There are now many examples of how nanotechnology has the potential to revolutionise IT and communications technology - but it might take some time yet, reports Martin Courtney.
So much changes so fast in the IT industry, but one thing never wavers: electronic components get continually smaller and more powerful, and are built to use less energy than their predecessors. Whispers of silicon approaching the peak of its abilities as the semiconducting material of choice for use in CPUs, RAM, storage and communication chips may be premature; but researchers the world over are nevertheless busy exploring how nanotechnology and new materials - atomic or molecular scale structures ranging from 1 to 100nm in scale - can provide the next generation of shrinkage required. One nanometre is roughly one-billionth of a metre, a human hair being typically 100,000nm thick. Manipulating different materials - from silicon to carbon and polymers to light - on that sort of scale is challenging, to say the least.
CPUs and silicon lead the way. Inevitably, it is the existing expertise and financial muscle of today’s IT and semiconductor companies, many of whom are already using 45nm manufacturing processes to create thousands of tiny switches on a single CPU, that are leading the way. All of which means a lot of nanoscale research continues to revolve around the use of silicon, metals and metal oxides.
This evolutionary, rather than revolutionary, approach - which tends to shun the idea of growing new ‘organic’ nanomaterials from scratch using chemical processes - is amply demonstrated by HP’s prototype file server processor, called a Nanostore. This uses a 3D stack of processor cores linked to a memristor - a little-used type of resistor first developed by HP in the 1960s in which the flow of electrical current in an IC is determined by the amount of charge that has previously flowed through it. By using two thin films of titanium oxide sandwiched by nanowire crossbars to control the electrons, HP estimates that within the next 10 years, just one of these chips could store a trillion bytes of memory.
Other researchers have proposed similar things using different materials like polymer and manganite, while companies such as hard-disk maker Seagate are investigating spin-based magnetisation processes that move electrons in different directions.
Elsewhere tiny pinpoints of light are being harnessed: IBM has built a nanoscale photonic communications switch based on a silicon electro-opto modulator which converts electrical signals into light and forwards them from one CPU core to another much faster than wired equivalents. Many companies, including Intel, HP, and IBM, are also looking at photonic interconnects - new ways of using short-range, high-bandwidth optical data transmission techniques within servers and other high-speed computers to replace copper wires and silicon circuits.
The large investment pool provided by silicon-derived revenue is also being used to explore more organic approaches to nanotechnology. As well as looking at complementary metal oxide semiconductor chips less than 16 nm wide, the Tera-scale Reliable Adaptive Memory Systems consortium, made up of Intel, Glasgow University, and other European academic institutions, is also exploring nanowire transistors, carbon nanotubes and other molecular electronics expected to be as small as 5nm in scale.
Nanochemists from Denmark and China have developed nanoscale electric contacts out of organic and tin oxide nanowires, crossing them together to build electric circuits. Much of this nanotechnology will inevitably find its way into the high-volume, low-margin market for PCs, mobile phones and other devices, but the race to build quantum computers is also on. It is hoped that nanoscale circuits will eventually form the basis of faster, more powerful systems that will ultimately be able to solve mathematical problems which current supercomputers cannot, or at least not in any reasonable length of time.
Simply by trying out the seemingly infinite variables and permutables within milliseconds, quantum computers could crack encrypted codes in seconds in comparison to contemporary mainframes, which could take years to perform the same job. They could also provide simulation and modelling applications which could be used for environmental monitoring, or oil discoveries for example, where current computers do not have the processing power, memory or speed to carry out the necessary calculations and data-gathering on this scale.
New types of faster, larger-capacity computer memory and solid state storage are also on the cards, enabled for the most part by using incredibly precise nanoscale lithography, photolithography and atomic layer deposition techniques to map tiny patterns onto silicon oxide and other materials.
Company Nantero is perfecting a proprietary non-volatile random access memory (RAM) technology called Nano-RAM, or NRAM, which relies on the mechanical position of carbon nanotubes deposited on a substrate layer of chip-like material to conduct electrons. NRAM has the potential to be much denser than DRAM or flash memory, meaning higher capacity, and also does not lose its charge (and therefore its data contents) when the device is powered off. Another firm, ZettaStore, is doing something similar with molecular RAM based on multi-porphyrin nanostructures, which can be oxidised and reduced (electrons removed or replaced) to store data.
Chains of polymer molecules
IBM has been working on ‘racetrack’ memory, which it hopes will eventually offer higher storage density than flash, similar to current hard-disk drives but with much faster read/write performance. Based on nanoscale permalloy wires that IBM expects to be able to scale down to around 50nm, this will use magnetic techniques like those in use by available low‑capacity magnetic RAM (MRAM).
Researchers at the universities of California and Massachusetts have also discovered a way to make chains of polymer molecules line up in perfect arrays across large surface areas by heating up sapphire crystals to temperatures of between 1,300 and 1,500°C which they say can provide storage densities of 10TB per square inch on whatever material it is deposited.
Nor are nanoscale storage developments restricted to solid state memory and storage media: Hitachi Maxell has partnered the Tokyo Institute of Technology to create new tape storage technology offering densities of 45 gigabits per square inch (a density which could potentially provide 50TB per tape if applied to LTO-5 format cartridges that currently offer 1.5TB of raw data capacity).
The technology uses the ‘facing targets sputtering method’ to arrange magnetic particles less than 10nm in size onto an ultra-thin nano-structured magnetic film, allowing it to fit more information into a much smaller area and raising the number of bits that can be stored.
Power consumption and battery life
New types of nanoscale CPUs and memory chips may well end-up using a fraction of the electricity as silicon-based equivalents. Combined with parallel development in the use of nanoscale materials in the lithium-ion batteries, which now power everything from laptops to electric cars, success here could see tomorrow’s portable and mobile devices last for significantly longer away from a power supply and recharge batteries in minutes rather than hours.
Professor Eric Pop at the University of Illinois is developing an ultra-low-power digital memory, which he anticipates could use 100 times less energy than equivalent flash memory devices, providing electronic devices with much longer battery life between charges. He is using carbon nanotubes instead of metal wires, creating bits by placing a small amount of the carbon-based phase change material in a nanoscale gap within the carbon nanotube, which can be switched on and off by passing small electrical currents through it.
Researchers at Stanford University have grown silicon nanowires on a stainless steel substrate which they estimate could have up to ten times the capacity of conventional lithium ion batteries. The silicon nanowires prevent silicon cracking when it swells as it absorbs lithium ions during charging then contracts as it discharges.
Coating the surface of the electrode with some type of nanoparticles, like phosphates or anodes composed of lithium titanate, can both increase battery capacity and decrease the time required to charge the battery, while also making them lighter and less flammable.
Over at Massachusetts Institute of Technology (MIT), researchers have developed a technique using carbon nanotubes as the anode and cathode in the lithium-ion battery, with a battery manufacturer called Contour Energy Systems having already licensed the technology to develop new Li-ion batteries. The company is harnessing fluorine-based materials for use in portable electronics, while NASA has commissioned Contour to develop batteries for manned space missions, for example, as well as rechargeable batteries for surface vehicles.
Hitachi Maxwell is also partnering Nagasaki University and Fuji Heavy Industries to develop nano-infused lithium with manganese instead of the usual cobalt element to create lithium-ion batteries with 20 times more power storage than current technology. Organic light-emitting diodes have already found their way into the latest television screens from the likes of Sony, Samsung, and LG, as well as smartphones, digital cameras, and other electronic devices which require thinner, lighter screens and low energy consumption - they reflect light so do not need a backlight and can preserve text or images on the display. A similar type of electrophoretic display using organic ink is already in use in e-readers.
Proof of concept
The next stage in this technology’s development looks to involve carbon nanotubes in sheets of graphene to provide less brittle, more flexible and even stretchable screens, which can then be used for a wider range of touchscreen-based applications, including foldable electronic newspapers and wearable computers.
The challenge is to develop new manufacturing processes that make large sheets of ultrathin graphene able to form full screen displays that are simple and cheap to make. Researchers at Sungkyunkwan University in South Korea have scaled-up an approach first adopted by chemist Rodney S Ruoff at the University of Texas by growing graphene on top of large sheets of copper foil, adding a thin adhesive polymer layer, then dissolving the copper backing before peeling off the adhesive polymer to leave a single graphene sheet, then treating the result of this with nitric acid to improve its conducting properties.
Scientists will also have to determine how to construct manufacturing equipment that can distribute nanotubes evenly and at a specific distance from each other across a large display to prevent the odd blank pixel, and differences in the brightness of neighbouring pixels, which can be visible to the human eye.
Factories may also need to find methods of operating at extremely low temperatures to prevent the display from melting during production, while further development of printed electronics able to map circuitry directly onto the graphene will help to keep production cost and complexity to a minimum.
Nanotechnology has immense, game-changing potential for the IT industry as a whole, but much of the work referenced is still at the proof of concept stage, with even working prototypes few and far between. Perfecting the technology in each case may take another five or ten years, at which point manufacturers may still have to build new fabrication facilities able to produce components in volumes which will prove cost effective, and profitable, in the long run.