ARTICLE pubs.acs.org/JPCC
Hole Injection Enhancement by a WO3 Interlayer in Inverted Organic Light-Emitting Diodes and Their Interfacial Electronic Structures Yoon Hak Kim,†,‡ Soonnam Kwon,§ Jong Hoon Lee,† Soon Mi Park,† Young Mi Lee,† and Jeong Won Kim*,†,‡ †
Korea Research Institute of Standards and Science, 209 Gajeong-ro, Daejeon 305-340, Korea Department of Nano Science, University of Science and Technology, 217 Gajeong-ro, Daejeon 305-350, Korea § Department of Materials Chemistry, Sejong Campus, Korea University, 339-700, Korea ‡
bS Supporting Information ABSTRACT: The interfacial energy level alignment of hole injection layers with WO3 insertion in an inverted organic lightemitting diode structure and its influence on electroluminescence have been studied using in situ X-ray, ultraviolet photoelectron spectroscopy, and device measurements. The hole injection barrier for N,N0 -bis(1-naphthyl)-N,N0 -diphenyl-1,10 biphenyl-4,40 -diamine (NPB) upon Al deposition was estimated to be 1.37 eV, which makes an Al anode unfavorable for hole injection. With a thin WO3 layer deposited onto the NPB, the NPB highest occupied molecular orbital (HOMO) level bends dramatically to 0.15 eV below the Fermi level. This NPB HOMO level and WO3 conduction band close to the Fermi level form a charge generation layer without an interfacial chemical reaction. This is why the WO3 interlayer dramatically helps charge injection even with a low work-function Al metal anode. Indeed, the electroluminescence and current efficiency from this structural device were greatly enhanced by 3 orders of magnitude compared with that without a WO3 interlayer.
1. INTRODUCTION Organic light-emitting diodes (OLEDs) have proven their readiness for commercialization in terms of lifetime and efficiency. In accordance with emerging new technologies, enhancement of light efficiency and extension of application fields are required. Particularly inverted structures, in which electron injection occurs at the bottom and hole injection on top, show crucial advantages due to their easy integration with Si-based driving circuits for active matrix OLEDs as well as a large open area for brighter illumination.1,2 To get better performance and process reliability, usually a proper buffer layer for carrier injection is needed. As a buffer material, several kinds of metal oxides, such as MoO3, V2O5, SiO2, and WO3, for OLED applications have been successfully utilized aiming at efficient hole injection properties.29 In the last few years, there has been a broad consensus that such metal oxide films play a role in the formation of a charge generation layer between the organic transport layer and the metal oxide.1014 This is reasonable because usual metal oxides have a high ionization potential and natural oxygen vacancies, which make an n-type material. Among them, we chose 2 nm of WO3 between NPB [N, N0 -bis(1-naphthyl)-N,N0 -diphenyl-1,10 -biphenyl-4,40 -diamine] and Al films. The WO3 thin film has shown its excellent properties, r 2011 American Chemical Society
such as relatively good conductivity, high transparency, and good chemical stability. Although the most intensively used metal oxide, MoOx, has shown various oxidation states depending on the interlayer structures,810,1517 the chemical state of WO3 film on organic films and its change under reactive metal deposition remain unclear. As such interfacial chemical states of metal oxide and organic film may influence the interfacial electronic structures, the simultaneous measurement of energy level alignment at the very interface is highly required to pick up detailed evidence for hole injection enhancement mechanism. Moreover, whether reactive and low work-function metals, such as Al, can be used as a useful anode in inverted OLED structures is to be questioned. Here, the interfacial energy-level alignment and chemical reaction as a function of film coverage have been measured by using in situ ultraviolet and X-ray photoelectron spectroscopy (UPS and XPS). Careful measurement of the valence band at the Al/WO3/NPB interfaces reveals that the WO3 interlayer substantially bends an underlying molecular orbital (MO) level and induces a charge generation layer. From comparative device Received: November 22, 2010 Revised: January 25, 2011 Published: March 18, 2011 6599
dx.doi.org/10.1021/jp111128k | J. Phys. Chem. C 2011, 115, 6599–6604
The Journal of Physical Chemistry C
ARTICLE
measurement, the insertion of the WO3 layer between NPB and the Al anode greatly enhances hole current density and electroluminescence (EL) efficiency. This observation clearly validates the use of a WO3 interlayer for hole injection even with a low work-function metal anode for inverted OLED applications.
2. EXPERIMENTAL SECTION Two inverted OLED structures fabricated here were Al/NPB on an ITO substrate and Al/WO3/NPB on an ITO substrate, which were used as the anode, hole injection layer (HIL), and hole transport layer (HTL), respectively. For cleaning the substrate, ITO-coated glass underwent UV ozone treatment at 3 Torr for 10 min prior to insertion into a vacuum chamber. After UV ozone treatment, it was promptly inserted into a system that is made of two deposition chambers and an analysis chamber. Using this system, samples could be transferred from one to another without air exposure. The base pressures of two deposition chambers for organic and metallic materials were 5.0 109 and 2.0 109 Torr, respectively, and that of the analysis chamber was 3.0 1010 Torr. NPB films were deposited on an ITO-coated glass by thermal evaporation at pressures less than 5.0 108 Torr, and its typical deposition rate was 0.1 Å/s. WO3 and Al layers were prepared below 5.0 108 Torr by thermal evaporation of WO3 powder and Al wire, respectively, using two different water-cooled Kundsen cells. Their typical deposition rates were 0.4 and 0.2 Å/s, respectively. Each film thickness was measured by a calibrated quartz crystal microbalance. Spectral measurements were performed in an UHV system by using a hemispherical electron energy analyzer (SES-100, VG-SCIENTA) with a He I (21.2 eV) gas discharge lamp for UPS measurement. Samples were biased to 10 and 5 V for the investigation of the secondary cutoff region and near the Fermi level, respectively. Meanwhile, for XPS spectra, a Mg KR (1253.6 eV) radiation was used as an excitation source. The OLED devices were fabricated onto a glass substrate that was precoated with 150 nm thick ITO with a sheet resistance of 20 Ω/sq. The ITO glass was precleaned using a conventional solvent cleaning method. The ITO surface was cleaned again with an UV ozone treatment just before the fabrication of the device. The currentvoltage characteristics of the OLEDs were measured using a source measure unit (Keithley 236). The EL spectra, luminance, and CIE color coordination were measured using a spectroradiometer (Photo Research PR650). 3. RESULTS AND DISCUSSION 3.1. Normal Structure. Figure 1a shows the secondary cutoff regions for Al/NPB/ITO films to measure work-function changes. The work function of the initial ITO film (4.34 eV) is decreased to 4.2 eV upon the deposition of 2 nm of NPB film. It is decreased further to 3.94 eV after 0.3 nm of Al deposition. Additional Al deposition increases the work function again, and it approaches 4.2 eV, which is close to the value of pure Al metal itself.18 Figure 1b is the valence band spectra near the Fermi level for the same sequence of films as in Figure 1a. The highest occupied molecular orbital (HOMO) onset position of NPB/ ITO is located at 1.05 eV, and it is increased by up to 0.2 eV as Al is deposited. Because the hole injection barrier is the energy difference between the NPB HOMO and the Fermi level, the onset position movement away from the Fermi level means the increase of the hole injection barrier from the Al anode to the
Figure 1. UPS spectra of various thicknesses of Al on 2.0 nm of NPB/ ITO glass. The arrows in (b) indicate HOMO onset positions.
Figure 2. XPS core level spectra of (a) N 1s and (b) Al 2p for Al/NPB/ ITO films. Peak positions are noted by arrows.
NPB layer. Except for the NPB HOMO change in its position and intensity, there is no evidence for a new state evolution, which takes place on other Al on organic interfaces.19,20 Above 1 nm of Al, the metallic Fermi edge can be observed. Thus, for the low work-function metals, such as Al, the hole injection to NPB would not be favorable. Figure 2a shows the N 1s core level spectra of NPB film as the Al thickness on it increases. Upon the deposition of 0.3 nm of Al, the original N 1s peak of NPB at 400.0 eV is shifted to 400.2 eV, but no further change is observed except for a slight broadening as the Al thickness increases. This N 1s shift in terms of its direction and magnitude is almost the same as the HOMO level shift observed in Figure 1b. Figure 2b exhibits the Al 2p spectra of the Al films on NPB under the same circumstances as in Figure 2a. There are two peaks at 73.1 and 75.4 eV, which are attributed to Al metal and oxidized Al, respectively. As the Al thickness increases, the intensity of the peak at 73.1 eV increases while the other remains very low. This is because the species for the higher binding peak at 75.4 eV is 6600
dx.doi.org/10.1021/jp111128k |J. Phys. Chem. C 2011, 115, 6599–6604
The Journal of Physical Chemistry C
ARTICLE
Figure 3. UPS spectra of various thicknesses of Al and WO3 on 2.0 nm of NPB/ITO-coated glass. The arrows in (b) and (c) indicate deep NPB MO and HOMO onset positions, respectively.
located mainly at the interface of Al/NPB or deeper sites. Some fraction of Al atoms may react with the ITO substrate through uncovered channels of NPB film. But, even with a very thin Al layer, the majority of Al atoms remains intact. This is consistent with the facts that there is no N 1s peak at the low binding energy side in Figure 2a and no gap state formation in Figure 1b. On the basis of these two observations in the N 1s and Al 2p core levels for the Al/NPB, it is believed that the Al atoms do not strongly diffuse into the NPB film but interact with the NPB film partially at the interface only. 3.2. WO3 Insertion. Figure 3ac shows UPS spectra for the measurement of work-function changes and the valence band for Al/WO3/NPB films deposited sequentially on the ITO surface. Each spectrum is rescaled for the best view. In Figure 3a, the work function of a 2 nm film of NPB is 4.2 eV and is increased gradually up to 6.44 eV as the WO3 thickness increases. However, as the Al is added on to 2 nm of the WO3 film, the work function is dramatically decreased back. Finally, the work function of the 2 nm Al film is measured to be 4.3 eV. The typical valence band spectrum of NPB exhibits MO positions in Figure 3b and the HOMO onset position at 1.06 eV in Figure 3c, respectively, which is very similar to the result of Figure 1b. The slight difference of 0.09 eV is due to the difference in two initial ITO work functions. However, as the WO3 film is deposited onto it, the whole MO levels, including the HOMO onset position, are moved toward the Fermi level, losing their intensities. Because the behavior of peak shift and intensity decrease in the deep MO level at 9.0 eV upon the WO3 deposition in Figure 3b is very similar to the spectral change in Figure 3c, it is confirmed that the attenuated peak near the Fermi level originated from the NPB HOMO rather than from the 5d of the WO3x film.21 Otherwise, the intensity of the marked peaks in Figure 3c should increase. On the other hand, the valence band maximum of the WO3 film is located at 3.15 eV in Figure 3b, which has been reported before.10 At the WO3 film thickness of 2 nm, the NPB HOMO onset position is located at 0.15 eV, which is shifted dramatically by 0.91 eV from the original position. Thus, by the addition of a WO3 interlayer between NPB and Al, the hole injection barrier to the NPB film is decreased to 0.15 eV. Figure 4a is the series of N 1s core level spectra of the NPB film upon the deposition of WO3 and Al in sequence. The initial N 1s
binding energy of the NPB on the ITO substrate is 400.0 eV, and it is shifted to 399.6 eV upon the deposition of 0.3 nm of WO3. Further deposition of WO3 or Al does not induce a substantial shift any more, but a little intensity attenuation. This negative binding energy shift is opposite to that shown in Figure 2a, where the Al deposition on the NPB gives rise to the positive shift. The signs of these opposite N 1s shifts are consistent with the ones observed in the directions of NPB HOMO level shifts in Figures 1b and 3c. Figure 4b shows the W 4f core level spectra of the WO3 film deposited on the NPB film and its changes upon Al deposition. The sharp doublet feature is separated by 2.18 eV. The W 4f7/2 binding energy of the WO3 films on the NPB is 36.2 eV, which corresponds to W6þ in the stoichiometric WO3 film.21,22 Its position and shape do not change as the WO3 film thickness increases. This indicates that the interfacial chemical reaction of the WO3 film and underlying NPB is negligible. This is also consistent with Figure 4a, where the N 1s does not give a new peak either by the WO3 deposition. However, the deposition of Al onto the WO3 film gives a strong change in that a series of reduced W atomic states (WO3x) appear on the low binding energy side of the W 4f spectra. Such a strong reaction between WO3 and Al keeps going as the Al thickness increases. Figure 4c shows the Al 2p spectra for the Al films deposited on WO3. At the Al thickness of 0.3 nm, the Al 2p shows the single binding energy of 75.1 eV, which is attributed to a fully oxidized state. Its intensity increases as the Al film thickness increases. At the Al/ WO3 interface, W and Al atoms exchange O atoms by the strong redox reaction. Only above 1 nm of Al film does a metallic Al 2p appear at 73.1 eV, and its intensity increases hereafter with the Al thickness. Thus, at the Al/WO3 interface, it is believed that there exists a transition layer that is composed of reduced W atoms and oxidized Al atoms. Figure 5 illustrates energy level diagrams for Al/NPB/ITO and Al/WO3/NPB/ITO interfaces. These diagrams are summarized results from the above Figures 1 and 3. The hole injection energy barrier from Al to the HTL (NPB) is increased by 0.22 eV in the case of the direct contact with Al. However, the hole injection barrier for the WO3-inserted structure is dramatically decreased to 0.15 eV. From this diagram, it is believed that the hole injection barrier from Al to NPB can be reduced by the 6601
dx.doi.org/10.1021/jp111128k |J. Phys. Chem. C 2011, 115, 6599–6604
The Journal of Physical Chemistry C
ARTICLE
Figure 4. XPS core level spectra of (a) N 1s, (b) W 4f, and (c) Al 2p on Al/WO3/NPB on ITO glass.
Figure 5. Comparison of energy level diagrams of NPB/Al and NPB/ WO3/Al interfaces. The insertion of the WO3 film dramatically bends the NPB HOMO, forming a charge generation layer.
insertion of high work-function material of WO3 film. This apparent change by the WO3 insertion will bring a high efficiency in hole injection for an inverted OLED. However, the situation is only true when the WO3 layer is thin enough, because the holes should tunnel through WO3. In fact, as the valence band maximum of WO3 is located at 3.15 eV, there is no allowable state in the WO3 film. As mentioned before, we have not observed any evidence of a gap state or a partially filled W 5d state. Instead, there is a strong band bending at the WO3/NPB interface, which makes the NPB HOMO level approach the WO3 conduction band. Assuming that the band gap of WO3 is 3.3 eV,10 the energy gap between the NPB HOMO and the WO3 conduction band minimum is only 0.3 eV. Thus, it requires a very small amount of energy for the NPB HOMO electrons to transfer to the WO3 conduction band. In the regime of a relatively thick WO3 film, the charge generation may primarily contribute to the current density enhancement. 3.3. Device Performance. Finally, we measured the current density and luminance from two different film structures, as in Figure 6. The current density measurement shows an apparent increase with the WO3 interlayer between Al and NPB. The turn-on voltage of the device with the structure Al(60)/ NPB(50)/Alq3(50)/Al(2 nm)/ITO is about 15 V. However, by the insertion of 3 nm of WO3, its value is remarkably
Figure 6. Comparison of device performance with (close circles) and without (open circles) the WO3 interlayer between Al and NPB: current density, current efficiency, and luminance. The legend shows the exact film structures used for these measurements.
decreased to 7 V. The current efficiency is also enhanced from 7 103 to 3 cd/A, which is the improvement by 3 orders of magnitude. The EL measured on both devices gives a dramatic difference as well. As a result, the improvement of the device performance by the WO3 interlayer is fully confirmed from this comparison, which is due to the large reduction of the hole injection barrier. 3.4. Hole Injection Enhancement Mechanism. The hole injection from the Al anode to NPB determines the current density measured in Figure 6. The high current density, low turnon voltage, and high EL efficiency by the insertion of the WO3 thin film show a strong merit for inverted OLEDs. This property of the WO3 film, like other metal oxides, is apparent in that the high work function of the WO3 interlayer induces a strong 6602
dx.doi.org/10.1021/jp111128k |J. Phys. Chem. C 2011, 115, 6599–6604
The Journal of Physical Chemistry C interface dipole by charge donation from an organic layer and gives a low hole injection barrier. Such a barrier-lowering mechanism has already been proposed based on MoO3 or MoO2 insertion from many other studies.5,911,15 The MoO3 usually gives oxygen vacancies by partial reduction of Mo atoms (MoO3x) as is deposited or once in contact with organic materials. This phenomenon is followed by the appearance of partially filled Mo 4d orbital states, giving rise to a gap state formation and n-doping.8,9 Surprisingly, this does not affect much on the interfacial dipole, meaning that oxygen vacancy formation has little to do with work-function change,8 or the device with the MoO3 interlayer is insensitive to air contamination.16,23 However, the WO3 film can be prepared relatively in its stoichiometric state,23 without substantial reaction with organic materials, such as NPB, as shown in Figure 4b. As a pure WO3 state in the absence of a gap state, it helps enhance hole injection, but the reduced form of WO3x does not.21 Thus, the partial reduction of WO3 and filling of W 5d orbital states is not likely a general mechanism for hole current enhancement. Rather than that, namely, a charge generation mechanism is more appreciable10,14 where the lowest unoccupied electronic state of the WO3 film is very close to EF so that the electrons from the NPB HOMO can be easily transferred to it at the interface. Through this charge generation layer, holes can be transported back in other direction. This charge separation mechanism can be applied exactly in the present inverted OLED structure. However, as the position of the NPB HOMO is very close to EF, the direct hole transfer from Al to NPB is not ignorable as long as the WO3 is thin enough. Indeed, the optimum thickness of the WO3 interlayer lies between 0.5 and 5 nm,21,2325 which might be thin enough for holes to tunnel through at a high field. In the case of the MO3 film, the optimum thickness of 0.75 nm is suggested for best performance.15 In the present data, the NPB HOMO level bending and work-function change upon the WO3 deposition on NPB look saturated at between 1 and 2 nm of thickness in Figure 3. At a higher thickness, charge transport may be limited by intrinsic metal oxide resistance and the charge generation rate. Thus, at a thin and optimum thickness, direct charge injection by tunneling might contribute to current density. Another important view is the opportunity to use the top Al anode because Al is a very common material widely used in the industrial field. Using a high work-function anode material, such as Au, the NPB HOMO level could also bend upward so that the hole injection barrier is lowered. Such a situation is exhibited in Figure S1 in the Supporting Information. Thus, in the case of the Au top anode, the advantage of the WO3 insertion layer no longer prevails. However, an expensive noble metal, such as Au, is not useful in a practical point of view.
4. CONCLUSION We have measured XPS and UPS for the studies in interfacial energy level alignment and chemical reactions on inverted OLED structures with and without a WO3 interlayer. The effect of WO3 insertion on the OLED device performances has been proven by the luminancecurrentvoltage measurements. In inverted OLED structures, we recognized that direct contact of Al metal to NPB is not a good choice. By the insertion of a thin WO3 film (HIL) as a buffer layer between NPB and Al, the improved device performance was achieved as expected from energy level alignment data. The major role of the WO3 film is to help charge transport, forming a charge generation layer between NPB (HTL) and WO3 (HIL). Although it makes a transition interlayer complex composed of AlWOx by the reaction with Al, the
ARTICLE
enhancement of current density and EL efficiency could be achieved. This is a clear demonstration that a low work-function and reactive metal, such as Al, can be applied to inverted OLEDs as long as a proper buffer layer is implemented.
’ ASSOCIATED CONTENT
bS
Supporting Information. Energy level diagrams for ITO/NPB/Au and ITO/NPB/WO3/Au interfaces. This material is available free of charge via the Internet at http://pubs.acs. org.
’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected].
’ ACKNOWLEDGMENT This work is supported by the Future-based Technology Development Program (Nano Fields) through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (2008-2004406, 20100023310). ’ REFERENCES (1) Gu, G.; Bulovic, V.; Burrows, P. E.; Forrest, S. R.; Thompson, M. E. Appl. Phys. Lett. 1996, 68, 2606. (2) Meyer, J.; Winkler, T.; Hamwi, S.; Schmale, S.; Johannes, H.-H.; Weimann, T.; Hinze, P.; Kowlasky, W.; Riedl, T. Adv. Mater. 2008, 20, 3839. (3) Ding, X. M.; Hung, L. M.; Cheng, L. F.; Deng, Z. B.; Hou, X. Y.; Lee, C. S.; Lee, S. T. Appl. Phys. Lett. 2000, 76, 2704. (4) Meyer, J.; Hamwi, S.; Schmale, S.; Winkler, T.; Johannes, H.-H.; Riedl, T.; Kowalsky, W. J. Mater. Chem. 2009, 19, 702. (5) You, H.; Dai, Y.; Zhang, Z.; Ma, D. J. Appl. Phys. 2007, 101, 026105. (6) Chu, T.-Y.; Chen, J.-F.; Chen, S.-Y.; Chen, C.-J.; Chen, C. H. Appl. Phys. Lett. 2006, 89, 053503. (7) Zhu, X. L.; Sun, J. X.; Peng, H. J.; Meng, Z. G.; Wong, M.; Kwoka, H. S. Appl. Phys. Lett. 2005, 87, 153508. (8) Kanai, K.; Koizumi, K.; Ouchi, S.; Tsukamoto, Y.; Sakanoue, K.; Ouchi, Y.; Seki, K. Org. Electron. 2010, 11, 188. (9) Lee, H.; Cho, S. W.; Han, K.; Jeon, P. E.; Whang, C.-N.; Jeong, K.; Cho, K.; Yi, Y. Appl. Phys. Lett. 2008, 93, 043308. (10) Kr€ oger, M.; Hamwi, S.; Meyer, J.; Riedl, T.; Kowalsky, W.; Kahn, A. Appl. Phys. Lett. 2009, 95, 123301. (11) Yi, Y.; Cho, S. W.; Kang, S. J. Appl. Phys. Lett. 2010, 97, 016101. (12) Hamwi, S.; Meyer, J.; Kr€oger, M.; Winkler, T.; Witte, M.; Riedl, T.; Kahn, A.; Kowalsky, W. Adv. Funct. Mater. 2010, 20, 1762. (13) Meyer, J.; Kr€oger, M.; Hamwi, S.; Gnam, F.; Riedl, T.; Kowalsky, W.; Kahn, A. Appl. Phys. Lett. 2010, 96, 193302. (14) Qi, X.; Li, N.; Forrest, S. R. J. Appl. Phys. 2010, 107, 014514. (15) Matsushima, T.; Murata, H. Appl. Phys. Lett. 2009, 95, 203306. (16) Meyer, J.; Shu, A.; Kr€oger, M.; Kahn, A. Appl. Phys. Lett. 2010, 96, 133308. (17) Kumaki, D.; Umeda, T.; Tokito, S. Appl. Phys. Lett. 2008, 92, 013301. (18) Kajita, S.; Nakayama, T.; Yamauchi, J. J. Phys.: Conf. Ser. 2006, 29, 120. (19) Lee, Y. M.; Park, Y.; Yi, Y.; Kim, J. W. Appl. Phys. Lett. 2008, 93, 123301. (20) Park, S. M.; Kim, Y. H.; Yi, Y.; Oh, H.-Y.; Kim, J. W. Appl. Phys. Lett. 2010, 97, 063308. 6603
dx.doi.org/10.1021/jp111128k |J. Phys. Chem. C 2011, 115, 6599–6604
The Journal of Physical Chemistry C
ARTICLE
(21) Son, M. J.; Kim, S.; Kwon, S.; Kim, J. W. Org. Electron. 2009, 10, 637. (22) Bussolotti, F.; Lozzi, L.; Passacantando, M.; Rosa, S. L.; Santucci, S.; Ottaviano, L. Surf. Sci. 2003, 538, 113. (23) Cheung, C. H.; Song, W. J.; So, S. K. Org. Electron. 2010, 11, 89. (24) Li, J.; Yahiro, M.; Ishida, K.; Yamada, H.; Matsushige, K. Synth. Met. 2005, 151, 141. (25) Meyer, J.; Hamwi, S.; Bulow, T.; Johannes, H.-H.; Riedl, T.; Kowalsky, W. Appl. Phys. Lett. 2007, 91, 113506.
6604
dx.doi.org/10.1021/jp111128k |J. Phys. Chem. C 2011, 115, 6599–6604
