Chapter 1 Introduction
1.6 Thesis Organization
In chapter 1, the overview of vacuum microelectronics and basic theory of field emission were first introduced. A description about the field emission displays and their cathode technologies were briefly addressed in following sections. Finally, the motivation of this thesis was mentioned before the thesis organization.
Chapter 2 shows the synthesis of nanotubes at low temperatures via atmospheric thermal CVD system in combination with multi-layered catalysts, and the functionality of different components in the catalysts is also discussed in this chapter.
In chapter 3, the field emission characteristics of nanotubes synthesized with multi-layered catalysts is improved by modifying the ratios of flow rates of the reaction gases.
Chapter 4 discloses a novel self-focusing gate structure and related processes, which could resolve the issue of electron beam spreading as the FEDs operate in triode configuration.
In chapter 5, a simple fabrication process consisting of wet etching and lift-off is proposed to create a submicron gap for planar edge emitters. Additionally, a quasi-planar structure of submicron gaps controlled via thin film deposition process is also demonstrated, and the field emission characteristics are further improved by a forming process so as to reduce the operation voltages.
Finally, the conclusions and recommendations for future research are described in chapter 6 and 7, respectively.
Chapter 2
Investigation of Carbon Nanotubes Synthesized at Low Temperatures Using Multi-layered Catalytic Films
In this chapter, carbon nanotubes were synthesized at 550℃ using multilayer catalysts which were composed of supporting layer, interlayer, and catalytic metal.
Supporting layer could effectively enhance the dispersion of catalytic nanoparticles and meanwhile avoid their agglomeration during the synthetic process. Interlayers not only promoted the distribution of nanoparticles but also function to enhance the precipitation of carbon atoms. Interlayers which had surface energy comparable to catalytic metal and positive of formation heat of carbide revealed to improve the growth of CNTs at low temperatures.
2.1 Introduction
Carbon nanotubes (CNTs), discovered by Iijima in 1991 [2.1], are promising candidates for emitter materials of cold cathodes due to their unique properties, such as high aspect ratio, high mechanical strength, chemical inertness and large current capability [2.2-2.4]. Wang et al. reported a low turn-on electric field of 0.8 V/µm from nanotube bundle emitters [2.5]; meanwhile, Zhu et al. demonstrated a very high current density of 4 A/cm2 from CNT field emitters [2.6]. In addition, Samsung
announced a matrix-addressable diode display using CNTs as emitters, showing the potential for field emission displays (FEDs) [2.7]. Therefore, field emission displays based on CNTs have attracted much attention in recent years [2.8-2.10]. For FEDs to be cost competitive with existing display technologies, inexpensive glass is commonly used as the substrates. Nowadays, several methods have been well developed for preparing CNTs on glass substrates at low temperatures, including screen printing [2.11], electrophoresis [2.12], and chemical vapor deposition (CVD) [2.13-2.17]. Among these methods, CVD appears to be a superior one due to its good uniformity and pixel resolution. Numerous CVD systems have been employed to directly grow CNTs on glass substrates, and they can be categorized into two types of instruments based on the ways by which the hydrocarbon gases are decomposed:
plasma-based CVD and thermal CVD. The plasma-based CVD systems, such as inductively coupled plasma CVD (ICP-CVD) [2.13], electron cyclotron resonance CVD (ECR-CVD) [2.14], and plasma-enhanced chemical vapor deposition (PECVD) [2.15], utilize a source of plasma to effectively decompose the hydrocarbon gases, thus enhancing the growth of nanotubes at low temperatures on glass substrates. In contrast, thermal CVD [2.16-2.17] imply that the decomposition of reaction gases is carried out with the aid of thermal energy supplied by the instrument. For the sake of different operation mechanism, the plasma-based systems revealed to be a better choice than thermal one due to the higher efficiency in synthesis of CNTs. It has been shown that the activation energy characterized for the growth of CNTs by PECVD was lower than that by thermal CVD [2.18-2.19] and that the growth limiting step was determined by the diffusion of carbon on catalysts: surface diffusion for PECVD and bulk diffusion for thermal CVD [2.20]. This indicates that effective decomposition of carbon precursors by plasma could enhance the growth rates at low temperatures due to a lower activation energy. Nevertheless, it is worth mentioning that a vacuum
environment is necessary for the purpose to ignite and sustain the plasma source. Thus an evacuation system should be equipped in plasma-based systems, that is, a higher cost for large area and low throughput. Moreover, it is not easy to well control the plasma source in large area because the parameters for sustaining plasma should be precisely provided to obviate a big issue in non-uniformity. Therefore, it is generally suggested that thermal CVD would be a favored method owing to its controllability and scalability. Despite the high feasibility, however, thermal CVD typically involves processing temperatures of over 600 to overcome the decomposition energy of ℃ carbon precursors, which is significantly higher than the softening point of glass substrates (580 ). Thus, in several studies, a two℃ -zone electric furnace has been employed, which consists of a high temperature zone for preheating the reaction gas to enhance its reactivity [2.21-2.23], hence facilitating the growth of CNTs at low temperatures. With the similar ideas, there were work employing hot filament assisted systems [2.24], which consisted of a filament heated at high temperatures (~1600 ), ℃ to pre-decompose the precursors. These techniques are not only complicated in the manufacturing processes but also thermally inefficient because of the high thermal budget.
According to work published, the type of catalyst also plays an important role in the growth of CNTs because the process is a catalytic reaction. Kamada et al.
indicated that a binary catalyst could synthesize CNTs well at low temperatures due to its high activity [2.25]. Park and coworkers also employed a ternary alloy as a catalyst for improving the crystallinity of CNTs synthesized at low temperatures [2.16-2.17].
Since thermal CVD utilizes thermal energy to assist the decomposition of carbon precursors, a highly active catalyst is essential for low-temperature synthesis of CNTs.
However, it is difficult to prepare these catalysts because the composition of elements should be carefully synthesized. In this chapter, multilayer catalysts were employed to
synthesize CNTs by thermal CVD at atmospheric pressure, and the morphologies and emission characteristics were investigated based on the compositions of multilayer catalysts. The components of multilayer catalysts were formed by sputtering in a simple manner. The functionality of each component of multilayer catalysts is discussed, and a growth mechanism is also constructed for clarifying and modeling the synthesis of nanotubes at low temperatures.
2.2 Experimental Procedures 2.2.1 Sample Fabrication
Multiwalled carbon nanotubes (MWNTs) were grown on Cr-coated N-type (100) silicon substrates with multilayer catalysts at atmospheric pressure by thermal CVD.
The multilayer catalysts formed of supporting layer, interlayer, and catalytic metal were sequentially deposited on Cr-coated substrates by magnetron sputtering (Ion Tech Microvac 450CB) at the pressure of 7.6×10-2 Torr at room temperature. Because the sputtering system consisted of three sputtering sources, the three components of multilayer catalysts could be sequentially prepared without breaking the vacuum environment. Cobalt and aluminum are used as the catalytic metal and supporting layer, respectively, while different kinds of metal were employed as interlayer, such as Ti, Cr, Pd, etc.
The atmospheric pressure thermal CVD system consists of a 2-in.-diameter horizontal quartz tube, an electric heating system, reaction gas supply and related mass flow controllers. Samples loaded into the quartz tube were heated to the predetermined temperature of 550℃ in a nitrogen flow for an oxygen-free ambience.
Prior to the CNT growth, hydrogen gas with a flow rate of 50 sccm was fed into the reaction tube for 5 min to reduce the catalyst metal to the metallic phase, meanwhile
transforming into nanoparticles. Then, carbon nanotubes were grown at designated temperatures with reaction gas, ethylene, at a flow rate of 75 sccm for 30 min. After that, samples were furnace-cooled to room temperature in nitrogen flow to fully exhaust the reaction and byproduct gases.
2.2.2 Material Analysis
The morphologies of the samples were characterized by scanning electron microscopy (SEM; Hitachi S-4700I). The fine internal structures of nanotubes and catalytic elements of nanoparticles were examined by high-resolution transmission electron microscopy (HRTEM; JEOL JEM-2000EX) and X-ray energy dispersive spectroscopy (EDS), respectively. The surface property of multilayer catalysts was characterized by X-ray photoelectron spectroscopy (XPS) using Al Kα source.
Field emission characteristics of CNTs were measured with a parallel diode-type configuration in a high-vacuum chamber with the pressure of 5×10-6 Torr. A glass substrate coated with indium tin oxide (ITO) and P22 phosphor (ZnS: Cu, Al) was used as the anode plate, and the gap between the cathode and the anode plate was set to be 160 µm. The emitting area was well defined as 0.25 mm2 by photolithography.
Anode voltages up to 1000 V were applied at intervals of 5 V with a source measure unit (Keithley 237) for the verification of field emission characteristics while the cathode was biased at 0 V.
2.3 Morphologies and Field Emission Characteristics of CNTs 2.3.1 Effect of Supporting Layer
Figure 2.1 showed the plane view SEM images of samples with catalysts after pretreatment process. In Fig. 2.1(a), the surface morphology of the sample with a
single layer of catalytic metal (20Co) did not significantly reveal the formation of nanoparticles after pretreatment. In contrast, as the underlayer, either Cr or Al, was employed, the formation of nanoparticles was remarkably enhanced, shown in Figs.
2.1(b), 2.1(c), and 2.1(d). Moreover, it was clearly that the sample with multilayer catalysts (20Co/20Cr/100Al) showed a more uniform morphology than others. By increasing the thickness of supporting layer from 2 nm to 20 nm, the morphologies of nanoparticles shown in Fig. 2.2 indicated that there was a smallest particle size for the sample with Al supporting layer of the thickness of 10 nm.
After the process of CNT growth, the SEM images, displayed in Fig. 2.3, clearly showed that since the nanoparticles were small and uniformly distributed after pretreatment, the samples exhibited a better morphology of nanotubes, especially for the sample with multilayer catalysts. The corresponding field emission characteristics of each sample after CNT synthesis were shown in Fig. 2.4. There was manifest correlation between morphology and emission characteristics of CNTs; that is, the sample with multilayer catalysts (20Co/20Cr/100Al) represented a better performance that was a field emission current density of 8.28 mA/cm2 at the electric field of 6 V/µm. The detail emission characteristics were summarized in Table 2.1.
According to the analysis of XPS examined on the sample surface which was conducted by growth process for 5 min (Fig. 2.5), there was information showing the formation of aluminum oxide. The spectrum revealed a significant 2p peak of oxidized-Al with a binding energy of 75.4 eV. It was conjectured that the thin aluminum layer transformed into aluminum oxide during the heating step due to the residual oxygen in the reaction chamber [2.26]. Jodin et al. [2.27] reported that catalysts combining with fumed alumina nanoparticles acting as a support could prevent agglomeration of catalytic particles during the CVD growth process; that is, the alumina support facilitated the uniform distribution of the catalytic nanoparticles
and preserved their small sizes for CNT growth. Moreover, as compared with a flat substrate coated with metal catalyst clusters, a buffer layer-supported catalyst had the advantage of a higher active surface area, thus more active sites for nucleation [2.28-2.29]. Accordingly, aluminum significantly revealed to be a good supporting layer for enhancing the growth of nanotubes.
2.3.2 Effect of Interlayer
As mentioned previously, the supporting layer of Al could effectively reduce the size of nanoparticles and resulted in a uniform dispersion of catalysts. However, there was a lack of information about the role interlayers played. In order to clarify the functionality of interlayer, various metallic metals were employed as interlayers interposed between supporting layer (Al) and catalytic metal (Co). The thickness of interlayer was determined to be 2 nm, while that of Al and Co were 10 nm and 2 nm, respectively. Figure 2.6 showed the SEM images of CNTs synthesized with different interlayer metals. It was obviously that the samples with interlayers of Cr and Ti exhibited the better morphologies as compared with other samples. The field emission characteristics of each sample were shown in Fig. 2.7, and the corresponding turn-on fields and current densities were listed in Table 2.2. Accordingly, the samples with Cr and Ti interlayers both had emission current density exceeding 10 mA/cm2, which was suitable for the application of FEDs [2.30].
The effect of interlayer thickness on morphology of CNTs was also investigated via samples with different interlayer thickness of Ti and Cr as the thickness of supporting layer and catalytic metal were fixed with 10 and 2 nm, respectively. The thickness of Cr interlayer was modified from 1 nm to 100 nm, and the corresponding SEM micrographs of samples after CNT growth were illustrated in Fig. 2.8. As the ratio of Co over Cr was close to 1~2, the CNTs showed a better morphology than
others. Similar effect was manifest for Ti interlayer according to the SEM images shown in Fig. 2.9. In other words, the optimal catalytic metal over interlayer ratio should be chosen close to 1~2, which would give rise to better field emission characteristics. The detail data of emission current and turn-on field with different interlayer thickness were illustrated in Fig. 2.10 and Table 2.3 for Cr, while in Fig.
2.11 and Table 2.4 for Ti.
Figure 2.12(a) showed a high-resolution TEM image of nanotubes grown with the multilayer catalyst (20Co/30Ti/100Al) at 550℃, and revealed a closed tip filled with catalytic metal particles and a multiwalled structure consisting of the wavy graphite sheets aligned parallel to the tube axis. The inner and outer diameters were about 8 nm and 27 nm, respectively. The outer graphitic sheets were usually less crystalline than the inner ones, and more defects were shown in the outer surface of nanotubes. The corresponding EDS analysis of the catalytic particle was shown in Fig.
2.12(b), confirming that the particle was composed of cobalt. Neither Al nor Ti peak appeared in the figure, and the Cu signal originated from the TEM microgrid. By the same token, similar information was found in the sample with Cr interlayer. The TEM and corresponding EDS analysis of nanotubes synthesized with multilayer catalysts (20Co/30Cr/100Al) were illustrated in Figs. 2.13(a) and 2.13(b), respectively. This result was similar to those of other literature in which binary catalysts were employed [2.31-2.32], and indicated that only Co particles directly participate in the growth of nanotubes.
In order to study in more detail and possible reaction of interlayer, samples with multilayer catalysts was conducted with CNT growth process for 5 min and then cooled down for XPS analysis. When the sample using Ti as interlayer was performed, the C1s spectrum in Fig. 2.14(a) revealed two different peaks at 281.5 and 285 eV, respectively. The peak at 281.5 eV was close to the signal of titanium carbides [2.33],
while the other peak at 285 eV corresponded to the binding energy of free carbon or CNTs. The C1s spectrum of sample with Cr interlayer shown in Fig. 2.14(b) displayed two carbon forms: a major peak at 285 eV and a minor peak at 283 eV, which was characterized as chromium carbide phases [2.34]. Conclusively, it was conjectured that the interlayer played an important role in two aspects for the low-temperature synthesis of nanotubes. First, the interlayer interposed between supporting layer and catalytic metal could be considered as another supporting layer other than Al. From the view point of surface energy, the materials with a large surface energy tended to merge into a large agglomeration for reduction in the surface area. According to the data of surface energy shown in Fig. 2.15, the interlayers could be classified into three categories as compared with the surface energy of Co: (1) surface energy smaller than that of Co, (2) surface energy close to that of Co, and (3) surface energy larger than that of Co. Since the difference of surface energies between interlayer and catalytic thin films would give rise to different morphologies of catalytic nanoparticles, the growth of nanotubes would be influenced by interlayers as well. As shown in Fig.
2.16, while the surface energy of interlayers were smaller than that of Co, the catalytic nanoparticles tended to merge into large ones for reduction in surface area, thus a stable state. On the contrary, while the surface energy of interlayers were larger than that of Co, the catalytic nanoparticles tended to form a film-like morphologies covering on the surface of interlayers for reduction in surface area of interlayer thin films. Both conditions mentioned above could not result in an uniform distribution of small nanoparticles. For this reason, interlayers whose surface energy close to that of catalytic metal (Co) would had a positive effect on the formation of catalytic nanoparticles. Conclusively, Ti and Cr whose surface energies close to that of Co seemed to be good interlayer candidates for the growth of nanotubes. Second, it was generally known that growth of nanotubes should incorporate the step of carbon
precipitation. It is proposed that the precipitation sites of carbon existed in the region of particles which is super-cooled because of the large driving force for the super-saturation. Therefore, the super-saturated carbon would precipitate at these sites to reconstruct the equilibrium. As described previously, the interlayer metal transformed into carbide phase during processing. According to the formation heat of carbides for various metals listed in Table 2.5, the values of formation heat could be classified into two categories: H＞0 and H＜0. The positive formation heat of carbide indicated that the reaction of carbide would consume thermal energy supplied from environment, that is, an endothermic reaction. On the contrary, negative formation heat means an exothermic reaction. While transforming into carbide phase, the interlayer of positive formation heat would consume thermal energy in the neighborhood and give rise to a reduction in temperature, thus a preferred precipitation sites of carbon. Accordingly, the interlayer, such as Ti, Ta, Cr, and Hf, would have the positive effect on synthesis of nanotubes. In combination of the aspects mentioned above, only Ti and Cr interlayer revealed to be candidates appropriate for low-temperature synthesis of CNTs, and this was correspondent to the results of experiments.
2.4 Growth Mechanism
Although the real growth mechanism of CNTs was still at issue, a possible mechanism for nanotubes synthesis with the multilayer catalysts was proposed.
Initially, the Co catalyst thin film broke up and transformed into nanoparticles, which minimized the surface energies during the heating process. These nanoparticles provided the nucleation and growth sites for nanotubes, and the size of these particles
dictated the diameter of nanotubes. From the results of experiment, the Al film acting as the supporting layer could remarkably enhance the formation of nanoparticles, and meanwhile improved the uniformity of catalytic particles by avoiding their mergence into large ones. Raw and decomposed carbon precursors absorbed on the surface of catalyst metal particles and dissociated into carbon atoms. The concentration of carbon in the catalyst particles increased until supersaturation was reached. Carbon atoms diffused along the concentration gradient via the surface and/or bulk of the metal particles and precipitated at the preferable nucleation sites to form the graphitic sheets. Since the incorporation of interlayer which had positive formation heat of carbides could enhance the precipitation of carbon, a positive driving force existed for the formation of graphite sheets, thus growth of nanotubes. Therefore, the multilayer catalysts showed an outstanding catalytic property for synthesis of CNTs at low temperatures.
2.5 Summary
2.5 Summary