As the revolution in information technology continues to boom, computer systems are having to cope with ever-increasing amounts of data. The solution to this overload may be to store it using light rather than electronics.
Current computer technology stores data in much the same way as a library: each item of information is placed in a specific location within a data bank, like books in a library. The drawback of this approach is that as the amount of data increases, so must the volume of the library.
But for several decades now, information engineers have been dreaming of an alternative, in which items of information are more densely packed by writing them on top of each other. In conventional computer memories, data is recorded in binary code, which can be written into an array of electronic or magnetic switches that are 'on' or 'off'.
This same code is used in light-based (optical) memories where it translates into a pattern of light and dark spots, like a black-and-white image on photographic film. The advantage of optical information storage is that the patterns, each representing a piece of information - say, a person's name or a bank account number - can be superimposed in the same image-recording medium.
In effect, the data bank is like a photographic film subjected to multiple exposures. But won't the images become hard to distinguish once several have been laid on top of each other? Not if each one is stored in three dimensions within a block of light-sensitive material, rather than in the two dimensions of photographic film. Not, in other words, if the image is a hologram.
The science of holograms was developed in the 1940s by Denis Gabor, a British-Hungarian physicist. Holograms are constructed using 'coherent' light, in which all the rays can be considered as oscillating in step with each other.
Gabor showed theoretically that if a coherent light beam is split in two, one half is shone on to an object and the scattered light beam is allowed to intersect the other half of the beam, the two beams will interfere to form a spatial pattern of light and dark that is a three-dimensional image of the object: a hologram.
Gabor's ideas were realised experimentally when the invention of lasers in the 1960s made it possible to generate coherent light. A hologram contains all the visual information from the original object, so one can see the front, back and sides.
Most people have seen plenty of evidence, in exhibitions and art shops, that holo graphy works. Putting holo grams to use for optical information storage with the data density and stability required by computers is, however, another matter entirely.
Not only does one need a material that will capture holo graphic images quickly and with high fidelity, but one must be able to read that information back out, to erase it and write over it, and to link the whole affair up to the electronic circuitry of a computer. These problems are now being solved.
To write digital electronic information into a light-sensitive material, the electronic signal is used to control arrays of liquid crystal shutters which chop a laser beam into a pattern of light and dark. This processed beam then writes the data as a holographic image into the storage medium. Like a film projector converting a film into light, a second laser beam is used to read the data back out. Highly sensitive light meters called charge-coupled devices (also used in astronomy for detecting extremely faint galaxies) convert the optical signal back into electronic form.
Can a system like this be used to process information at speeds and data densities comparable to electronic computers? This was the challenge taken up by the Stanford team, led by Dr Lambertus Hesselink. In August they announced their success in developing a holo graphic storage system that writes and reads digital information when hooked up to a standard computer hard drive.
The performance of this prototype was not impressive: it was slow and could handle fewer data bytes than a floppy disk. Moreover, after reading the information out several times, the memory became corrupted.
Dr Hesselink's team is optimistic about improving the performance; it has set up a company, Optitek, and aims to have a commercial system in two to three years. The Stanford team is not alone in believing that holographic storage could solve the data storage problem. Other companies are pursuing the idea, too. The biggest obstacle is the holographic storage medium itself.
Dr Hesselink used the recording medium that has been the mainstay of optical technology for years: a crystalline material called lithium niobate. Hopes for the future are pinned on new materials made from organic polymers which will allow higher storage densities, faster writing speeds and greater ease of processing.