generation in OLEDs is an inevitable issue because they are operated by current-driving, which can lead to catastrophic damage in an OLED display. This heating effect is roughly proportional to the display size because of the integration of heat from each pixel. Although an improved efficiency by utilizing triplet emission could reduce heat generation in part, it is theoretically impossible to completely prevent it.^1–4 Hence, one of the most realistic possibilities is to invent organic semiconducting materials with an intrinsically high thermal stability.^2,3

On this account, a lot of organic semiconducting materials with high glass transition temperatures (T_g) have been developed, and some are actually used in commercial OLED displays.^5–10 Most of this research was based on the well-known multilayer device structure in which organic layers are composed of a hole injection layer (HIL), a hole transport layer (HTL), an emission layer (EML), a hole blocking layer (HBL), and an electron transport layer (ETL).^2,3 However, when it comes to the mass production of OLED displays, reducing the number of and/or thickness of the organic layers is of crucial importance for cost-reduction in order to compete with conventional liquid crystal displays (LCDs). To achieve this goal, a new concept is needed, without a big shift from the existing multilayer structure.

In this work, as the first step for a new concept, we have set our sights on the ETL in the multilayer structure. To date, tris(8-hydroxyquinolinato)aluminium (Alq₃) has been widely used, though a couple of new materials have been recently reported which improve on the electron mobility of Alq₃.^11,12 However, in order to achieve the desired performance level, the thickness of a conventional ETL should be typically 20–50 nm, which gives rise to an extension of process time. For example, the process time increases by 100 s for a 50 nm thick ETL when the deposition rate is 0.5 nm s⁻¹. Hence, we have invented a new method which features organic subnanolayers (<1 nm thick) that replace the conventional ETL. The organic subnanolayers were prepared by thermal evaporation of bis-[1,10]phenanthrolin-5-yl-bicyclo[2,2,2]oct-7-ene-2,3,5,6-tetracarboxylic diimide (Bphen-BCDI) that was synthesized from corresponding dianhydride and amine moieties. Results showed that the deposited Bphen-BCDI molecules formed nanoclusters (not a layer), and the incorporation of Bphen-BCDI subnanolayer (0.5 nm thick), instead of Alq₃, greatly improved the efficiency of OLEDs (about 2-fold) compared with a device without the Bphen-BCDI subnanolayer.

Results and discussion

As depicted in Scheme 1, the Bphen-BCDI compound was synthesized from [1,10]phenanthrolin-5-ylamine (PTA) and bicyclo[2,2,2]oct-7-ene-2,3,5,6-tetracarboxylic dianhydride (BCDA). A two step reaction was carried out in order to obtain a high yield: (1) Formation of bis-[1,10]phenanthrolin-5-yl-bicyclo[2,2,2]oct-7-ene-2,3,5,6-tetracarboxylic diacid (Bphen-BCDA) in the presence of the base catalyst triethylamine (TEA) at a low temperature; (2) Thermal conversion (imidization) of Bphen-BCDA to Bphen-BCDI at a reflux temperature of the solvent.¹³ To increase the purity of Bphen-BCDI, the final product powders were subjected to vacuum sublimation after drying.¹⁴ The purified Bphen-BCDI was a white powder. A computer simulation result showed that the energy minimized 3D structure of the Bphen-BCDI molecule is a twisted W-shape (Scheme 1 and ESI† Fig. S1).

Scheme 1 Synthesis of Bphen-BCDIvia its precursor (Bphen-BCDA) from the corresponding amine (PTA) and dianhydride (BCDA): The intermediate state related with TEA is not shown here. The energy minimized 3D structure is shown in the bottom-right corner.

The thermal stability of Bphen-BCDI (purified by vacuum sublimation) was examined using differential scanning calorimetry (DSC) and thermogravimetric analysis (TGA). As shown in Fig. 1a, the T_g (onset) of Bphen-BCDI was measured at around 232 °C, which is higher than that of Alq₃ (∼173 °C).^2,15 However, its melting point was not measured, though the temperature was increased up to 375 °C (see ESI† Fig. S2). The initial thermal decomposition (1 wt% loss) temperature of Bphen-BCDI powders was measured as ∼310 °C under nitrogen flow. These two results prove that Bphen-BCDI has outstanding thermal stability.

Fig. 1 Thermal, optical, photoelectronic and electrochemical characteristics of Bphen-BCDI: (a) Thermograms of DSC and TGA (inset: weight loss (L_W)), (b) OD and PL spectra of the thin films, and (c) PE yield and CV.

The optical absorption and emission spectra of Bphen-BCDI thin films are shown in Fig. 1b. The absorption onset was observed at 325 nm (see OD in Fig. 2b), giving an optical band gap energy of 3.8 eV. Upon excitation at a wavelength of 270 nm, the Bphen-BCDI film showed blue photoluminescence (PL), indicating that the twisted W-shape structure might prevent the Bphen-BCDI molecules from packing closely with each other: We note that almost no PL was measured from the highly ordered and compactly stacked imides and/or polyimides owing to a strong tendency to make charge transfer complexes.¹⁶ This is also evidenced from the absence of any ground state complexes, monitored by extending the wavelength up to 1100 nm (ESI† Fig. S3).

Fig. 2 J–V–L characteristics of OLEDs with subnanolayers of Bphen-BCDI (filled squares: 0 nm, open circles: 0.3 nm, filled triangles-upward: 0.5 nm, and open triangles-downward: 0.7 nm): The flat energy band diagram of the OLEDs is shown in the inset of top panel, while the current efficiency (η_C) as a function of voltage is given in the inset of bottom panel.

Next, in order to know the highest occupied molecular orbital (HOMO) energy of Bphen-BCDI, a photoelectron spectrum was obtained, as displayed in Fig. 1c. From the onset point the HOMO energy level was calculated to be 5.7 eV, which was confirmed from the cyclic voltammetry (CV) measurement that exhibits the related oxidation potential at around 1.03 V. The oxidation peak at around 0.23 V is considered to have originated from the carbonyl groups (in the imide rings) whose excitation (band gap) energy might be higher than the present PE measurement energy level (>6.5 eV) such that its PE yield was not detected here. From the optical band gap and HOMO energy values, the lowest unoccupied molecular orbital (LUMO) energy level was calculated to be ∼1.9 eV.

Based on the obtained energy band information of Bphen-BCDI, we designed the layer structure of the OLEDs as shown in the inset diagram of Fig. 2 (top panel). Here, considering the high-lying LUMO level (1.9 eV) of Bphen-BCDI, which can disturb the smooth electron injection and/or transport from the cathode, we decided to control its thickness to below 1 nm (see the device structure in the ESI† Fig. S4). As seen from the current density–voltage (J–V) characteristics, the device without the Bphen-BCDI layer exhibited the highest current density at a fixed voltage (we note that the standard deviation of device data is about 5% on the basis of good pixels in our laboratory condition). This indicates that the presence of a subnanometer thick Bphen-BCDI layer did indeed impede the charge transport inside the devices. Here it is noteworthy that the current density marginally increased as the Bphen-BCDI layer became thicker, from 0.3 to 0.5 nm. A further increase to 0.7 nm reduced the current density again. This result implies the existence of an optimum thickness of Bphen-BCDI, which is quite similar to the influence of LiF layer thickness.^2,3,17

The thickness effect of the Bphen-BCDI layer was more pronounced in the luminance–voltage (L–V) characteristics. As shown in the bottom panel of Fig. 2, the device with the 0.5 nm thick Bphen-BCDI layer exhibited the highest luminance, a more than 2-fold increase at 14 V, compared to the device without the Bphen-BCDI layer. However, the devices with 0.3 or 0.7 nm thick Bphen-BCDI layers showed lower luminance than the device without the Bphen-BCDI layer. As a result, the device with the 0.5 nm thick Bphen-BCDI layer exhibited a current efficiency (η_C) of ∼18 cd A⁻¹ at 9.5 V (∼1500 cd m⁻²), which is about a 2-fold improvement compared to the device without the Bphen-BCDI layer (∼8.5 cd A⁻¹ at 9.5 V (∼950 cd m⁻²)) (see also the η_C–J relation in the ESI† Fig. S5). Here we note that this efficiency, achieved using the 0.5 nm thick Bphen-BCDI layer, is nearly within reach of our best efficiency of the standard control OLED with the Alq₃ ETL fabricated in our laboratory, suggesting a further improvement is obtained by controlling each layer.¹⁷ At a higher voltage (15 V), the difference of current efficiency between these two devices became much bigger (12 cd A⁻¹versus 3.3 cd A⁻¹). As expected, the devices with 0.3 or 0.7 nm thick Bphen-BCDI layers showed generally lower efficiency. The J–L relationship is given in the ESI† Fig. S6.

Considering the J–V and L–V characteristics according to the BPhen-BCDI thickness, we can assume that the BPhen-BCDI nanoclusters (at optimum thickness) play a dual role of electron injection (into the BAlq layer) and hole trapping (for hole carriers passed from the BAlq layer without blocking) in the present OLED structure. The improved electron injection is evidenced by the greatly (more than 2-fold) enhanced luminance as the luminance should be lowered if the electron injection was interrupted by the presence of the BPhen-BCDI nanoclusters (0.5 nm thick). This assumption further supports the evidence that the hole trapping phenomenon (of hole carriers leaked from the BAlq layer) is mainly responsible for the lowered current density (but slightly increased voltage) of the OLED with the 0.5 nm thick BPhen-BCDI layer (nanoclusters) compared to the control device. However, it is noteworthy that, as the BPhen-BCDI thickness increased, the electron injection became worse again (and the voltage increased) because the luminance was lower for the devices with the BPhen-BCDI layers than the control device.

Finally, we tried to understand how the Bphen-BCDI layer was formed on the surface of HBL (BAlq), and how this shape could result in such a high efficiency. As shown in the scanning electron microscope (SEM) images (Fig. 3a, b), it is observed that the thermally evaporated Bphen-BCDI molecules did not form a film (layer), but made nanoclusters with an apparent (horizontal) size of less than 100 nm. Further analysis using an Auger electron microscope (AEM) showed that the Auger signal spots for aluminium (Al) in the BAlq molecules were obviously reduced in a random fashion over the entire area, and not limited to the sites where the nanoclusters reside, after deposition of Bphen-BCDI (Fig. 3c, d). This indicates that some much smaller nanoclusters that are not detected with SEM were formed in the presence of the big nanoclusters. This is also evidenced from the carbon–AEM images which display quite a different carbon atom distribution after deposition of Bphen-BCDI. We note that the similar big nanoclusters (<100 nm in horizontal size) were detected from the carbon–AEM images (Fig. 3e, f). The detailed composition analysis data of samples with and without the Bphen-BCDI layer (0.5 nm) are given in the ESI† Fig. S7 and S8. The atomic force microscope (AFM) images detected much smaller nanoclusters, with a size of ∼50 nm or less, on the surface of Bphen-BCDI deposited samples (see circled parts in Fig. 3h). In particular, the phase mode AFM examination confirmed that the surface of BAlq was covered with the Bphen-BCDI nanoclusters (Fig. 3g, h, images on the right). The surface roughness analysis also showed that the 0.5 nm thick Bphen-BCDI deposition increased the surface roughness from 0.39 nm to 0.72 nm (ESI† Fig. S9).

Fig. 3 Surface images of the Bphen-BCDI (0.5 nm thick) coated sample (b,d,f,h: ITO/HIL/HTL/EML/BAlq/Bphen-BCDI) and control sample (a,c,e,g: ITO/HIL/HTL/EML/BAlq): (a,b) SEM images, (c,d) AEM images focusing on aluminium (Al_2/4) atoms, (e,f) AEM images focusing on carbon (C₁) atoms (arrows indicate big nanoclusters similar to those observed in SEM image (b)), and (g,h) AFM images (left: height mode; right: phase mode), where the arrow and circles represent a region of small associated nanoclusters.

Conclusions

Bphen-BCDI was prepared via its amic acid precursor (Bphen-BCDA) which was made from PTA and BCDA in solution. The purified (sublimed) Bphen-BCDI powders showed an excellent thermal stability (T_g = 232 °C), while the Bphen-BCDI thin film exhibited E_g = 3.8 eV, HOMO energy = 5.7 eV, and LUMO energy = 1.9 eV. Blue PL was measured from the Bphen-BCDI film, which was assigned to suppressed CT complex formation, due to less compact molecular packing owing to its twisted W-shaped structure. The OLED with the 0.5 nm thick (according to the calibrated thickness monitor) Bphen-BCDI layer exhibited ∼24 000 cd m⁻² at 15 V, which was a more than 2-fold higher luminance than the device without the Bphen-BCDI layer, while its current efficiency reached ∼18 cd A⁻¹ (a more than 2-fold higher value). However, other thicknesses of Bphen-BCDI resulted in lower performance. The improved device performance was attributed to the hole trapping and electron injection role of the 0.5 nm thick BPhen-BCDI layer (nanoclusters) in the present device structure. The electron and atomic microscope images revealed that the thermally evaporated Bphen-BCDI molecules formed nanoclusters with various sizes less than 100 nm, not a layer (film). Finally, we expect that the present approach using Bphen-BCDI nanoclusters can pave a way for the size reduction of all the surrounding organic layers except the key layer (i.e., EML) in OLEDs.

Experimental

Synthesis

Bphen-BCDI was synthesized in solution: 0.9307 g (3.75 mmol) of BCDA was dissolved in 50 mL of N,N-dimethylacetamide (DMAc) under a nitrogen atmosphere. A catalytic amount of TEA was added to the solution and then the solution was subject to vigorous stirring. Next, 1.680 g (8.625 mmol) of PTA was slowly added to the solution upon stirring at room temperature. After keeping the reaction at room temperature until no more change in solution colour was observed, the reaction temperature was increased up to the boiling point of DMAc. At this condition (160–170 °C) the reaction mixture was refluxed for 36 h, leading to a further change in the solution colour. To terminate the imidization reaction, the reaction mixture was cooled to room temperature. The precipitate in the solution was filtered off and washed with DMAc several times, followed by washing with alcohols (methanol/ethanol). The final solids were dried in a vacuum oven. The synthesis yield was ∼80%. In order to ensure increased purity, the dried product was purified using a vacuum sublimation technique at 300–320 °C (heating stage temperature: the temperature inside glass sublimation kit might be lower than this temperature) and ∼10⁻³ Torr (the exact pressure was not measured). The purified product (Bphen-BCDI, chemical formula: C₃₆H₂₂N₆O₄) was characterized using elemental analysis (EA, measured (calculated) atoms: C = 71.5% (71.8%); N = 13.8% (13.9%); H = 3.7% (3.7%); O = 11% (10.6%)), matrix–assisted laser desorption–time of flight–mass spectroscopy (MALDI-TOF-MASS, measured mass = 602.719; calculated mass = 602), Fourier transform-infrared spectroscopy (FT-IR), and Fourier transform-nuclear magnetic resonance spectroscopy (FT-NMR) (see ESI† Fig. S10–S13).

Thin film and device fabrication

Prior to the film and device fabrication, all substrates were subject to wet (acetone and isopropyl alcohol) and dry (UV-ozone) cleaning processes. To measure the optical and photoelectron spectra, the purified Bphen-BCDI powders were thermally evaporated onto quartz substrates or ITO-glass substrates, leading to ∼50 nm thick Bphen-BCDI film. The OLED devices were fabricated by sequential thermal deposition of 4,4′,4′′-tris(N-(2-naphthyl)-N-phenyl-amino)triphenylamine (2TNATA, 40 nm, HIL), N,N′-bis(naphthalen-1-yl)-N,N′-bis(phenyl)benzidine (NPB, 30 nm, HTL), 4,4′-bis(carbazol-9-yl)biphenyl (CBP) doped with 5% tris(2-phenylpyridine)iridium(III) [Ir(ppy)₃] (30 nm, EML), BAlq (10 nm, HBL), Bphen-BCDI (0–0.7 nm), LiF (1 nm, EIL), and Al (100 nm) using a thermal evaporation chamber. All devices were stored inside a nitrogen-filled glove box before measurement. For the measurement of AEM and AFM images, the samples with organic layers without the LiF and Al electrodes were used.

Measurements

The thermal transition behaviour of the Bphen-BCDI powders was measured using DSC (TA4000, TA Instruments), while the thermal stability (degradation) was examined using TGA (TG/DTA320, Seiko). The optical absorption and emission spectra were measured using a UV-visible spectrophotometer (Optizen 2120+, Mecasys) and fluorescence spectrometer (FP-6500, JASCO). The HOMO energy was measured using a photoelectron spectrometer (AC-2, Riken Keikki). The cyclic voltammograms were measured using a potentiostat (273A model, EG&G) at a sweep rate of 100 mV s⁻¹: The supporting electrolyte was 0.1 M TBAP (Tetrabutylammonium perchlorate) and the concentration of Bphen-BCDI was 1 × 10⁻³ M in chlorobenzene. A platinum (Pt) wire and disk were employed as a counter and a working electrode, respectively. A silver (Ag) wire was used as a reference electrode. The J–V–L characteristics of the OLEDs were measured using a specialized OLED measurement system equipped with an electrometer (Keithley 238) and a spectroradiometer (CS-1000S, Konica Minolta). The SEM and AEM images were measured using an Auger electron microscope equipped with an SEM module (PHI700, ULVAC-PHI), while the AFM images were measured using a scanning probe microscope (Nanoscope IIIa, Digital Instruments).

Acknowledgements

This work was financially supported by Korean government grants (Priority Research Center Program_2009-0093819, Pioneer Research Center Program_2010-0002231, NRF_20090072777, KETEP-2008-N-PV08-J-01-30202008, NRF_20100004164) and DGIST Basic Research Program.

Notes and references

1. C. W. Tang and S. A. VanSlyke, Appl. Phys. Lett., 1987, 51, 913.
2. Y. Kim and C. S. Ha, Advances in Organic Light-Emitting Device, Trans Tech Publications, Switzerland, 2008.
3. L. S. Hung and C. H. Chen, Mater. Sci. Eng., R, 2002, 39, 143.
4. M. A. Baldo, D. F. O'Brien, Y. You, A. Shoustikov, S. Sibley, M. E. Thompson and S. R. Forrest, Nature, 1998, 395, 151.
5. D. F. O'Brien, P. E. Burrows, S. R. Forrest, B. E. Koene, D. E. Loy and M. E. Thompson, Adv. Mater., 1998, 10, 1108.
6. Y. Kim, J. Lee and S. Kim, Adv. Mater., 1999, 11, 1463.
7. J. P. Chen, H. Tanabe, X. C. Li, T. Thoms, Y. Okamura and K. Ueno, Synth. Met., 2003, 132, 173.
8. J. Li, D. Liu, C. S. Li, H. L. Kwong and S. Lee, Chem. Mater., 2005, 17, 1208.
9. Y. Kim, K. Han and C. S. Ha, Macromolecules, 2002, 35, 8759.
10. Y. Shirota and H. Kageyama, Chem. Rev., 2007, 107, 953.
11. S. J. Su, T. Chiba, T. Takeda and J. Kido, Adv. Mater., 2008, 20, 2125.
12. H. Tsuji, C. Mitsui, T. Sato and E. Nakamura, Adv. Mater., 2009, 21, 3776.
13. D. Wilson, H. D. Stenzenberger and P. M. Hergenrother, Polyimides, Chapman and Hall, New York, 1990.
14. H. Klauk, Organic Electronics: Materials, Manufacturing and Applications, Wiley VCH, Germany, 2006.
15. Y. Kim, J. G. Lee, K. Han, H. K. Hwang, D. K. Choi, Y. Y. Jung, J. H. Keum, S. Kim, S. S. Park and W. B. Im, Thin Solid Films, 2000, 363, 263.
16. M. K. Ghosh and K. L. Mittal, Polyimides: Fundamentals and Applications, Marcel Dekker, Inc., New York, 1996.
17. Y. Kim, Nanotechnology, 2008, 19, 355207.

---

Footnote

† Electronic Supplementary Information (ESI) available: 3D structure of Bphen-BCDI, DSC thermogram, UV-visible spectrum, OLED structure and materials used, current efficiency of OLED as a function of current density, luminance as a function of current density, Auger electron spectra, 3D AFM images, FT-IR spectrum, MALDI-TOF-MASS spectrum, and NMR spectra of Bphen-BCDI. See DOI: 10.1039/c0nr00496k/

---