Considering the challenges that confront the development of a computer technology on the molecular scale, the idea of logic gates or memory cells on an atomic scale seems pure fantasy. Believe it or not, such a proposal was put forth by none other than Richard Feynman when he said in 1959 that "the principles of physics do not speak against the possibility of maneuvering things atom by atom. It would be possible for a physicist to synthesize any substance that the chemist writes down." Feynman saw the manipulation of individual atoms as an inevitable extension of technology. He pointed out that the first large and crude machines were used to build smaller and finer machines, which in turn were used to construct still smaller devices, until it would shrink to a microscopic scale. This would continue until the limits of matter are reached. Feynman's proposal of machines building ever-smaller machines became known as the top-down approach.

With the emergence of several new revolutionary concepts lately in the field of computers, each claiming to herald in an era of tiny, super fast computers in a scale undreamt of, Feynman's  hypothesis is becoming a reality. Time can attest whether these developments, still at an embryonic stage, would really usher in a new era of extremely minute, super fast processors altering existing methodology of basic computational techniques using silicon chip technology.

Recently computer scientists claimed to have achieved great progress towards constructing a tiny, super-fast computer, which they are calling as Molecular Computers. They say that such computers built on a crystalline structure, will soon replace those based on silicon chips and ultimately make it possible to have a computer so small as a grain of salt.

What more are these scientists claiming? They say that these will require far less power than current computers, and will be able to store voluminous data permanently. They will do away with the need to erase files, and also be immune to attacks from computer viruses, crashes and other forms of destruction. A team from Hewlett Packard and the University of California at Los Angeles says its work could ultimately lead to computers 100 billion times faster than today's most powerful PCs. Their findings have been published in Science magazine.

As of now, computers are based on silicon chips. The information they carry is etched onto them with beams of light, and it is becoming increasingly difficult to do this precisely on chips getting smaller by the day. The ability to shrink these ICs (Integrated Circuits) - the on-off switches that are the basis of computing is limited by the wavelength of light.

Researchers claim they have devised a way to construct circuits using a chemical process. This should make them as small as a molecule. They believe the process would lead to components infinitely smaller than today's smallest transistor. Smaller transistors consume less power and generally switch on and off more quickly. They can also be produced in greater quantities without raising costs of production.

Phil Kuekes, a computer scientist at Hewlett Packard in Palo Alto claims " we have actually built the very simplest gates used in computers - logic gates - and they work." James Heath, a UCLA chemistry Prof. who has worked on the project, said his team created molecular logic gates that switches "on" and "off" by changing electrical voltage that represent "bits" of information. They have achieved this by creating a new compound, called rotaxane, which grows in a crystalline structure.

The team of Heath and Kuekes, who wrote in the journal SCIENCE recently, said that rotaxane molecules, sandwiched between metal electrodes, functioned as logic gates. A crystal can absorb information in the form of an electrical charge, and organize it more efficiently. The chips made using this molecular technology can be as small as a grain of sand. However, it can have additional capabilities, like switching the TV to your favorite channel, and eliminating the necessity for input device like the mouse.

The next step would be to structure the chip. Instead of etching the structure onto the surface, as is currently done with silicon chip technology, it will de downloaded electrically. The team claimed that they could download all the complexity by wire attached to a bigger computer. But there're a few hurdles yet. Currently available wires are too big - much bigger than the rotaxane molecules - to do this. Efforts are being made to shrink the wires until they're the same diameter as the molecules, and then the miniaturized technology will become a reality.

They also said that it might be possible to use carbon nanotubes - long thin tubes made of pure carbon. Also known as "Bucky tubes," they are no thicker than most molecules.

This same team has recently announced that they made the largest "defect tolerant" computer ever and called it the Teramac.