Memristor
A memristor, short for memory resistor, is a two-terminal passive electrical component recognized as the fourth fundamental circuit element, alongside resistors, capacitors, and inductors. Memristors uniquely combine properties of resistors and nonvolatile memory, allowing them to retain the last charge that flowed through them even when power is turned off. This ability to remember information without a continuous electrical flow opens up possibilities for revolutionary advancements in computing technology, potentially eliminating the need for traditional storage solutions like hard drives.
The concept of memristors was introduced by physicist Leon Chua in 1971, leading to years of research that culminated in the successful manufacturing of functional memristors in the mid-2000s. Their operation can be compared to a flexible pipe that adjusts its diameter based on the direction of electrical flow, enabling dynamic resistance that persists when the current stops. Moreover, memristors have shown promise in mimicking neuronal functions, which could pave the way for developments in artificial intelligence. Despite their potential, practical application is hindered by challenges in manufacturing and high costs, as researchers continue to explore suitable materials like titanium dioxide. The future of memristors may hold significant implications for the evolution of computing and AI technology.
On this Page
Subject Terms
Memristor
A memristor, or memory resistor, is a nonlinear, two-terminal passive electrical component thought to be the fourth fundamental electrical circuit component alongside resistors, capacitors, and inductors. Memristors are similar to resistors in that they can create and maintain a safe flow of electrical current across a device. Unlike resistors, however, a memristor can remember the last charge that flowed through it. Because they are also nonvolatile, which means they can remember charges when there is no active current or voltage, memristors are able to store information when the device they are a part of is turned off. Many experts believe that memristors have the potential to revolutionize the science of computing, but manufacturing such advanced circuit components has proven difficult. Although tech giant Hewlett-Packard (HP) succeeded in building a working memristor in 2008, high manufacturing costs and other challenges have slowed the development of memristors that can be used for practical purposes.
Background
The idea of a memristor was formally introduced in 1971 by Leon Chua, a physicist at the University of California, Berkeley. Chua's breakthrough paper on the subject was the culmination of more than fifty years of research and observation conducted by various scientists who tried to determine whether the concept of memristance was a real phenomenon. Through his work, Chua laid the groundwork for a definitive confirmation of that hypothesis.
Chua originally envisioned the memristor as a nonlinear, passive two-terminal electrical component that connected electrical charge and magnetic flux. In the years that followed, this definition was expanded to include any type of nonvolatile memory based on resistance switching, which is a physical phenomenon that occurs when a type of insulator called a dielectric suddenly changes its resistance in the presence of an electrical field or current.
In initially describing the memristor, Chua proposed that it should be considered the fourth fundamental circuit element. The existing fundamental circuit elements include a trio of electrical components called resistors, capacitors, and inductors. A resistor opposes the flow of the current and is used to control a circuit. A capacitor is used for storing electrical energy. An inductor stores energy in the form of a magnetic field. The memristor is most similar to the resistor, but it has properties that none of the other components can duplicate individually or together in any combination. The most important of these properties is the ability to store information even without a continuous flow of electricity. Like a transistor, the memristor relies on a flow of electrons to work. The difference is that while resistors rely on this electron flow alone, memristors couple the electrons themselves with electrically charged atoms called ions. These ions enable memristors to store and retain information even when the flow of electrons stops.
After Chua published his paper, scientists spent many years trying to build a working memristor. This effort finally began to pay off when a controlled memristor was successfully manufactured in 2005. A more important breakthrough came when HP finally built the first fully functional memristor three years later.
Overview
In a sense, a memristor works like a water pipe that expands and contracts in diameter depending on the direction in which water flows through it. When water is flowing in one direction, the pipe expands to allow the water to move faster. If the direction of flow changes, the pipe contracts and forces the water to move slower. When the water is turned off, the pipe retains the same diameter until it is turned back on again. Like this hypothetical pipe, a memristor increases the flow of electrical current in one direction, decreases it in the other, and maintains its resistance value when the flow of electricity is stopped.
The memristor's ability to retain information even when the flow of electricity is interrupted is the most important factor in its potential technological value. This unique property would theoretically make it possible to build a computer that could be turned on and off like a lightbulb without any loss of data. In this way, the memory systems made possible by memristors could entirely eliminate the need for traditional hard drives. Moreover, memristors could even end the so-called "silicon era" of computing. Given that only a limited number of transistors can be installed on a computer chip and given the fact that such chips have already been miniaturized to nearly the maximum possible extent, a time will undoubtedly come when it becomes necessary to migrate away from silicon-based computing. The memristor could provide a means for making this important transition.
In addition to possibly becoming the successor to silicon-based computing, memristors may play an important role in the further development of artificial intelligence. This is because scientists have found that memristors may be able to mimic the function of neurons in the human brain. Neurons are electrically excitable cells that receive, process, and transmit information through electrical and chemical signals. While a neuron's activity often is determined by the signals it receives in real-time, it has a sort of biochemical short-term memory that allows it to be easily reactivated if it has already been activated in the recent past. Researchers have found that certain types of memristors can effectively do the same thing when metal is allowed to diffuse within the device itself. This type of memristor technology could potentially be used to create machines that can think and learn the way humans do.
The challenge in unlocking the memristor's technological potential lies in producing a functional memristor that can serve a multitude of practical purposes. In reality, this challenge is much more difficult than it may appear. Perhaps the biggest obstacle for scientists to overcome is finding best material to use in the manufacture of memristors. Laboratory researchers working for tech industry giants such as IBM and HP have identified titanium dioxide as one of the best possible materials to use in making memristors, but others are being studied as well. Another major obstacle in the way of memristor implementation is the high costs involved in their manufacture.
Bibliography
Beckett, Jamie. "Demystifying the Memristor: Proof of Fourth Basic Circuit Element Could Transform Computing." HP, Apr. 2008, www.hpl.hp.com/news/2008/apr-jun/memristor.html. Accessed 16 Oct. 2017.
Greenemeier, Larry. "HPE Debuts Its Next-Gen Computer—sans Much-Anticipated Memristors." Scientific American, 16 May 2017, www.scientificamerican.com/article/hpe-debuts-its-next-gen-computer-sans-much-anticipated-memristors/. Accessed 16 Oct. 2017.
Mellor, Chris. "Never-Never Chip Tech Memristor Shuffles Closer to Death Row." The Register, 28 Jun. 2016, www.theregister.co.uk/2016/06/28/memristor‗moves‗closer‗to‗death‗row/. Accessed 16 Oct. 2017.
"Memory Resistor (Memristor)." Techopedia,www.techopedia.com/definition/1744/memory-resistor-memristor. Accessed 16 Oct. 2017.
Mills, Allison. "Memristors: Making a New Generation for Digital Memory and Computation." Michigan Tech, 2 Feb. 2016, www.mtu.edu/news/stories/2016/february/memristors-making-new-generation-for-digital-memory-computation.html. Accessed 16 Oct. 2017.
Prisco, Jacopo. "So Long, Transistor: How the 'Memristor' Could Revolutionize Electronics." CNN, 2 Mar. 2015, www.cnn.com/2015/02/26/tech/mci-eth-memristor/index.html. Accessed 16 Oct. 2017.
Rouse, Margaret. "Memristor." TechTarget, whatis.techtarget.com/definition/memristor. Accessed 16 Oct. 2017.
Timmer, John. "Memristor That Forgets Makes a Good Model Neuron." ArsTechnica, 28 Sep. 2016, arstechnica.com/science/2016/09/memristor-that-forgets-makes-a-good-model-neuron/. Accessed 16 Oct. 2017.