Properties and characterization of carbon-nanotube-based transparent conductive coating

Transcription

1 Properties and characterization of carbon-nanotube-based transparent conductive coating C. M. Trottier P. Glatkowski P. Wallis J. Luo Abstract Transparent and electrically conductive coatings and films have a variety of fast-growing applications ranging from window glass to flat-panel displays. These mainly include semiconductive metal oxides such as indium tin oxide (ITO) and polymers such as poly(3,4-ethylenedioxythiophene) doped and stabilized with poly(styrenesulfonate) (PEDOT/PSS). In this paper, we show alternatives to ITO and conducting polymers, using single-wall carbon nanotubes (SWNT). These CNT-based technologies offer conducting substrates having a broad range of conductivity, excellent transparency, neutral color tone, good adhesion, abrasion resistance, and flexibility. Additional benefits include ease of both processing and patterning. This paper reports on optoelectronic properties and structure characterization of these materials. Keywords Single-wall carbon nanotube (SWCNT), transparent conductor, light transmittance, sheet resistance films, coatings, display applications. 1 Introduction Transparent conductors are an essential component in many optoelectronic devices, including LCDs, OLEDs, touch screens, and photovoltaics. Indium tin oxide (ITO) has been the preferred choice for four decades.1 The author and co-workers have discovered that highly transparent conductive films can be formed when carbon-nanotube (CNT) dispersions are applied at a thickness of <100 nm.2 T h e optoelectronic properties improve dramatically with increased CNT purity and degree of dispersion. Today, the highest-quality CNT films result in 90 97% visible-light transmittance and Ω/䊐 sheet resistance very close to the optoelectronic performance of sputtered ITO and suitable for many flat-panel-display applications (Invisicon, Eikos, Ink.). Highly transparent and conductive coatings are formed by applying specially formulated, purified SWCNT dispersions onto polyethylele terephthalate (PET) or glass substrates. The CNT coating is protected by the addition of polymer binders. The selection of polymer binders is critical for maximizing the optoelectronic performance of the coating. The binders also optimize adhesion, abrasion resistance, and flexibility. Characterization of the CNT coating thickness and structures presents significant challenges. Microscopic methods, including AFM, SEM, and TEM, are used to provide some basic guidance for understanding the structure property relationship and aid in future product FIGURE 1 CNT raw material: TEM image of CNT powder prior to purification (left) and after purification (right). The authors are with Eikos, Inc., 2 Master Dr., Franklin, MA 02038; mtrottier@eikos.com. Copyright 2005 Society for Information Display /05/ $1.00 Journal of the SID 13/9,

2 FIGURE 2 FESEM micrograph of CNT film at 250 Ω/. design. This new category of transparent conductor has remarkable potential for versatile applications in areas including (but not limited to) flat-panel displays, touch screens, flexible displays, and ESD coatings. 2 CNT substrate materials A purification process was employed to prepare CNT dispersions. In this process, arc-discharge-produced CNT powder (soot) is acid refluxed to separate the amorphous carbon, metal catalyst, and other contaminates from the carbon nanotubes. The result of the purification process is shown in Fig. 1. These micrographs were produced using a JEOL 1200EX transmission electron microscope (TEM) with digital image capture and illustrates the efficiency of removing contamination from SWCNT. 3 Morphology FIGURE 3 CNT film (250 Ω/ ): SPM image (2-µm scan size) showing morphology of CNT film. RMS roughness, ~6.8 nm. To observe the morphology of CNT coatings, CNT ink (Invisicon, Eikos,Inc.)wassprayedontoa175-µm PET sheet to a sheet resistance of 250 Ω/. The open network produced is shown in Fig. 2 using a Field Emission Amray 3600 scanning electron microscope (SEM) and scanning probe microscopy (SPM) using a Digital Instruments NanoScope with 2-µm scansize(fig.3).themicrographs show a propensity of 1 2-nm-diameter SWCNTs to rope together to form a continuous network. Although, the length of an individual rope cannot be determined, the bundledrope diameter is approximately nm. As a comparison, ellipsometry analysis measured a nominal coating thickness of approximately 30 nm and an RMS surface roughness of 7 nm for a 500-Ω/ coating. These ropes or bundles are intertwined with 2-D orientation on the thin-film plane to form a relatively dense layer with open interstices. Some impurities, mainly residual carbon catalyst shells from the CNT manufacturing process, can still be observed in the applied film. With further continuous improvement in purification and selective separation of metallic and semiconductivetype CNT, one can expect that optoelectronic performance of CNT electrodes will eventually surpass sputtered ITO. 4 Opto-electronic performance Three different transparent conductor materials were evaluated: (1) ITO sputtered onto glass substrate and PET substrates, (2) poly(3,4-ethylenedioxythiophene) poly(styrenesulfonate) wet-coated onto polyester film substrate (PE- DOT/PSS), and (3) CNT spray coated onto glass and PET substrates. The ITO films were sputtered at a 30-nm (±10 nm) thickness onto 0.7-mm OA10 NIHON-DENKI- GLASS non-alkali glass substrate (similar to Corning 1737). The sheet resistance of the ITO layer was approximately 200 Ω/. Note that a proprietary antireflective coating was used to optimize visible-light transmittance. ITO films sputtered onto 175-µm PET, however, contained no antireflective coating. The PEDOT/PSS 3 films were wet-coated (Meyer rod method) to 24 µm (wet film thickness) by Bayer onto 175-µm polyester (PET) film. The sheet resistance of the PEDOT layer was approximately 250 Ω/. ThisPEDOT sample is representative of a high conductivity grade (BAYTRON FHC). The CNT films were spray coated by using a method by Eikos onto a 0.7-mm OA10 NIHON- DENKI-GLASS non-alkali glass substrate and 175-µm polyester (PET) film. The sheet resistance of the CNT layer was varied from 200 to 650 Ω/ and contained a melamine/acrylic coat. 760 Trottier et al. / Carbon-nanotube-based transparent conductive coating

3 FIGURE 4 Optical Performance. Visible-light transmittance of ITO, PEDOT, and CNT films. FIGURE 6 Color measurements. CNT films exhibit much closer to neutral color (closest to origin) than ITO and PEDOT. The visible-light transmittance of ITO, PEDOT, and CNT films is shown in Fig. 4. All transmittance values were measured using a Perkin-Elmer Lambda 3B UV-vis spectrophotometer and refer to the transparent conductor layer only (with substrate effect removed). The CNT film displays high transparency across the complete visible-light spectrum. In comparison, ITO has a maximum transparency in the range of nm, at the expense of significantly lower transparency at other wavelengths. For the same level of conductivity, current CNT films show somewhat lower transparency at 550 nm compared to the peak transmittance of ITO. However, CNT films exhibit significantly higher transparency across the entire visible-light spectrum. CNT films are also much more transparent than PEDOT. The performance gap between CNT and ITO will be minimized by further product optimization. Over the years, the optoelectronic performance of our CNT technology (Invisicon ) has improved, as shown by the technology curves in Fig. 5. An advantage to the CNT coatings is the ability to tailor the sheet resistance over a large resistive range from one order of magnitude up to as high as 10 orders of magnitude (ESD). For most display applications, neutral color is desired. ColormeasurementsusingaMacBethColorimeterwitha D65 illuminant and a 10 Observer confirm that CNT films are much closer to neutral color than both ITO and PEDOT, which show their characteristic yellow and blue hues, respectively (see Fig. 6). FIGURE 5 Optoelectronic performance of carbon-nanotube transparent conductive coatings. 5 Mechanical and chemical durability To evaluate the mechanical and chemical functionality of the CNT coatings, test specimens were made via spray coating a purified CNT dispersion onto heat-stabilized PET film. A binder coat of melamine/acrylic was applied by dip coating, followed by air drying and curing at 135 C for 5 minutes. The CNT/binder was ~75 nm thick with a sheet resistance of ~650 Ω/. Cyclic loading tests were conducted using a Roll Fatigue Tester 1 (mandrel diameter, 19.1 mm). Samples were precision cut to mm. Testing was at 0.7% strain amplitude, 1.25 Hz, and 25 C, and the resistance was measured continuously throughout the experiment. As shown in Fig. 7, Invisicon CNT coating showed <0.5% change in resistance after 2500 cycles, whereas ITO control samples showed >2% change after only 1000 cycles. The difference is even more dramatic when one compares the rate of change in resistance. From cycles, the slope of the ITO curve is more than Journal of the SID 13/9,

4 FIGURE 9 Compatibility with LCD CF Process. CNT films exhibit excellent chemical and heat resistance. FIGURE 7 Cyclic testing of Invisicon CNT coating on 175-µm PET compared to ITO on PET. (Courtesy of Dr. Jose Vedrine, Brown University.) FIGURE 8 Minimat tensile testing machine at 0.1-mm/min strain rate, in uniaxial tension comparing CNT and ITO-coated PET. (Courtesy of Dr. Jose Vedrine, Brown University.) 10 larger than the slope for the Invisicon CNT coating throughout 2500 cycles. The degradation in ITO resistance during flex testing is attributed to cracking of the ITO film. As flex cycling continues, these cracks continue to grow, ultimately leading to catastrophic failure (open circuit). At these strain levels, this failure mechanism is not observed for Invisicon CNT coatings. ThesameCNT-coatedPETsamplesweretestedat 25 C in a Minimat tensile testing machine at 0.1-mm/min strain rate, in uniaxial tension, up to 18% strain. Samples were cut into traditional dog bones (25 mm long, 3.5 mm wide).theresistancewasmeasuredin-situ using a digital multimeter. Below 1% strain, there appears to be fixture slack, as evidenced by the near-zero change in both measured stress and resistance, as shown in Fig. 8. Between 1 5% tensile strain, the CNT-coated PET film behaves elastically. Above 5% strain, there appears to be plastic deformation in the PET substrate, which dominates the electrical-resistance response. However, even after 18% tensile strain, only 14% change in resistance was observed. Note that ITO-coated PET has been extensively investigated by Cairns and co-workers 4 andboutenandco-worker. 5 They report that the onset of cracks in the ITO film occurs at ~2% tensile strain, with ITO failing catastrophically before 3% tensile strain is reached (resistance change >20,000%). To further evaluate the robustness of CNT films, the change in transparent CNT electrode visible-light transmittance after exposure to chemical and heat treatments commonly used in display manufacturing, was measured. Overall, the CNT film performed quite well, except immersion for 30 minutes in 5% NaOH solution. This alkaline test is very challenging for many organic coatings. However, the CNT films exhibited high resistance to alkaline attack. But, since the NaOH solution is able to penetrate through the CNT film, the film/glass interface is readily attacked by this aggressive alkaline solution, resulting in delamination of the CNT film. The CNT film exhibited excellent resistance to strong acid, organic solvents, and high-temperature exposure (250 C). This is consistent with the expected stability ofcntmaterialsandsummarizedinfig Trottier et al. / Carbon-nanotube-based transparent conductive coating

5 6 Conclusion This study confirms that transparent CNT electrodes can be a viable alternative to ITO for display applications, offering ease of processing (wet coating) and neutral color and is well-suited for shielding electrodes. As optoelectronic properties improve, CNT films can be expected to also be used as common electrodes for a variety of display applications. Acknowledgments The authors would like to acknowledge the guidance and testing support provided by Dr. Ueyama from Toppan R&D, Paul Johnson at the University of Rhode Island for TEM analysis, Jose Vendrine for cyclic testing at Brown University, and Graphic Utilities for color measurements. References 1 B Lewis and D Paine, Applications and processing of transparent conducting oxides, MRS Bull, (2000). 2 D Arthur, P Glatkowski, P Wallis, and M Trottier, Flexible transparent circuits from carbon nanotubes, SID Symposium Digest Tech Papers 35, 582 (2004). 3 L B Groenendaal, F Jonas, D Freitag, H Pielartzik, and J R Reynolds, Poly(3,4-ethylenedioxythiophene) and its derivatives: past, present, and future, Adv Mater 12, No. 7, 481 (2000). 4 D R Cairns, R P Witte, D K Sparacin, S M Sachsman, D C Paine, and G P Crawford, Strain-dependent electrical resistance of tin-doped indium oxide on polymer substrates, Appl Phys Lett 76, No. 11, 1425 (2000). 5 P C P Bouten, P J Slikkerveer, and Y Leterrier Mechanics of ITO on plastic substrates for flexible displays, in Flexible Flat Panel Displays, ed. G P Crawford (John Wiley & Sons, Ltd., 2005). Paul J Glatkowski is Vice President of Eikos, Inc. He is the inventor of Invisicon transparent conductive coatings and holds numerous patents in this and related carbon-nanotube technologies. Philip Wallis is technical Director at Eikos, Inc. He holds a Ph.D. in surface chemistry and is an expert in the area of inks. He is the holder of several patents related to ink and ink-flow-control systems. C. Michael Trottier is a Senior Scientist at Eikos, Inc. He holds a Ph.D. in chemical engineering from the University of Rhode Island and is an expert in organic and inorganic coating technologies. He has contributed to numerous papers on the characterization of single-walled carbon nanotubes and piezoelectric composites. Journal of the SID 13/9,