Editorial Feature

Memristors - Nanoelectronic Components

Memristors can be defined as a fourth class of electrical component, in addition to the capacitor, inductor and the resistor, which exhibit distinct characteristics at the nanoscale. Memristors, or memory resistors, are a kind of passive circuit element that enables an association between the time integrals of voltage and current across a two terminal element. The time-dependent  component, which is unique in passive components, allows the history of the voltages applied to the device to be accessed via a series of minute read charges.

Until the first prototype was developed by HP Labs under Stanley William, memristance as a characteristic of a real material was virtually unknown. At length scales longer than the nanoscale, the memristance effect is overridden by other field and electronic effects - hence all the prototype memristors developed to date have used nanomaterials.

Presently HP, IBM, Samsung, HRL and several other research labs are investigating memristors made of titanium dioxide or niobium dioxide, and there are few other memristor materials still in the early stages of development.

Memristors will enable the development of artificial neural networks which will be able to model the activity of brain cells much more accurately than existing transistor-based processors.

Memristors will enable the development of artificial neural networks which will be able to model the activity of brain cells much more accurately than existing transistor-based processors. Image Credits: Photos.com

Spiking Activity Enabled by Memristors

Computing hardware comprises a number of binary switches that can be either on or off. The human brain, on the other hand, is not binary - each neuron exhibits brief spikes of activity and encodes data based on the timing and pattern of these spikes.

Because of this fundamental difference, modeling of neurons using computer hardware has to date been very difficult and limited. In the last few years, however, researchers have been working on a way of fabricating a chip that behaves more like a neuron, using a circuit made of capacitors and memristors. The signal spikes the chip generates tend to be more regular when compared to neurons, but future versions may be able to make these more variable.

The key to fabricating these neuron-like devices is known as a Mott insulator. Mott insulators have an unusual electrical property - their structure should allow them to conduct electricity according to conventional band theory, but when measured they are found to be insulators at normal temperatures due to electron-electron interactions. Upon heating, these electronic interactions are overcome at a critical transition temperature, allowing them to conduct normally.

Niobium dioxide (NbO2) is one of the few materials which behaves like a Mott insulator - in the work done by HP Labs in 2012, the heat required to affect the transition into the conductive state was generated from the resistance of the material itself when a voltage is applied. When the voltage is removed, the subsequent cooling returns the material to its resistive state.

To obtain neuron-like spiking behavior, a simplified neuron model based on proteins that allowed transmission of electrical signals was considered. The firing of a neuron causes sodium channels to open, allowing ions to rush into a nerve cell and modifying the relative charges inside and outside the membrane. In response to these changes, potassium channels open and allow ions to move out and restore the charge balance. The charge flow then stops, and various pumps begin restoring the former ion balance.

The circuit developed by the HP researchers included two units, one representing sodium channels and the other potassium channels. Each unit included a capacitor parallel to a memristor. When arranged correctly, spikes of activity are produced on exceeding a given voltage threshold. This device was termed a neuristor.

The NbO2 neuristor is an effective proof-of-concept, but it has a number of limitations, including high power consumption and incompatibility with existing chip manufacturing techniques. HP Labs and other research groups are confident that it will be possible to find a material and architecture that solves these problems, allowing larger-scale research and early applications of memristors and neuristors in computing.

Types of Memristors Discovered

The other types of memristors discovered include the following:

  • Titanium Oxide Memristor: This device developed by HP in 2008 includes a thin titanium dioxide film sandwiched between two thick electrodes, one Pt and the other Ti. The titanium dioxide film has two layers one with a slight depletion of oxygen atoms. These oxygen vacancies behave as charge carriers implying the depleted layer has a lower resistance than the non-depleted layer. The application of an electrical field results in drifting of oxygen vacancies, changing the boundary between low resistance and high resistance layers. The film resistance as a whole depends on how much charge has been transmitted in a particular direction which can be reversed when the direction of current is changed. The HP device is considered a nanoionic device because it enables rapid ion conduction at the nanoscale.

  • Polymeric Memristor: In 2012, researchers attempted to create neural synaptic memory circuits with organic ion-based memristors. The article titled, “Modelling Neural Plasticity with Memristors,” was published in the IEEE Canadian Review. The synapse circuit showed long-term potential for learning and inactivity-based forgetting. A light pattern was stored and recalled later using a circuit grid.

  • Ferroelectric Memristor: The ferroelectric memristor is based on a thin ferroelectric barrier placed between two metallic electrodes. The ferroelectric material polarization is switched by applying a negative or positive voltage across the junction leading to large resistance variations over two orders of magnitude. The polarization switching is not abrupt but gradual. The ferroelectric memristor has two key benefits: firstly It is possible to tune ferroelectric domain dynamics so that memristor response can be engineered and secondly the variations in resistance are due to electronic phenomena rendering the device more reliable.

Applications of Memristors

The main application areas of memristors are neural networks (computers which model the behaviour of the human brain). The shift in fundamental computing architechture will have benefits for tasks such as signal processing and CMOS logic computation, and may also enable the development of brain-computer interfaces.

It will be possible to integrate Williams' solid-state memristors into devices known as crossbar latches, that could substitute transistors in future computers, occupying a much smaller area.

They can also be made into solid-state, non-volatile memory that would enable higher data density when compared to hard drives with access times similar to DRAM replacing both components.

In 2011, researchers from North Carolina State University announced the development of a biocompatible hydrogel-based memristor array which was designed to be used as part of a brain-computer interface.


At a conference in 2012, Stan Williams from HP stated that commercial memristor hardware will be available at the earliest by the end of 2014. The company has partnered with flash memory manufacturer Hynix to produce the devices.

Once memristor-based processors are commercially available, it will enable a whole new field of computing to open up. The brain-like computations which these devices will enable will transform our ability to model brains and other biological systems, and provide novel routes to solving complex non-linear computing problems such as stock market predictions, image processing, and language recognition in a more efficient manner.

Sources and Further Reading

Will Soutter

Written by

Will Soutter

Will has a B.Sc. in Chemistry from the University of Durham, and a M.Sc. in Green Chemistry from the University of York. Naturally, Will is our resident Chemistry expert but, a love of science and the internet makes Will the all-rounder of the team. In his spare time Will likes to play the drums, cook and brew cider.


Please use one of the following formats to cite this article in your essay, paper or report:

  • APA

    Soutter, Will. (2019, May 15). Memristors - Nanoelectronic Components. AZoNano. Retrieved on May 26, 2024 from https://www.azonano.com/article.aspx?ArticleID=3215.

  • MLA

    Soutter, Will. "Memristors - Nanoelectronic Components". AZoNano. 26 May 2024. <https://www.azonano.com/article.aspx?ArticleID=3215>.

  • Chicago

    Soutter, Will. "Memristors - Nanoelectronic Components". AZoNano. https://www.azonano.com/article.aspx?ArticleID=3215. (accessed May 26, 2024).

  • Harvard

    Soutter, Will. 2019. Memristors - Nanoelectronic Components. AZoNano, viewed 26 May 2024, https://www.azonano.com/article.aspx?ArticleID=3215.

Tell Us What You Think

Do you have a review, update or anything you would like to add to this article?

Leave your feedback
Your comment type

While we only use edited and approved content for Azthena answers, it may on occasions provide incorrect responses. Please confirm any data provided with the related suppliers or authors. We do not provide medical advice, if you search for medical information you must always consult a medical professional before acting on any information provided.

Your questions, but not your email details will be shared with OpenAI and retained for 30 days in accordance with their privacy principles.

Please do not ask questions that use sensitive or confidential information.

Read the full Terms & Conditions.