Unleashing the potential of carbon nanotubes in various devices

Carbon nanotubes have great attributes. Depending on their geometry, they can combine semiconducting or metallic levels of conductivity with an incredibly high aspect ratio, a phenomenal current density of 109 A cm-2 and great strength – this allotrope of carbon has the highest Young’s modulus of any solid. In addition, carbon nanotubes can support ballistic electron transport, and they have a very high thermal conductivity of up to 4000 W/m.K and an inert nature that stems from the nature of the bonds. Given this great set of intrinsic characteristics, one would expect that carbon nanotubes, which were first discovered in 1991, to have spawned a multi-million dollar industry. But that’s not the case – their deployment is essentially limited to three commercial applications, which all use the bulk form of this material. Today, these tubes are strengthening carbon electrodes in lithium batteries; forming strong composite materials used in the blades of wind turbines; and creating strong, lightweight golf clubs and tennis rackets with eye-watering price tags.  To unleash the potential of non-bulk forms of carbon nanotubes, which could create and enhance many different types of device, the European Commission has invested € 5.3 million in a €8.4 million, three-year ‘Technotubes’ project involving 12 academic and industrial partners. They are the University of Cambridge, Aixtron, Philips, imec, Thales Research and Technology, Thales Electron Devices, Cambridge CMOS sensors, Fritz Haber Institute, the Technical University of Berlin, the Technical University of Denmark, ETH Zurich and CNR-Trieste. This recently completed programme that was led by John Robertson from the Department of Engineering at the University of Cambridge had several goals. These included the construction of the first equipment for production of carbon nanotubes on 300mm substrates, and the development of: High-conductivity interconnects for improving microelectronics; time-modulated cold-cathodes for X-ray scanners and tomography; and enhanced surfaces for microfluidics. Scaling production When the project kicked off in May 2009, Aixtron’s reactor portfolio included commercial tools for depositing carbon nanotubes on substrates up to 150 mm in diameter. This company’s primary task in the Technotubes programme has been to develop a new tool for nanotube growth on a 300 mm platform that meets the needs of the other players in the project. Speaking at the project’s closing meeting, Ken Teo, Director of Nanoinstruments at Aixtron, explained that they have developed a reactor for 300 mm wafers with a robot-loading system and a suite of monitoring tools – it is fitted with multiple infrared pyrometers, an in-situ camera and analysis ports. This reactor uses C2H2, C2H4 and CH4 as the gas sources; argon and nitrogen as inert sources; and hydrogen and ammonia as reducing gases (see Figure 1 for details). Like the MOCVD process, the substrate is heated from below, and growth of material proceeds via decomposition of source gases on the wafer surface. Figure 1. One of the elements of the Technotubes project has been the design of a reactor for forming carbon nanotubes on 300 mm substrates. Initially, engineers at Aixtron developed a 300 mm ‘R&D’ tool – a blueprint for the commercially launched BM 300 that requires manual loading and has no heating of the substrate from above – before they went on to build a production reactor. This tool, the BM 300T (see Figure 3), is in the process of internal qualification. It has a wafer temperature uniformity of 4 °C, and it has been used to produce various forms of carbon nanotubes (see Figure 2). This includes single-walled variants that could find application in sensors and interconnects and are produced with anneal and growth times of 10 minutes and 2 minutes, respectively. Trim the anneal time to 3 minutes and extend the deposition time to 5 minutes and it is possible to make multi-wall carbon nanotubes that could find deployment in X-rays sources and interconnects. Straight, vertical, carbon nanofibres that could be used in X-ray sources, cell probes and microfluidics, can also be formed in the BM 300T with another set of growth conditions.   Figure 2. Different types of nanotube can be formed through adjustments to anneal times, growth times and growth temperatures. Figure 3. Robot loading is one of the features of Aixtron’s BM 300T, which is in the process of internal qualification. Teo wrapped up his presentation by providing an estimate of the production prices for carbon nanotube manufacture. His calculations are based on three production machines that would share a robot-handling system, pumps and infrastructure. If the life of the tool is ten years, utility is 85 percent and capital expenditure is €5.05 million, the cost-per-wafer and cost-per-inch for a substrate coated in carbon nanotubes will be € 21 and €0.17, respectively. Aiding Moore’s law Seven partners in the Technotubes project are working together to develop deposition processes for various aspects of microelectronics. Carbon nanotubes are viewed as promising materials for making through-silicon vias and interfaces with great thermal properties. According to the International Technology Roadmap for Semiconductors, fast-forward ten years and the current density in copper could be limiting further scaling of transistor sizes in cutting-edge logic circuits. Single-walled carbon nanotubes are a very attractive alternative to copper because they have a theoretical current density limit that is more than two orders of magnitude higher, and their high aspect ratio makes them suitable for filling vias. Carbon nanotubes can also make an impact in the silicon foundries of the 2020s if they are formed with a density of at least 3 x 1013 cm-2, because this can ensure a sufficiently low electrical resistance. Partners in the Technotubes project have got to within 50 percent of that density with three different approaches: Improving the catalyst support layer; using a patent-pending, multi-cycle catalyst deposition technology; and employing catalyst carburisation, a process pioneered by Toshiba. Ideally, carbon nanotube growth processes should be applicable to a range of substrates. Deposition on conductive substrates – such as CoSi2, TiN and tantalum – is needed to form interconnects, while growth on copper leads to good thermal interfaces. The European team of researchers has shown that carbon nanotubes can be created on conductive platforms with a high areal density by depositing a catalyst film. This is restructured into islands that initiate tube growth. One method for doing this is to create a CoxFy film on CoSi2 by reactive ion etching in SF6, followed by reduction in hydrogen gas to form cobalt spheres that define the locations of the nanotubes. Efforts on 200 mm silicon-wafer processing of carbon nanotubes in the microelectronics work package have led to an integrated process flow for carbon-nanotube vertical interconnects, which are built on metal contacts and attached to top metal contacts (see Figure 4). Another aspect of this strand of the Technotubes project has involved the use of carbon nanotubes as a very high thermal conductivity interface between high-power GaN LEDs and their packaging materials. Working together, Philips Lumileds and the University of Cambridge have developed carbon nanotube heat spreaders on copper.   Figure 4. Researchers in the Technotubes programme developed processes to form contacts with carbon nanotubes. X-ray sources Philips was also involved in another aspect of the Technotubes project, which focused on the development of carbon nanotube cathodes for X-ray tubes. These sources of radiation, which are used in radiography equipment, generate X-rays via the collision of high-energy electrons into an anode. Conventional X-ray tubes, which feature a thermionic cathode and a rotating anode in an evacuated tube, consume a tremendous amount of energy. Typical currents, voltages and powers are 1A, 150 kV and 120 kW, respectively. Requirements for next-generation X-ray tubes include higher operating currents, plus a smaller spot size for the X-ray source that can improve spatial resolution. It is also possible to realise gains in temporal resolution with higher peak powers and faster switching times, and if a moving source is replaced by multiple ones with fast switching times, mechanical movement can be eliminated from X-ray imaging systems. Switching from conventional cathodes to those based on carbon nanotubes can address all these areas. As part of the Technotubes project, researchers have targeted the development of carbon nanotubes arrays delivering 1 A cm-2. These must be fabricated by a scalable, industrially viable growth process that enables the production of nanotube cathodes at a cost that is competitive with the incumbent technology. Efforts directed at these goals have revealed that heating the underlying substrate from room temperature to 350 °C increases the current density, lowers the threshold voltage and improves current stability. This team has also incorporated its carbon nanotube emitter arrays in ‘non-medical’, low-power, transmissive anode X-ray tubes. Images of light bulbs have been taken using total emission currents of 0.1 mA to 0.2 mA (see Figure 5).   Figure 5. Cathodes constructed from carbon nanotubes have been used in X-ray imaging tools. Improving fluid flow Another team within the Technotubes programme has been developing carbon nanotube-based microfluidic components and surfaces. These could find application in three rapidly growing, multi-million-dollar markets: The $3 billion chromatography market that is growing at 10 percent per year; the $1.2 billion electrophysiology sector that is increasing in value by 6 percent per year and involves drug screening, implants and diagnostics; and the $0.5 billion market for ultra-capacitors for providing fast energy storage. This market, which will expand as renewables are used for a greater proportion of electricity generation, is tipped to grow at 18 percent per year. In the chromatography industry, nanotubes can be inserted in the channels of microfluidic chromatography devices to enable different types of molecules to be separated by the speed they travel through the tubes. With this approach, the researchers have separated amines and gone on to build a prototype chromatography chip system. When conventional microelectrode cells are used in electrophysiology, they provide a poor interface between the probes and the specimen, leading to poor measurements and even death of the cell. Thanks to the very small diameter of carbon nanotubes, they cause far less damage. If the density of the tubes is too high, they cannot penetrate the cell, so researchers in the Technotubes project have developed a process to form nanotubes on an array of titanium columns spaced about 2 µm apart. One of the requirements for ultra-capacitors for energy storage is a high surface area to combine a high power density with a high energy density. These demands can be met with high-density nanotube forests, formed by direct growth on aluminium foil. Efforts in this direction have created a dense array of 25 µm-high tubes on 10 µm-thick aluminium foil (see Figure 6). This work, plus that of all the other projects in the Technotube programme, showcases the potential of carbon nanotubes for making electronic devices. If commercial success follows, the revenue generated in medical, electronic and power industries could soon overshadow today’s sales of the bulk form of carbon nanotubes. Figure 6. The combination of large surface area and low weight make carbon nanotubes, which can be deposited on aluminium foil, an attractive candidate for building ultra-capacitors.