Development of triode-type carbon nanotube field-emitter arrays with suppression of diode emission by forming electroplated Ni wall structure J. E. Jung, a),b) J. H. Choi, Y. J. Park, c) H. W. Lee, Y. W. Jin, d) D. S. Chung, S. H. Park, and J. E. Jang Samsung Advanced Institute of Technology, P.O. Box 111, Suwon, Korea 440-600 S. Y. Hwang, T. Y. Ko, Y. S. Choi, S. H. Cho, C. G. Lee, and J. H. You Samsung SDI, 575 Shin-Dong, Paldal-Gu, Suwon, Korea 442-731 N. S. Lee Department of Nano Science and Technology, Sejong University, 98 Gunja-Dong, Gwangjin-Gu, Seoul, Korea 143-747 J. B. Yoo Department of Materials Engineering, Sungkyunkwan University, 300, Chunchun-Dong, Jangan-Gu, Suwon, Korea 440-746 J. M. Kim e) Samsung Advanced Institute of Technology, P.O. Box 111, Suwon, Korea 440-600 Received 4 April 2002; accepted 24 June 2002; published 3 February 2003 Triode-type field-emitter arrays were developed by screen printing a photosensitive paste including single-walled carbon nanotubes. Ni wall structure NWS was electroplated to form a thick gate to suppress diode emission induced by strong electric strengths due to an anode potential and to focus electron beams to their destined color subpixels. It was observed in computer simulations, as well in experiments that the NWS with the optimum thickness was effective in reducing the diode emission and enhancing electron-beam focusing by modifying electrical potentials around the carbon nanotube emitters. Our fully sealed field-emission display panel using the field-emitter arrays with the NWS demonstrated full color moving images without serious diode emission and with satisfactory color separation. 2003 American Vacuum Society. DOI: 10.1116/1.1516181 I. INTRODUCTION Carbon nanotubes CNTs have recently demonstrated a great potential to be applied to field-emitter materials. 1 6 Our group has been developing several types of triode structure field-emission displays FEDs using single-walled CNTs. 7 9 We have focused upon screen printing as low-cost, simple, and scalable fabrication technologies for CNT cathode plates with a uniform and reliable emission over a large area at low operation voltages. Among the several triode-type emitter structures that we have developed, a normal gate triode structure CNT-FEDs seems to be one of the most promising for an application of CNTs to FED emitters in terms of operation voltages as well as FED device performance. This structure is similar to the Spindt-type FED, but has a larger gate hole diameter due to excellent field-emission characteristics of CNTs. On the field-emitter array FEA templates fabricated by the same processes as those of the Spindt-type FEAs, CNT emitter dots were formed inside gate holes by screen a Electronic mail: jejung@sait.samsung.co.kr b Also at: Department of Materials Engineering, Sungkyunkwan University, 300, Chunchun-Dong, Jangan-Gu, Suwon, Korea 440 746. c Also at: Department of Vacuum Science and Technology, Sungkyunkwan University, 300, Chunchun-Dong, Jangan-Gu, Suwon, Korea 440-746. d Also at: Department of Materials Engineering, Sungkyunkwan University, 300, Chunchun-Dong, Jangan-Gu, Suwon, Korea 440-746 e Also at: Samsung SDI, 575 Shin-Dong, Paldal-Gu, Suwon, Korea 442-731 and The National Creative Research Initiatives Center for Electron Emission Source. printing of a photosensitive paste containing CNTs and subsequent back side exposure of an UV light. 10 A dot of CNT emitters with a diameter of 20 m was defined for each gate hole with a diameter of 30 m. For this emitter structure, a driving voltage of gate electrodes was about 60 V. During the operation of such a triode structure, there are two problems to be solved to realize normally operated FED devices. One is to suppress a diode emission from the CNT emitters which is induced by anode voltages. Generally speaking, electron emission from emitters has to be controlled only by relative magnitudes of gate voltages to cathode voltages, not by anode biases. CNT emitters with extremely low threshold electric fields, however, can emit electrons even in their off state when anode voltages exceed a certain level. If high anode voltages beyond this level are required for sufficient luminance from phosphors, such a diode emission phenomenon cannot be avoided. The other problem is to focus electron beams onto their designated phosphor subpixels. For this structure with gate hole diameters much larger than those of the Spindt-type FEAs, electrons launched from CNTs would spread in wide angles. To suppress the diode emission as well as to focus electron beams during the triode operation of the FEAs, an electroplated Ni wall structure NWS was applied to CNT-FEAs. The NWS, which is simply a thick gate electrode, is expected to modify electrical potential distributions around dots of CNT emitters in the direction of decreasing the 375 J. Vac. Sci. Technol. B 21 1, JanÕFeb 2003 1071-1023Õ2003Õ21 1 Õ375Õ7Õ$19.00 2003 American Vacuum Society 375
376 Jung et al.: Development of triode type carbon nanotube 376 FIG. 2. Schematic model for simulating electric fields, electrical potentials, and electron-beam trajectories for FEAs with NWS where symbols, 1, 2, and x, y, designate locations and directions, respectively. FIG. 1. Schematics of fabrication processes of CNT-FEAs with the NWS: a FEAs with the dots of CNT emitters inside gate holes defined by a back side UV light exposure through the holes of an a-si layer, b PR patterning in the way that each PR dot covers three gate holes in a running direction of cathode electrodes, c Ni electroplating in a dipping bath by applying the voltage of 3 V and the current density of 3 A/cm 2 between a bath electrode and the gate electrodes of the CNT-FEAs, and d PR stripping leading to the final CNT-FEA structure with the NWS. electric-field strengths of anode voltages as well as reducing the divergence of electron beams. II. EXPERIMENT AND SIMULATION Schematics of fabrication processes of a CNT-FEA with the NWS are shown in Fig. 1. After patterning UV light transparent indium tin oxide cathode electrodes, an amorphous-silicon a-si mask layer, which would serve as an embedded back side exposure mask for photolithography of CNT dots, a 1.5 m thick SiO 2 insulation layer and a 0.3 m thick Ni gate layer were coated by plasma-enhanced chemical vapor deposition and electron-beam evaporation, respectively. Photoresist PR with the thickness of 6 m was spin coated and then photolithographed to produce holes with the diameter of 20 m. The Ni gate layer was then wet etched, resulting in a gate hole diameter of about 26 m. Subsequently, a reactive ion etching with a CF 4 /O 2 gas mixture was performed to eliminate the SiO 2 insulation layer and the mask layer through the Ni gate holes. The diameters of holes through the insulation layer and the mask layer were mostly determined by the PR pattern due to its anisotropic etching behavior. Dots of CNT emitters inside the gate holes were formed by screen printing of the photosensitive paste including single-wall CNTs and a subsequent back side UV light exposure through the holes of the a-si mask layer underneath the SiO 2 insulation layer. Then, the CNT dots were fired in an N 2 ambient. The aforementioned processes led to the fabrication of a CNT-FEAs shown in Fig. 1 a, which was described in detail elsewhere. 10 On such a CNT-FEA, patterning 30 m thick PR was performed for electroplating nickel, as given in Fig. 1 b. Each PR patterned dot encompassed three gate holes in a direction of cathode electrodes. The Ni electroplating was carried out in a dipping bath under the voltage of 3 V and the current density of 3 A/cm 2 between a bath electrode and the gate electrodes of the CNT-FEAs, which corresponded to a deposition rate of 0.6 m/min. Ni was deposited through the PR patterned areas as presented in Fig. 1 c. PR stripping produced the final CNT-FEA structure with the NWS shown in Fig. 1 d. The deposition time was varied to obtain different thicknesses of the NSW layer. OPERA-3D Vector Fields, Ltd. utilizing a finite-element method was applied to simulate electric-field and electrical potential profiles around the dots of CNT emitters with and without the NWS for given electrical bias conditions. In order to address focusing characteristics, electron-beam trajectories were also calculated for different gate voltages. A schematic model is given in Fig. 2. Diameters of a gate hole, an insulation layer hole, and a dot of CNT tips are 26 m, 20 m, and 15 m, respectively. The thicknesses of a gate electrode, an insulation layer, and a dot of CNT tips are 0.3 m, 1.5 m, and 1.8 m, respectively. An anode bias is 2 kv and the gap between FEAs and an anode is 1.1 mm. III. RESULTS AND DISCUSSION Figure 3 a shows a typical structure of the CNT-FEAs with the NWS, observed by scanning electron microscopy. Each set of three dots of CNT emitters is surrounded by the NWS and each body of the NWS encloses three sets of the CNT dots. A discrete distribution of the NWS is expected to relieve a stress which might be generated by the thick NWS layer. Such a design of the NWS was optimized by considering physical and electrical characteristics of the device as well as the compatibility of fabrication processes. Figure 3 b presents an image around the CNT dot which is sur- J. Vac. Sci. Technol. B, Vol. 21, No. 1, JanÕFeb 2003
377 Jung et al.: Development of triode type carbon nanotube 377 FIG. 3. SEM images of the CNT-FEAs with the NWS: a each set of three CNT dots surrounded by the NWS, b a magnified image around a CNT dot, and c erect CNT emitters on a dot. rounded by the SiO 2 insulator and then by a gate electrode as concentric circles. In Fig. 3 c, many erect CNTs are observed inside each dot. The main materials that substitute as-prepared CNT paste are CNT powder, glass frit, and organic vehicles. Since firing the dots of CNTs, which was carried out before the PR processing for electroplating the NWS burns out organic binders premixed into a CNT paste, the subsequent PR processing including a stripping process Fig. 1 d should not chemically or mechanically interfere with the dots of a CNT paste. Figure 4 shows electric-field distributions resulting from the anode biasing on the CNT-FEAs with and without NWS in the 1-x direction given in Fig. 2. In order to figure out diode emission effects occurring by anode voltages, electricfield strengths were calculated while restraining triode emission, i.e., the potential of the gate electrode to which the NWS was electrically connected was set to the ground 0 V. The cathode and anode voltages were 0 V and 2 kv, respectively, for a gap of 1.1 mm between the cathode and anode plates. Since the CNT dot was assumed as a conductor disk, a calculated situation might be deviated from a reality. Nevertheless, it can be seen that the grounded NWS Fig. 4 b significantly decreases the electrical-field strengths due to anode potentials around the dot compared to those without NWS Fig. 4 a. The electric fields are concentrated at the sharp edges of the NWS Fig. 4 b, but no diode emission was observed from these sites up to the anode voltage of 4 kv in the measurement using the FEA template with the NWS where the CNT dot emitters were not formed. Figure 5 presents calculated electric-field strengths near the surface of the CNT dots for different Ni wall heights along the x and y directions at positions 1 and 2 defined in Fig. 2. An electrical boundary condition to the electrodes was the same as that of Fig. 4. The electric-field strengths do not vary significantly with positions and directions. The most important parameter that affects the electric fields of an anode voltage on the surface of the CNT dots is the Ni thickness. For example, the 25 m thick NWS gives rise to lower electric-field strengths by an order of magnitude than that without NWS. It is noted that the electric-field strengths are quite uniform over the dots along the x and y directions, but a little higher around the edges of the CNT dots. This suggests that electrons would be emitted with a higher probability at the edges rather than at the central areas of the CNT dots. Conclusively, the diode emission from the CNT dots caused by anode voltages can be drastically suppressed by increasing the NWS thickness. Figure 6 a shows an emission image of the FEAs with the 20 m thick NSW operated in a triode mode at the gate and anode voltages of 60 V and 2 kv, respectively. This image was taken with a gate bias duty of 1/120 and a frequency of 100 Hz for the cathode electrode bias of 0 V. Diode emission was not observed up to the anode voltages of 3 kv at the fixed gate bias of 60 V. For the FEAs without JVST B-Microelectronics and Nanometer Structures
378 Jung et al.: Development of triode type carbon nanotube 378 FIG. 4. Distributions of electrical-field strengths for CNT-FEAs a with the NWS and b without the NWS along the 1-x direction of Fig. 2. The cathode and gate electrodes were set to ground 0 V while biasing 2 kv to the anode for the cathode-to-anode gap of 1.1 mm. FIG. 5. Distributions of electric-field strengths on the CNT-FEAs with different NWS thickness calculated at the locations of 0.5 m above the surface of the CNT dot for every 1.5 m along the a 1-x, b 1-y, c 2-x, and d 2-y directions defined in Fig. 2. The cathode and gate electrodes were set to the ground 0V while biasing 2 kv to the anode for the cathode-to-anode gap of 1.1 mm. J. Vac. Sci. Technol. B, Vol. 21, No. 1, JanÕFeb 2003
379 Jung et al.: Development of triode type carbon nanotube 379 FIG. 6. Emission images of CNT-FEAs with the 20 m thick NWS at the gate and anode voltages of a 60Vand2kVand b 0 V and 3.8 kv, respectively. The spacing between lower and upper electrode was 1.1 mm and the cathode electrodes were set to 0 V. This image was observed with a duty ratio of 1/120 and a frequency of 100 Hz. The anode currents were 0.93 ma and 0.002 ma in a and b, respectively, during the operation. NWS, diode emission began to occur at the anode voltages of about 1 kv. The FEAs with NWS were more resistant to electrical arcing than the FEAs without NSW. We suppose that strong peak currents of ions instantly induced by electrical arcing inside a FED panel would be absorbed by the thick NWS metal layer and dissipated away as heat without any damage to the FEAs. Even for the FEAs with the NWS, however, some hot spots produced by diode emission began to be observed at the anode bias of 3.5 kv for the electrically grounded cathode and gate electrodes. Figure 6 b presents such a diode emission phenomenon observed at the cathode, gate, and anode electrode voltages of 0 V, 0 V, and 3.8 kv, respectively. To avoid diode emission at such higher anode voltages, it is considered that the NWS thickness should be increased. Since the thick NWS modifies electrical potentials around the gate holes, trajectories of electrons emitted from the CNT dots inside the gate hole would be affected by the NWS. Figure 7 shows electron-beam trajectories of the CNT-FEAs without the NWS calculated for different gate voltages. In this simulation, cathode and anode electrodes were electrically biased by 0 V and by 2 kv, respectively. The electronbeam trajectories were calculated at the both edges and the center of a CNT dot where electron beams were assumed to be fired at angles of 30, 0, and 30 relative to the surface normal at the height 0.5 m above the surface of the CNT dot. The initial energy of the electron was set to the electrical potential value at the position where it was launched. Electron-beam trajectories for the CNT-FEAs with the 20 m thick NWS are given in Fig. 8. The same assumptions as those of Fig. 7 are applicable to this case but an existence of the 20 m thick NWS. Irrespective of an existence of the NWS, the equipotential lines around the CNT dots exhibit a convex shape from the point of view of the CNT dot, which tends to spread out the electron beams there. For the FEAs FIG. 7. Electron-beam trajectories of the CNT-FEAs without the NWS calculated for different gate potentials of a 10 V, b 20 V, c 30 V, and d 60Vat the cathode and anode electrode biases of 0 V, and 2 kv, respectively. The electron-beam trajectories were simulated at the both edges and the center of a CNT dot where electrons were assumed to be launched at angles of 30, 0, and 30 at the height 0.5 m above the surface of the CNT dot. JVST B-Microelectronics and Nanometer Structures
380 Jung et al.: Development of triode type carbon nanotube 380 FIG. 8. Electron-beam trajectories of the CNT-FEAs with the NWS calculated for different gate potentials of a 10 V, b 20 V, c 30 V, and d 60 V at the cathode and anode electrode biases of 0 V and 2 kv, respectively. The electron-beam trajectories were simulated at the both edges and the center of a CNT dot where electrons were assumed to be launched at angles of 30, 0, and 30 at the height 0.5 m above the surface of the CNT dot. without the NWS, this convex shape of the electrical potential distributions is maintained with larger radii of curvature at higher elevations. As shown in Fig. 8, however, the NWS produces a concave shape of equipotential lines below and even above the heights of the NWS around the gate holes. Thus, the NWS clearly modifies the electrical potential distributions around the gate holes so that the electron beams are less spread out. In Fig. 8, the radii of curvature of the concave potential lines are larger at higher gate voltages. This induces wider spreading of the electron beams with higher gate voltages, resultantly increasing a possibility of cross talk between R, G, and B color subpixels. Thus, a panel structure should be optimized to achieve color separation so that the emitted electron beams arrive at their corresponding phosphor subpixels under the operation conditions. An optimization of the panel structure should include designing a gate hole diameter, a NWS height, a diameter and height of the CNT dot, a cathode-to-anode gap, etc., under the operation conditions. Figure 9 shows a video running full color moving image of our fully vacuum-sealed, 5 in. diagonal CNT-FED with the 20 m thick NWS. The spacing between the cathode and anode plates was 1.1 mm. This image was observed at the gate and anode electrode biases of 60 V and 2 kv, respectively, with a duty ratio of 1/120 and a frequency of 100 Hz. The CNT-FED device was electrically operated by matrixaddressable circuitry with a 256 gray scale for each color. Although the driving gate voltage was around 60 V, R, G, and B color separation of the CNT-FEAs with the 20 m thick NWS was found to be improved in comparison with that of the CNT-FEAs without the NWS. It seems that the NWS works as physical electrical apertures prohibiting the electron beams from diverging. Since the structural parameters such as the thickness and horizontal dimensions of the NWS affect electron-beam trajectories by modifying electrical potentials as well as physical blocking of electrons, a structural optimization is under investigation. IV. CONCLUSION The NWS was successfully implemented into the CNT- FEAs by Ni electroplating. Our calculations suggested that the thick NWS surrounding a set of CNT dots with the triode FIG. 9. Full color moving image of a fully vacuum-sealed, 5 in. diagonal CNT-FED with the 20 m thick NWS. The spacing between lower and upper electrode was 1.1 mm. This image was observed at the upper electrode and anode biases of 60 V and 2 kv, respectively, with a duty ratio of 1/120 and a frequency of 100 Hz. The anode current was 0.64 ma during the operation. J. Vac. Sci. Technol. B, Vol. 21, No. 1, JanÕFeb 2003
381 Jung et al.: Development of triode type carbon nanotube 381 structure considerably reduced electric-field strengths given by an anode potential near the surface of the CNT dots, resulting in a decrease of possibility that diode emission occurred by an anode voltage. It was found that diode emission was experimentally suppressed up to the anode voltages of 3 kv in the CNT-FEAs with the 20 m thick NWS, while diode emission began to occur at the anode voltages of 1 kv in the FEAs without the NWS. The simulations of electronbeam trajectories for the FEAs with the NWS showed that emitted electrons spread out, as the gate biases increased. Electrons were well focused up to the gate voltages of 20 V. Our fully sealed CNT-FEDs with the NWS demonstrated full color moving images with satisfied color separation as well as with the suppression of diode emission. 1 P. G. Collins and A. Zettl, Appl. Phys. Lett. 69, 1969 1996. 2 Q. H. Wang, T. D. Corrigan, J. Y. Dai, and R. P. H. Chang, Appl. Phys. Lett. 70, 3308 1997. 3 Z. F. Ren, Z. P. Huang, J. W. Xu, J. H. Wang, P. Bush, M. P. Siegal, and P. N. Provencio, Science 282, 1105 1998. 4 J. M. Bonard, J. P. Salvetat, T. Stockli, W. A. de Heer, L. Forro, and A. Chatelain, Appl. Phys. Lett. 73, 918 1998. 5 O. M. Kuttel, O. Groening, C. Emmenegger, and L. Schlapbach, Appl. Phys. Lett. 73, 2113 1998. 6 W. Zhu, C. Bower, O. Zhou, G. Kochanski, and S. Jin, Appl. Phys. Lett. 75, 873 1999. 7 W. B. Choi, D. S. Chung, J. H. Kang, H. Y. Kim, Y. W. Jin, I. T. Han, Y. H. Lee, J. E. Jung, N. S. Lee, G. S. Park, and J. M. Kim, Appl. Phys. Lett. 75, 3129 1999. 8 N. S. Lee, D. S. Chung, J. H. Kang, H. Y. Kim, S. H. Park, Y. W. Jin, Y. S. Choi, I. T. Han, N. S. Park, M. J. Yun, J. E. Jung, C. J. Lee, J. H. You, S. H. Jo, C. G. Lee, and J. M. Kim, Jpn. J. Appl. Phys., Part 1 39, 7154 2000. 9 D. S. Chung, W. B. Choi, J. H. Kang, H. Y. Kim, I. T. Han, Y. S. Park, Y. H. Lee, N. S. Lee, J. E. Jung, and J. M. Kim, J. Vac. Sci. Technol. B 18, 1054 2000. 10 D. S. Chung, S. H. Park, Y. W. Jin, J. E. Jung, Y. J. Park, H. W. Lee, T. Y. Ko, S. Y. Hwang, J. W. Kim, M. H. Yoon, C. G. Lee, J. H. You, N. S. Lee, and J. M. Kim, Proceedings of the IVMC 01, 2001, p. 179. JVST B-Microelectronics and Nanometer Structures