Forty years after it was predicted, it has finally arrived: the memristor, the fourth fundamental element of electronic circuits. Researchers predict that this will enable future computers to boot substantially faster than now.
All computer owners know it: it takes forever and a day before your computer starts. Why does it take that long? Because the working memory of present-day computers consists of so-called dynamic RAM (Random Access Memory) that loses its contents when power is cut off. Every time one restarts a computer, all fundamental operating instructions and data must be loaded in memory. That takes time.
A new electronic component, the memristor or memory resistor will change all that. This thing remembers the state it was in before power was cut off. That saves quite a bit of time, especially when starting the computer. In this week's Nature, researchers of the Hewlett-Packard Lab in Palo Alto, California, describe their experiments with a prototype of the memristor.
The existence of the memristor was already predicted in 1971 by -still young- electronics engineer Leon Chua who worked at the University of Calfornia, Berkeley. He was, apparently quite litterally, monkeying around with the complex mathematics that describe how the four fundamental variables of electronic systems (electricity, voltage, charge and magnetic flux) behave in the three basic components of electronics that were known at the time: resistor, capacitor and induction element. Combining these three elements in different ways, makes it possible to create electronic circuits of different types.
Using mathematical logic, Chua came to the conclusion that there had to be a fourth element, an element with the rather interesting property that it remembers the charge that passed through it before power was cut off. He called this theoretical fundamental element a "memristor", an abbreviation of memory resistor. However, is was unclear how the element had to look and what it's physical properties should be.
Now, nearly forty years later, Stan Williams and his colleagues from HP Labs are describing a prototype of Chua's memristor. They discovered it quite by accident, during experiments to make nano switches, tiny switches, with minuscule amounts of titanium oxide. The behaviour of the nano-elements seemed unexpectedly bizarre and after some puzzling Williams and his colleagues determined that they looked very much like Chua's memristors.
The memristors built by the group consist of a sandwich of titanium oxide (a semiconductor) and two extremely thin layers of platinum. They can't be seen with the naked eye: they have a thickness of 5 nanometres or approximately one ten thousandth of the thickness of a human hair. This creates a lot more space on a chip considering that the smallest electronics components today are about 45 nanometres, and the limits of the miniaturization are slowly coming nearer, since there is not really any possibility to go below 20 nanometres with current technology. This means that Moore's Law would no longer be valid. Of course, it has never been valid in the first place, it was nothing more than a prediction that happens to be vaguely correct if one looks not too closely, not unlike the prophecies of the Bible or Nostradamus.
Thanks to the memristor, miniaturization should be able to continue for a while longer. Williams and his colleagues already used it to build a functioning transistor that is smaller than conventional transistors by connecting two memristors. But that's not all. According to Williams, the memristor enables us to create fundamentally different elements than the current three that already exist. "It gives us a much larger palette to choose from," says Williams.
In an accompanying commentary in the same issue of Nature, nano-expert James Tour writes that it'll probably take a while before memristors will actually become common in the electronics industry. Alternatives for the "good old transistor" are often looked at with some suspicion by those involved. The industry will first want to see some large-scale, well-functioning examples, writes Tour. "But once that is done, the race for smaller components will accelerate."
Leon Chua who is approaching retirement is very happy, "I did not dare to hope that I would live long enough to see this actually happening," he told the press agency AP.
From Stanley Williams et al: "
The Missing Memristor Found" in Nature, May 1, 2008 and James Tour: "
The Fourth Element" in Nature, May 1, 2008.
