Carbon Nanotube
Final Report
Deanna Zhang
Chuan-Lan Lin
May 14, 2003
Introduction
Carbon nanotubes are formed from pure carbon bonds. Pure carbons only have two covalent bonds: and , the former constitutes graphite and the latter constitutes diamond. Figure 1 shows the structure of graphite and diamond.
Graphite Diamond
Figure 1- Structure of graphite and diamond [5]
is a strong bond within a plane but weak between planes. is composed of one S orbital and two P orbitals. When more bonds come together, they form six-fold structures, like honeycomb pattern, which is the plane structure, the same structure as graphite. Graphite is stacked layer by layer so it’s only stable for one single sheet. That’s why graphite is used in pencils. Viewing these layers perpendicularly shows the honeycomb patterns of graphite. Wrapping these patterns back on top of themselves, joining the edges, and close one end while leave one end open, we form a tube of graphite, which is a Nanotube [5].
Two types of nanotubes exist in nature: single-wall nanotube and multi-wall nanotube. Figure 2 shows the structure of single-wall nanotube and multi-wall nanotube.
Single Nanotube Multiwall Nanotube
Figure 2- Structure of single-wall and multi-wall carbon nanotubes [5].
Single-wall nanotubes has only one single layer and with a diameter of 1 to 5 nm [5]. The properties of SWNT are more stable than MWNT so it is more favorable. MWNT is the first nanotube discovered in 1991. MWNT is a little bigger than SWNT because MWNT has about 50 layers. MWNT’s inner diameter is from 1.5 to 15 nm and the outer diameter is from 2.5 nm to 30 nm. SWNT have better defined shapes of cylinder than MWNT, thus MWNT has more possibilities of structuredefects and its nanostructure is less stable [5]. Most researchers focus on SWNT and develop applications based on SWNT due the physical stability of SWNT.
Properties of Carbon Nanotubes
Electrical Properties
Carbon nanotubes have some distinct electrical properties. One of the important properties of carbon nanotube is that it can exhibit the characteristics of a metal or a semiconductor [6]. Specially, the energy gap is determined by the rolling direction of nanotube.
Figure3-Diagram showing rolling direction of nanotube [6].
In Figure 3, Ch is Mamada vector connecting two crystallographical equivalent sites, and and are the unit vectors of the unit cell. Ch= +, whereand are integers. The nanotube is formed by connecting A and A’ point. There is a simple rule to determine if the nanotube acts as a metal or a semiconductor: if (+)/3 = integer, nanotube acts as a metal otherwise it acts as a semiconductor [6] Due to the miniscule size of the nanotube, the electrical conductivity is measured by the four point probe method to determine the sheet resistance. The four point probe method measures the resistivity of any semiconductor material [1]. Figure 4 shows the setup of the four point probe method. At the same time, nanotubes have been studied to make switches and transistors, which would be much smaller than the silicon chips currently used. The wires made by nanotubesare capable of currents that are 100 times greater than metal wires, making nanotubes useful in the production of flat panels [6].
Figure 4-Four Point Probe method setup. The sheet resistivity Rs= [1].
Additional properties of nanotubes:
Nanotubesare the strongest fibersthat are currently known. A single-wall nanotube can be up to 100 times stronger than that of steel with the same weight. The Young’s Modulus of SWNT is up to 1TPa, which is 5 times greater than steel (230 GPa) while the density is only 1.3 g/cm^3 [6]. That means that materials made of nanotubes are lighter and more durable. There may also be other applications due to these properties, such as car bumpers, and strong wires. Nanotubes also have a very high aspect ratio. The lengths of nanotubes are usually around 1 µm, while the diameter for SWNT is only 1 nm (50 nm for MWNT) [6]. This property makes nanotubes useful for tips and nanowires. In addition, the thermal conductivity (2000W/m.K) is five times greater than that of copper (400W/m.K) [6].
History of Carbon Nanotubes:
The discovery of the carbon nanotube is based off of the discovery of the Buckyball. Buckyballs, also know as , are made up of 60 carbon atoms,arrangedin the shape of a sphere. Figure 4a shows the Buckyball structure. The exact chemical name of is Buckminsterfullerenebut is often referred to as fullerene or Buckyball [5]. The discovery of the Buckyball was accidental. While researchingRadioastronomy, scientists found unusuallong molecules that had not been synthesized in laboratory. The concentration of these long molecules was much higher than anyone expected. Scientists attempted to synthesize these long carbon chains in the laboratory. They used a powerful laser to evaporate small amounts of graphite into a hot cloud of particles, then cooled the cloud with a stream of helium gas [5]. This allowed the atoms to condense into clusters. They then used a mass-spectrometer to identify the molecular mass. The result was a molecular mass of 720amu, which corresponds to 60 carbon atoms (12amu x 60=720amu) [5]. Because the amount synthesized was too little to allow a structure analysis, they hypothesized that 60 carbon atoms had formed themselvesinto a sphere. Scientists around the world duplicated these results, creatingBuckyballs in order to perform structural analysis. From these experiments, the scientists finally proved the existence of the sphericalstructure.
Figure 4a: Buckyball structure Figure 4b: Nanotubes viewed under
of [5] electron microscope [5]
Using carbon arc-evaporation synthesis, more Buckyballs were created. When the carbon arc power supply was changed to DC current instead of AC, nanotubes were created. In 1991, a Japanese Scientist, Sumio Iijima, first observed and reported the existence of entirely carbon-made nanotubes. He found that the graphite structures contained nanoparticles and nanotubes, which were later identified as multilayer nanotubes [5]. Figure 4b shows these nanotubes as viewed under an electron microscope.
Timeline of the discovery of Nanotubes:
When / Who / Events1940s / German Chemist Otto Hahn / When trying to create heavier atoms by arc carbon method of neutron, Hahn reported the existence of carbon chains. Because he was interested in only metal atoms, the research of carbon chains was not continuing.
1970s / England Chemists Harry Kroto and Dave Walton / They were synthesizing long carbon chains to make something like gas cloud in the galaxy.
1985 / Kroto and his American colleague, Rick Smalley / They collaborated a project to simulate conditions of red giant stars in the laboratory.
Late 1980s / Scientists around the world / Buckyball was synthesized and confirmed as
1991 / Japanese Scientist, Sumio Iijima / Discovery of multi wall carbon nanotubes
1993 / S, Iijima and T, Ichihashi / Synthesis of single wall carbon nanotubes
Begin to use Laser ablation method
1995 / A.G. Rinzler / Research of nanotubes as field emitters
1996 / Professor Robert F. Curl, Jr., Rice University, Houston, USA, Professor Sir Harold W. Kroto, University of Sussex, Brighton, U.K., and Professor Richard E. Smalley, Rice University, Houston, USA, / Awarded 1996 Nobel Prize in Chemistry for the discovery of Buckyball
1998 / Development of HiPCO, CVD methods
1999 / Samsung Company / Demonstrated Flat Panel display prototype (4.5”, full-color) using nanotube as filed-emission source.
2001 / IBM research group / The first computer circuit composed of only one single carbon nanotube was announced
2001 / M. Kociak / Intrinsic superconductivity of carbon nanotubes
Fabrication of Carbon Nanotubes
The physical mechanisms and reactions that cause the formation of carbon nanotubes are still unknown. However, four methods currently exist to fabricate these nanotubes. The four methods are all gas-phase processes that start with a source of carbon that is evaporated from a surface [3]. The four techniques used produce carbon nanotubes are the carbon arc-discharge method, the laser vaporization technique, chemical vapor deposition, and high-pressure carbon monoxide method.
Carbon Arc-Discharge
The arc-discharge method produces good quality multi-wall and single-wall nanotubes. This process produces nanotubes at a greater rate than the pulsed laser vaporization technique. This technique utilizes two graphite electrodes to generate an arc by a high dc current. After arc discharging for a period of time, a carbon rod builds up at the cathode. Carbon nanotube bundles and amorphous carbon both form at the cross section of the rod. Figure 8 shows the setup of the arc-discharge method. This method does not produce clean results due to the existence of these amorphous carbons. Helium gas is present to increase the speed of carbon deposition. Some parameters that are critical in this process are the pressure of the helium, the temperature, and the dc current. Efficient cooling is necessary to form homogenous deposition of carbon nanotubes [9].
The group at University of Montpellier, France perfected the arc-discharge method by using yttrium and nickel metal catalysts. The group produced carbon nanotubes from vapors of carbon containing a small amount of nickel and yttrium catalysts. “An electric arc vaporizes an anode containing the catalysts” [4]. An electrical current of 100 amps and 35 volts provide the energy to generate the discharge. The flow system is controlled via gas pressure and is pumped by a mechanical vacuum pump [4].
Figure 8- Carbon Arc-Discharge Method
Pulsed Laser Ablation or Vaporization
The pulsed laser vaporization (PLV) of graphite in the presence of an inert gas and catalyst forms single-walled carbon nanotubes. Without the presence of these catalysts, the vaporized graphite would form buckyballs instead. Buckyballs are formed in nature when both ends of the graphite sheet closes. However, in the presence of a catalyst, the graphite layers stays open on one end [4].
The PLV of carbon containing metal catalysts produces the purest SWNT. Some factors that determine the amount of carbon nanotubes produced are the amount and type of catalysts, laser power and wavelength, temperature, pressure, type of inert gas present, and the fluid dynamics near the carbon target [4]. The PLV method reduces the amount of amorphous carbon contaminates [3]. The PLV of carbon is an expensive method that avoids the high electric fields involved in the arc-discharge method [9].
The NASA Johnson Space Center Carbon Nanotube Project uses two lasers to impinge on a composite graphite and metal catalyst target. These two lasers are Neodymium: Yttrium-Aluminum-Garnet pulsed. The frequency of pulse operation is 60 Hz, which is very fast. The lasers operate at infrared and green wavelengths at 300 mJ. The laser is focused onto a carbon targets containing one atomic percent each of the catalysts cobalt and nickel. The carbon target is ablated in a 5 cm flow tube pressurized with argon at 500 Watt, in an oven at 1473 K [4]. Figure 9 shows the setup. The Argon gas carries the vapors from the high temperature chamber into a water-cooled copper collector positioned downstream [9]. The nanotubes self-assemble from the carbon vapors and condense on the walls of the flow tube. The nanotubes produced have a very narrow distribution of diameters [9]. The yield is about 0.3=0.4 grams per hour [4].
Figure 9- Pulsed Laser Vaporization Apparatus [4]
Chemical Vapor Deposition
The chemical vapor deposition (CVD) method deposits hydrocarbon molecules on top of heated catalyst material. The catalyst is usually predeposited on a substrate. Metal catalysts dissociate the hydrocarbon molecules. Figure 10 shows the apparatus of the CVD process. The CVD method produces both single-wall and multi-wall nanotubes. The CVD process uses hydrocarbons as the carbon source. Hydrocarbons flow through the quartz tube where it is heated at a high temperature. The dissociation of the hydrocarbons breaks the hydrogen carbon bond, producing pure carbon molecules. At high temperatures, the carbon form carbon nanotubes. Growth of SWNTs also occurs at a higher temperature than MWNTs [7]. Some advantages of the CVD are low power input, lower temperature range, relatively high purity and possibility to scale up the process [9].
Figure 10-Apparatus of the CVD process [9]
The main focus in the CVD process is the prepattern of the substrate with a catalyst. Selective growth is achieved through prepatterning the substrate with a metal catalyst. One of the ways of prepatterning the substrate is to use standard lithography techniques. In this process, a photoresist is deposited onto the substrate. The resist is exposed and developed, creating a pattern of resist on the surface. Metal catalysts are deposited on the film using the resist pattern as a mask. The catalyst is lift off with the unexposed substrate through wet etching. The photoresist not covered by the catalyst is etched away. The CVD growth of the carbon nanotubes exists on the catalysts. Figure 11 shows the prepatterning process [7].
Figure 11-Prepatterning Process [7]
High Pressure Carbon Monoxide Method
The high pressure carbon monoxide (HiPCO) method can produce large quantities of carbon nanotubes. The high yield demonstrates the high potential of this method for bulk production of SWNTs. “Catalysts for SWNT growth form in situ by thermal decomposition of ion pentacarbonyl in a heated flow of carbon monoxide at pressures of 1-10atm and temperatures of 800-1200 C” [2]. Previous methods for growing CNTs using hydrocarbons as source have resulted large quantities of amorphous carbon and graphitic deposits due to the thermal breakdown of hydrocarbons at high temperatures [2]. The amorphous carbon overcoating would have to be removed in subsequent steps. The HiPCO method uses carbon monoxide as the carbon feedstock and Fe(CO)5 as the iron-containing catalyst precursor. The SWNT yield and the diameter of the nanotubes produced can vary over a wide range determined by the condition and flow-cell geometry [2].
The HiPCO method produces SWNTs by flowing CO mixed with a small amount of Fe(CO)5 through a heated reactor [2]. Figure 12 shows the setup of the HiPCO method. “The products of the Fe(CO)5 thermal decomposition react to produce iron clusters in gas phase. These clusters act as nuclei upon which SWNTs nucleate and grow” [2]. The solid carbon is formed through CO disproportionation, also known as the Boudouard reaction:
CO+CO C(s)+CO2
This reaction occurs catalytically on the surface of the iron particles. The iron particles “promote the formation of the tube’s characteristic graphitic carbon lattice” [2]. The flow tube has a thick quartz wall and is contained within the furnace. The rate at which the reactant gases are heated determines the amount and quality of the SWNTs produced. The CO and Fe(CO)5 gases are maintained at a low temperature initially through a water-cooled injector. This low temperature is maintained so that rapid heating can occur inside the furnace [2]. This apparatus of the HiPCO method yielded high quantities of SWNTs.
Figure 12-HiPCO apparatus [2]
Applications of Carbon Nanotubes
Field Effect Transistors
Carbon nanotubes are a new form of carbon with unique electrical and mechanical properties. Varying the diameter of the cylinder can control the band gap of semiconducting nanotubes. The width of the diameter of the nanotube is inversely proportional to the size of the band gap [16]. Semiconducting nanotubes can be used to build molecular field-effect transistors (FETs) while metallic nanotubes can be used to build single-electron transistors [17].
Single wall carbon nanotube transistors are electronic devices based on a single rolled-up sheet of carbon atoms. When the first SWNT transistors were formed, they could not operate at room temperature. However, recent research has developed SWNT transistors that do operate at room temperature. Early nanotube FETs used a non-local back-gate with the nanotube side-bonded to noble metal electrodes [16]. This setup gave large contact resistance and poor characteristics [16]. The recent CNFETs are now built with top-gate geometry and resemble more conventional silicon CMOS devices. Figure 13 shows the layout of the carbon nanotube transistor along with its I-V characteristics. The SWNT transistors consist of a semiconducting carbon nanotube about 1 nm in diameter bridging two closely separated metal electrodes a top a silicon surface coated with SiO2 [16]. Applying an electric filed to the silicon via the gate electrode turns on and off the flow of the current across the nanotube by controlling the movement of charge carriers onto it.
Figure 13- Carbon nanotube transistor layout and its I-V characteristics [17]
The research group at IBM has studied methods of producing single- and multi-wall carbon nanotubes to be used as field-effect transistors. The SWNTs were formed via laser ablation of graphite doped with cobalt and nickel catalysts [17]. The SWNTs were ultrasonically treated with H2SO4/H2O2 solution to clean the NTs. The MWNTs were formed by arc discharge evaporation methods [17]. Figure 13 shows the schematic of the NT device. The field transistor consist of either an individual SWNT or MWNT bridging two electrodes deposited on a 140 nm thick gate oxide films on a doped SI wafer. The Au electrodes were generated through electron beam lithography.
The behavior of the NT transistors is similar to that of a p-channel MOSFET. The I-V characteristic of the NTs was measured at room temperature. Figure 13 shows the I-VSD of a device consisting of a single SWNT with a diameter of 1.6 nm. The source-drain current decreases strongly with increasing gate voltage indicates that the NT device operates as a filed effect transistor and also that the transport through semiconducting SWNT is dominated by positive carriers [17]. The source of these carriers is possibly from the carrier concentration inherent to the NT or these carriers can be the majority carriers that were injected at the gold-nanotube contacts.
The CNFETs have several important differences from the conventional semiconductor transistors. One of the difference is that carbon nanotube is one-dimensional which greatly reduces the scattering probability [16]. As a result, the devices may operate in the ballistic regime. Another difference is that the nanotube conducts essentially on the surface where are the chemical bonds are saturated and stable [16]. The stable bond indicates that there is no need for careful passivation of the interface between the nanotube channel and the gate dielectric. The MOSFET devices do not have such an interface between the silicon/silicon dioxide interfaces. A third difference is that the Schottky barrier at the metal-nanotube contact is the active switching element in an intrinsic nanotube device [16]. The switching behavior involves mostly the contacts as oppose to the bulk of the nanotube.