MAGNETIC RAM

There are several flavors of MRAM and one has seen limited commercial use already. MRAM offers a number of attractive features, including the fact that it is non-volatile, enabling devices such as a PC or mobile phone to boot up in little or no time. Several companies are working on MRAM technologies, and suggestions are that while the data densities might not be as good as with other technologies, cost per bit could be very low.

HARD DRIVES AND TAPES

Hard drives currently on the market for PCs have capacities in the tens of gigabytes and data densities of a few tens of gigabytes per square inch with 100 Gbits/sq. in. expected in the next generation. The use of magnetic nanoparticles offers the potential of terabyte drives, as does the patterned media approach being pursued by IBM and General Electric. Interestingly, this latter approach is being pursued using nanoimprinting technology. Fuji announced late in 2001 new magnetic coating promising 3-gigabyte floppy disks.

MEMORY AND STORAGE

We have noted that part of the difficulty in creating processors with nanotubes or molecular electronics relates to complexity. Data storage structures are far less complex than processors and many new technologies are converging on this area, promising commercialization in five years or less. Information storage requirements continue to grow but vary in nature from one application to another and can be approached in many ways. Magnetic disks in computers have been increasing their capacity in line with Moore's law, and have a market at the moment of over $40 billion. The other type of information storage common to all computers is DRAM (dynamic random access memory). DRAM provides very quick access but is comparatively expensive per bit. Magnetic disks can hold much more information but it takes much longer to access the data. Also, DRAM is volatile—the information disappears when the power is switched off. The trade-offs between access speeds, cost, storage density and volatility dictate the architecture of computers with respect to information storage. While hard disk technology continues to offer increasing data densities and lower costs per bit, a number of nanotechnologies are promising new types of RAM that are non-volatile and could have enough capacity to make disk storage unnecessary for applications such as personal computers. Companies are forecasting commercial products within two to four year time frames. How much penetration each technology will achieve in the variety of areas in which storage is used depends on the complex interplay of factors that have led to the current division of data storage technologies, but it would certainly be surprising if the consequences weren't disruptive for the industries involved.

Quantum Computing

In the much longer term quantum computing, offers staggering potential by virtue of the ability to perform simultaneous calculations on all the numbers that can be represented by an array of quantum bits (qubits). The scale at which quantum effects come into play, the atomic scale, argues for a requirement for nanoscale structures and quantum dots come up regularly in discussions of quantum computing. Primary applications would be in cryptography, simulation and modeling. The realization of a quantum computer is generally believed to be a long way off, despite some very active research. Funding in the area is thus still largely that provided for pure research, though some defense department money has been made available.

Spintronics

Magnetism is dictated by the direction of spin of electrons and increasing research into the use of this property has led to the coining of the term spintronics. The read heads of disk drives already exploit electron spin in an effect called giant magneto Resistance, as does MRAM (see later), which has already seen limited commercial production. An effect called ballistic magneto resistance has recently been demonstrated to have the capability of producing read heads that can deal with storage densities of a terabit per square inch—ten times the density expected in the next generation of hard drives. Commercial application of spintronics in electronics is farther away but the promise is there—a Canadian group recently created a transistor that was switched by the spin of a single electron confined in a quantum dot.

Molecular Nanoelectronics

Organic molecules have also been shown to have the necessary properties to be used in electronics and a single atom transistor was even demonstrated recently. Devices made of molecular components would be much smaller than those made by existing silicon technologies and ultimately offer the smallest electronics theoretically possible without moving into the realm of subatomic particles. The issues of connectivity, and thus mass production, apply to molecular electronics too, but choosing your molecule carefully does offer the potential of using self-assembly (discussed earlier) to create structures, an approach that could offer great economies.