Vacuum-deposited, nonpolymeric flexible organic ... - OSA Publishing

172

OPTICS LETTERS / Vol. 22, No. 3 / February 1, 1997

Vacuum-deposited, nonpolymeric flexible organic light-emitting devices G. Gu, P. E. Burrows, S. Venkatesh, and S. R. Forrest Advanced Technology Center for Photonics and Optoelectronic Materials, Department of Electrical Engineering and Princeton Materials Institute, Princeton University, Princeton, New Jersey 08540

M. E. Thompson Department of Chemistry, University of Southern California, Los Angeles, California 90089 Received September 11, 1996 We demonstrate mechanically f lexible, organic light-emitting devices (OLED’s) based on the nonpolymetric thin-f ilm materials tris-(8-hydroxyquinoline) aluminum (Alq3) and N , N 0 -diphenyl-N , N 0 -bis(3-methylphenyl)110 biphenyl-4, 40 diamine (TPD). The single heterostructure is vacuum deposited upon a transparent, lightweight, thin plastic substrate precoated with a transparent, conducting indium tin oxide thin f ilm. The f lexible OLED performance is comparable with that of conventional OLED’s deposited upon glass substrates and does not deteriorate after repeated bending. The large-area (,1-cm2 ) devices can be bent without failure even after a permanent fold occurs if they are on the convex substrate surface or over a bend radius of ,0.5 cm if they are on the concave surface. Such devices are useful for ultralightweight, f lexible, and comfortable full-color f lat panel displays.  1997 Optical Society of America

Because of the widespread application of vacuumdeposited, small-molecule-based heterostructure organic light-emitting devices (OLED’s) to f lat panel displays,1 many research groups have studied their physics2,3 and their performance4,5 and have explored various device architectures and applications that are diff icult or impossible to realize with inorganic semiconductors.6,7 The devices are composed of thin organic thin films creates the potential to fabricate f lexible OLED’s, provided that appropriate substrates are available. These devices can be used for lightweight, portable, roll-up displays8 or conformable displays that can be readily attached to windows, windshields, or instrument panels. Polymer-based f lexible OLED’s have been demonstrated.8,9 In that research it was assumed that f lexibility is unique to polymers,9 and to date the potential of making f lexible OLED’s from small-molecule organic materials has not been explored to our knowledge. In this Letter we show that a conventional small-molecule-based OLED can be successfully grown on a nonconducting polymer substrate. This opens the potential for low-cost mass production of f lexible displays, including roll-to-roll processing. The ability to achieve a highly f lexible displays vacuum-deposited molecular organic materials depends on two factors: The molecular bonds responsible for the mechanical properties of the thin films comprising the OLED must be tolerant of the stress applied to the structure on bending, and the substrates must be suff iciently f lat and uniform that mechanical defects are not formed during growth or f lexing. In the first case, virtually all organic materials used in vacuum-deposited OLED’s are held together by highly f lexible van der Waals bonds. We showed that the bonding of aromatic molecules similar to those used in OLED’s is highly compressible.10 For example, we found10 that the compressibility of the van der Waals-bonded naphthalene-based molecu0146-9592/97/030172-03$10.00/0

lar crystal NTCDA has a roughly 20-times-higher compressibility than most ductile metals11 such as In or Al. Hence molecular materials should also be suff iciently ductile to undergo signif icant stress without cracking. The second consideration, that the substrates used are suff iciently f lat, is thus one of the issues addressed here. We demonstrate what we believe to be the first mechanically f lexible OLED based on the vacuumdeposited, nonpolymeric materials tris-(8-hydroxyquinoline) aluminum (Alq3) and N , N 0 -diphenyl-N , N 0 -bis(3-methylphenyl) 1-10 biphenyl-4, 40 diamine (TPD). Rather than using a thin polyaniline film as the f lexible transparent hole-injecting electrode,9 here we grow the OLED’s on indium tin oxide (ITO) contacts predeposited directly upon a plastic substrate. The predeposited ITO (Ref. 12) has both conductivity and transparency superior to polyaniline.9 We find that the devices grown on the plastic substrates have eff iciencies comparable with conventional vacuumdeposited OLED’s grown on glass and that they are mechanically robust. The inset of Fig. 1 shows the structure of the f lexible OLED. The substrate is a 175-mm-thick, transparent polyester sheet precoated with a transparent, conducting ITO thin film.12 The sheet resistance of the ITO thin film is 60 Vysquare, and the transparency of the coated substrate is ,80% throughout the visible spectrum.12 Because the total thickness of the organic films comprising the OLED heterostructure is only ,100 nm, we investigated the surface roughness of these substrates. Figure 2 shows an atomic force microscope image of the ITO (top) substrate surface [Fig. 2(a)] as well as of the polyester (bottom) substrate surface [Fig. 2(b)]. We find that the ITO surface has a rms roughness of only 1.8 nm, whereas the polyester surface is somewhat rougher, with a rms value of 2.8 nm. There is some variation from substrate to substrate, although we did not find ITO sur 1997 Optical Society of America

February 1, 1997 / Vol. 22, No. 3 / OPTICS LETTERS

Fig. 1. Current versus voltage characteristics of a 1-mmdiameter, vacuum-deposited, f lexible OLED before bending and after repeated bending and of a conventional OLED on glass. Inset, the structure of the vacuum-deposited f lexible OLED.

face regions with a rms roughness exceeding 3.6 nm. In either case the substrates are suff iciently smooth (i.e., the height of the surface features is a small fraction of the total device thickness) such that no signif icant damage should be incurred by the OLED heterostructure on growth or bending. Before the deposition of the organic films, the substrate is ultrasonically cleaned in detergent for 2 min, then rinsed with DI water. Next, it is rinsed in 2propanol at room temperature for 2–3 min, then boiled in 2-propanol, again for 2–3 min, followed by drying in filtered nitrogen. A 80-nm-thick layer of TPD is deposited by thermal evaporation in a vacuum of ,4 3 1027 Torr, followed by deposition of an 80-nmthick layer of Alq3. The top electrode consists of a

173

150-nm-thick layer of Mg –Ag and a 50-nm-thick Ag cap deposited through a shadow mask. A conventional device on an ITO-precoated glass substrate was simultaneously fabricated for comparison. The sheet resistance and transparency of the ITO-precoated glass substrate are 20 Vysquare and ,90%, respectively. Returning to Fig. 1, we show the current–voltage (I –V) characteristics of a 1-mm-diameter f lexible device before bending, after repeated (4–5 times) bending over a small radius of curvature (,0.5 cm), and the conventional device on a glass substrate. All the I –V curves follow the power-law dependence of current on voltage discussed previously.2,3 At lower voltages, I ~ V , indicating ohmic behavior, whereas at higher voltages, the curves follow I ~ V m11 with m ­ 7, suggestive of trap-limited conduction typical of OLED’s. The power-law dependence is observed for at least 4 orders of magnitude change in current in the high-current region. There is no obvious change in the I–V characteristics after the device is repeatedly f lexed. The turn-on voltages (the voltages at which the currents due to ohmic and trap-limited conduction are equal) of the three curves are almost identical (, 6.5 V), whereas the leakage current at low voltages of the f lexible device is less than that of even the conventional device and is not increased after bending. This indicates that the ITO film precoated upon the f lexible substrate is suff iciently uniform and that the organic layers are suff iciently compressible such that current-shunt paths between the top and bottom contacts are not induced after bending, even for very thin-f ilm (,160-nm) molecular organic heterostructures. The light output power versus current characteristics of the f lexible device before and after bending and of the conventional device are shown in Fig. 3. The external quantum eff iciency of the f lexible device is

Fig. 2. Atomic force microscope (AFM) images of a typical ITO-coated polyester substrate film: surface, ( b) polyester ( bottom) substrate surface.

(a) ITO (top) substrate

174

OPTICS LETTERS / Vol. 22, No. 3 / February 1, 1997

Fig. 3. Light output power versus current characteristics of a 1-mm-diameter, vacuum-deposited, f lexible OLED before bending and after repeated bending and of a conventional OLED on glass.

area device were also studied. If on the convex side of a curved substrate, the device can be bent without failure even after a permanent fold occurs in the polyester film. If on the concave side the device remains operational when bent over a radius of curvature down to 0.5 cm. At smaller radii, cracks propagate through the device, and current-shunt paths are created between bottom and top contacts after further bending. When ITO-precoated substrates are similarly bent, the same cracking phenomenon is observed, from which we infer that the cracks occur in the ITO rather than in the OLED itself. In conclusion, we have fabricated a vacuumdeposited, van der Walls molecular, nonpolymeric f lexible OLED, using an ITO-precoated transparent polyester film as the substrate. We have shown that an ITO thin film, when precoated on a pliable substrate, provides a f lat, highly transparent, conductive, f lexible contact suitable for OLED applications. This hole-injecting ITO-coated substrate should work equally well with polymeric OLED’s. In addition, performance similar to that achieved in this Letter is expected if nonpolymeric devices are vacuum deposited on polymeric, transparent hole-injecting contacts such as polyaniline, which can be used if even greater f lexibility is required in certain applications. The authors thank Universal Display Corporation and DARPA. References

Fig. 4. Photograph of an array of nine unpackaged 1 cm2 vacuum-deposited, nonpolymeric f lexible OLED’s. One device in contact the probe arm is operating in air in a well-illuminated room at normal video display brightness (,100 cdym2 ).

0.20% and that of the conventional device is 0.14%. In both cases the efficiency is determined from light emitted only in the forward scattering direction, which considerably underestimates the actual external quantum eff iciency but is nevertheless useful for comparing devices.13 Once again, we find that the quantum efficiency of the f lexible device is not affected by repeated bending. The fact that there is no appreciable change in either the I –V or light output power versus current characteristics after the device is f lexed indicates that the ITO contact, the organic layers, and the alloy top contact are not significantly affected by bending even over a small radius of curvature. Large-area (,1 cm2 ) devices, also fabricated by similar methods, are shown in Fig. 4. As for the smaller devices, the large devices can also be bent over radii of ,0.5 cm without apparent degradation. That these larger areas can be achieved suggests that f lexible OLED’s can be used in large, roll-up, or conformable f lat panel displays. This, along with the fact that the ITO-precoated substrate is available on large spools,12 suggests that f lexible, OLED-based displays can be mass manufactured on a roll-to-roll basis by use of suitable volume growth technologies such as organic vapor phase deposition.14 Failure modes of the large-

1. S. R. Forrest, P. E. Burrows, and M. E. Thompson, Laser Focus World 31,(2) (1995). 2. P. E. Burrows and S. R. Forrest, Appl. Phys. Lett. 64, 2285 (1993). 3. P. E. Burrows, Z. Shen, V. Bulovic, D. M. McCarty, S. R. Forrest, J. A. Cronin, and M. E. Thompson, J. Appl. Phys. 79, 7991 (1996). 4. C. Tang, S. VanSlyke, and C. Chen, J. Appl. Phys. 65, 3610 (1989). 5. T. Wakimoto, R. Murayama, K. Nagayama, Y. Okuda, H. Nakada, and T. Tohma, SID 96 Digest (Society for Information Display, Santa Ana, Calif., 1996), p. 894. 6. G. Gu, V. Bulovic, P. E. Burrows, S. R. Forrest, and M. E. Thompson, Appl. Phys. Lett. 68, 2606 (1996). 7. P. E. Burrows, S. R. Forrest, S. P. Sibley, and M. E. Thompson, ‘‘Transparent light-emitting devices,’’ submitted to Nature, (London). 8. P. Yam, Sci. Am. 273, 83 (1995). 9. G. Gustaffson, G. M. Treacy, Y. Cao, F. Klavertter, N. Colaneri, and A. J. Heeger, Synth. Metals 57, 4123 (1993). 10. Y. Zhang and S. R. Forrest, Phys. Rev. Lett. 71, 2765 (1993). 11. C. Kittel, Solid State Physics, 4th ed. (Wiley, New York, 1971), p. 143. 12. Southwall Technologies, Inc., 1029 Corporation Way, Palo Alto, Calif., 94303, Part No. 903-6011. 13. D. Z. Garbuzov, S. R. Forrest, A. G. Tsekoun, P. E. Burrows, V. Bulovic, and M. E. Thompson, J. Appl. Phys. 80, 4644 (1996). 14. P. E. Burrows, S. R. Forrest, L. S. Sapochak, P. Fenter, T. Buma, V. S. Ban, and J. L. Forrest, J. Cryst. Growth 156, 91 (1995).