Thesis Organization

The overview of vacuum microelectronics and basic principles of field emission theory was described in chapter 1. Moreover, the properties and applications of CNTs were presented clearly. Several papers about CNTs growth at low temperature for FED application were depicted in chapter 1. Then we referred to the motivation in this chapter.

The experimental procedures were revealed in chapter 2. First, we grew CNTs using multilayer catalysts on silicon substrate at low temperature. Then, the improvement of gate controlled triode structure was proposed. Finally, CNTs were grown on glass substrate for FED application.

Results and discussion were summarized in chapter 3. Then, we accomplished many important results including, (1) SEM images, (2) TEM images, (3) EDS analysis, (4) Raman analysis, (5) XRD analysis, and (5) Field Emission Measurement.

Finally, the conclusions and future researches are provided in chapter 4.

Chapter 2

Experimental Procedures

2.1 Introduction

First of all, we used several metals as the multilayer catalysts for CNT synthesis and found the most fitting ones for CNT growth at low temperature. Then we fabricated triode structure on silicon substrate and diode structure on glass substrate to prove the superiority of novel multilayer catalysts in CNT-FED. Finally, we accomplished many kinds of analysis for the above mentioned. The scheme of the whole experimental procedures was shown in Fig.

2-1.

2.2 CNTs Grown using Multilayer Catalyst Films at Low Temperature 2.2.1 Forward Arrangement

A (100) n-type silicon wafer was prepared for the substrate. After the RCA clean and lithography processes, we defined three kinds of patterns (squares of 1000、100、10 um²) for CNT field emission arrays. A 2000Å Cr layer was deposited by dual E-gun evaporation (JAPAN ULVAC EBX-10C) as the cathode between the substrate and catalysts. We investigated three multilayer catalyst films, including Co/Cr/Al, Ni/Cr/Al, and Fe/Cr/Al metal films.

Each of three multilayer catalyst films were deposited layer by layer at 0.1A/s (deposition rate) using sputtering system (ENGLAND Ion Tech Microvac 450CB). After the

deposition of the catalysts into the patterns as described above, the photoresist was lifted off in acetone solution. At last, the samples were already accomplished for CNT growth.

2.2.2 CNT Synthesis

The samples with different catalysts were transferred into a thermal CVD chamber (Fig.

2-2). First, the catalyst film must be pretreated to form nano-particles with H2. It is more difficult for catalyst film to become particles without plasma treatment. Then, the hydrocarbon gas source was added to synthesize CNTs. In this instrument, there were two types of hydrocarbon gases: CH4 and C2H4. Because of the differences of chemical reactivity, we applied C2H4 (more reactive) to synthesize CNTs in general. The whole experimental process was presented schematically in Fig. 2-3.

To synthesize CNTs, several experiments were proceeding. In experiment A, B, C, D, and E, we grew CNTs with different catalysts at 500 ℃ and there will be an exhaustive investigation in the following section of results and discussions. In experiment D and E, we compared different composition proportion of Co, Cr, and Al. In particularly, experiment F, G, and H is investigated at 550℃, 600℃, and 650℃, respectively. The experimental parameters of CNT synthesis were described as follows:

Experiment A

The catalysts’ nano-films were pretreated at 500℃ for 5 minutes with 50 sccm H2, and for 5 minutes with 1000sccm N2, respectively.

Experiment B、C、D、E

Prior to CNT growth, the catalysts’ nano-films were pretreated at 500℃ for 5 minutes with 50 sccm H2, and for 5 minutes with 1000sccm N2, respectively. Then 5 sccm C2H4 was added to

grow CNTs at 500℃ for 30 minutes.

Experiment F、G、H

The growth recipe is the same as experiment B, except the growth temperature. The experiment F, G, and H is investigated at 550℃, 600℃, and 650℃, respectively. Table 2-1 was presented clearly the experimental process.

2.3 MIM Triode Structure on Si Substrate 2.3.1 Structure Fabrication

The fabrication procedures of the triode structure for CNT-FED were shown in Fig. 2-4.

A (100) n-type silicon wafer as the substrate was cleaned by RCA clean. As shown in Fig.

2-4(a), 2000A Cr as the cathode, 1 um SiO2, and 2000A Cr as the gate were deposited layer by layer using the E Gun, Plasma Enhanced Chemical Vapor Deposition(PECVD), and E Gun, respectively. Then there are two masks in the lithography process as Fig. 2-4(b). One whose shape is stripe is to define the gate region and isolate the neighbor devices. The other one whose shape is square is to define the catalyst metal deposition region. As described in Fig.

2-4(c), the gate and SiO2 were etched in the wet etching, respectively. With the previously patterned photoresist layer as the shadow mask, 100A Al, 20A Cr, and 20A Co were deposited on the patterned Cr cathode by Sputter (Fig. 2-4(d)). Finally, the Al and catalyst layers on photoresist were removed by the lift-off method as presented in Fig. 2-4(e), and transferred into the thermal CVD chamber for CNT growth immediately.

2.3.2 CNT Synthesis

Since the triode structure had been fabricated, we synthesized CNTs in thermal CVD

chamber with Co/Cr/Al as the catalysts at relatively low temperature (500℃). By means of the change of the C/H ratio, we tried to increase more active carbon radicals for catalyst particles and synthesized CNTs at 500℃. The experimental parameters were described as follows:

Experiment I:

Prior to CNT growth, the catalysts’ nano-films were pretreated at 500℃ for 10 minutes with 100 sccm H2 and for 5 minutes with 1000sccm N2, respectively. Then 75 sccm C2H4 and 1000sccm N2 was added to grow CNTs at 500℃ for 15 minutes.

Experiment J:

Prior to CNT growth, the catalysts’ nano-films were pretreated at 500℃ for 10 minutes with 100 sccm H2 and for 5 minutes with 1000sccm N2, respectively. Then 75 sccm C2H4, 20 sccm H2 and 1000sccm N2 was added to grow CNTs at 500℃ for 8 minutes.

The relative parameters of thermal CVD process were presented in Table 2-2.

2.4 CNTs Grown on Glass Substrate 2.4.1 Sample Preparation

A glass substrate, soda-lime glass, was prepared for Diode structure. The continuous process was shown completely in the Fig. 2-5. A 2000A Cr was deposited as a cathode electrode by sputtering system (Fig. 2-5 (a)). Then we dined the catalyst film region in the lithography process as shown in Fig. 2-5(b). The 20A Co / 20A Cr/ 100A Al were deposited by sputter system (Fig. 2-5(c)). Finally, the photoresist was removed and the diode structure on glass substrate was transferred to thermal CVD chamber for CNT growth.

2.4.2 CNT Synthesis

The sample of patterned catalyst film on glass substrate was loaded into the thermal CVD chamber for CNT growth. In order to avoid the melt of soda-lime glass and contamination of the chamber, we set a silicon wafer under the glass substrate. Then we synthesized CNTs as described in the following:

Experiment K:

Prior to CNT growth, the catalysts’ nano-films were pretreated at 500℃ for 5 minutes with 50 sccm H2 and 1000sccm N2, respectively. Then 75 sccm C2H4 and 1000sccm N2 was added to grow CNTs at 500℃ for 60 minutes.

2.5 Analysis

Surface morphology and internal structure of CNTs synthesized in our experiments were characterized by a Hitachi S-4700I high-resolution field-emission scanning electron microscopy (SEM) and high-resolution transmission electron microscopy (HR-TEM) respectively. Simultaneously which kinds of elements were detected by the Energy Dispersive X-Ray (EDS). In addition, Raman spectra revealed the two peaks of CNTs (graphite and defect) and XRD analysis revealed which elements around the CNT bottom. The electric characteristics were measured by Keithley 237 (Fig. 2-6) in a 10^-6 torr chamber. We also applied an ITO glass with phosphor coating on it as the anode to observe the luminescent image.

Chapter 3

Results and Discussion

3.1 CNTs Grown Using Multilayer Catalysts at Low Temperature 3.1.1 Effect of Al Thickness (Exp. A, B)

It is generally believed that the formation of catalyst nanoparticles is necessary before CNT growth. In order to form these nanoparticles of catalyst, we proceeded a procedure before CNT growth─ pretreatment. The pretreatment method is to balance the surface energy (catalyst/gas), interface energy (catalyst/substrate), and interior energy (body free energy) of catalyst’s nanofilm at designated temperature, and then to achieve the purpose of nanoparticles’ formation. The nanoparticles can not only help the CNT growth have better quality but also control the size and morphology of CNTs. Therefore, the formation of nanoparticles plays an important role in CNT synthesis for device application.

The formation of nanoparticles has two important purposes. One is lower melting point as catalyst film thickness decreases [3.1]. In the Fig. 1-51, nanoparticle indeed had the lowest melting point (1057K) compared to nanowire (1315K) and nanofilm (1536K) as the size is 2nm. The other one is high activity to decompose the hydrocarbon gases because of the nano-phenomenon as the catalyst film thickness decreases. Generally, the smaller catalyst particles have higher surface energies and thus promote nucleation and growth processes of CNTs.

From the Fig. 3-1(a), (b) and (e), we could find out that the particles are more uniform

if 100A Al film exists in the multilayer film. From the Fig. 3-1(c) to (f), the thickness of Co and Cr was 20A and the thickness of Al was varied from 20A to 200A. Generally, the density of nanaparticles was increasing as the Al thickness increased. Therefore, P. Oelhafen et al.

investigated that in the pretreatment pure Al buffer layer could transform into Al2O3 from XPS analysis and enhanced the CNT growth [3.2]. In Taek Han et al. investigated spherical Al2O3 particles were formed during the heating prior to the CNT growth, preventing catalyst clusters from agglomerating which subsequently resulted in the formation of thin CNTs [3.3].

In the Fig. 3-1(f), the catalyst particles merged together because Al buffer layer thickness (200A) was too much. With the Al thickness increased, catalyst particles became bigger and bigger. But, with the Al thickness (100A), we could significantly find out that diameter of catalyst particles are 10 to 30nm close to its multilayer film thickness. In summary, the results of SEM images are same as the investigation of In Taek Han et al. With the optimum Al thickness, it could not only transform into Al2O3 in the pretreatment without merging together but also support catalyst particles on Al2O3.

In the CNT growth at 500ْC (Fig. 3-2), the morphology of CNT growth using 20Co/20Cr/100Al performed very well and its length was 1.7 um. In the Fig. 3-2(a), no CNTs were grown because Al buffer layer could not support catalyst particles and let them merge together. In the Fig. 3-2(b) and (d), the morphology was very similar because Al could not support catalyst particles very well. Furthermore, we investigated the effect of Al thickness on the graphite crystallization by Raman analysis (Fig. 3-3). The G peak and D peak located on 1580 cm^-1 and 1350 cm^-1. The results showed that Al wouldn’t affect the graphite crystallization of CNT growth and IG/ID ratio made no difference. In the Fig. 3-4, the TEM and EDS analysis were investigated for CNTs using 20Co/20Cr/100Al in the Exp. B. Then, the outer graphite layer of CNT from TEM image was multi-wall structure and catalyst particle stuffed in the middle of CNT. In addition, in the EDS analysis for the catalyst particle,

only Co element existed in the catalyst particle. The result meant that only Co took part in the CNT growth except Al and Cr. The role of Co is a catalyst for CNT growth, but Cr and Al play the supporting role for CNT growth. Finally, the field emission measurements (Fig.

3-5(a)) for the Exp. B, CNTs using 20Co/20Cr/100Al had superior field emission characteristics. Its turn on field is 3.81V/um and anode current is 8mA/cm². The F-N plot and field emission characteristics were showed in the Fig. 3-5(b) and Table 3-1, respectively.

3.1.2 Effect of Catalyst Metal (Exp. C)

The kinds of catalyst metals play an important role on the CNT growth by Thermal CVD.

The functions of catalyst metals generally can be separated into two parts. One is to decompose hydrocarbon gases and the other one is to let carbon atoms diffuse in the catalyst metals. For optimal growth in typical chemical vapor deposition procedures, metal catalysts need to exhibit sufficient carbon solubility, rapid carbon diffusion, and limited carbide formation. The pure catalyst metals as Fe, Co, and Ni are always used for CNT growth. In Exp. C, we investigated the effect of catalyst metals on the multilayer catalyst films with the same thickness, such as (1)20A Co/20A Cr/100A Al, (2)20A Fe/20A Cr/100A Al, and (3)20A Ni/20A Cr/100A Al. The SEM images of each three samples were showed in the Fig. 3-6. The morphology of CNTs using Fe multilayer catalyst film was worse compared to Ni and Co multilayer catalyst film, respectively. This result is very similar to some paper researches. Fe has the highest melting point (1808K) than Ni (1726K) and Co (1768K). Therefore, Fe isn’t suitable to CNT growth at low temperature (500 ْC). As for Ni and Co multilayer catalyst film, the sample of Co had more uniform and aligned CNTs and more activity for CNT growth without amorphous carbon deposition. In many paper researches, the CNT growth has no difference using Ni or Co catalyst metals because of their similar melting point and carbon solubility. But based on paper research [3.4], Co had the smaller activation energy (0.4ev)

than Ni’s one(0.5ev) for CNT growth at low temperature. We supposed that the activation energy of CNT growth plays a critical role for CNT growth at low temperature. So Co multilayer catalyst film had the better performance than Ni’s one. Then field emission measurement was showed in the Fig. 3-7. The Co’s sample possess the superior field emission characteristics including the turn on field (3.81 V/um) and the anode current (8 mA/cm²) compared to Ni’s one and Fe’s one. From the proof of field emission results, the Ni’s sample had too much amorphous carbon deposition so that screening effect resulted in the poor field emission characteristics. Fe’s sample had no field emission characteristics because it is not suitable for CNT growth at low temperature (500 ْC). In order to show the excellent ability for CNT growth, we investigated the luminescent image of Co’s sample. In the Fig. 3-8, the green light was emitted on the phosphor anode at 6.25 V/um from the 1 mm² CNT pattern.

This result represents the low temperature display could be realized in future. Besides, as shown in Fig. 3-9, the uniformity of luminescent images was achieved at only 3.75 v/um. Its results revealed CNTs using multilayer catalyst films could be realized in the backlight unit applications.

3.1.3 Effect of Cr Film and Different Co/Cr Ratios on The CNT Growth (Exp. D, E)

The effect of Cr film on multilayer catalyst films will be discussed in this section. In the Fig. 3-10, we compared five different samples after CNT growth, including (a) 20A Co, (b) 20A Co/100A Al, (c) 20A Co/10A Cr/100A Al, (d) 20A Co/20A Cr/100A Al, and (e) 20A Co/50A Cr/100A Al. As the Fig. 3-10(a), it apparently showed that pure Co catalyst had less activity for CNT growth at 500ْ C. The morphology had extreme differences like graphite pastes and well CNTs in the Fig.3-10 (b) and (d), respectively. The above three results proved

Cr film is very essential to CNT growth at 500ْ C. With Cr film thickness increased, the optimum Cr thickness (20A) was achieved. The CNT tips had much graphite agglomerating, the CNTs were more aligned and uniform, and no CNT existed in the Fig.3-10 (c), (d), and (e), respectively. The above results revealed the Cr film is dispensable in the multilayer catalyst films. From the paper researches [3.5-6], the Cr film could enhance adhesion between CNTs and substrate due to the formation of interface bonds. BesidesLee et al. [3.7] have recently reported that the addition of Cr as co-catalyst can lower down the carbon nanotube growth temperature using Co-Ni alloy catalyst. Furthermore, in the Table 1-5, the heat formation of transition metal carbides of Co and Cr is close to -0.22ev and 0.555ev, respectively. So in the

CNT growth Cr prefer to transform into carbide type more than Co does.

Christian P. Deck et al.[3.8] investigated that carbide formation inhibits 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. From the proof of XRD analysis (Fig. 3-11) using 20Co/10Cr/100Al in Exp. D, we could find out Cr carbide (Cr3C2 and Cr23C6) existed onto the CNT bottom. In summary, Cr could enhance adhesion between CNT and substrate and Cr carbide could disperse carbon atoms uniformly so that CNT could be grown well when Co achieved carbon saturation to precipitate carbon atoms.

The above has mentioned each function of multilayer catalyst films. The EDS analysis (Fig. 3-4(b)) revealed CNT growth was dominated by Co catalyst. Cr could enhance adhesion between CNT and substrate and Cr carbide could disperse carbon atoms uniformly so that CNT could be grown well. Al film could transform into stable Al2O3, which could prevent catalyst particles from merging together and increased the number of active sites. In the Exp.

E, we investigated the effect of Co thickness on the same under multilayer catalyst films, including (a) 10A Co/20A Cr/100A Al, (b) 20A Co/20A Cr/100A Al, (c) 50A Co/20A

Cr/100A Al, and (d) 100A Co/20A Cr/100A Al. The melting point of multilayer catalyst films decreased as its thickness decreased [3.1]. In particularly, K.K. Nanda et al. [3.9] publish the model for the relationship between melting point and particle sizes….Co metal for example Fig.1-56. The size of catalyst particle was close to 2nm and the relative melting point was reduced to 0.45 Tm (522.6ْ C). In the Fig. 3-12(a), Co thickness was 10A and it should synthesize the well CNTs better than 20A Co’s sample. We supposed that in the same growth recipe 10Co could not satisfy the amounts of carbon atoms and the results apparently revealed the CNTs were shorter and sparse with the lower CNT growth rate compared to the 20A Co’s sample. As for 50A Co and 100A Co, they disobey the theory of lower melting point so that the graphite sheets were deposited at 500ْ C because bigger and lower active catalyst particles formation. The field emission measurements of Exp. D and E were showed in the Fig. 3-13.

The field emission characteristics in Exp. D,E was showed in the Table 3.3. The 20A Co/20A Cr /100A Al had the best field emission characteristics and the second was 10A Co/20A Cr /100A Al.

3.1.4 Effect of Growth Temperature (Exp. B, F,G,H)

In order to prove the effect of growth temperature on the surface diffusion of CNT growth mechanism, we synthesized CNTs at 500ْ C, 550ْ C, 600ْ C, and 650ْ C in the Exp. B, F, G, and H, respectively. From the morphology of CNT cross-section images (Fig. 3-14), the length of CNTs were 1.63um, 3.25um, 6.25um, and 12.7 um in the Exp. B, F, G, and H, respectively. With growth temperature increased, the growth rate of CNTs and carbon diffusion ability was enhanced so that the density of CNTs increased. The field emission measurements were showed in the Fig. 3-15. The field emission characteristics were showed in the Table. 3.4. We could significantly find out the field emission properties got better and better because the graphite crystallization and length of CNTs increased as growth

temperature increased. Especially, the anode current at 6V/um was achieved to 8mA/cm², 13mA/cm², 32mA/cm², and 22mA/cm² in the Exp. B, F, G, and H, respectively. In addition, TEM and EDS analysis was investigated in the Fig. 3-16 and Fig. 3-17, respectively. The TEM images showed whole CNTs were multi-walled structure. The EDS analysis revealed only Co took part in the CNT growth. The Arrhenius plot of ln(growth rate/[C]sat) vs 1/T (K) can be used to derive the activation energy as displayed in Fig. 3-18. The data fit well to a linear function, providing the activation energy of 0.37ev. This value is close to the surface diffusion energy of carbon in bulk Co (0.4ev). In the comparison with paper researches[3.4,3.9], the resulted revealed multilayer catalyst films which overcame the lower CNT growth rate in thermal CVD indeed had extreme activity to be grown CNTs.

3.2 Gated Triode Structure (Exp. I, J)

The promising multilayer catalyst films (20A Co/ 20A Cr/ 100A Al) had been investigated (i.e. the morphology and field emission characteristics of CNTs ). Further, we fabricated the low temperature triode structure in order to realize low voltage driving, high resolution and full gray-scale imaging using multilayer catalyst films (20A Co/ 20A Cr/ 100A Al). In the past years, many researches had been investigated the triode structure [3.10] (poly Si/SiO2/cathode), but their fabrication temperature was higher than 550ْ C whose temperature is the melting point of sodalime glass substrate. Although they could achieve the excellent field emission properties, their methods had no values on the display applications. In our current study, the Cr gate/SiO2/Cr cathode triode structure was realized and its fabrication

The promising multilayer catalyst films (20A Co/ 20A Cr/ 100A Al) had been investigated (i.e. the morphology and field emission characteristics of CNTs ). Further, we fabricated the low temperature triode structure in order to realize low voltage driving, high resolution and full gray-scale imaging using multilayer catalyst films (20A Co/ 20A Cr/ 100A Al). In the past years, many researches had been investigated the triode structure [3.10] (poly Si/SiO2/cathode), but their fabrication temperature was higher than 550ْ C whose temperature is the melting point of sodalime glass substrate. Although they could achieve the excellent field emission properties, their methods had no values on the display applications. In our current study, the Cr gate/SiO2/Cr cathode triode structure was realized and its fabrication