Cold Cathode Structures and Materials for Field Emission Displays

most promising applications of flat panel displays, which can overcome the drawbacks of TFT-LCD. The features of FEDs are small volume, low power consumption, fast response speed, color performance similar to CRT, large view angles, and low cost. Otherwise, FEDs have some drawbacks, such as the uniformity of brightness, high breakdown voltage, the package spacer charging, short life time, and vacuum package, which are all serious problems.

In the past 30 years, so many applications of FEDs were demonstrated by the academia and noted laboratories. Table 1.3 showed the comparison of the FEDs. Based on the emitting methods, we could separate the FEDs into two groups: direct and fractional emitters. The CNTs and Spindt type can emit electrons directly from the emitters, so they belong to the

direct emitters. Moreover, the BSD, MIM, and SED emit electrons form dispersed nano-particles of the thin films, and they belong to the fractional emitters.

A. Spindt-Type Field Emitters

Since 1960s, the concepts of FEDs were demonstrated, and many scientists contributed their innovation to the FED applications. Ken Shoulders published the first micro-electric device based on the F-N tunneling theory [1.26]. However, Capp Spindt who successfully fabricated and operated the FEA (Field Emission Array) contributed the semiconductor technology to the vacuum microelectronics, called the Spindt-type cold cathode [1.7]. The Spindt-type is micro-meter scale emitters and possess self aligned metal pyramids and gates [1.27]. The Fig 1-4 was Spindt-type triode structure fabricated by Capp Spindt. From then on, many applications of FEDs were demonstrated by the academia and noted laboratories. The Spindt-type has some advantages, such as high thermal stability, high emission reliability, low cost, and larger area displays of many tip arrays. In addition, the most importance is Spindt-type can be easily fabricated in the general semiconductor fabrication technology.

However, the most popular Spindt-type material is silicon tip emitters. Fig 1-5 is silicon tip diagrams. The most desirable characteristics of silicon tips are the low operation voltage, emission stability, long durability, and high emission current density. These properties are strongly dependent on the tip curvature radius (related to the field enhancement factor), emission material (determines the work function), and surface conditions. According to Eq.(1-1), the methods to enhance the emission currents are increasing the work function of cathode materials and increasing the field emission factor (β). We had to decrease the work function of silicon tips and coat conduction layers, such as DLC (Diamond Like Carbon), carbon films, Mo….etc [1.9-1.11,1-29]. Conduction layers can decrease the silicon work function and enhance the emission efficiency. However, there are some existing drawbacks of Spindt type field emitters when fabricating Spindt type FED such as (1) high gate driving

voltage required; for a Spindt type field emission triode with 4 µm gate aperture, the driving voltage is typically more than 60 V, which results in the high cost of the driving circuits. To reduce the gate driving voltage, (1) frontier lithography technologies such as E beam lithography must be applied to reduce the gate aperture to the sub-micron level and (2) huge and expensive high vacuum deposition system is required during fabricating large area Spindt type FED.

Some research groups had successfully fabricated commercial FED products based on Spindt type field emitters such as Futaba, Sony/Candesent, Futaba and Pixtech[1.30], the products above mentioned companies were shown in Fig.1-6.

B. BSD Field Emitters

BSD (Ballistic Electron Surface Emitting Device) was developed by Panasonic and Tokyo University of Agriculture and Technology. The operation diagram of BSD was showed in the Fig 1-7. The PPS (porous poly-Si) of cathodes of BSD have micro-crystal particles, which have a thin oxide film on itself. The offset voltages are applied between the anode and cathode, the electrons emitted by the cathode inject into the PPS, the electrons accelerated by the PPS interactions gain high energies and emit, and the emission efficiency of electrons is about 2%. Because the external voltages are focused on the oxide surface of the micro-crystal particles, then the thin film surface forms strong electrical fields, and the electrons can be emitted. However, the phenomenon called “Ballistic Electron Conduction”. In addition, the high active electrons are emitted from the cathode in the vertical direction without the deviation adjustment. The BSD has the advantages, such as simple structures, but it is very sensitive to the ion bombardments and has serious drawbacks like the poor emission efficiency.

C. MIM Field Emitters

MIM (Metal Insulator Metal) technology was investigated by Hitachi and LGE, but they didn’t publish the detailed technology researches. According to the paper survey, MIM technology is the hot electron generation in the dielectric. The features of MIM are the low degrees of electron dispersion, low driving voltages, and high current density, but it is easily affected by the ion bombardments. However, the structure and fabrication of MIM are very complicated.

D. SCE Field Emitters

Canon and Toshiba co-operated the whole new 36 inches planar panel display called

“SCE (Surface Conduction Emitter)” and published a plan for large productions in Sep 14 2004 [1.31]. Fig 1-8 was the base structure of SCE. At SID 2005, Canon showed the SCE display specifications, including 400 cd/m² peak luminance, high contrast about 10000:1, response speed below 1ms, and panel thickness about 7.3mm. By the way, it showed the superior emission stability

As the Fig 1-8, the mechanism of emission of SCE is the external voltage is applied on the PdO particles located in the extreme narrow gaps, the electrons emit from the extreme narrow gaps based on tunneling effect, and electrons hit onto the phosphors of anodes.

Moreover, the PdO thin film was coated by the ink-jet methods, and its thickness was about 10nm. The currents passed through the PdO thin film and about 2% electrons emited into the vacuum. Its emission efficiency compared to other field emission emitters is much better. The features of SCE are the simple structure, low cost and large panel display. But, in accord with the other FEDs, some essential properties like stable emission currents, low driving voltages, low cost fabrications, and longer operation life are still needed to be resolved.

1.3 The Promising Field Emission Technology – Carbon Nanotubes

Since the discovery of carbon nanotubes (CNTs) by Iijima in 1991, [1.32] CNTs have attracted considerable interests because of their unique physical properties and many potential applications [1.33]. CNTs have numerous potential applications in nanoelectronics, nanometer-scale structural materials, hydrogen storage, field-emission devices, and so on.

Among these applications, CNTs seem to be very promising as electron emitters for field-emission displays (FEDs).

1.3.1 Structures of Carbon Nanotubes

Carbon nanotubes are cylindrical molecules consisting of single or multi-graphite layers.

Unlike other covalent matters, carbon shows various structures. The common allotropes of carbon are (1) graphite, (2) diamond, and (3) fullerene (Fig. 1-9). Among them, graphite is a planar structure of carbon in sp² bonding. As the scale of graphite is close to the nano scale, it could curl and form a hollow tube. According to the number of layers in the tube wall, we classify the carbon nanotubes into two kinds: (1) single-walled (SWNTs) and (2) multi-walled carbon nanotubes (MWNTs) (Fig. 1-10).

A. Single-walled Carbon Nanotubes

The curling direction of graphite decides the characteristics of single-walled carbon nanotubes, like metal or semiconductor. Based on the curling direction, we can classify SWNTs into three kinds: (1) armchair, (2) zigzag, (3) chiral. The armchair SWNTs reveal metallic property. The zigzag and chiral SWNTs possess both of metallic and semiconductor properties.

B. Multi-walled Carbon Nanotubes

Since the distance between layers in MWNTs must satisfy 0.34nm, the curl direction of

every layer may be different. MWNTs present different morphology by various growing parameters. In terms of appearance, we can approximately categorize MWNTs into several types: (1) standard tube, (2) spiral, and (3) “Y” model. If we further investigate the microstructure of the MWNTs, we can observe that there are different structures such as hollow-liked, bamboo-liked and fish-bone-liked MWNTs.

1.3.2 Physical and Chemical Properties of Carbon Nanotubes

Carbon nanotubes have several superior characteristics, such as high thermal conductivity (2000 W/m.K), good heat resistance, low chemical reactivity, and high aspect ratio [1.34]. These properties are suitable for a field-emission source. In addition, carbon nanotubes have an excellent mechanical property, especially for the high Young modulus about 1.25T Pa [1.35]. But, CNTs have drawbacks, such as semiconductor/metal types [1.36], and uniformity growth issues. As the Fig 1-11, carbon nanotubes growth can be classified into two types: tip growth model and base growth model. There are many explanations about carbon nanotubes growth mechanism. Generally speaking, the growth mechanism is VLS (Vapor Liquid Solid). First, catalyst films transform to the particles because of thermal force or plasma. Formation of catalyst particles enhances the diffusion of carbon in the metal alloy.

Second, the hydrocarbon precursors are decomposed from the reaction gases and carbon sources dissolve into the catalyst particles until the carbon saturation. Finally, carbons diffuse from the catalyst center to its surface and precipitate to form carbon SP2 bonding. Then, through the VLS, we can synthesize the carbon nanotubes.

1.3.3 The Synthesis Methods of Carbon Nanotubes

Carbon nanotubes (CNTs) have been extensively investigated for the synthesis using arc

discharge, laser vaporization, pyrolysis, solar energy, and plasma-enhanced chemical vapor deposition (CVD), for its unique physical and chemical properties and for applications to nanoscale devices. However, common methods of CNT synthesis include: (1) arc-discharge [1.37], (2) laser ablation [1.38], (3) thermal CVD [1.39-1-41], and (4) plasma enhanced CVD [1.42-1-43].

The laser ablation can synthesize pure carbon nanotubes in high fabrication temperature, but large scale display panel can not be fabricated in the high fabrication temperature above the melting point of glass substrate. The arc discharge can synthesize carbon nanotubes in shorter fabrication times, but it has some issues, such as (1) poor purity, (2) hard to control growth orientations of carbon nanotubes, and (3) poor emission uniformity. Compared to laser ablation and arc discharge, using CVD for carbon nanotube growth has some features, such as (1) high purity carbon nanotubes, (2) selective growth only for catalyst metal, (3) controlling growth direction, and (4) much suitable to semiconductor fabrication procedure. However, with the display technology trend, it is the time for large panel display. We need to synthesize carbon nanotubes using CVDs on large panel substrates. For this reason, carbon nanotube growth at low temperatures is unavoidable, but the purity of carbon nanotubes at low temperatures is poor. So, we still try some methods to increase growth rates at low temperatures, such as (1 ) multilayer catalysts, (2) plasma CVDs, (3) post-treatment, and so on.

1.3.4 Applications of Carbon Nanotubes

Since the discovery of CNTs in 1991, CNTs had attracted much attention for their unique physical and chemical properties. Their extensively potential applications lead them to become a super star of nano technology, which cover: (1) Chemical sensor [1.44], (2) IR

detector, (3) Nano-conducting Wire, (4) Vehicles for Hydrogen Storage [1.45], (5) Field Effect Transistor [1.46], (6) Field Emission Display (FED), (7) Probe of AFM and etc.

In the wide-ranging applications of CNTs, FED arouses researchers’ interest particularly.

In virtue of the superior field emission characteristics, CNTs are applied to the emitting source of cold cathode. The advantages of FED are its low response time, wide view angle, high brightness, high working temperature range and well combination with mature phosphor technology. However, a major problem needs to be solved in this field. It is not allowed to effectively analyze CNTs on a flat panel at relatively lower temperature (<500°C) and this barrier restriction obstructs the development of CNT-FED so far.

1.4 Paper Review

1.4.1 Overview of CNTs Growth at Low Temperature

Carbon nanotubes are the most important materials for field emission displays (FED) due to their superior emission characteristics as described in the previous section 1.3. Therefore, CNTs are attributed to be one of the most attractive of potential field emitters for flat-panel displays. How to synthesize high purified carbon nanotubes is a great challenge. Some critical factors like catalyst metals, hydrocarbon gases, and growth methods extremely affect CNT growth.

Various synthetic methods such as arc discharge [1.47], laser vaporization [1.48], pyrolysis, and plasma-enhanced [1.49-1.51] or thermal chemical vapor deposition CVD [1.52-1.53] were employed. Synthesis of well-aligned CNTs on a large area is necessary for one of various applications, electron emitter of field emission displays. The arc discharge and the laser vaporization techniques can produce the large amount of CNTs, but it is very difficult to control the alignment and size. These techniques also require purification process

to separate pure CNTs from other particles. Many research groups have employed the CVD method for the purpose of large scaled production of CNTs. It is shown that the CVD technique can synthesize the CNTs with high purity, high yield, selective growth, and good vertical alignment.

The synthesis of carbon nanotubes can be divided into non-catalytic and catalytic methods. In the catalytic method, nickel, iron and cobalt are the only three transition metals that can be used as pure-metal catalysts for carbon nanotube growth. In the considerable reports regarding carbon nanotube synthesis, nickel, iron and cobalt are used either separately in different methods or together as a composite catalyst. The study clearly shows that the catalyst strongly influences not only diameter, growth rate, etc., but also morphology and microstructure [1.54]. In the growth reaction of CNTs, the diffusion of carbon in the catalyst metal has been believed to be the rate-determining step. We also know that the solubility of carbon in a metal is temperature dependent [1.55-1.56]. As theTable 1.4, Fe, Co, and Ni are the most suitable catalyst metals compared to other metals for the CNT growth.

According to researches by Siegal et al., four hydrocarbon gases were studied: C2H2, C2H4, and C2H6, which have a C–C triple, double, and single bond, respectively, and CH4

[1.57]. The chemical reactivity of these compounds is related to their heats of formation, obtained from the CRC Handbook of Physics and Chemistry, and shown in Fig. 1-12. In general, the greater bond numbers, the more reactive. Therefore, the reactivity is C2H2>

C2H4> C2H6. With negative heats of formation, both ethane and methane are the least reactive, with ethane slightly less than methane.

Molecular dynamic simulation results using a potential energy surface method had been published by Ding et al. [1.58] showing that these iron clusters must reach supersaturated carbon levels before nanotube formation begins. This saturation had also been described by Kuznetsov with the use of phase diagrams [1.59]. The rapid diffusion of carbon in iron

allowed for rapid carbide formation and quick graphite precipitation. With cobalt and nickel, meta-stable carbides (Co3C, Co2C, and Ni3C [1.60]) formed immediately following saturation of carbon in the solid solution, and as additional carbon diffuses into the catalyst, graphite precipitates out, forming a nanotube. Furthermore, carbide formation inhibited nanotube growth from the pure metal, carbides could be used as potential catalysts to encourage graphite precipitation, provided the carbide particles were appropriately sized and carbon diffusion was sufficiently fast. For optimal growth in typical chemical vapor deposition, metal catalysts need to exhibit (1)sufficient carbon solubility, (2)rapid carbon diffusion, and (3)limited carbide formation[1.61]. Table 1.5 was summarized for the many kinds of carbide formation.

Considering the direct growth of CNTs on the substrate, thermal CVD is often used because the damage to the substrate is small. The reaction temperature of thermal CVD is generally higher than 700℃ in earlier experiments. There are two critical purposes to grow CNTs at higher temperature. One is to gain uniform catalyst nano-particles after pretreatment, and the other is to obtain fine graphite structure of CNTs.

Recently, the direct growth of CNTs on glass substrate by thermal CVD at low temperature has been researched for the fabrication of field emission displays. For practical application, soda lime glass is one of the most preferable materials for the cathode substrate of filed emission displays because of its cheapness. Although using soda lime glass substrate indeed costs down, the melting point of 570℃ restricts the processing temperature when growing CNTs by thermal CVD. Some methods like (1) external treatments or (2) two-stage thermal CVD are employed to enhance CNT growth at low temperature. Yih-Ming Shyu et al.

synthesized CNTs using Fe-Ni alloys by the thermal CVD at 400ْC [1.62-1.63]. As the Fig.

1-13, they introduced C2H2 into the pre-treatment region in order to increase carbon solubility of catalyst particles so that they successfully synthesize CNTs at 400ْC. As the Fig. 1-14,

Takeda et al. had found that preheating the reaction gas C2H2 at 700℃ enhanced its reactivity and contributed to the effective decomposition of the gas on the surfaces of catalyst particles heated even at 450℃ [1.64-1.65]. They could synthesize high graphite CNTs even better than CNT grown at high temperature by two stage temperature zones. Plasma-enhanced CVD is also the popular method for the CNT growth because plasma energy compared to thermal CVD (Fig. 1-15) can more effectively decompose catalyst metals to form the particles [1.66-1.68]. So, its methods are more suitable to CNT growth at low temperature. But, it still has drawbacks, including plasma damages to substrates, vacuum operation, high cost, and non-uniformity issues. For a summary, thermal CVD is most promising method for CNT growth.

1.4.2 Theoretical Background for CNTs at Low Temperature 1.4.2.A Effects of Catalyst Film Thickness on The Melting Point

The most importance for CNT growth is low temperature fabrication. As described in the previous section 1.4.1, there are so many methods to be employed for the low temperature fabrication. However, if we could effectively lower the melting point of catalyst metals, we would grow CNTs at much lower temperature and did not damage the graphite quality of CNTs. For examples, Takagi in 1954 [1.69] demonstrated for the first time that ultrafine metallic particles melt below their corresponding bulk melting temperatures. Moreover, the metallic nanocrystals could exhibit not only melting point depression, but also superheating, depending on their surrounding environment. When the thickness of a thin film reaches monolayer level, the thin film may melt at a much lower temperature than its bulk melting point. Recently, a model, free of any adjustable parameter, for the size-dependent melting for nanocrystals was introduced by the use of Mott’s expression for the vibrational entropy of

melting [1.70] and the Lindemann’s criterion for melting [1.71]. In light of Lindemann criterion, the melting point decreases abruptly when the catalyst was close to nano-sizes (Fig 1-16). Q. Jiang et al. investigated several thin films, including Pb, In, Sn, and Fe thin films [1.72]. As the Fig 1-17, taking the Fe for examples, and we can find out that the melting temperature of a crystalline thin film decreases as its thickness decreases. In other paper survey [1.73-1.74], they investigated the surface tension between the catalyst thin films and substrates and also proposed the same result on the lower melting point as the thin film thickness decreased.

1.4.2.B Effects of Buffer Layers on CNT Growth

The use of buffer layers between catalyst and silicon substrate in growth of carbon nanotubes by chemical vapor deposition (CVD) has the two effects of both increasing the efficiency of the growth process by avoiding undesired chemical interaction between catalyst and substrate, and of altering the catalyst-support interaction, therefore modifying the characteristics and growth rate of CNTs. In sum, using good buffer layers can effectively make catalyst films form the catalyst particles without merging together and control the CNT density as its thickness increases. In the paper surveys, the comparative efficiency of buffer layers of Al, Al2O3, TiN and TiO2 in the CVD growth of CNTs was investigated [1.75-1.76]^. Making use of an experimental setup for CNT growth combined with in situ photoelectron spectroscopy analysis, we found that CNTs did not grow whenever pure Al layers were employed, in opposition to the results published by Delzeit et al.. However, when we used the corresponding oxide Al2O3 as buffer layer CNT growth was successful. Furthermore, Al2O3

The use of buffer layers between catalyst and silicon substrate in growth of carbon nanotubes by chemical vapor deposition (CVD) has the two effects of both increasing the efficiency of the growth process by avoiding undesired chemical interaction between catalyst and substrate, and of altering the catalyst-support interaction, therefore modifying the characteristics and growth rate of CNTs. In sum, using good buffer layers can effectively make catalyst films form the catalyst particles without merging together and control the CNT density as its thickness increases. In the paper surveys, the comparative efficiency of buffer layers of Al, Al2O3, TiN and TiO2 in the CVD growth of CNTs was investigated [1.75-1.76]^. Making use of an experimental setup for CNT growth combined with in situ photoelectron spectroscopy analysis, we found that CNTs did not grow whenever pure Al layers were employed, in opposition to the results published by Delzeit et al.. However, when we used the corresponding oxide Al2O3 as buffer layer CNT growth was successful. Furthermore, Al2O3