Characteristic features of new electron-multiplying channels in a field emission display

Size: px

Start display at page:

Download "Characteristic features of new electron-multiplying channels in a field emission display"

Julian Branden Patterson
5 years ago
Views:

1 Characteristic features of new electron-multiplying channels in a field emission display Whikun Yi, Taewon Jeong, Sunghwan Jin, SeGi Yu, Jeonghee Lee, and Jungna Heo The National Creative Research Initiatives Center for Electron Emission Source, Samsung Advanced Institute of Technology, P.O. Box 111, Suwon , Korea Jibeom Yoo Department of Material Engineering, Sungkyunkwan University, Chunchun Dong 300, Jangan-Ku, Suwon, Korea J. M. Kim a) The National Creative Research Initiatives Center for Electron Emission Source, Samsung Advanced Institute of Technology, P.O. Box 111, Suwon , Korea Received 30 April 2001; accepted 1 October 2001 We have developed a microchannel plate MCP to increase the efficiency of a field emission display FED by introducing new materials and process technologies. A substrate was constructed with alumina, and channel walls of pore arrays of the MCP were deposited with thin films using electroless plating and sol-gel process. The resulting MCP has been evaluated with a high input current source from a continuous electron beam and from Spindt-type field emitters. Some features of the MCP were also characterized in terms of brightness, anode current, input and output pulse, and focused luminescent spot, in a FED. With the MCP between the cathode and the anode of a FED, the brightness and anode current increased four- to fivefold due to electron multiplication through an array of pores in the device. In addition, the fabricated microchannel plate was found to prevent spreading of electrons emitted from the cathode tips, thus improving both display resolution and picture quality American Vacuum Society. DOI: / I. INTRODUCTION Field emission displays FEDs typically include a planar substrate with an array of projecting field emitters cathode, 1,2 which are usually conical projections. A conductive extraction grid gate is positioned in close proximity to the emitters, and is driven by a voltage of about V. 3 The grounded emitters are then selectively activated by an electric field between them and the extraction grid. The FED includes a display screen, which also serves as an anode, mounted parallel to the cathode array and positioned on the other side of the cathode array relative to the gate see Fig. 6 a. The anode screen is composed of a glass plate coated with a transparent conductive material, which is biased to between 1 and 5 kv. A cathodoluminescent layer, i.e., phosphor, covers the exposed surface of the anode and it is this, when struck by electrons, that is responsible for the light emission. The brightness of light emitted from the phosphor screen depends upon the density and kinetic energy of electrons that strike a given area of the screen. The kinetic energy of the electrons is derived from the voltage difference between the anode and the cathode. Typically, to achieve a high current density, and therefore high brightness in a FED, the electric field that extracts electrons from the emitter must be in the range of 10 7 V/cm. 1 However, this high field strength invariably leads to cathode tip damage by erosion, for example, cathode sputtering by positive ions generated from residual II. FABRICATION A MCP is a wafer-like glass array consisting of millions of microscopic channel electron multipliers Fig. 1. The channel matrix is fabricated from lead silicate glass and processed to optimize the secondary electron emission SEE yield of the surface and to render the channel walls semiconducting. Charge replenishment to the channels is accomplished by deposition of metal electrodes on the front and rear surfaces. Used alone or in cascade, MCPs allow signal gains of due to avalanche multiplication of seconda Author to whom correspondence should be addressed; electronic mail: jongkim@sait.samsung.co.kr gases inside a FED. 4 Such erosion makes the emitted beam unstable in a short time and impairs long term reliability. In 1994, Yu et al. 5 reported cathode deterioration of a FED at high current densities due to ohmic power dissipation, electron static stress, and ion etching from residual gases in the evacuated space. They also compared long term and short term current instabilities during electron emission at high current densities. In addition, another important problem involves the spreading of electrons spatial broadening of the beam as they are emitted from the cathode tips, which causes the cathodoluminescent area to exceed the desired size of the unit domain, a pixel. Consequently, the light emitted from the area may bleed into an adjacent pixel, reducing resolution and picture quality. In this work a specially developed electron multiplying microchannel plate MCP has been tested with various input currents from an electron gun, and then inserted into a FED where it was used to overcome both of the problems mentioned above J. Vac. Sci. Technol. B 19 6, NovÕDec Õ2001Õ19 6 Õ2247Õ5Õ$ American Vacuum Society 2247

2248 Yi et al.: Characteristic features of new electron-multiplying channels 2248 FIG. 1. Cutaway view and basic operation of a microchannel plate MCP. ary electrons.

With a bias potential applied to the electrode faces, a uniform electrostatic field is generated in each channel.

2 2248 Yi et al.: Characteristic features of new electron-multiplying channels 2248 FIG. 1. Cutaway view and basic operation of a microchannel plate MCP. ary electrons. When an incident electron or radiation strikes the input side of the MCP, electrons are emitted from the surface. With a bias potential applied to the electrode faces, a uniform electrostatic field is generated in each channel. This field accelerates the electrons along the channel until they strike the wall, generating secondary electrons. Repetition of this action down the length of the channel creates an avalanche of electrons corresponding to the incident electron or radiation. The tape casting method 6 was used to build the body of the MCP. Dispersed alumina slurry was passed through a doctor blade to produce a greensheet. Micron-sized holes were made through the blank sheets using a computerprogrammed puncher UHT MP7150. The perforated layers of greensheets were aligned, stacked, and laminated in a flatplate press. The stacked green laminate was baked at 1600 C to remove organic materials. Using this process, we produced the body of the MCP, with 2 5-mm-deep pores of 170 m diameter on 220 m centers Fig. 2. FIG. 2. Photograph of our newly fabricated MCP and its pore arrays. FIG. 3. X-ray diffraction pattern of Cu/Al 2 O 3 at 1030 C. This pattern shows both of the peaks from CuAl 2 O 4 and Al 2 O 3. Within a single capillary channel, the wall acts as a continuous dynode that supports avalanche multiplication of electrons when a bias potential is applied across its length. The wall is coated with an electron emissive layer, which is formed upon a lower resistive layer. Functionally, the emissive layer is responsible for generating secondary electrons upon primary electron bombardment and the resistive layer facilitates the replenishment of the electron deficiency so produced in the emissive layer. Electroless copper plating 7 and successive heat treatments formed a resistive layer of CuAl 2 O 4, on the channel wall of each MCP pore. Figure 3 shows the x-ray diffraction patterns of Cu/Al 2 O 3 at 1030 C. Both of substrate Al 2 O 3 and newly formed CuAl 2 O 4 are shown at this temperature. This spectrum does not exhibit any copper oxide peaks such as Cu 2 O or CuO. It is believed that copper oxide formed at the initial stage of heat treatment is converted into CuAl 2 O 4 at the final temperature as a result of the reaction between Cu and bulk Al 2 O 3. 8 According to our experimental results, electron multiplication occurred well with either CuAl 2 O 4 or copper oxide as a resistive layer. The performance of each layer depends on the resistance measured through the electrodes located on both sides of the MCP device. The resistance of CuAl 2 O 4 was better controlled and more reproducible than the copper oxide. The emissive layer (SiO 2 ) was formed upon the resistive layer using a sol-gel process, 9 which was performed using an aqueous solution of tetraethyl orthosilicate TEOS containing alcohol and a small amount of HCl. Successive thermal oxidation at 450 C was then used to remove the organic materials, and form the SiO 2 layer. To verify the existence of remaining organic materials and/or the resulting SiO 2 layer, thermal gravimetric analysis TGA and infrared IR spectra were recorded. 10,11 According to the TGA graph of the TEOS-containing solution, it did not show any peaks above 350 C, which means only the SiO 2 film remained on the surface after thermal treatment above 350 C. The IR spectrum also confirmed the formation of SiO 2 with a vibrational peak at 1070 cm 1, which is attributed to Si O Si asymmetric bond stretching mode of SiO 2 Fig. 4. Furthermore, no J. Vac. Sci. Technol. B, Vol. 19, No. 6, NovÕDec 2001

3 2249 Yi et al.: Characteristic features of new electron-multiplying channels 2249 FIG. 4. IR spectrum of SiO 2 prepared from TEOS-containing precursors followed by thermal treatment at 450 C for 2 h. vibrational peaks were found between 3000 and 2800 cm 1 in the IR spectrum, which is attributed to C H stretching modes being associated with the presence of alkoxy groups of TEOS. The optimum thickness of SiO 2 and CuAl 2 O 4 layer was found to be Å and 1 2 m, respectively, for the MCP to exhibit the highest performance. Finally, both faces of the MCP were chemically mechanically polished and coated with metal in an electron beam evaporator. III. EXPERIMENT Two different thickness MCPs were used in this experiment, i.e., 2 and 5 mm. For the case where the MCP had 2-mm-deep pores with 170 m circular openings on 220 m centers, aspect ratio and open area ratio are around 12 and 47%, respectively. For the 5 mm MCP, the aspect ratio was around 30. The electron multiplying factor of the MCP was examined with a high current electron source such as an electron beam from a commercial electron gun Kimball EFG-7 or field emitter arrays of a FED. Throughout the experiments, the 2 mm MCP was evaluated with a FED and the 5 mm MCP with an electron gun. The continuous electron beam from the electron gun was focused onto one side of the MCP, and the output electrons passing through pore arrays were captured and counted with a positively biased stainless-steel anode holder by varying the voltage applied between the metal electrodes of the MCP (V MCP ) inset of Fig. 5. The MCP was also characterized in terms of brightness and degree of focusing in a FED. The same field emission single-integrated microtips and phosphor coated anode plates were used throughout this work. The integrated molybdenum tips were produced using a lithographic technique and controlled evaporation Spindt type at our laboratory, using a four step fabrication procedure described previously. 12 Schematic diagrams of the driving circuits for field emission measurements are shown in Fig. 6 a. A square wave pulse of width 130 s, amplitude 80 V, and a duty ratio of 150 was applied to the gate. For electron multiplication, a dc voltage of 200 and 800 V was applied to the bottom (V BOT ) and top (V TOP ) faces of the MCP, respectively. FIG. 5. Amplification yield vs the voltage applied across the two electrodes of the MCP (V MCP ) for three input currents, i.e., 0.05, 0.29, and 0.75 A. Inset shows the schematic diagram of measuring system. IV. RESULTS AND DISCUSSION First, an electron gun was used to measure the electron multiplying factor, i.e., signal gain for the 5-mm-thick MCP. This MCP had 5-mm-deep pores with 170 m circular openings on the 220 m centers, thus the aspect ratio and the open area ratio were around 30 and 47%, respectively. The continuous electron beam from the electron gun was focused onto one side of the MCP, and the output electrons passing through the pore arrays were captured and counted with a positively biased stainless-steel anode holder by varying the voltage applied on the metal electrodes of the MCP (V MCP ) inset of Fig. 5. Three values of input current from the electron gun were used: 0.05, 0.29, and 0.75 A. As V MCP was increased to 2400 V, the signal gain was found to be 700, 250, and 100, respectively, for the three input currents. This implies that the lower input current corresponds to a higher signal gain. Judging from the fact that the signal gain of a MCP is directly proportional to the SEE yield of the emissive layer here, SiO 2 film, the lower or higher input current generates the higher or lower SEE yield upon bombarding FIG. 6. a Schematic circuit diagrams of a FED in operation with and without the MCP at normal condition, V g 80 V, V A V, V BOT 200 V, and V TOP 800 V. b Brightness and anode current change of a FED as a function of V TOP of the MCP (V A 1200 V). Here, the arrows at the y axes indicate the brightness and anode current measured for the case without the MCP 24 cd/m 2 and 2.8 A, respectively. JVST B-Microelectronics and Nanometer Structures

4 2250 Yi et al.: Characteristic features of new electron-multiplying channels 2250 the SiO 2 film. This can be easily explained by the deeper positive charging effect for the case of the higher input current. Upon the incidence of higher input electron current, the SiO 2 surface will be charged more positively, since the replenishment of electrons from metal electrode to SiO 2 surface through the resistive layer is rather difficult compared with the case for the lower input current. Thus, a more positively charged SiO 2 sample can recapture more secondary electrons leaving from the surface, leading to a decrease of SEE and signal gain. However, when we compared the output current coming through all the pores of the MCP, its value was equal to 35, 72, and 70 A for the three input currents, i.e., 0.05, 0.29 and 0.75 A. In other words, higher signal gain does not always indicate higher output current. Conventional MCPs typically allow electron multiplication factors of for input currents in the pa range. 13 The total resistance of those kinds of MCPs is of the order of 10 3 M. When a bias potential of 1000 V is applied between two electrodes of a MCP, the total bias current through the array of pores is about 1 A, from Ohm s law (I V/R). If the input current and multiplication factor are 1 pa and , respectively, the output current through the MCP will be 0.1 A, i.e., 10% of the total bias current on the MCP. In the case of a relatively higher input current such as A in our system, the resistance of an MCP should be adjusted to a lower value than 10 3 M to obtain a high enough bias current to supply electrons for amplification within the MCP. 14,15 From the measurement of V MCP and i MCP 2400 V and 2200 A, respectively during our experiments, the resistance of MCP (R MCP ) was estimated to be several M. Conversion efficiency of this MCP with A output current (i A ) and 2200 A total bias current (i MCP ), is 1.6% 3.3%. Conventional MCPs fabricated for amplifying input currents in the range of a few pa are inappropriate for use in a FED device, since the typical current density from the cathode tips of a FED is several A/cm 2. 2 In our test with a commercial MCP, which was composed of hollow channels of lead silicate glass, electron multiplication from a FED cathode was minimal. Our newly fabricated MCP was inserted into a FED as shown in Fig. 6 a, then the anode current and brightness were measured simultaneously as the bias voltage across the MCP ( V MCP V TOP V BOT ) was increased. As shown in Fig. 6 b, both current and brightness increase linearly through the given ranges. The anode current and brightness measured at the same experimental condition without the MCP were 2.8 A and 24 cd/m 2, respectively, which are indicated by the arrows on the y axis in Fig. 6 b for comparison. Consequently, both the anode current and brightness increase 4 5 times at V MCP 1200 V with a newly fabricated MCP. Figure 7 shows the electron field emission pattern as monitored by a phosphor screen with and without the MCP. To make the comparison as direct as possible, only the upper half of the 5.2 in. full cathode panel shown had MCP amplification. The brightness increased from 15 cd/m 2, without the FIG. 7. Electron field emission of a 5.2 in. full-cathode panel as monitored by a phosphor screen with the MCP inserted in the upper half only, using the circuit and voltages shown in Fig. 6 a. MCP lower half of Fig. 9, to 45 cd/m 2. This value of brightness was obtained at anode voltage of 1000 V (V A ); the kinetic energies of electrons arriving at the anode plate are the same, regardless of the presence of the MCP. If we regard phosphor brightness as the multiplication of the number of incident electrons and their kinetic energy, then the brightness enhancement is mainly attributable to the amplification effect caused by secondary electrons from the MCP. The input and output pulses were detected with an oscilloscope for the cases with and without a MCP. Figure 8 shows the shape of input pulse at the gate electrode and the corresponding output pulses at the anode with and without a FIG. 8. Input pulse a at the gate, and the corresponding output pulse at the anode with b and without c the MCP, which are measured by an oscilloscope with respect to time. J. Vac. Sci. Technol. B, Vol. 19, No. 6, NovÕDec 2001

5 2251 Yi et al.: Characteristic features of new electron-multiplying channels 2251 gate and the bottom face of the MCP and is amplified through the channels before reaching the cathodoluminescent materials. These experimental results are evidence verifying that our newly fabricated MCP is quite suitable to enhance the efficiency of a FED by focusing electron beams. FIG. 9. Luminescent spot changes on increasing the bottom voltage the voltage on the cathode-facing side of the MCP, i.e., V BOT a 60, b 100, c 200, d 300 V, and on increasing the top voltage the voltage on the anode-facing side of the MCP, i.e., V TOP e 800, f 1000, g 1200 V using the circuit and voltages shown in Fig. 6 a. MCP. The input pulse a, which was originally a square wave form with its width of 130 s, tends to decay with a small tail because of leakage currents from the emitter tips to the gate lines. Since the output signal is measured across a resistor connected in series with the anode plate, the shape of the output pulse represents the voltage drop of the anode with respect to the resistor during the bombardment by pulsed electrons. The higher voltage drop with a MCP is a result of electron multiplication. In fact, the peak area ratio for the case with and without the MCP, 4.4, is in the same range with the brightness ratio and the anode current ratio of Fig. 6, i.e., 4 5. The x-axis location for the maximum peak height is almost the same for two output peaks within the experimental error. In addition, the calculated decay time for the peak b in Fig. 8, 260 s, is a little shorter than that for the peak c, 290 s. This result means more sensitive response at the anode screen for a given input pulse through the MCP. Although the reason is not clear yet, this is a positive result for circuit operation of a FED. The result above also implies that the interaction between the electrons emitted from cathode tips and the emissive layer of a MCP, even it exists, has no effect on delaying or obliterating the initial information arrived by the electron pulse. Focusing experiments were conducted using the newly fabricated MCP, and their results are shown in Fig. 9. Six pixels corresponding to an area 1650 m 750 m were chosen for the test. As the bottom voltage of the MCP was increased, the luminescent spot became smaller, brighter, and more clearly focused Figs. 9 a 9 d. Meanwhile, as the top voltage of the MCP was increased from 800 to 1400 V, the spot became brighter with little change in size. This implies that the electron beam is being focused between the V. CONCLUSION We have developed a MCP to improve the efficiency of a FED with new materials and process technologies. A substrate was constructed with alumina, and channel walls of pore arrays of the MCP were deposited with thin films using electroless plating and sol-gel process. The resulting MCP has been evaluated with high input current sources including a continuous electron beam and Spindt-type field emitter array. In the case of a A incident beam from an electron gun, the signal gain of the MCP (AR 30) increased times. With the fabricated MCP (AR 12) between the cathode and the anode of a 5.2 in. sized FED, the brightness and anode current increased four- to fivefold by electron multiplication through an array of pores in the device. From the comparison of output pulses of a FED with and without the MCP, the decay time was a little shorter in the case with the MCP than that without the MCP. In addition, the fabricated MCP prevented spreading of electrons emitted from the cathode tips, thus improving both display resolution and picture quality. ACKNOWLEDGMENTS This work was supported by the Korean Ministry of Science and Technology, through the National Creative Research Initiative Program. 1 C. A. Spindt, I. Brodie, L. Humphrey, and E. R. Westerberg, J. Appl. Phys. 47, A. Ghis, R. Meyer, P. Rambaud, F. Levy, and Th. Lerou, IEEE Trans. Electron Devices ED-38, C. G. Lee, H. Y. Ahn, B. G. Park, and J. D. Lee, J. Vac. Sci. Technol. B 14, R. Smith, J. Phys. D 17, M. L. Yu, B. W. Hussey, H. S. Kim, and T. H. Chang, J. Vac. Sci. Technol. B 12, J. U. Knickerbocker, Am. Ceram. Soc. Bull. 71, O. B. Dutkewych, U.S. Patent No. 4,695, September T. Fujimura and S. I. Tanaka, Acta Mater. 46, M. Langlet, D. Walz, P. Marage, and J. C. Joubert, Thin Solid Films 221, M. Asomoza, M. P. Dominguez, S. Solis, and T. Lopez, Mater. Lett. 33, M. Langlet and C. Vautey, J. Eng. Mater. Technol. 13, J. M. Kim, H. W. Lee, Y. S. Choi, N. S. Lee, J. E. Jung, J. W. Kim, W. B. Choi, Y. J. Park, J. H. Choi, Y. W. Jin, W. K. Yi, and N. S. Park, J. Vac. Sci. Technol. B 18, J. L. Wiza, Nucl. Instrum. Methods 162, W. Yi, S. Jin, T. Jeong, J. Lee, S. Yu, Y. Choi, and J. M. Kim, Appl. Phys. Lett. 77, W. Yi, T. Jeong, S. Jin, S. Yu, J. Lee, and J. M. Kim, Rev. Sci. Instrum. 71, JVST B-Microelectronics and Nanometer Structures

Development of triode-type carbon nanotube field-emitter arrays with suppression of diode emission by forming electroplated Ni wall structure

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.