Access
This article is part of Nature's premium content.
Published online 30 April 2008 | Nature | doi:10.1038/news.2008.789
News
Found: the missing circuit element
Described in 1971, made in 2008: 'memristors' promise a computer revolution.
High-school physics students grappling with the delights of capacitors, inductors and resistors will be groaning into their exercise books. Electronics experts in California have finally succeeded in proving the existence of a fourth fundamental unit of electronic circuits: the 'memristor'.
To read this story in full you will need to login or make a payment (see right).
Comments
Reader comments are usually moderated after posting. If you find something offensive or inappropriate, you can speed this process by clicking 'Report this comment' (or, if that doesn't work for you, email redesign@nature.com). For more controversial topics, we reserve the right to moderate before comments are published.
Prof. Chua had been a very important figure in the field of electronics. Although much honored by organizations like the IEEE, he has yet to receive the "basic honor" in the modern world -- his own Wikipedia entry. Hopefully, this news entry would spur someone to write it.
This is likely the next big thing in small as far as infotech is concerned. It not only will have ramifications in computer/informations science but will likely contribute to neural networks. With coupling to hybridization of DNA multiplexed arrays it could lead to machines that mimic 'thinking'. Then again as Dr. Chua says of his own work, it just may be too weird or heretical. It is also possible DNA will not even be needed since it could lead to less robustness of the system's dynamics. Kudos Prof. Chua and Stan Williams' team!
Nature magazine always let me a bit disappointed when it comes to electronics. You can't go on saying that a fourth element "was lacking", and that the resistor, capacitor and inductor are not a fundamental basis for electronic circuits. You must know very little about electronics to dare saying something like that. The memristor is certainly interesting, but it must be presented in the adequate context. It is extremely important to stress that we are talking about a phenomenon that can only be appreciated in NON-LINEAR circuits. When you talk about fundamental laws of basic electronics, one immediately supposes we are talking LINEAR circuits... At fist, the overall explanations of how this element behaves seemed to me just like some kind of transimpedance. I also imagined immediately that there should exist an active RLC circuit to implement it, and I was just left thinking "What is actually new here? Why I am not able to appreciate this discovery, am I dumb?" If it is a PASSIVE device that behaves like something that you can only get with and ACTIVE RLC circuit, that is the news. This is a crucial information. This, and the fact that it is not a linear system. Or else you make all electrical engineers seem like a bunch of losers, spending years working with a crippled set of theories. Can you imagine dozens of students of basic electronics answering to their teachers right now, saying "no, you are wrong, you moron, there is also a fourth element, I read today at Nature..."? More than all of this, it should be stressed that this new device does something that you CAN implement with other technologies, but requiring something larger, or that consumes much more energy. We must compare it to other ways to get the same thing, and not talk like something similar to a new kind of quark or state of matter was found. One thing I doubt much is that this new device will allow us to compute things that we can't compute right now. Computability is a very serious issue, that must not be taken lightly. We can't let people have the impression that someday we might find a new electronics element that will permit us to build an ANN that will overcome Gödel's incompleteness theorem, for example, or solve the entscheidungsproblem. Perhaps compute some things faster and with less energy. For example, it can certainly simulate itself better than any model can. But that talk at Nature podcast about having a standard computer at one side, and a revolutionary memristor neural computer doing the magic that brains do and that computers would not be able to do, really pissed me off. There is a huge number of researchers today not only working on different circuits to try to do whatever the brain does, but trying to do it in the "traditional" computer. To say that this technology will be a revolution in computing and in artificial neural network building is a very heavy claim. So now, not only electrical engineers spent decades without a "fundamental" element, but all artificial neural network researchers were also losing their times in flawed technologies? I don't think so. I am certainly curious to see what will come from this discovery, but let's be a bit more scientific please. H Tse: I strongly suggest you to create an account at Wikipedia. We need interested contributors to make sure that no article gets left behind.
I agree with Nicolau Werneck's comment. I don't think memristor's role is like resistor, capacitor and inductor simply because either R, L or C has a specific way of linking voltage, current and/or time, no matter whether it is a function or a constant. (Actually in theories C is a constant, just in practical measurements it could be a function of V. And what the function looks like has nothing to do with the identity of capacitor.) But in the paper the authors mentioned that if memristance is a constant it's identical to resistance. This somehow breaks the existing symmetry of "fundamental elements", if the memristor, which links I & V in the same way as R and its feature can only be expressed through a function of q, is regarded as the 4th cousin.
I agree with Nicolau Werneck's comment. I don't think memristor's role is like resistor, capacitor and inductor simply because either R, L or C has a specific way of linking voltage, current and/or time, no matter whether it is a function or a constant. (Actually in theories C is a constant, just in practical measurements it could be a function of V. And what the function looks like has nothing to do with the identity of capacitor.) But in the paper the authors mentioned that if memristance is a constant it's identical to resistance. This somehow breaks the existing symmetry of "fundamental elements", if the memristor, which links I & V in the same way as R and its feature can only be expressed through a function of q, is regarded as the 4th cousin.
This reminds me of schroedinger's cat. Are we talking about a basic circuit element or something that is a combination of elements that are predicated on quantum physics rather than macro physics? Without knowing more detail, such as when you assess the value of the memrister, how much have you changed the value by that very same measurement.
The problem here is that people are not mentioning that the device is non-linear. The linear, time-invariant electronics taught in high school and universities is not very different from Euclidean geometry in terms of being complete and well-known. No high-school books about electronics will ever be rewritten by the discovery of a non-linear electronic device (unless it's a kind of transistor, not a "basic", "fundamental" passive component). The completeness of the electronic holy (and linear) trinity: resistor, capacitor and inductor, can be easily seen considering that they are just special cases of the more general concept of inductance. I sure admire Chua, he is one great hero and explorer of the harsh and fascinating lands of non-linear circuits (I find it *very* strange that nobody mentioned he is the creator of the famous, important and chaotic Chua's circuit), the HP folks are great too, and I do think that basic electronics books should dedicate some space to talk about complicated non-linear things such as the memristor. But there are some exaggerated claims being made here. For example, to say that the voltage is changing the "actual resistance" is certainly questionable, and makes the understanding more difficult. For all I know, the resistance that is changing is the one in a linear small-signal model of the circuit, not unlike a diode with a DC level applied to it. And a "resistor with memory" is a very nice description of a capacitor with a resistor in series, or in parallel, so that description is very far from displaying the true memristiveness of the entity they studied, which I am sure is something very interesting, that this article here did not explain at all.
Obviously, in the previous comment I meant "impedance", not "inductance". The holy trinity: resistors, inductors and capacitors, are three and one at the same time. They are all special cases of the more general and complex-valued impedance.
After reading this article, I was really surprised what can be published under such a brand as Nature. I had to check whether this was not some leftover from the 1 April issue. From the scientific point of view, there is no need for further comments after nice remarks by Nicolau Werneck. On the other hand, the controversial tone of the article actually opened my eyes and I would like to forward the message to all my colleagues working in the field of ferroelectric materials and ferroelectric random access memory (FRAM) technology: Adopting the logic of the article, the great guys working hard in the companies successfully commercializing FRAM memory chips for more than twenty years are actually developing the fifth fundamental circuit element – the memcitor. Clearly, a memcitor is a capacitor with memory. Who is going to rewrite the high-school textbooks first?