2-1-1 Sample fabrication and CNTs synthesis
In the beginning, the process procedure of the conventional two samples was depicted in (Fig. 2-1). P-type silicon wafers (100) with low resistance (1~10 Ω/cm2) were utilized as the substrates in our experiment. To distinguish these two samples with different surface morphology for the sake of convenience, they were referred to as sample (A) and sample (B), respectively. So as to ensure the cleanness of the silicon substrate and to prevent the effect of particle contamination at the beginning of our experiment, the two conventional samples were both carried out with the RCA clean process (Fig. 2-1 (a)).
After the RCA clean process, a 4-nm-thick Cobalt layer was deposited on both sample (A) and sample (B) by magnetron sputtering (Ion Tech Microvac 450CB) at the pressure of 7.6×10-2 torr at room temperature. This Cobalt layer would transform into the catalyst nanoparticles which are the growth site of CNTs afterwards.
Moreover, sample (B) would deposit an extra 1-nm-thick Ti capping layer to generate the different surface morphology from sample (A) after the CNTs growth. After the metal deposition, both the conventional samples were loaded into the chamber of thermal-CVD. The atmospheric pressure thermal CVD system we used here was consisted of a 2-in.-diameter horizontal quartz tube, an electric heating system, a reaction gas supply and several related mass flow controllers as shown in (Fig. 2-2).
26
Samples loaded into the quartz tube would pass through some high temperature process in different gases environment to grow CNTs. The parameters of thermal CVD to grow the CNTs were illustrated in (Fig. 2-3) and (Table 2-X). At the start, samples would be heated to the predetermined temperature of 600℃ in a nitrogen flow rate of 1000 standard cubic centimeter per minute (sccm) for an oxygen-free ambience. Prior to the CNTs growth, hydrogen gas with a flow rate of 300 sccm was fed into the reaction tube for 10 min to reduce the catalyst metal to the metallic phase, meanwhile transform into nanoparticles. It is worth noting that the size and uniformity of the nanoparticles are related to the temperature, the flow rate of reducing gas and the characteristics of catalyst metal. After the pretreatment step, the chamber would be heated again under nitrogen flow rate of 1000 sccm to 700℃. Then, CNTs were grown at this temperature with hydrogen, nitrogen and ethylene at a flow rate of 300 sccm, 500 sccm and 100 sccm for designated growth time. After that, samples were furnace-cooled to room temperature in nitrogen flow rate of 5000 sccm to fully exhaust the reaction and byproduct gases.
What’s more, after the investigation of the gas breakdown characteristics of sample (A) and sample (B), a CNTs-based film synthesized from the co-deposited catalyst structure was proposed to improve the gas breakdown stability, and this sample was referred to as sample (C). Its schematic flowchart for the fabrication was depicted in (Fig. 2-4). At first, a 10-nm-thick aluminum supporting layer was
deposited on the substrate. Then a 1-nm-thick titanium interlayer and a 4-nm-thick Co-Ti co-deposited layer were deposited sequentially. The effect of Al supporting layer and Ti interlayer to the grown CNTs would be discussed in chapter 3. All of these metal layer were deposited also 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 multi-metal
27
layers could be sequentially prepared without breaking the vacuum environment.
After the catalyst structure was prepared, sample (C) would be loaded into the furnace and go through the same thermal CVD processes like sample (A) and sample (B) to grow the CNTs. After that, its gas breakdown characteristics and surface morphology before and after the stability test would be compared with the conventional two samples as well.
Moreover, in order to ameliorate the issue of high-voltage operation, the pattern technology was used to redistribute the local electrical filed. To compare the
difference, we fabricate two types of samples just like the method utilized above. One is the CNTs film sample synthesized from the co-deposited catalyst structure
mentioned above. The other is pillar arrays of vertical aligned CNTs bundles synthesized also from the co-deposited catalyst structure. With the purpose of the convenience in their comparison, we referred the former to film sample and the latter to pattern sample. Their detailed fabrication processes were depicted in (Fig. 2-4) and (Fig. 2-5), firstly, a lithography processes was used to define circle patterns (Fig. 2-6) for CNTs pillar arrays with 50 µm in diameter, and the spacing of the center of each circle was 80 µm , that is, the pillar-to-pillar spacing was 30 µm. After the lithography processes, Co-Ti (40 Å)/Ti(10 Å) /Al (100 Å) as catalyst structure were deposited on the silicon substrate by magnetron sputtering, and the lift-off technique using acetone was utilized to transfer the circle patterns onto the substrate. Finally, both of the film sample and the patterned sample were loaded into the chamber of thermal CVD to undergo the same processes (Fig. 2-3) to grow the CNTs. However, in this time, the parameter of CNTs growth time was adjusted to control the height of CNTs.
Therefore, pillar array of vertical aligned CNTs bundles with different R/H ratio were obtained in our experiment.
28
2-1-2 Material analysis and Breakdown characteristics measurement
The micrographs of samples were taken by Scanning electron microscope (SEM) to see the surface morphologies (Fig. 2-7). Besides, the high resolution transmission electron microscope (HRTEM; JEOL JEM-2000EX) (Fig. 2-8) and Raman spectra (Fig. 2-9) were also applied to analyze the structure and crystallinity of the CNTs.
To obtain the gas breakdown characteristics, all the samples with different catalyst structures were loaded into a vacuum chamber to measure the discharge current. The gas ionization measurement setups were depicted in (Fig. 2-10). In the chamber, an Indium Tin Oxide (ITO) was utilized as the anode electrode and CNTs as the cathode electrode with a distance (d) of 160~170µm. Besides, the vacuum degree of the whole chamber was controlled by a turbo pump and a rotary pump. In the application of gas ionization sensor measurement, the chamber would be firstly evacuated to high vacuum (10-6 torr) and then it switched to the rotary pump to control the vacuum degree. Then, the target gas was fed into the chamber and the measurement was conducted under different gas pressures.
At the first step, we mainly discussed the gas breakdown characteristics between different surface morphology and catalyst structure. Therefore, the measurements were conducted only in the nitrogen environment and the discussion would focus on their gas breakdown characteristics under different gas pressure.
Moreover, a high-voltage source measurement unit, Keithley 237, was applied on the anode and a high-current source measurement unit, Keithley 238, was applied to the cathode (sample substrate). The samples were measured from 0V to several hundreds of volts until the gas breakdown occured and a high discharge current appeared (the discharge current is limited at 10mA for equipment protection).
29
2-1-3 Stability test
In the stability test, the samples were measured for 1000 cycle times where one cycle was defined as applying voltages from 0V to1000V with 10V as a unit, and the discharge current was limited at 10mA when gas breakdown occurs for the sake of preventing damages to equipments. All the experimental data were record and
analysis by computers. After the stability tests, the micrographs of samples were taken by SEM to see their surface morphologies.
2-1-4 Gas sensing measurement
The same setups like (Fig. 2-10) were used in the gas sensing measurement.
However, different gases like carbon dioxide, oxygen, air and argon were fed into the chamber in replace of nitrogen we used in above, and their Paschen’s curve would be depicted, correspondingly. Finally, the change of breakdown voltages was measured in different concentration of mixed gas environment of air and target gas.