Process parameters of SWNTs growth in MP-CVD

Different parameter can result in distinct properties of SWNTs e.g.

morphology, growth mechanism, tube number density, structure, etc.

(a)Catalyst：

Different catalysts own distinct properties and the growth of SWNTs can be affected. If a catalyst particle is tightly sintered, it becomes harder to melt during reduction or pretreatment process. In other words, to form nano-scaled liquid droplets which provides the embryo of SWNTs is more difficult, so the same catalyst leads to the growth of nano-sheet instead of nanotube with lower growth temperature.[Lin-2002-922] The ways of depositing catalyst also cause influences as well.

(b)Growth temperature：

Besides the above-mentioned influence of catalyst reduction, temperature can also affect the diffusion rate of carbon atoms inside the catalyst. With higher growth temperature, the diffusion rate within catalyst becomes quickly and direct affects the growth rate, length and tube number density of CNTs. Also, the precursor gases decompose amount can be different with different temperature.

It changes the mixing concentrations of precursors. In addition, the crystalline of graphene layer becomes better with higher growth temperature.

(c)Gas

This condition includes precursor type, flow rate and gas ratio. In literature, many carbon source and reducing gas has been used. [Yun-2003-6789] [Lin-2002-922]

Different gas types have distinct pyrolysis temperature and different

bombardment effect in plasma environment.

(d)Working pressure：

When the inlet and pumped out gas flow reach a steady state, the pressure in chamber is called working pressure. The main effects of this parameter are the plasma behavior. With higher pressure, the mean free path of radical decrease but collision probability increase. The plasma induced self-bias will change with different pressure.

(e)Growth time：

Growth time usually changes the length of SWNTs. But after catalyst is poisoned, the growth time won’t affect it any more.

2.5 Analyses of SWNTs

(a) Raman spectroscopy analysis

(1)Measurement principle[Rao-1997-187]

Raman spectroscopy provides information about molecular vibrations that can be used for sample identification and quantitation. The Raman effect arises when a monochromatic light (laser) shines on a sample and excites molecules in the sample, which subsequently scatter the light. While most of this scattered light is at the same wavelength as the monochromatic light, some (±0.0001%) is scattered at a different wavelength. This in elastically scattered light is called

Raman scatter. It results from the molecule changing its molecular motions. The energy difference between the monochromatic light and the Raman scattered light is equal to the energy involved in changing the molecule’s vibrational state.

This energy involved in changing the molecule’s vibrational state. This energy difference is called the Raman shift. Several different Raman shifted signals will often be observed; each being associated with different vibrational or rotational motions of molecules in the sample. Figure 2.21 shows the schematic diagram of state change of Raman scattering.

Fig. 2.21 Schematic diagram of state change of Raman scattering[Rao-1997-187]

(2)Raman shifts of SWNT[Raravikar-2002-235424]

SWNT has 15 or 16 Raman-active modes, but some of their Raman scattering are difficult to detect, so only 7 Raman-active modes can be observed in measurement, as shown in Fig. 2.22.

Figures 2.23(a) and 2.23(b) show a typical Raman spectrum of SWNT sample taken at room temperature. The two prominent peaks with a peak position of 182 and 264 cm^-1 in Fig. 2.23(a) are the two RBM (radial breath mode，A1g) peaks. The two peaks from Fig. 2.23(b) at peak positions ~1590 and

~1350 cm^-1 belong to the G-band (tangential stretching mode，E2g) and D-band ( from amorphous carbon，A1g), respectively.

Figures 2.24(a), 2.24(b) and 2.24(c) show the G-band, D-band vibrational modes and the RBM’s in SWNTs respectively. In graphite and CNTs, the G-band Raman vibrational modes are present due to the in-plane vibrational movement of carbon atoms, which involves a combination of stretching and bending of the carbon-carbon (C-C) bonds. The disorder-induced D-band Raman vibrational mode, which is a highly dispersive spectral feature, is also present in these materials due to the collective in-plane vibrational movement of atoms towards and away from the center of the hexagons formed by the covalently (sp²) bonded carbon atoms. Therefore, the D-band mode involves stretching and bending of C-C bonds.

The RBM is a unique feature in the Raman spectrum of SWNTs and involves a collective vibrational movement of the carbon atoms towards and away from the central axis of a SWNT. The RBM oscillations are associated with a periodicity imposed on a graphene sheet by wrapping it into a finite-size (small diameter) tube. Consequently, the associated RBM Wavelength and frequency are directly related to the perimeter of the nanotube. Base on this relationship, as the diameter of the nanotube increases, the RBM frequency shifts to lower wave numbers. For larger and, particularly, MWNTs, the RBM frequency becomes very small and, at the same time, the intensity of the radial breathing mode decreases and ultimately becomes undetectable by Raman spectroscopy measurements. Hence Dresselhaus et al. [Dresselhaus-2002-2043]proposed that the frequency of the RBM are：

ω = ä(cm^-1 nm)/d(nm) (2-7)

where ä for the Si /SiO substrate is experimentally found to be 248 cm^-1nm for isolated SWNTs, ω is the peak position, d is the SWNT diameter and RBM is independent of chiral angle. Furthermore, Tuinstra et al. [Tuinstra-1970-1126] proposed that the Raman spectra of single crystals of graphite shows only one mode at 1575cm^-1 and as the increase of defects and disorder, D-band will show, so the IG/ID ratio can be used to determine the graphitized degree of SWNTs.

Fig. 2.22 Raman-active normal mode eigenvectors and frequencies for a (10,10) nanotube [Rao-1997-187]

Fig. 2.23 Raman spectra of SWNTs [Raravikar-2002-235424]

Fig. 2.24 Schematic diagram of Raman vibrational modes of CNTs (a) G-band mode (b)D-band mode (c)Radial breath mode [Raravikar-2002-235424]

(b) High-resolution transmission electron microscope (HRTEM) [Flahaut-2000-249]

HRTEM is the most direct method to analyze the structure of SWNT, however it is not easy to obtain a clear image of one SWNT. The main reasons are the structure of SWNT will be destroyed easily by high energy electron beam and electron scattering are not apparent because of SWNT is composed of so few carbon atoms. Furthermore, SWNT is hard to be built on stilts to avoid the background interference from copper grid and SWNTs dispersion are also an important problem. Figure 2.25 shows the HRTEM images of (a) SWNTs (b) SWNT bundles (c) DWNT (d) MWNT.

Fig. 2.25 (a) SWNTs (b) SWNT bundles (c) DWNT (d) MWNT [Flahaut-2000-249]

Fig. 2.26 STM image of various chiral angles SWNTs [Cees Dekker]

(c) Scanning tunneling microscope (STM)

STM can directly observe surface carbon atoms of SWNTs, providing information of carbon atoms site and arrangement. So we can measure length and chiral of SWNTs, as shown in Fig. 2.26.

2.6 Applications of SWNTs

Many researchers and engineers have been devoted to combine the CNTs with living. There are a lot of possible applications of CNTs products such as FED, field effect transistor (FET), hydrogen storage, etc. Until now, lots of prototypes of these applications have been published. Thus, it is believed that more and more commercial products will be published soon in the feature.

(a) Electron field emission elements:

The electron field emission elements, as implied by the name, utilized the field emission properties of CNTs. Among all of them, the closest to our life is FED. It is a next generation display after plasma display panel (PDP) and liquid crystal display (LCD) technologies. The theorem of formation of image is to use CNTs as cathode, then applies the potential between cathode and anode.

Electrons emits from cathode to anode with phosphors which generate illumination. Ultra thin, wider view angle, superior brightness and low operation power are main advantages of FED. Samsung corporation had been public the

4.5” FED prototype in 1999 [Fig. 2.27].[Choi-1999-3129] And the electron source like SEM filament[Chow-1992-1] or X-ray tube[Yue-2002-355] can also employ the CNTs as electron emitters, which possess longer life, small energy spreading and power-saving significantly. Another field emission application related to general public is cathode-ray tube (CRT) lighting elements. The original has been published in 1998 by Ise Electronics corporation, Japan.[Saito-1998-L346] The fabricated CRTs are of a triode type, consisting of a cathode (nanotubes field emitter arrays), a grid and an anode (phosphor screen) [Figs. 2.28(a) and 2.28(b)]. The maxima brightness with anode at 200μA is 64000 cd/cm². Stable electron emission, adequate luminance and long life (over 10000 hours) are demonstrated. It can be applied to a giant outdoor display or ultra-high quality color CRT displays.

Fig. 2.27 FED display at color mode with red, green, and blue phosphor column.

[Choi-1999-3129]

(a) (b)

Fig. 2.28 Schematic drawing (a) and physical object (b) of a longitudinal cross section of a CRT fluorescent display with a field emission cathode composed of carbon nanotubes. [Saito-1998-L346]

(b)FET

FET is a very important electronic device in history. The overwhelming majority of FET is silicon or Ⅲ-Ⅴ based just because these materials are semiconductors. But some of CNTs also have semiconducting properties, it makes researchers want to fabricate the CNTs based FET. In 1998, Sander reported the room-temperature transistor based on a single SWNTs FET.

[Sander-1998-49] Fig. 2.29 shows the I-V curve of the CNT-FET. In 2001, Derycke in IBM corporation prepared both p-type and n-type nanotubes transistors to build the first nanotubes-based logic gates: voltage inverters [Figs. 2.30(a) and

2.30(b)].[Derycke-2001-453] Surely, it still have lots of complicated problems to mass production above-mention devices, but these results have told that the nano-electronics is not hollow words any more.

Fig. 2.29 Two probe I-V_bias curve for various values of the gate voltage from a CNTs-based FET. [Sander-1998-49]

(a) (b)

Fig. 2.30 (a) Atomic Force Microscopy (AFM) image shows the design of the voltage inverter. (b) Characteristics of the resulting intra-molecular voltage inverter. [Derycke-2001-453]

(c) Lithium intercalation

The basic principle of rechargeable lithium batteries is electrochemical intercalation and de-intercalation of Li in both electrodes. An ideal battery has a high-energy capacity, fast charging time and long cycle time. The capacity is determined by the Li saturation concentration of the electrode material. The SWNTs have shown to possess both high reversible and irreversible capacities

[Gao-1999-153].

(d) Hydrogen storage material

Face to possible energy-crisis of gasoline, people has started to find the substitution methods for many years. Fuel cell was considered to have potential among all of solutions. Once it does be generated, its use as a fuel that creates neither air pollution nor greenhouse gas emissions. But it needs a huge hydrogen storage capability material. SWNTs just can play this role.SWNTs can absorb higher hydrogen than conventional materials. A H₂ uptake of 4.2 weight %, which corresponds to a H/C atom ratio of 0.52, was obtained by these SWNTs with an estimated purity of 50 weight %. Also, ~80% of the adsorbed H₂ can be released at room temperature. These results indicate that SWNTs are highly promising for H₂ adsorption even at room temperature [Liu-1999-1127]. The hydrogen storage mechanisms of CNTs are still not well known yet, and these properties usually occur at high pressure or low temperature environment. It remains

impossible to apply on commercial product so far.

(e) Composite materials

The SWNTs may be used as reinforcements in high strength, Low weight, and high performance composites due to their excellent mechanical properties. A main advantage of using SWNTs for structural polymer composites is that SWNT reinforcements will increase the toughness of the composites by absorbing energy during their highly flexible elastic behavior. Other advantages are the low density of the nanotubes, an increased electrical conduction and better performance during compressive load, or induced high thermal conductivity reinforced material.

(f) Other applications

Atomic Force Microscope (AFM) is employed to obtain the surface morphologies and roughness. It uses a probe scanning the surface of sample, and an incident laser beam irradiates the arm of probe reflecting to a detector which passes signals to computer and draws the images. In order to obtain a high resolution images, the tip must be ultra thin, extremely sharp and high strength.

General type of AFM tip is made of Si₃N₄. The first article that utilized SWNTs as AFM tip was reportedin 1996[Fig. 2.31]. [Dai-1996-147] From Fig. 2.32, one can clearly see the SWNTs tip shows the better image resolution. At the same time, SWNTs with excellent mechanical properties can make the damage ratio of tip

decrease as low as possible. It has been some commercial products of SWNTs AFM tip at present.

In 1998, Wong demonstrated that CNTs tip can be used for chemical and biological discrimination.[Wong-1998-52] Another possible application applied to biotechnology of medicine carriers are developing as well. In the feature, people will easily get to know what disease that we get. Also, people can use CNTs filled with drags injecting into body, then induce it to the proper position relaxing the medicine to destroy the etiology without hurting normal cell nearby.

Fig. 2.31 SWNT attached to the pyramidal tip of a silicon cantilever for AFM.

[Dai-1996-147]

(a) (b)

Fig. 2.32 (a)Tapping mode AFM image of a 400-nm-wide, 800-nm-deep trench taken with a bare pyramidal tip. (b) The image taken with a nanotubes attached to the pyramidal tip with the same specimen.[Dai-1996-147]

Chapter III

Experimental Methods

3.1 Flow chart

Fig. 3.1 Flow Chart of the experiment p-type Si (100) substrate

Catalyst precusor coating by PVD (CoCrPtOx, CoCrOx, Co)

H-plasma pretreatment by MPCVD

AlON buffer layer coating by PVD

Field emission

XPS EDX

Scanning local laser heating pretreatment

Substrate recycling

Figure 3.1 shows the experimental flowchart for the fabrication and analyses of the catalyst precursor-assisted SWNTs. First, the CoCrPtOx films to act as catalyst precursor were deposited by physical vapor deposition (PVD) and buffer layers were prepared by DC reactive sputter. Then, two type of pretreatment were used. (1) H-plasma pretreatment was performed in microwave plasma chemical vapor deposition (MP-CVD) system to carry out reduction of oxidized CoCrPtOx film on silicon wafer. The specimens were subsequently heated up to grow SWNTs in an appropriate CH4/H2 atmospheres for several minutes at appropriate chamber pressure. (2) Local laser heating pretreatment was operated using 659 nm pump laser to heat numerous spots on substrate surface up to ~700 ^oC, and then transferred specimens to MPCVD chamber to synthesize SWNTs on Si substrate. The morphologies of the pretreated catalyst precursor were studied by scanning electron microscopy (SEM). The size and distribution conditions of catalyst particles after pretreatment on the silicon wafer were characterized from grinding cross-section and plane view samples by transition electron microscopy (TEM). Additionally, X-ray photoelectron spectroscopy (XPS) was employed to characterize binding energy of CoCrPtOx

layer at as-deposited and after H-plasma pretreatment to analyze the self-assembly mechanism. The morphologies, microstructures and bonding structures of the as-grown SWNTs were investigated by SEM, TEM and Raman

spectroscopy with a 632.8 nm He-Ne laser, etc. The field emission measurements (J-E) of the specimens were conducted by the simple diode configuration and performed in high vacuum. The oxidation resistance properties of as-grown SWNTs was measured by thermal gravimetric analysis (TGA).

3.2 Raw materials (a) Substrates:

Silicon wafer [P-type (100)]

(b) Source gases:

Hydrogen gas (purity 99.9995%) Jian Ren Chemical Co.

Methane gas (purity 99.999%) San Fu Chemical Co.

Oxygen gas (purity 99.9995%) Jian Ren Chemical Co.

Nitrogen gas (purity 99.998%) Jian Ren Chemical Co.

Argon gas (purity 99.9995%) San Fu Chemical Co.

(c) Target:

CoCrPt Co 57.08 %, Cr 10.97 %, Pt 31.95 % CoCr Co 67.19%, Cr 32.81%

Al purity 99.999%

3.3 Strategy of self-assembly CoCrPtO_x catalyst precursor

As shown in Fig. 2.12, the oxidized phase of PtOx is unstable and can be easily reduced back to metallic state with very fine size when temperature is approximately 500 °C [Kim-2003-1701]. Hence, it has been recently used in optical storage media and nonvolatile memory, as shown in section 2.2. Additionally, Co is the good element for carbon species dissolubility and Cr2O3 are revealed to suppress the grain growth of Ni-Cr alloy effectively [Shaijmon-2005-192]. Combining these unique properties, oxidized film of CoCrPt deposited by PVD is acted as the catalyst precursor to fabricate vertically well-aligned SWNTs and synthesize SWNTs with low process temperature on Si substrate in this study.

3.4 Catalyst precursor and buffer layer deposition procedures (a) Catalyst precursor

The CoCrPtOx thin films which were as catalysts were coated on a (100)-oriented p-silicon wafer with and without buffer layer using a pure CoCrPt target by PVD (Helix 6-gun) in a mixed argon and oxygen atmosphere. The argon and oxygen ratio for depositing was 10:30 (sccm/sccm) and the deposited thickness of CoCrPtOx films were 1, 2, 3, 5, 10 nm.

(b) Buffer layer

AlON which was reported as the most efficient buffer layer [Wang-2005-1906]

was used in this study. AlON film was deposited on a (100)-oriented p-silicon wafer by DC reactive sputtering (Unaxis Cube Trio) using an Al target in a gas mixture of oxygen, nitrogen and argon. The deposited thickness of AlON film was 10 nm which was the best thickness to assist catalyst particles in distributing, as shown in section 2.2[Wang-2005-1906].

3.5 Microwave plasma chemical vapor deposition system (MPCVD)

The schematic diagram of MPCVD system is shown in Fig. 3.2. The main components of the system can be divided into six parts: the microwave generator, wave guides, reaction chamber, gas flow controller, gas pressure controller and pumping system. The microwave generator of microwave source system (Frequency 2.45 GHz, Power 1.3 kW) was produced by Tokyo electronic Corp.

Ltd. The reaction chamber contains quartz tube (inner: 47 mm, outer: 50 mm, China Quartz Corp. Ltd), stainless chamber, stainless holder and rotary pump (Hitachi Corp. Ltd). As Fig. 3.3 shown, sample holder is manufactured by stainless steel, it can bear high working temperature and reduce vacuum pollutions while plasma working. The upper electrode that was made by stainless steel is connected to the DC power supply output. The upper electrode (ground) and substrate holder (negative) were employed to applying the substrate negative bias. The substrate temperature is measured by thermal couple

which equipped in the holder. Mass flow control (MKS model 247) system is used to regulate the flow rate of reacting gas while depositing. Besides, the flow rate controller with different range (1-10 sccm, 10-100 sccm) of flow rate of mass flow control system can be properly adjusted for different gas (Ar, NH3, H2, CH4, C2H2). The low pressure (0.1~100 Torr) of chamber can be detected by thermal couple of vacuum gauge and absolute pressure gauge (MKS Baratron).

The work pressure of chamber can be regulated stably by throttling valve. The degree of the throttling valve was controlled by APC controller (MKS model 263). There is no external heater system equipped on MPCVD. The plasma is used to heat the substrate as the heat source. Cooling cycle system is made up of the refrigerator with closed cooling water and the conduit.

Fig. 3.2 Schematic drawing of MPCVD system

Fig. 3.3 Schematic drawing of MPCVD reactor

3.6 Pretreatment methods

3.6.1 H-plasma pretreatment by MPCVD

After catalyst precursor and buffer layer deposition, the specimen was transferred in air to the chamber of MPCVD which was pumped down to its base pressure. Hydrogen plasma was utilized to activate the CoCrPtOx film and change surface morphology of it. The purpose of this part is to find out the optimum conditions of H-plasma pretreatment to obtain the small average size distribution and appropriate density of catalyst nano-particles. The most optimum CoCrPtOx catalyst precursor pretreatment conditions were: microwave power 600 W, working pressure 30 Torr, sample temperature 580 ^oC, H2 flow 100 sccm and process time 10 minutes. Finally, we draw three straight lines on SEM micrograph and calculate the average particle size by the ratio of line length to particle numbers.

3.6.2 Scanning local laser heating pretreatment

For scanning local laser heating pretreatment, catalyst precursor-deposited sample was transferred to the chamber of two laser static tester (Tueopicts, 633-nm-wavelength cw model diode laser and 659-nm-wavelenth duration mode diode laser). The 633-nm laser was used to monitor the reflectivity change to focus another 659-nm laser. Then using 659-nm pump-laser to heat numerous spots on substrate surface simultaneously to carry out reduction of oxidized

CoCrPtOx film up to ~ 700 ^oC and then well-distributed fine nano-particles were

CoCrPtOx film up to ~ 700 ^oC and then well-distributed fine nano-particles were