A new device will enter designs

With HP researchers recently claiming to prove the existence of a new basic passive element—memristor—for electronic circuits, designers will soon have one more passive component to work with apart from the traditional resistor, capacitor, and inductor. Memristor, a two-terminal circuit element, maintains a functional relationship between the time integrals of current and voltage, which results in resistance variation in accordance with the device’s memresistance function. This variation has the potential of saving data. This component makes it possible to design systems whose memories are not lost when the power goes off, and that consume far less power. Memristor will enable a system such as a computer to start instantly without needing to boot.

MAKING DRAM OBSOLETE

As memristor has properties that cannot be duplicated by the three traditional passive components or any combination of these three, it will open up new avenues in electronic design. HP believes that memristor will make possible IC designs that greatly enhance the performance and energy efficiency of data centers as well as PCs. This groundbreaking component will make DRAM obsolete. Memristors will enable design of systems that remember and link a series of events in a way identical with that a human brain recognizes patterns. This can radically change design considerations in security and facial recognition technologies, enabling recognition of biometric features and appliances to learn from experience.

I would call for some caution in evaluating the basic significance of the new invention. How far is it appropriate to label this device as a fundamental passive component like R, L, or C? All fundamental elements are linear devices, but memristor is a non-linear two terminal device, and a perfectly linear memristor is equivalent to an ideal resistor. So some designers are less than enthused in calling memristor a fundamental element. They also point out that a fundamental circuit element is characterized by a specific relation between current, voltage, and time, and have a specific unit. But memristor relates current and voltage in the same way as resistance, and only a certain application or function can establish its identity in a clear-cut expression.

Those who support memristor as the fourth basic passive component say that such a component is characterized by a specific relation between two of the following: voltage, current, charge and/or magnetic flux, without any direct relationship with time. It is unimportant whether the parameters linking the two are a constant or special function. Memristor couples charge and magnetic flux and thus satisfies the criterion of a basic element.

TRUE MEMRISTOR?

Not all designers are convinced that this device reported by HP is a true full-fledged memristor. Some have said that it is just one version of Resistive Random Access Memory (RRAM). RRAM switches at a certain threshold voltage in contrast with memristor which recognizes no threshold voltage. The memresistance is a function of charge and slowly changes its stage irrespective of the magnitude of the applied current, and attains its extreme conductivity both at the high and low end when the current is present for a sufficient period. This issue has not been sufficiently reported in the HP paper claiming the first memristor.

I think whether memristor is a fundamental passive component or not is less important than the fact that a new device is on the verge of entering designs and providing designers with the kind of felicity they have been long waiting for. Since Leon Chua propounded the memristor theory in 1971, there has been much debate about the feasibility of such a product. The HP device, even if not a full-fledged memristor, meets designer needs. As more companies join the memristor bandwagon, a better evaluation and evolution of the product will be possible and controversies will end.