- Email: [email protected]
Study of the influence of the dielectric layer thickness in a CNT-FED
ARTICLE IN PRESS
Ultramicroscopy 107 (2007) 844–848 www.elsevier.com/locate/ultramic
Study of the influence of the dielectric layer thickness in a CNT-FED Xing Su, Lifang Zhang, Wei Lei, Xiaobing Zhang Department of Electronic Engineering, Southeast University, Nanjing 210096, People’s Republic of China
Abstract An experimental investigation is carried out to study the influences of the dielectric layer thickness variation on the field emission characteristics and luminance distribution in a CNT-FED fabricated by screen-printing. Two steps are contained in the investigation: (1) the dielectric layer thickness fluctuations are presented with an ultrasonic thickness gauge, and (2) a simulation model is constructed to study the corresponding influences of the dielectric layer thickness fluctuations on the field emission characteristics and luminance uniformity on the screen. Our findings indicate that the dielectric layer thickness fluctuations are mainly larger than 5 mm, which mean the dielectric layer thickness fluctuation is an important cause of the non-uniform luminance distribution according to the analysis results from our simulation model. From the simulation results, we also determine the tolerance of the dielectric layer thickness in a CNT-FED to achieve uniform luminance and spot size on the screen. r 2007 Elsevier B.V. All rights reserved. PACS: 61.46.+w; 79.70.+q; 85.60.Pg Keywords: CNT-FED; Field emission; Dielectric layer; Film uniformity
1. Introduction FED is considered to be a promising flat panel display, due to its excellent performance such as high luminance, wide viewing angle, CRT-like color gamut and low power consumption. Many prototypes of FED device were firstly realized by conical microtips of Mo or Si, which are called Spint-type field emitters [1]. In 1991, the carbon nanotubes (CNTs) were discovered by Iijima [2]. Since the application of CNTs as field emission cathode by De Heer [3] in 1995, the technologies of cathode fabrication based on CNTs have been rapidly developed. Due to the rather low cost, the screen-printing method would be an interesting option for fabricating large size CNT-FEDs. However, FEDs fabricated by screen-printing CNT-paste usually suffer from a non-uniform luminance distribution [4,5]. This non-uniform luminance distribution may be caused by a poor dispersion of CNTs into the printing paste. Besides, the fringe field effect is serious near the edge of the CNT cathode, so it influences the emission uniformity of Corresponding author. Fax: +86 25 8336 3222.
E-mail address: [email protected] (W. Lei). 0304-3991/$ - see front matter r 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.ultramic.2007.02.017
the CNT cathode. Apart from this reason, the triode dimensions, such as fluctuations of the diameter of aperture and of the thickness of the dielectric layer, can also generate a non-uniform luminance distribution. In practice the triode parameters, such as the thickness of the dielectric layer, etc., may differ from the design values and show tolerance ranges. These fluctuations or tolerance ranges will affect the display performance [6,7]. So it is important to analyze the sensitivity of the structure parameters on the display performance in CNT-FEDs, which are fabricated by screen-printing method. In this paper, the influences of variations in the dielectric layer thickness on the field emission performance and the luminance distribution are described. 2. Experiment and results The fabrication process of a CNT triode with screenprinting method has been depicted in Fig. 1. As shown in Fig. 1(a)–(e), the cathode plate consists of cathode electrode, CNT emitter, dielectric layer and gate electrode, which are printed and sintered sequentially. The diameter of the gate aperture is 1 mm and the diameter of CNTs
ARTICLE IN PRESS X. Su et al. / Ultramicroscopy 107 (2007) 844–848
845
Fig. 2. Phosphor screen image when the same voltage is applied to all pixels. The anode voltage and gate voltage are 2000 and 150 V, respectively.
Fig. 1. Fabrication process of a CNT triode with screen-printing method.
cathode dot is 0.8 mm of our triode. To minimize leakagecurrent between the gate electrode and cathode electrode, the dielectric layer has been printed and sintered three times. As a result, the thickness of the dielectric layer is about 30 mm. As shown in Fig. 1(f) and (g), the anode plate is fabricated separately: an ITO layer is coated on the anode plate and a layer of the CRT-phosphors (P22-type commercial green phosphor) is printed on top of the ITO layer. The cathode plate is separated using ceramic slices from the anode plate, as shown in Fig. 1(h). The screen-printing paste is the mixture of CNT powders and vehicle. The CNT powders are synthesized by chemical vapor deposition method, while the vehicle is made of ethyl cellulose and a-terpineol. The raw CNT powders are purified and dispersed by ball-milling and filtering processes before being incorporated into the paste [8]. The diameter and length of the CNTs are respectively 10–20 nm and 0.5–10 mm. When the voltage applied to the gate electrode is high enough, electrons can emit from the CNTs cathode, which are accelerated by the anode voltage and bombard the phosphor layer. Fig. 2 shows the luminance image when the anode voltage is 2000 V and the gate voltage is 150 V. In this luminance image, every spot corresponds with a CNTs film in a single hole in the dielectric layer. From Fig. 2 it can be seen that the luminance distribution is not good due to the non-uniform field emission, since we assume that the effect of thickness variation of the phosphor layer can be neglected as is the case in CRTs. The dispersion of CNTs into the printing paste is likely an important factor influencing the emission uniformity.
Various ways have been used to disperse the CNTs in the printing paste: ultrasonic vibration, three milling roll dispersion and diverse detergents. However, the nonuniformity of the luminance is almost the same in these experiments. Consequently, it is assumed that other not well-controlled parameters in our experiments may affect the luminance uniformity as well. As shown in Fig. 1, the gate electrodes of the triode structure are located on top of the dielectric layer. The distance between the CNTs and the gate electrode affects the local field at the top of the CNTs. If the thickness of the dielectric layer is not uniform, the emission of a CNTs cathode array is not uniform either. To find out the effects of the dielectric layer thickness on the luminance uniformity, we first need to investigate the fluctuation of the dielectric layer thickness. The dielectric layer thickness in different regions is measured with an ultrasonic thickness gauge of Krautkramer, type CL3 DL. Its accuracy is smaller than 1 mm, so it can be used to get a good estimate of the dielectric layer thickness. Fig. 3 shows the fluctuation of the dielectric layer thickness obtained from these measurements. In this figure a cubic interpolation method is adopted to get a smooth representation of the discrete data. From these measurement results it can be seen that the fluctuation of the dielectric layer thickness is larger than 5 mm. Fig. 4(a) and (b) are SEM images of cross-sections of the dielectric layer in two different regions, which show some bulges in the dielectric layer. The height of bulge is larger than 5 mm. A few bulges or concaves have been observed, confirming once more that the uniformity of the dielectric layer thickness is not good. These SEM observations are similar to the fluctuation measured with the ultrasonic thickness gauge.
ARTICLE IN PRESS 846
X. Su et al. / Ultramicroscopy 107 (2007) 844–848
first. A finite difference method is used to calculate the distribution of the electric field in three-dimensional (3D) space [7]. Neglecting space charge effects, the electrostatic potential can be calculated from Laplace equation by q2 j q2 j q2 j þ þ ¼ 0, q2 x q2 y q2 z
(1)
where j is the potential of a node. The CNT is simplified as a cylinder with a dome cap. To fit arbitrary boundaries the mesh is suggested to be irregular. After the calculation of the potential distribution, the electric field is estimated by interpolation with high order polynomials [9,10]. By using the Fowler–Nordheim equation the emission current density from a single tip Jtip can be expressed as Fig. 3. Fluctuation of the dielectric layer, reconstructed from the measurements.
J tip ¼ at2 ðyÞf1 E 2tip exp½bvðyÞf3=2 =E tip ,
(2)
where f is the local work function and Etip the local field at the tip, a (E1.54 106 A eV V2) and b ð 6:83 107 eV3=2 V cm1 Þ are the Fowler–Nordheim constants, 4 and y ¼ cE1/2 eV V1/2 cm1/2, v(y) tip /f, where cE3.79 10 2 and t (y) are electric field-dependent elliptical functions. It is generally assumed that the approximations t2(y) ¼ 1.1 and v(y) ¼ 0.95y2 may be applied for most cathodes at operational conditions [9]. After the calculation of the emission current from a single CNT, the total emission current of the CNTs array can be obtained with the following equation: I¼
n X
J i Si ,
(3)
i¼1
Fig. 4. SEM images of cross-sections of the dielectric layer in two different regions, which show some bulges (as the arrows indicated) in the dielectric layer. The height of bulge is larger than 5 mm.
3. Modeling and discussion From the above observations and measurements, it is obvious to assume that the fluctuation of the dielectric layer thickness influences the luminance uniformity if the triode structure is fabricated with the screen-printing method. This section presents some analysis results by means of a numerical simulation method of the sensitivity of the emission and luminance as a function of dielectric layer thickness. Only the influence of the fluctuation of the dielectric layer thickness has been analyzed; other parameters, such as the aperture diameter and the density of CNTs, are regarded as the constants. To calculate the field emission current, the electric potential and field distribution of the triode is calculated
where I is the total emission current from the CNTs array, Ji and Si are the local emission current density and the emission area associated with the ith CNT, respectively. However, the computation of the current emitted from the cathode according to Eqs. (1)–(3) is complicated and quite slow. One complication is due to the fact that not all CNTs in the CNT layer, which is fabricated by screenprinting, are vertically aligned to the cathode substrate. Another complication is that a part of the emission current could come from defects of the CNTs [11,12]. Consequently, it is quite difficult to use Eqs. (2) and (3) to estimate the current density of a printed CNTs cathode film accurately. In this study we have used a simple two steps method to calculate the emission current in a CNTs triode [13], e.g., in step 1 a CNT cathode is fabricated with the screen-printing method and in step 2 the emission characteristic of this CNTs cathode film is measured in a diode. Fig. 5 shows the J–E curve obtained in the experiment. The current density of a triode can now be estimated from the J–E curve of Fig. 5 and the electric field strength in front of the cathode, which has been calculated separately with the numerical method outlined before. In order to evaluate the effect of the fluctuation of dielectric layer thickness on the emission characteristics, triode structures with dielectric layer thickness of 25, 30 and 35 mm are studied. The diameter of the holes in the
ARTICLE IN PRESS X. Su et al. / Ultramicroscopy 107 (2007) 844–848
847
Fig. 5. Experimental J–E curve for a CNT-film cathode.
Fig. 6. Emission current plotted versus gate voltage at different dielectric layer thickness. Anode voltage is 2000 V.
dielectric layer in our simulation is 100 mm. The emission current is calculated as a function of gate voltage for the different dielectric layer thicknesses at an anode voltage of 2000 V, as shown in Fig. 6. It can be seen clearly from this figure that the emission current decreases substantially when the dielectric layer thickness increases; moreover, the thinner the dielectric layer, the greater the influence of the dielectric layer thickness fluctuation. The simulation results of the potential distribution and trajectories are shown in Fig. 7. The fourth order Runge–Kutta method is used to calculate the electron trajectories [9]. Besides this, the focus performance of the electron beam and the spot size on the screen are also presented in this study. From Fig. 7 it can be seen that the electron beam is under-focused when the dielectric layer thickness is 25 mm. When the dielectric layer thickness increases, the electron lens effect near the gate aperture becomes stronger and consequently, the electron beam is
Fig. 7. Simulation results of the electric field distribution, trajectories and spot size when the dielectric layer’s thickness is, respectively, 25 mm (a), 30 mm (b) and 35 mm (c). Diameter of the hole is 100 mm, and the anode voltage and gate voltage are 2000 and 150 V, respectively.
focused stronger. When the dielectric layer thickness is 35 mm, the electron beam is over-focused and the spot size becomes large. From the simulation results presented above, it can be clearly seen that the emission current of a triode fabricated from CNTs by screen-printing is very sensitive to the dielectric layer thickness. If the dielectric layer thickness varies with 10%, the corresponding variation of the emission current is about 30%. Consequently, the luminance uniformity on the screen is also influenced by the uniformity of the thickness of dielectric layer. Since, to ensure the pixel-to-pixel luminance variation o3% and the corner-to-corner luminance variation o30% in a large-size display device [6], the pixel-to-pixel fluctuation of the
ARTICLE IN PRESS 848
X. Su et al. / Ultramicroscopy 107 (2007) 844–848
dielectric layer thickness must be o1 mm and the long range fluctuation should be o5 mm, respectively. 4. Conclusions It has been demonstrated that thick-film deposition techniques, such as the screen-printing technique, yield a certain amount of variation in thickness. Such variation in gate thickness could potentially affect field emission current uniformity. In this paper, an experimental investigation is carried out to study the influences of the dielectric layer thickness variation on the field emission characteristics and luminance distribution in a CNT-FED fabricated by screen-printing. Our findings indicate that the dielectric layer thickness fluctuations are mainly larger than 5 mm, which mean the dielectric layer thickness fluctuation is an important cause of the non-uniform luminance distribution according to the analysis results from our simulation model. From the simulation results, we also determine the tolerance of the dielectric layer thickness in a CNT-FED to achieve uniform luminance and spot size on the screen. The simulation results show that the tolerance of the dielectric layer thickness should be o5 mm. Acknowledgments The project is supported by National Key Basic Research Program 973 (2003CB314702, 2003CB314706),
Foundation of Doctoral Program of Ministry of Education (20030286003) and Program for New Century Excellent Talents in University. The PPD department of LG.Philips Display Company provides the simulation tool. We express them great thanks for this help. We also like to thank Dr. Daniel den Engelsen in Philips, for his help in preparing this paper.
References [1] S. Itoh, M. Tanaka, T. Tonegawa, J. Vac. Sci. Technol. B 22 (2004) 1362. [2] S. Iijima, Nature 354 (1991) 56. [3] W.A. De Heer, A. Chatelain, D. Ugarte, Science 270 (1995) 1179. [4] S.Z. Deng, Z.S. Wu, N.S. Xu, J. Chen, Ultramicroscopy 89 (2001) 105. [5] K.F. Chen, K.C. Chen, Y.C. Jiang, L.Y. Jiang, Y.Y. Chang, M.C. Hsiao, L.H. Chan, Appl. Phys. Lett. 88 (2006) 193127. [6] C. Xie, IEEE Trans. Nanotechnology 3 (2004) 404. [7] W. Lei, B. Wang, H. Yin, J. Vac. Sci. Technol. B 17 (1999) 1575. [8] S.K. Kang, J.H. Choi, J.H. Park, J.H. Han, J.-B. Yoo, J.W. Nam, C.K. Lee, J.M. Kim, J. Vac. Sci. Technol. B 22 (2004) 1345. [9] R.G. Forbes, J. Vac. Sci. Technol. B 17 (1999) 534. [10] D.W. Jenkins, IEEE Trans. Electron Devices 40 (1993) 665. [11] Y. Chen, D.T. Shaw, L. Guo, Appl. Phys. Lett. 76 (2000) 2469. [12] A. Sawada, M. Iriguchi, W.J. Zhao, C. Ochiai, M. Takai, J. Vac. Sci. Technol. B 21 (2003) 362. [13] W. Lei, X. Zhang, G. Yang, M. Xie, J. Vac. Sci. Technol. B 24 (2006) 962.
Ultramicroscopy 107 (2007) 844–848 www.elsevier.com/locate/ultramic
Study of the influence of the dielectric layer thickness in a CNT-FED Xing Su, Lifang Zhang, Wei Lei, Xiaobing Zhang Department of Electronic Engineering, Southeast University, Nanjing 210096, People’s Republic of China
Abstract An experimental investigation is carried out to study the influences of the dielectric layer thickness variation on the field emission characteristics and luminance distribution in a CNT-FED fabricated by screen-printing. Two steps are contained in the investigation: (1) the dielectric layer thickness fluctuations are presented with an ultrasonic thickness gauge, and (2) a simulation model is constructed to study the corresponding influences of the dielectric layer thickness fluctuations on the field emission characteristics and luminance uniformity on the screen. Our findings indicate that the dielectric layer thickness fluctuations are mainly larger than 5 mm, which mean the dielectric layer thickness fluctuation is an important cause of the non-uniform luminance distribution according to the analysis results from our simulation model. From the simulation results, we also determine the tolerance of the dielectric layer thickness in a CNT-FED to achieve uniform luminance and spot size on the screen. r 2007 Elsevier B.V. All rights reserved. PACS: 61.46.+w; 79.70.+q; 85.60.Pg Keywords: CNT-FED; Field emission; Dielectric layer; Film uniformity
1. Introduction FED is considered to be a promising flat panel display, due to its excellent performance such as high luminance, wide viewing angle, CRT-like color gamut and low power consumption. Many prototypes of FED device were firstly realized by conical microtips of Mo or Si, which are called Spint-type field emitters [1]. In 1991, the carbon nanotubes (CNTs) were discovered by Iijima [2]. Since the application of CNTs as field emission cathode by De Heer [3] in 1995, the technologies of cathode fabrication based on CNTs have been rapidly developed. Due to the rather low cost, the screen-printing method would be an interesting option for fabricating large size CNT-FEDs. However, FEDs fabricated by screen-printing CNT-paste usually suffer from a non-uniform luminance distribution [4,5]. This non-uniform luminance distribution may be caused by a poor dispersion of CNTs into the printing paste. Besides, the fringe field effect is serious near the edge of the CNT cathode, so it influences the emission uniformity of Corresponding author. Fax: +86 25 8336 3222.
E-mail address: [email protected] (W. Lei). 0304-3991/$ - see front matter r 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.ultramic.2007.02.017
the CNT cathode. Apart from this reason, the triode dimensions, such as fluctuations of the diameter of aperture and of the thickness of the dielectric layer, can also generate a non-uniform luminance distribution. In practice the triode parameters, such as the thickness of the dielectric layer, etc., may differ from the design values and show tolerance ranges. These fluctuations or tolerance ranges will affect the display performance [6,7]. So it is important to analyze the sensitivity of the structure parameters on the display performance in CNT-FEDs, which are fabricated by screen-printing method. In this paper, the influences of variations in the dielectric layer thickness on the field emission performance and the luminance distribution are described. 2. Experiment and results The fabrication process of a CNT triode with screenprinting method has been depicted in Fig. 1. As shown in Fig. 1(a)–(e), the cathode plate consists of cathode electrode, CNT emitter, dielectric layer and gate electrode, which are printed and sintered sequentially. The diameter of the gate aperture is 1 mm and the diameter of CNTs
ARTICLE IN PRESS X. Su et al. / Ultramicroscopy 107 (2007) 844–848
845
Fig. 2. Phosphor screen image when the same voltage is applied to all pixels. The anode voltage and gate voltage are 2000 and 150 V, respectively.
Fig. 1. Fabrication process of a CNT triode with screen-printing method.
cathode dot is 0.8 mm of our triode. To minimize leakagecurrent between the gate electrode and cathode electrode, the dielectric layer has been printed and sintered three times. As a result, the thickness of the dielectric layer is about 30 mm. As shown in Fig. 1(f) and (g), the anode plate is fabricated separately: an ITO layer is coated on the anode plate and a layer of the CRT-phosphors (P22-type commercial green phosphor) is printed on top of the ITO layer. The cathode plate is separated using ceramic slices from the anode plate, as shown in Fig. 1(h). The screen-printing paste is the mixture of CNT powders and vehicle. The CNT powders are synthesized by chemical vapor deposition method, while the vehicle is made of ethyl cellulose and a-terpineol. The raw CNT powders are purified and dispersed by ball-milling and filtering processes before being incorporated into the paste [8]. The diameter and length of the CNTs are respectively 10–20 nm and 0.5–10 mm. When the voltage applied to the gate electrode is high enough, electrons can emit from the CNTs cathode, which are accelerated by the anode voltage and bombard the phosphor layer. Fig. 2 shows the luminance image when the anode voltage is 2000 V and the gate voltage is 150 V. In this luminance image, every spot corresponds with a CNTs film in a single hole in the dielectric layer. From Fig. 2 it can be seen that the luminance distribution is not good due to the non-uniform field emission, since we assume that the effect of thickness variation of the phosphor layer can be neglected as is the case in CRTs. The dispersion of CNTs into the printing paste is likely an important factor influencing the emission uniformity.
Various ways have been used to disperse the CNTs in the printing paste: ultrasonic vibration, three milling roll dispersion and diverse detergents. However, the nonuniformity of the luminance is almost the same in these experiments. Consequently, it is assumed that other not well-controlled parameters in our experiments may affect the luminance uniformity as well. As shown in Fig. 1, the gate electrodes of the triode structure are located on top of the dielectric layer. The distance between the CNTs and the gate electrode affects the local field at the top of the CNTs. If the thickness of the dielectric layer is not uniform, the emission of a CNTs cathode array is not uniform either. To find out the effects of the dielectric layer thickness on the luminance uniformity, we first need to investigate the fluctuation of the dielectric layer thickness. The dielectric layer thickness in different regions is measured with an ultrasonic thickness gauge of Krautkramer, type CL3 DL. Its accuracy is smaller than 1 mm, so it can be used to get a good estimate of the dielectric layer thickness. Fig. 3 shows the fluctuation of the dielectric layer thickness obtained from these measurements. In this figure a cubic interpolation method is adopted to get a smooth representation of the discrete data. From these measurement results it can be seen that the fluctuation of the dielectric layer thickness is larger than 5 mm. Fig. 4(a) and (b) are SEM images of cross-sections of the dielectric layer in two different regions, which show some bulges in the dielectric layer. The height of bulge is larger than 5 mm. A few bulges or concaves have been observed, confirming once more that the uniformity of the dielectric layer thickness is not good. These SEM observations are similar to the fluctuation measured with the ultrasonic thickness gauge.
ARTICLE IN PRESS 846
X. Su et al. / Ultramicroscopy 107 (2007) 844–848
first. A finite difference method is used to calculate the distribution of the electric field in three-dimensional (3D) space [7]. Neglecting space charge effects, the electrostatic potential can be calculated from Laplace equation by q2 j q2 j q2 j þ þ ¼ 0, q2 x q2 y q2 z
(1)
where j is the potential of a node. The CNT is simplified as a cylinder with a dome cap. To fit arbitrary boundaries the mesh is suggested to be irregular. After the calculation of the potential distribution, the electric field is estimated by interpolation with high order polynomials [9,10]. By using the Fowler–Nordheim equation the emission current density from a single tip Jtip can be expressed as Fig. 3. Fluctuation of the dielectric layer, reconstructed from the measurements.
J tip ¼ at2 ðyÞf1 E 2tip exp½bvðyÞf3=2 =E tip ,
(2)
where f is the local work function and Etip the local field at the tip, a (E1.54 106 A eV V2) and b ð 6:83 107 eV3=2 V cm1 Þ are the Fowler–Nordheim constants, 4 and y ¼ cE1/2 eV V1/2 cm1/2, v(y) tip /f, where cE3.79 10 2 and t (y) are electric field-dependent elliptical functions. It is generally assumed that the approximations t2(y) ¼ 1.1 and v(y) ¼ 0.95y2 may be applied for most cathodes at operational conditions [9]. After the calculation of the emission current from a single CNT, the total emission current of the CNTs array can be obtained with the following equation: I¼
n X
J i Si ,
(3)
i¼1
Fig. 4. SEM images of cross-sections of the dielectric layer in two different regions, which show some bulges (as the arrows indicated) in the dielectric layer. The height of bulge is larger than 5 mm.
3. Modeling and discussion From the above observations and measurements, it is obvious to assume that the fluctuation of the dielectric layer thickness influences the luminance uniformity if the triode structure is fabricated with the screen-printing method. This section presents some analysis results by means of a numerical simulation method of the sensitivity of the emission and luminance as a function of dielectric layer thickness. Only the influence of the fluctuation of the dielectric layer thickness has been analyzed; other parameters, such as the aperture diameter and the density of CNTs, are regarded as the constants. To calculate the field emission current, the electric potential and field distribution of the triode is calculated
where I is the total emission current from the CNTs array, Ji and Si are the local emission current density and the emission area associated with the ith CNT, respectively. However, the computation of the current emitted from the cathode according to Eqs. (1)–(3) is complicated and quite slow. One complication is due to the fact that not all CNTs in the CNT layer, which is fabricated by screenprinting, are vertically aligned to the cathode substrate. Another complication is that a part of the emission current could come from defects of the CNTs [11,12]. Consequently, it is quite difficult to use Eqs. (2) and (3) to estimate the current density of a printed CNTs cathode film accurately. In this study we have used a simple two steps method to calculate the emission current in a CNTs triode [13], e.g., in step 1 a CNT cathode is fabricated with the screen-printing method and in step 2 the emission characteristic of this CNTs cathode film is measured in a diode. Fig. 5 shows the J–E curve obtained in the experiment. The current density of a triode can now be estimated from the J–E curve of Fig. 5 and the electric field strength in front of the cathode, which has been calculated separately with the numerical method outlined before. In order to evaluate the effect of the fluctuation of dielectric layer thickness on the emission characteristics, triode structures with dielectric layer thickness of 25, 30 and 35 mm are studied. The diameter of the holes in the
ARTICLE IN PRESS X. Su et al. / Ultramicroscopy 107 (2007) 844–848
847
Fig. 5. Experimental J–E curve for a CNT-film cathode.
Fig. 6. Emission current plotted versus gate voltage at different dielectric layer thickness. Anode voltage is 2000 V.
dielectric layer in our simulation is 100 mm. The emission current is calculated as a function of gate voltage for the different dielectric layer thicknesses at an anode voltage of 2000 V, as shown in Fig. 6. It can be seen clearly from this figure that the emission current decreases substantially when the dielectric layer thickness increases; moreover, the thinner the dielectric layer, the greater the influence of the dielectric layer thickness fluctuation. The simulation results of the potential distribution and trajectories are shown in Fig. 7. The fourth order Runge–Kutta method is used to calculate the electron trajectories [9]. Besides this, the focus performance of the electron beam and the spot size on the screen are also presented in this study. From Fig. 7 it can be seen that the electron beam is under-focused when the dielectric layer thickness is 25 mm. When the dielectric layer thickness increases, the electron lens effect near the gate aperture becomes stronger and consequently, the electron beam is
Fig. 7. Simulation results of the electric field distribution, trajectories and spot size when the dielectric layer’s thickness is, respectively, 25 mm (a), 30 mm (b) and 35 mm (c). Diameter of the hole is 100 mm, and the anode voltage and gate voltage are 2000 and 150 V, respectively.
focused stronger. When the dielectric layer thickness is 35 mm, the electron beam is over-focused and the spot size becomes large. From the simulation results presented above, it can be clearly seen that the emission current of a triode fabricated from CNTs by screen-printing is very sensitive to the dielectric layer thickness. If the dielectric layer thickness varies with 10%, the corresponding variation of the emission current is about 30%. Consequently, the luminance uniformity on the screen is also influenced by the uniformity of the thickness of dielectric layer. Since, to ensure the pixel-to-pixel luminance variation o3% and the corner-to-corner luminance variation o30% in a large-size display device [6], the pixel-to-pixel fluctuation of the
ARTICLE IN PRESS 848
X. Su et al. / Ultramicroscopy 107 (2007) 844–848
dielectric layer thickness must be o1 mm and the long range fluctuation should be o5 mm, respectively. 4. Conclusions It has been demonstrated that thick-film deposition techniques, such as the screen-printing technique, yield a certain amount of variation in thickness. Such variation in gate thickness could potentially affect field emission current uniformity. In this paper, an experimental investigation is carried out to study the influences of the dielectric layer thickness variation on the field emission characteristics and luminance distribution in a CNT-FED fabricated by screen-printing. Our findings indicate that the dielectric layer thickness fluctuations are mainly larger than 5 mm, which mean the dielectric layer thickness fluctuation is an important cause of the non-uniform luminance distribution according to the analysis results from our simulation model. From the simulation results, we also determine the tolerance of the dielectric layer thickness in a CNT-FED to achieve uniform luminance and spot size on the screen. The simulation results show that the tolerance of the dielectric layer thickness should be o5 mm. Acknowledgments The project is supported by National Key Basic Research Program 973 (2003CB314702, 2003CB314706),
Foundation of Doctoral Program of Ministry of Education (20030286003) and Program for New Century Excellent Talents in University. The PPD department of LG.Philips Display Company provides the simulation tool. We express them great thanks for this help. We also like to thank Dr. Daniel den Engelsen in Philips, for his help in preparing this paper.
References [1] S. Itoh, M. Tanaka, T. Tonegawa, J. Vac. Sci. Technol. B 22 (2004) 1362. [2] S. Iijima, Nature 354 (1991) 56. [3] W.A. De Heer, A. Chatelain, D. Ugarte, Science 270 (1995) 1179. [4] S.Z. Deng, Z.S. Wu, N.S. Xu, J. Chen, Ultramicroscopy 89 (2001) 105. [5] K.F. Chen, K.C. Chen, Y.C. Jiang, L.Y. Jiang, Y.Y. Chang, M.C. Hsiao, L.H. Chan, Appl. Phys. Lett. 88 (2006) 193127. [6] C. Xie, IEEE Trans. Nanotechnology 3 (2004) 404. [7] W. Lei, B. Wang, H. Yin, J. Vac. Sci. Technol. B 17 (1999) 1575. [8] S.K. Kang, J.H. Choi, J.H. Park, J.H. Han, J.-B. Yoo, J.W. Nam, C.K. Lee, J.M. Kim, J. Vac. Sci. Technol. B 22 (2004) 1345. [9] R.G. Forbes, J. Vac. Sci. Technol. B 17 (1999) 534. [10] D.W. Jenkins, IEEE Trans. Electron Devices 40 (1993) 665. [11] Y. Chen, D.T. Shaw, L. Guo, Appl. Phys. Lett. 76 (2000) 2469. [12] A. Sawada, M. Iriguchi, W.J. Zhao, C. Ochiai, M. Takai, J. Vac. Sci. Technol. B 21 (2003) 362. [13] W. Lei, X. Zhang, G. Yang, M. Xie, J. Vac. Sci. Technol. B 24 (2006) 962.