Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
He helped invent a a whole new category of tech called "software-as-a-service," where companies rent their software over the Internet, instead of install it on computers.
The concept almost died at birth. Shortly after Salesforce.com launched, the Internet bubble burst.
Benioff persisted, however, and today every major software company is chasing him.
Along the way, Benioff has become one of the most flamboyant, generous, larger-than-life personalities in the tech industry.
Me. Hmmm “software as a service” Remember our first p r about that concept about 12 ? years ago.? And Marc invented it. Looks like we’ve been following him ( the leader) for quite some time now. I’d be willing to bet he’s chomping at the bit to buy us. …. We wait.
Doc. I agree. Remember Wade was in a joint venture with Salesforce and it fell apart like all the other jv’s he was involved with way back when. . Who could do business with Wade? No one. Marc B lives here on the Big Island. He is an all around good guy. Buying fire trucks and even helicopters for our fire departments and donating a lot of land for the homeless and low income housing etc. I’m hoping Len and his men are talking to him about a new licensing agreement. Or maybe working on a deal for him to buy v ? It would be a good fit. And he’s a good guy.
I have always thought that adobe was infringing. jfyi
Jedi. I wonder why Dan deleted all of my “A LAUGHABLE PLACE” posts? I just started a new one.
Weird? Or? Did I touch a nerve? We shall see.
No time limit on Texas Supreme Court to decide if they will hear his appeal. He lost in district court and lost in Dallas court of appeals I think they can see right through this clown. If not. I loose all faith in our court system. THUMP! I’m thinking no more than a month? But only they know. Probably depends a lot on their case load.
Yep Doc. We’re covered til 2035 ! Relax. Go to the beach. Get a shave ice. Visit Akaka Falls and cool off in the rainforest.
On January 20, 2014, Vertical Computer Systems, Inc.. (the “Company”) filed a Current Report on Form 8-K (the “Original 8-K”) to, among other things, report events pertaining to its patents. The Company is filing this Current Report on Form 8-K/A to amend the Original 8-K to correct certain references concerning U.S. Patent No. 6,826,744 and certain continuation patents.
ITEM 8.01 OTHER EVENTS.
On January 14, 2015, Vertical Computer Systems, Inc., (the “Company”) received notice from the United States Patent and Trademark Office (the “USPTO”) that a continuation patent would be issued by the USPTO on February 3, 2015. The continuation patent will be issued as United States (“U.S.”) Patent No. 8,949,780 and is a continuation of U.S. Patent No. 7,716,629, which is a continuation patent of U.S. Patent No. 6,826,744.Collectively, they form the underlying patents (the “Patents”) under the Company’s SiteFlash™ and SiteFlash-derived products
In addition, the Company has retained the law firm of Davidoff Hutcher & Citron LLP on a contingency basis to act in connection with the licensing of the Patents,
Collectively, these Patents are a system and method for generating computer applications in an arbitrary object framework. The method separates content, form, and function of the computer application so that each may be accessed or modified separately. The method includes creating arbitrary objects, managing the arbitrary objects throughout their life cycle in an object library, and deploying the arbitrary objects in a design framework for use in complex computer applications.
Wade replied with an p d f of a letter from court with his wrong address on it. But ….. he or anyone and everyone is responsible for notifying the court of your correct current address. Imo. He’s toast. These Supreme Court justices will take 1 look at the judgements against him upheld by a majority of the Dallas court of appeals and say. Bam ! NEXT!
In the study, the team programmed a Bayesian neural network on an assembly of 75 crossbar arrays, each containing 1,024 memristors, complemented by CMOS periphery for in-memory computing. The experiment involved a two-layer Bayesian neural network trained to differentiate nine classes of heart arrhythmia from electrocardiogram (ECG) recordings.
The experimental results were compelling, with the memristor-based Bayesian neural network nearly matching the best performances of a software neural network in terms of both aleatoric (area under the curve of 0.91) and epistemic uncertainty evaluation (area under the curve of 0.99), albeit with a slight reduction in raw accuracy. This performance sharply contrasted conventional neural networks, which exhibit little ability to recognize unknown data and evaluate uncertainties.
That’s amazing stuff! Now add these new memristors to quantum computers , LLM’s and AI. Bam!
A stochastic process can have many outcomes, due to its randomness, and a single outcome of a stochastic process is called, among other names, a sample function or realization.[28
Doc. Yes. And my key takeaway from this article…
This performance sharply contrasted conventional neural networks, which exhibit little ability to recognize unknown data and evaluate uncertainties.
Hmmm The ability to recognize unknown data and. …. Wait for it… “evaluate uncertainties “. Or Arbitrary objects!
Like I said before. The Tec is finally catching up to our patent..
Chinese scientists unlock potential of memristor semiconductor building block that could boost artificial intelligence, self-driving cars and more
World’s first fully system-integrated memristor chip could make artificial intelligence smarter and up to 75 times more efficient, researchers say
Advances could lead to AI that is capable of more human-like learning, with implications for how smart devices and autonomous driving work
Topic | Science
Follow this topic and get notified the next time we publish content about Science.
Victoria Bela
Victoria Bela
Published: 10:07pm, 11 Oct, 2023
The world’s first fully system-integrated memristor chip has been unveiled by a team of Chinese scientists who believe it could not only make artificial intelligence smarter, but also more time and energy efficient. While the semiconductor has yet to leave the lab setting, it could allow for the development of AI that is capable of more human-like learning, which could have implications for the way smart devices and autonomous driving work, according to the researchers. “Learning is highly important,” for edge intelligence devices, the research team from Tsinghua University said in their study released in the journal Science on September 15, referencing devices that process data internally with technology like AI. The advancement is the latest in a series of Chinese semiconductor innovations announced since US-imposed export controls and sanctions restricted the supply of advanced chips and chip-making equipment to the nation.
Advertisement
Advertisement
The development of a new microchip announced last month by Huawei made a splash in the media, and called into question whether China had the expertise to advance semiconductors without US technology. China’s new memristor chip is a significant step forward in developing such technology, according to the researchers.
“Memristor-based computing technology has recently received considerable attention because of its potential to overcome the so-called ‘von Neumann bottleneck’ of conventional computing architecture,” Yury Suleymanov, an associate editor at Science, said in his editor’s summary of the paper, referring to the computational limitation set by the separation of memory and processing.
China At A Glance
By submitting, you consent to receiving marketing emails from SCMP. If you don't want these, tick here
SIGN UP
By registering, you agree to our T&C and Privacy Policy
A resistor is an element of a circuit that is able to limit the flow of energy by providing resistance to electrons flowing through.
China downplays reports of Taiwanese firms helping Huawei’s semiconductor ambitions
11 Oct 2023
A memristor – a contraction for memory resistor – is able to remember the most recent value of current that was passed through it when it was turned on, meaning future resistance can depend on prior history, according to Nature Electronics. As such, it is capable of improvement-based learning, or maintaining pre-acquired knowledge when something new is learned.
This differs from transfer learning, which focuses on moving to a new set of data and can sacrifice the accuracy of prior data, according to the paper. To train artificial neural networks, which mimic how human neurons pass on data in the brain, conventional hardware requires a great deal of energy and time to move data between the computing and memory units.
The study is an important step towards future chips with high energy efficiency and extensive learning capabilities
Wu Huaqiang and Gao Bin
Memristor-based computing is able to reduce the energy required for a task by allowing the learning to occur on-chip with no external memory source.
Several studies have investigated memristors, but they still used additional external processors, the team said in their paper. The researchers produced a chip capable of conducting complete on-chip improvement learning, and proposed a learning architecture for it.
They said they were able to demonstrate on-chip learning with multiple tasks, including image classification, motion control, and audio recognition.
How artificial intelligence will redefine the future of work
20 Jun 2022
In one demonstration, the motion control of a model car designed to pursue a laser light was better at finding the light in a dark environment and would lose track of it in a brighter environment. But after implementing learning, it was able to find the light equally well in both environments.
The accuracy in the bright environment greatly increased, while accuracy in the dark environment did not decrease with the addition of the new ability. “The memristor-based neuro-inspired computing chip could facilitate the development of edge AI devices that could adapt to new scenes and users,” the researchers said in their paper.
With further research on the learning architecture, the team said it believed it could enable on-chip learning that is 75 times more energy efficient than current machines designed for AI processing.
Still, challenges remain in the research and development of such chips – in particular, engineering the chips for large-scale integration. It could take time for the technology to make it out of the laboratory, Wu Huaqiang and Gao Bin, professors at Tsinghua and leads of the research team, admitted to Chinese-language news site Science Times. “The study is an important step towards future chips with high energy efficiency and extensive learning capabilities,” the researchers said, adding they hoped their findings “will accelerate the development of future smart edge devices”
International Researchers Move the Needle on Memristor Technology
December 19, 2023 by Jake Hertz
Academics across the globe have made concurrent breakthroughs in a new form of non-volatile memory.
Memristors, or memory resistors, are an emerging form of non-volatile memory considered a viable alternative to traditional memory storage. These devices act as switches that can recall their electric state even after their power is turned off, allowing them to behave like human neurons—both computing and storing data.
Memristors are a hot research topic for their numerous unique properties, including scalability to lower process nodes, which enable smaller, faster, and more energy-efficient devices.
Cea-Leti's wafer-level memristor crossbar arrays
Cea-Leti's wafer-level memristor crossbar arrays using a prober. Image (modified) used courtesy of Cea-Leti
Now, as the demand for more sophisticated and power-efficient computing grows, memristors hold promise in various applications, including artificial intelligence, neuromorphic computing, and high-density data storage. In recent months, several international research teams have published findings about memristors that may help push this technology from lab to market.
The University of Rochester: Strain-Engineered Phase-Change Memristors
Researchers at the University of Rochester recently developed high-performance, bipolar phase-change memristors based on strain-engineered multilayer molybdenum ditelluride (MoTe2).
The strain-based memristor
The strain-based memristor. Image used courtesy of Nature
These devices, distinguished by their bipolar phase-change characteristics, represent a significant leap from conventional filamentary conduction mechanisms. The study successfully utilizes process-induced strain engineering to induce a semimetallic-to-semiconducting phase transition in MoTe2.
This transition is key to achieving the memristor's remarkable switching characteristics, including low voltages of 90 mV, high on/off ratios of 10^8, swift switching times of 5 ns, and prolonged retentions over 10^5 seconds. Such performance metrics indicate the potential of these memristors for applications in robust, fast, low-power, and scalable memory and neuromorphic computing systems.
The device's performance can also be finely tuned by varying a single process parameter: the contact metal film force, a product of film stress and thickness. This control over the device's switching voltage and on/off ratio underlines the memristor's adaptability and potential for diverse applications.
Cea-Leti: Using Memristors for Bayesian Neural Networks
In France, a group of Cea-Leti researchers announced that they have successfully integrated memristor hardware into artificial intelligence systems. Specifically, the team leveraged the inherent probabilistic nature of memristors to effectively execute the computationally intensive tasks of Bayesian neural networks.
Filamentary memristor
Filamentary memristor in the back end of the line of Cea-Leti's hybrid memristor/CMOS process. Image used courtesy of Nature
In the study, the team programmed a Bayesian neural network on an assembly of 75 crossbar arrays, each containing 1,024 memristors, complemented by CMOS periphery for in-memory computing. The experiment involved a two-layer Bayesian neural network trained to differentiate nine classes of heart arrhythmia from electrocardiogram (ECG) recordings.
The experimental results were compelling, with the memristor-based Bayesian neural network nearly matching the best performances of a software neural network in terms of both aleatoric (area under the curve of 0.91) and epistemic uncertainty evaluation (area under the curve of 0.99), albeit with a slight reduction in raw accuracy. This performance sharply contrasted conventional neural networks, which exhibit little ability to recognize unknown data and evaluate uncertainties.
Tsinghua University: 2D-Material-Based Memristors
Finally, in China, a team at Tsinghua University developed a new system-integrated memristor computing-in-memory chip the researchers claim is the world's first of its kind to support efficient on-chip learning. This chip, still in the laboratory phase, is expected to significantly advance developments in fields such as artificial intelligence, autonomous driving, and wearable devices, among others
The memristor-based integrated circuit
The memristor-based integrated circuit. Image used courtesy of Tsinghua University
The chip is notably energy efficient, consuming only 3% of the energy of an application-specific integrated circuit (ASIC) system for on-chip learning. According to the team, this innovation started with research on memristors for storage applications and took about 11 years to develop to this point.
Memristors Gain International Steam
Memristors converge computing and memory within a single device, opening up new frontiers for energy-efficient, high-speed, and intelligent computing solutions. As international researchers continue to explore and refine memristor technology, they pave the way for more advanced and efficient computing systems that could define the next generation of electronics.
International Researchers Move the Needle on Memristor Technology
December 19, 2023 by Jake Hertz
Academics across the globe have made concurrent breakthroughs in a new form of non-volatile memory.
Memristors, or memory resistors, are an emerging form of non-volatile memory considered a viable alternative to traditional memory storage. These devices act as switches that can recall their electric state even after their power is turned off, allowing them to behave like human neurons—both computing and storing data.
Memristors are a hot research topic for their numerous unique properties, including scalability to lower process nodes, which enable smaller, faster, and more energy-efficient devices.
Cea-Leti's wafer-level memristor crossbar arrays
Cea-Leti's wafer-level memristor crossbar arrays using a prober. Image (modified) used courtesy of Cea-Leti
Now, as the demand for more sophisticated and power-efficient computing grows, memristors hold promise in various applications, including artificial intelligence, neuromorphic computing, and high-density data storage. In recent months, several international research teams have published findings about memristors that may help push this technology from lab to market.
The University of Rochester: Strain-Engineered Phase-Change Memristors
Researchers at the University of Rochester recently developed high-performance, bipolar phase-change memristors based on strain-engineered multilayer molybdenum ditelluride (MoTe2).
The strain-based memristor
The strain-based memristor. Image used courtesy of Nature
These devices, distinguished by their bipolar phase-change characteristics, represent a significant leap from conventional filamentary conduction mechanisms. The study successfully utilizes process-induced strain engineering to induce a semimetallic-to-semiconducting phase transition in MoTe2.
This transition is key to achieving the memristor's remarkable switching characteristics, including low voltages of 90 mV, high on/off ratios of 10^8, swift switching times of 5 ns, and prolonged retentions over 10^5 seconds. Such performance metrics indicate the potential of these memristors for applications in robust, fast, low-power, and scalable memory and neuromorphic computing systems.
The device's performance can also be finely tuned by varying a single process parameter: the contact metal film force, a product of film stress and thickness. This control over the device's switching voltage and on/off ratio underlines the memristor's adaptability and potential for diverse applications.
Cea-Leti: Using Memristors for Bayesian Neural Networks
In France, a group of Cea-Leti researchers announced that they have successfully integrated memristor hardware into artificial intelligence systems. Specifically, the team leveraged the inherent probabilistic nature of memristors to effectively execute the computationally intensive tasks of Bayesian neural networks.
Filamentary memristor
Filamentary memristor in the back end of the line of Cea-Leti's hybrid memristor/CMOS process. Image used courtesy of Nature
In the study, the team programmed a Bayesian neural network on an assembly of 75 crossbar arrays, each containing 1,024 memristors, complemented by CMOS periphery for in-memory computing. The experiment involved a two-layer Bayesian neural network trained to differentiate nine classes of heart arrhythmia from electrocardiogram (ECG) recordings.
The experimental results were compelling, with the memristor-based Bayesian neural network nearly matching the best performances of a software neural network in terms of both aleatoric (area under the curve of 0.91) and epistemic uncertainty evaluation (area under the curve of 0.99), albeit with a slight reduction in raw accuracy. This performance sharply contrasted conventional neural networks, which exhibit little ability to recognize unknown data and evaluate uncertainties.
Tsinghua University: 2D-Material-Based Memristors
Finally, in China, a team at Tsinghua University developed a new system-integrated memristor computing-in-memory chip the researchers claim is the world's first of its kind to support efficient on-chip learning. This chip, still in the laboratory phase, is expected to significantly advance developments in fields such as artificial intelligence, autonomous driving, and wearable devices, among others
The memristor-based integrated circuit
The memristor-based integrated circuit. Image used courtesy of Tsinghua University
The chip is notably energy efficient, consuming only 3% of the energy of an application-specific integrated circuit (ASIC) system for on-chip learning. According to the team, this innovation started with research on memristors for storage applications and took about 11 years to develop to this point.
Memristors Gain International Steam
Memristors converge computing and memory within a single device, opening up new frontiers for energy-efficient, high-speed, and intelligent computing solutions. As international researchers continue to explore and refine memristor technology, they pave the way for more advanced and efficient computing systems that could define the next generation of electronics.
This part stuck out.
“The scientists rewrote Bayesian equations so a memristor array could perform statistical analyses that harnesses randomness—a.k.a. “stochastic computing.”
“That harness randomness”. ?
Or arbitrary objects?
744/629 is very,very valuable. Wouldn’t you agree Doc? All the pieces seem to be finally catching up to the patent. Now, the more breakthroughs with memristors, the more valuable 744/629 will be. Imo
Memristors Run AI Tasks at 1/800th Power
These brain-mimicking devices boast tiny energy budgets and hardened circuits
CHARLES Q. CHOI
04 JAN 20234 MIN READ
brain shaped computer wiring against a blue glowing background
ISTOCK
Memristive devices that mimic neuron-connecting synapses could serve as the hardware for neural networks that copy the way the brain learns. Now two new studies may help solve key problems these components face not just with yields and reliability, but with finding applications beyond neural nets.
Memristors, or memory resistors, are essentially switches that can remember which electric state they were toggled to after their power is turned off. Scientists worldwide aim to use memristors and similar components to build electronics that, like neurons, can both compute and store data. These memristive devices may greatly reduce the energy and time lost in conventional microchips shuttling data back and forth between processors and memory. Such brain-inspired neuromorphic hardware may also prove ideal for implementing neural networks—AI systems increasingly finding use in applications such as analyzing medical scans and empowering autonomous vehicles.
However, current memristive devices typically rely on emerging technologies with low production yields and unreliable electronic performance. To help overcome these challenges, researchers in Israel and China fabricated memristive devices using a standard CMOS production line. The resulting silicon synapses the team built boasted a 100 percent yield with 170- to 350-fold greater energy efficiency than a high-performance Nvidia Tesla V100 graphics processing unit when it came to multiply-accumulate operations, the most basic operation in neural networks.
“Memristors are a highly promising lead to provide low-energy-consumption artificial intelligence.”
—Damien Querlioz, Université Paris-Saclay
The scientists developed the new devices using the floating-gate transistor technology used in commercial flash memory. Whereas conventional floating-gate transistors have three terminals, the new components have only two. This greatly simplified fabrication and operation and reduced their size. In addition, memristors have only binary inputs and outputs, eliminating the need for the large, energy-hungry, analog-to-digital and digital-to-analog converters often used in neuromorphic hardware, says study senior author Shahar Kvatinsky, an associate professor of electrical and computer engineering at the Technion-Israel Institute of Technology, in Haifa.
The new devices displayed high endurance, operating past more than 100,000 cycles of programming and erasing by using voltage pulses. In addition, they showed only moderate device-to-device variation and are projected to possess long data retention times of more than 10 years.
The researchers use an array of about 150 of these components to implement a kind of neural network that operated using only binary signals. In experiments, it could recognize handwritten digits with roughly 97 percent accuracy. Kvatinsky says that this work “is just a start—a proof of concept and not a whole integrated chip or a large neural network. Integration and scaling up is a major challenge.”
In another study, a team of French researchers investigated using memristors for the statistical computing technique known as Bayesian reasoning, in which prior knowledge helps compute the chances that an uncertain choice might be correct. Its results are fully explainable—unlike many nearly inscrutable AI computations—and it can perform well when there is little available data, as it can incorporate prior expert knowledge. However, “it is just not obvious how to compute Bayesian reasoning with memristors,” says study coauthor Damien Querlioz, a research scientist at CNRS, Université Paris-Saclay.
Implementing Bayesian reasoning using conventional electronics requires complex memory patterns, “which increase exponentially with the number of observations,” says neuromorphic scientist Melika Payvand at the Institute of Neuroinformatics in Zurich, who did not participate in either study. However, Querlioz and his colleagues “found a way of simplifying this,” she notes.
Memristor AI “excels in safety-critical situations, where high uncertainty is present, little data is available, and explainable decisions are required.”
—Damien Querlioz, Université Paris-Saclay
The scientists rewrote Bayesian equations so a memristor array could perform statistical analyses that harnesses randomness—a.k.a. “stochastic computing.” Using this approach, the array generated streams of semi-random bits at each tick of the clock. These bits were often zeroes but were sometimes ones. The proportion of zeroes to ones encoded the probabilities needed for the statistical calculations the array performed. This digital strategy uses relatively simple circuitry compared to nonstochastic methods, all of which reduces the system’s size and energy demands.
The researchers fabricated a prototype circuit incorporating 2,048 hafnium oxide memristors on top of 30,080 CMOS transistors on the same chip. In experiments, they had the new circuit recognize a person’s handwritten signature from signals beamed from a device worn on the wrist.
Bayesian reasoning is often thought of as computationally expensive with conventional electronics. The new circuit performed handwriting recognition using 1/800th to 1/5,000th the energy of a conventional computer processor, suggesting “that memristors are a highly promising lead to provide low-energy-consumption artificial intelligence,” Querlioz says.
The new device also can instantly flick on and off, suggesting it can work only when needed, to conserve power. Moreover, it is also resilient to errors from random events, making it useful in extreme environments, the researchers say. All in all, the new circuit “excels in safety-critical situations, where high uncertainty is present, little data is available, and explainable decisions are required,” Querlioz says. “Examples are medical sensors or circuits for monitoring the safety of industrial facilities.”
A future direction for the Bayesian circuits might include machines that collect multiple types of sensory data, Payvand says, which might include autonomous vehicles or drones. If the machine is not confident about predictions made based off one sense, it could boost its confidence by analyzing data from a different sense, she notes.
A key obstacle Bayesian systems face “is their scalability to larger problems or networks,” Payvand cautions. Querlioz notes that the team have designed a considerably scaled-up version of their device “that is currently being fabricated.” He notes their circuit is currently specialized for certain types of Bayesian computation, and they want to create more adaptable designs in the future.
To a certain degree, both studies use randomness—Querlioz and his colleagues use it for statistical analysis, while Kvatinsky and his collaborators have their neural networks sample data at random intervals to avoid the kinds of errors that can occur when sampling data a limited number of times.
“These approaches pair very well with the inherited randomness of memristor devices,” says Giacomo Pedretti, a senior AI research scientist at Hewlett Packard Labs, who did not take part in either study. It would be “very interesting” to try to use the inherent noise in these electronics to generate controlled randomness “rather than implementing costly digital pseudorandom number generators,” he says.
Both studies appeared 19 December in the journal Nature Electronics.
This article appears in the March 2023 print issue.
Computer Science | Physics
New memristor design could be a game-changer for AI and big data
by Andrey Feldman | Feb 22, 2024
Computers based on memristors promise significant energy savings and improved accuracy in large-scale computing.
Abstract image of a computer circuit.
Advanced technologies such as neural networks have found extensive application in image recognition, big data processing, financial analysis, and many other fields. However, training them demands significant computational resources and energy consumption, posing challenges for their widespread use and further development.
The problem comes down to bottlenecks imposed by traditional computing systems built using transistors, which keep their memory and processing units separate, requiring power-intensive and performance-degrading data transfers between them. Perhaps an even greater disadvantage is that their memory requires constant power to store information, which further increases energy consumption.
To solve this, researchers propose an alternative: the memristor. “Memristors, also known as memory resistors, are switches that can ‘remember’ their previous electric state, even after power is switched off,” said Desmond Loke, professor at the Singapore University of Technology and Design, in an email.
“Memristors can be used to create devices that store data because they have the ability to drastically reduce the time and energy required for data transmission between memory and processors in traditional microchips,” he continued. “They might be perfect for constructing neural networks, artificial intelligence systems for medical scan processing, and enabling driverless vehicles.”
In a study published in Advanced Physics Research, Loke and his colleagues investigated neural network training using new memristor design that stores information in a unique material made up of germanium, tellurium, and antimony.
This substance exists in an amorphous phase, but when exposed to an electric current and a change in temperature, it exhibits ordered crystalline regions. Depending on the properties of the pulse, such as its intensity and signal shape, the number and size of these regions changes, consequently affecting the material’s electrical and optical properties, which are preserved after the current is turned off. In this way, information can be recorded in the memristor and read from it by repeatedly applying an electric current.
In total, the authors could create fifteen, clearly distinguishable stages of crystallization in the memristor, each corresponding to specific information necessary for training the neural network. “We used [the] memristors to simulate a neural network,” Loke explained. “In trials, the neural network recognized handwritten numerals with an accuracy of over 96%.”
This accuracy is an improvement on other memristor designs by more than 5%, say the team, indicating a promising trajectory for this type of memristor computing technology. But there are still significant hurdles to overcome before this technology can be implemented into real-world computing systems.
“System integration and scale pose considerable challenges,” Loke explained. Nevertheless, the researchers remain hopeful, envisioning a future where memristors might play a pivotal role in training large neural networks. They anticipate smaller, more potent, and dramatically more energy-efficient units compared to conventional computers if the technology can be developed further.
“Currently, this work is just a proof of concept,” Loke concluded. “The next step is to develop a fully integrated circuit and a big neural network.”
Reference: Kian-Guan Lim et al, Toward Memristive Phase-Change Neural Network with High-Quality Ultra-Effective Highly-Self-Adjustable Online Learning, Advanced Physics Research (2024). DOI: 10.1002/apxr.202300085
Feature image credit: AcatXIo on Pixabay
…maybe SBV/V already have it?🤞
Naw. She still has 39 k on the side from scholarships she can use for travel expenses. She got 89,475 of the 94,475 full tuition from institutional grants! Only has to pay 5 k a yr.
Good question They have sensors on everything else these days. O t. Yale, Duke, Brown all accepted ! Dartmouth waited. Receiving. 89k of the 94 k cost of Yale X 4 years ! The best deal of the 3 🎉🎉🎉🎉😎🤙Go Bulldogs! KAATN’s
Curt_Hopkins
Memristor tech helps new device store ranges instead of bits and bytes
Curt_Hopkins ?10-05-2021 08:00 AM
layerssized.jpgA group of scientists mostly from Hewlett Packard Labs have published a paper in Nature Communications that outlines a “novel analog content addressable memory (CAM) based on emerging memristor devices for fast look-up table operations.”
Tree-based machine learning performed in-memory with memristive analog CAM was authored by Giacomo Pedretti, Catherine E. Graves, Sergey Serebryakov, Xia Sheng, and Martin Foltin, all of Labs; along with Can Li and Ruibin Mao of the University of Hong Kong; and John Paul Strachan of the Peter Grunberg Insitut.
The accelerator
According to the paper “(t)ree-based machine learning techniques, such as Decision Trees and Random Forests, are top performers in several domains…(h)owever, these models are difficult to optimize for fast inference at scale without accuracy loss in von Neumann architectures due to non-uniform memory access patterns.”
To solve for this problem, lead author Pedretti and the rest of the team “propose for the first time to use the analog CAM as an in-memory computational primitive to accelerate tree-based model inference.” They used an efficient mapping algorithm leveraging the new analog CAM capabilities such that each root-to-leaf path of a Decision Tree is programmed into a row, making this process much faster.
But what the team did exactly was to design a new cell, based on the memristor.
“Thanks to this new device, instead of storing digital information we can store a continuous value,” says Pedretti. “What we store now are ranges and we give an input which is made of continuous values.”
The reason why
“Right now, explainability is the focus of our team,” says Pedretti. “Think of it as a first step towards an ‘accelerator for explainable AI.’”
According to Pedretti, tree-based algorithms are more explainable than neural networks and deep learning. Those working in artificial intelligence have been developing huge, very complex models. But when people have to use the products of those models for industry or government, they often want to know not just what, but why. Why, for example, is a given image classified as a traffic light. Knowing that in situations where proof or law is in question, can be as important as the classification itself.
“Imagine a situation in the financial or insurance industry, where you need to explain why you are not giving money to somebody,” says Pedretti. “Or in a clinical context, you may be required to explain why you gave, or withheld, a drug to a given patient.”
Labs’ CAM-accelerated, tree-based model does not just create desired product, it does so while allowing users to understand how and a specific decision was reached.
An arrival
“The way we designed the stack, from materials to circuit and architecture, based as it is on memristor technology, has allowed us to create an efficient decision-making engine,” says Pedretti. “It has also allowed us to make it in such a way that it is explainable. But in addition to that, it is good.”
What Pedretti means is that they are able to process the data in a massively parallel way, so the accelerator does not have to wait for the first-step decision in the tree to be made before moving on to the rest of the branches. They infer the likely next step, looking at the tree altogether.
The movie “Arrival” and the story it was based on, “The Story of Your Life,” are science fiction. But Fermat’s Theory of Least Time, the mathematical lynchpin of the story, is not. Nor is the idea that you may bring into practice alternate ways of looking at time, life, and certainly computer architecture, ways that do not restrict you to doing one thing as a time.
The Natives are getting restless over at Vcsy-chat site. We’re all tired. Worn out really. Pissed off to put it mildly. But what can we do other than trust in our BoD at this point? Pretty much, nothing. I want to believe Len and his men have a plan. I think everyone needs to relax. Take a chill pill and go to the beach or forest or desert or whatever relaxes you. Not much longer and we’ll all hear some thing. .
And Yes. I think Milks should be sued for whatever we can sue him for and make him suffer the penalties. All in due time.
I read the article. It’s a good thing our Markman Hearing was over with years ago now. It seems the Chinese are pulling out all the stops. SBV/V should be careful. What’s next?
THUMP! ( sound of my forehead hitting my desk)
You and over 1,000 of us feel the same way. Somethings gotta give soon. We’ve all been waiting patiently for a long long time. Our new great communicator is no better than Wade . It looks like we all may have been had , again. I hope I’m wrong. Once the TSC rules in our favor and they don’t put out a ton of good news and we don’t hear anything about all these secret nda’s Johnny quan told us were coming , that will be the last straw imo
Why would we let all of our adversaries have a secure communications app that no one could break into ? We wouldn’t. But now that the cats out of the bag after 15 years we may get paid. Finally. We shall see.
Doc. Probably. I thought about that years ago. Ploinks would never work right for average Joe. Only the military. Makes sense. Only plausible explanation imo
A lot of talk on Vcsy chat site about reverse splits etc. No reverse split and no sale of Corp to anyone other than a well know established huge corporation for no less than half stock and half cash. No way a sale to some newbie Corp. That would be a huge red flag. Just sayin. We should all be on guard for anything and everything. Very little communication from your corporation is never good. A little “ hang in there just a little longer shareholder's “ would go a long way. But we get crickets. Pathetic. Sad. Disgusting really.
I think Wade will try to file a sur reply to ours after the fifteen days . April 8 if working days. Then we’ll have 15 ? Days to reply? But let’s hope if Wade tries that they deny his motion and end this b s fiasco once and for all and the J D and the IRS finally cuff him. I’m old and tired. I want justice. And I will not be screwed again.
Won’t let me be post the link. Our Lawyer submitted proof to court In black and white clearing up some lies. Wade has 15 days to respond. Then they will rule if they will even hear the case. I seriously doubt it. It’s easy to see what a crook he is/was. Not much longer now imo. I doubt there will be any more delays.
By George! I think you got it! Or at least I hope you do. Isn’t today Wade’s last day to respond? Hard to fight against court transcripts. lol. He’s toast imo o t . Lehua won District and State American Legion Constitutional Speech contest again this year! On to Nationals again.At Hillsdale College this year. She will represent Hawaii ! 25k. 22.5 20k third place ! 🎉🎉🎉🎉
lol. Thanks for that Doc. It would be a great Grad present. She also just won, last night, the 86 annual American Legion Constitutional Speech contest for The Big Island again this year! Going to State comp again which she won last year and competed in Nationals in Indy and came within 1 point of being in finals. She knows the ropes and is ready to win it this year! 25 k first 22.5 and 20 k third. Everything falling into place. Waiting sucks. The silence sucks. But I hope it will all pay off soon. If not? Oh how many torches will be lit! lol We wait.
Latest Headlinesupdated 12:48pm EDT
Researchers Develop New Machine Learning Method for Modeling of Chemical Reactions
Mar. 7, 2024 — Researchers have used machine learning to create a model that simulates reactive processes in organic materials and ...
New Method Measures the 3D Position of Individual Atoms
Mar. 5, 2024 — Since more than a decade it has been possible for physicists to accurately measure the location of individual atoms to a precision of smaller than one thousandth of a millimeter using a special type ...
Network of Quantum Sensors Boosts Precision
Mar. 4, 2024 — Quantum sensor technology promises even more precise measurements of physical quantities. A team has now compared the signals of up to 91 quantum sensors with each other and thus successfully ...
Umbrella for Atoms: The First Protective Layer for 2D Quantum Materials
Mar. 1, 2024 — As silicon-based computer chips approach their physical limitations in the quest for faster and smaller designs, the search for alternative materials that remain functional at atomic scales is one of ...
Scientists Make Nanoparticles Dance to Unravel Quantum Limits
Mar. 1, 2024 — The question of where the boundary between classical and quantum physics lies is one of the longest-standing pursuits of modern scientific research and in new research, scientists demonstrate a novel ...
Photonic Memristor for Future Computing: A Perspective
Jing-Yu Mao, Li Zhou, Xiaojian Zhu, Ye Zhou,* and Su-Ting Han*
Photonic computing and neuromorphic computing could address the inherent limitations of traditional von Neumann architecture and gradually invalidate Moore’s law. As photonics applications are capable of storing
and processing data in an optical manner with unprecedented bandwidth and high speed, two-terminal photonic memristors with a remote optical control of resistive switching behaviors at defined wavelengths ensure the benefit of on-chip integration, low power consumption, multilevel data storage, and a large variation margin, suggesting promising advantages for both photonic and neuromorphic computing. Herein, the development of photonic memristors is reviewed, as well as their application in photonic computing and emulation on optogenetics-modulated artificial synapses. Different photoactive materials acting as both photosensing and storage media are discussed in terms of their optical-tunable memory behaviors and underlying resistive switching mechanism with consideration of photogating and photovoltaic effects. Moreover, light-involved logic operations, system- level integration, and light-controlled artificial synaptic memristors along with improved learning tasks performance are presented. Furthermore, the challenges in the field are discussed, such as the lack of a comprehensive understanding of microscopic mechanisms under light illumination and a general constraint of inferior near-infrared (NIR) sensitivity.
1. Introduction
With the gradual failing of Moore’s law and the constraint of the von Neumann bottleneck, a number of the revolutionary com- puting technologies, including photonic computing and neuro- morphic computing, are exploited to fill the subsequent power vacuum. Following the theoretical concept of the “memristor” proposed by Leon Chua and the first physical demonstration
J.-Y. Mao, Prof. Y. Zhou
Institute for Advanced Study
Shenzhen University
Shenzhen 518060, P. R. China
E-mail: yezhou@szu.edu.cn
Dr. L. Zhou, Prof. S.-T. Han
College of Electronic Science & Technology
Shenzhen University
Shenzhen 518060, P. R. China
E-mail: sutinghan@szu.edu.cn
Dr. X. Zhu, Prof. S.-T. Han
Department of Electrical Engineering and Computer Science University of Michigan
Ann Arbor, MI 48105, USA
The ORCID identification number(s) for the author(s) of this article can be found under https://doi.org/10.1002/adom.201900766.
in 2008,[1] rapid progress has been made in the growth of memristive devices.[2] The storage layer of the memristor can be dynamically reconfigured under external stimuli, such as electric fields,[3] magnetic fields,[4] or light illumination,[3] leading to variations in local conductivity and memory effects. The creation and anni- hilation of a conductive filament with either abrupt (binary) or gradual (analog) processes induce a wealth of device behav- iors in this seemingly simple device, suggesting that the memristor holds great potential to facilitate the development of both photonic computing and biologically inspired computing.[5–10]
In general, photonics are capable of storing and processing data in an optical manner with unprecedented bandwidth and high speed.[11–14] Photonic computing utilizes on-chip optical interconnections to substitute electric wires in connecting memories, and central processing units (CPU) are proposed to mitigate the von Neumann bottleneck with low power consumption and high speed in data com- munication.[15–17] Two-terminal photonic
memristors are able to be directly integrated onto processor chips, which subsequently improve the efficiency of the optical information shuttling.[18,19] On the other hand, with the failing of Moore’s law, continuously scaling down to ensure integration of the increased number of memory cells is hindered by fabri- cation complexity and photolithography.[20,21] Exploring multi- level memristor cells with high storage density is an alternate method with the challenges shifting significantly to achieve separable states with high stability.[22] In photonic memristors, the light signal is employed as the extra terminal of the mem- ristor devices to ensure a large memory window and variation margin of multiple storage levels.[23]
The human brain operates in a highly parallel and efficient fashion through a densely interconnected network of synapses and neurons.[24] The information transferred between the neu- rons is assisted by the synapses, which regulate their connection strength or synaptic plasticity to correspond to neural activities via chemical flux.[25,26] This plasticity is largely responsible for forming the basis of memory and learning inside the brain.[24] Motivated by the human brain, neuromorphic computing is expected to address the inherent limitations of conventional von Neumann architecture as well.[27] It is therefore indispen- sable to construct biorealistic synaptic elementary devices with rich spatiotemporal dynamics for the low power neuromorphic hardware.[28] In biological systems, utilizing light to modulate
DOI: 10.1002/adom.201900766
Adv. Optical Mater. 2019, 7, 1900766 1900766 (1 of 15) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
www.advopticalmat.de
neural and synaptic functions, including activated/deactivated light-sensitive proteins, ion channels, agonists/antagonists, receptors, ion pumps, or ligands considered optogenetics, offers a higher degree of spatiotemporal resolution compared to traditional chemical and electrical approaches.[29–32] Inspired by the optogenetics, photonic memristors suggest an effective approach to modulate the synaptic plasticity for future neuro- morphic computing as well.[33,34]
In this Progress Report, we focus on the recent advances in photonic memristors and their applications in photonic com- puting and optogenetics-tunable artificial synapses. Photonic memristors will be discussed in terms of light-involved physical mechanisms of photonic memristors, including the photovoltaic effect-mediated Schottky barrier, photovoltaic effect-induced formation/annihilation of conductive filaments, photogating effect, and photoinduced chemical reaction/ conformation change. Additionally, the light-induced memory performance of photonic memristors is presented. Light- involved functional systems integrated with memristors are also demonstrated to indicate novel photonic computing appli- cations. Afterwards, optogenetics-tunable synaptic functions and neuromorphic systems emulated by photonic memristors are under discussion. In the end, we give a brief summary of the existing challenges hindering the development of light- involved applications and envision the future prospects of this novel physical modulation channel of light.
2. Physical Mechanism of Photonic Memristors (Resistive Switching Devices)
Normally, resistive switching (RS) devices are operated under electrical stimuli with binary resistance transitions between a high resistance state (HRS) and a low resistance state (LRS). In some cases, an intermediate resistance state (IRS) occurs, which involves multiple resistance states instead of two with an enhanced memory window or logic states for multilevel data storage. The resistance transition from HRS (LRS) to LRS (HRS) is defined as a SET (RESET) process, as shown in the middle part in Figure 1. In an RS memory device, the active materials (also known as switching medium) are sandwiched between the top and bottom electrodes to form a metal/insu- lator/metal (MIM) structure, implying its critical potential for the integration arrays with an ultrasmall device area (down to few nanometers), an ultrahigh density of scaled integration of memory elements, an extremely high switching speed, and low power consumption for the RS event. With various physical processes developing over time, the resistance changes occur in the form of either gradual or abrupt variations, suggesting rich behaviors with a seemingly simple structure.[35–37] It is worth noting that the RS event is not only dependent on the active materials but also dominated by the choice of electrodes and the junction area at the interface between active materials and electrodes. Therefore, in consideration of the physical form of the conductive channel, the RS behaviors can be divided into two types, filamentary type and interface type, both of which are related to the physical reconfiguration of ion conditions via the redistribution of ions driven by an electric field.[38] As the most common type, filamentary-based RS behavior has features
of highly localized and restricted filament shapes, resulting in the variation in the local resistivity and high current ratio (high LRS current) due to the formation of an extremely conductive filament bridging two electrodes.[39] Electrochemical metal- lization memristors (ECM) are associated with the switching behavior caused by the migration and redistribution of (active metal such as Cu and Ag) cations and redox reaction.[38,40] On account of the thin active layer, a high electric field is gener- ated by a modest amount of applied voltage for the migration of ions while adjusting the ionic conformation within the active material serving as a solid dielectric electrolyte.[41] In addition, RS behavior in memristors may also stem from the conduc- tive nanochannel formed by the growth and clustering of silver nanoparticles instead of charged ions, as was systematically
Jing-Yu Mao is currently a postgraduate student in the Institute for Advanced Study, Shenzhen University, under the supervision of Prof. Ye Zhou. His research interests focus on the design and fab- rication of memory devices including resistive switching memory and flash memory, as well as their application in artificial neural networks.
Ye Zhou is an IAS Fellow in the Institute for Advanced Study, Shenzhen University. His research interests include flex- ible and printed electronics, organic/inorganic semicon- ductors, surface and interface physics, nanostructured mate- rials, and nanoscale devices for technological applications, such as logic circuits, data storage, energy harvesting, photonics, and sensors.
Su-Ting Han is an asso-
ciate professor at Shenzhen University and a visiting associate professor at The University of Michigan. She received her M.Sc. degree
in Analytical Chemistry from Hong Kong Baptist University and her Ph.D. degree in Physics and Materials Science from City University of Hong Kong. Her research interests
include functional electronic devices and flexible, stretch- able, and wearable electronics.
Adv. Optical Mater. 2019, 7, 1900766 1900766 (2 of 15) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advopticalmat.de
Figure 1. Physical mechanisms of phototunable resistive switching devices. a) Sketch diagrams of positively and negatively charged ion drift to the opposite direction in perovskite under positive bias, resulting in the p–i–n junction within CH3NH3PbI3 film. Reproduced with permission.[47] Copyright 2014, Springer Nature. b) Schematics of a light gated memristor with a structure of ITO/CeO2–x/AlOy/Al. Reproduced with permission.[11] Copyright 2017, American Chemical Society. c) Rupture of conductive filament in MoSe2/Bi2Se3 nanosheets film under NIR illumination. Reproduced with permission.[58] Copyright 2019, Wiley-VCH. d) Facilitated filament rupture in CH3NH3PbI3 planar device under light illumination. Reproduced with permission.[34] Copyright 2018, American Chemical Society. e) Schematic diagram of the flash transistor-like structure of a MoS2/h-BN/graphene heterostructure. Reproduced with permission.[45] Copyright 2018, Wiley-VCH. f) Formation of conductive path in carbon quantum dots-silk device. Reproduced with permission.[67] Copyright 2018, Wiley-VCH. g) Transitions among different resistance states and corresponding threshold voltages of donor-bridge-acceptor compound (DBA) memory device. Reproduced with permission.[81] Copyright 2012, American Chemical Society. h) Schematic illustration of the BMThCE-based memory device as well as the chemical structures of the light-induced transition between o-BMThCE and c-BMThCE. Reproduced with permission.[85] Copyright 2017, Wiley-VCH.
studied and characterized by the Yang group.[26,42,43] For ECM, the severe challenge lies in the precise modulation of the mor- phology and controlled growth of metal filaments; additionally, high operation current must be addressed. Similar to ECM, valence change memristors (VCM) have the switching event driven by the movement and redistribution of internal ions, such as negatively charged oxygen ions (or positively charged oxygen vacancies) in oxides.[38,40] Nevertheless, the electrodes applied here are normally inert electrodes, such as Au or
Pt, without electrochemical activity, and active species, such as oxygen vacancies, play a dominant role in RS behavior. In addition, the reaction and gathering of oxygen vacancies at interfacial areas in the case of Ti or Al electrodes may also con- tribute to the RS operation.[44] Under an electric field, the device experiences a resistance transition from HRS, which results from the insulating property of stoichiometric oxides, to LRS as a conductive channel for electron transport formed by the field-assisted migration of oxygen vacancies. In stark contrast
Adv. Optical Mater. 2019, 7, 1900766 1900766 (3 of 15) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
www.advopticalmat.de
to oxides (or chalcogenides) used in ECM, the initially insu- lating near-stoichiometric oxides as active materials in VCM are involved in the chemical process to form nonstoichiometric products along with changes in the valance state of the metal element after RS operation. Furthermore, the generation of oxygen vacancies may give rise to the deformation of oxides and increasing amounts of trap sites (defects) for charge carriers that also lead to the trap-mediated switching event. The major problem for VCM is the operation stability, since repeated switching cycles may lead to the deterioration of reliability.
The microscopic mechanism of the RS under light illumi- nation becomes more complicated when taking photovoltaic, photogating effects, and photoinduced chemical reactions into consideration when the fundamental operation mechanism is reliant on the electrically induced dynamic motions mentioned above. With the incoming light, RS memory parameters no longer stay the same, where the operation current, including IHRS and ILRS, and SET/RESET voltages can be properly tuned depending on the wavelength and the intensity of light. More- over, RS under the light application of limiting conditions possesses the virtue of diversified operation modes, encom- passing photoinduced SET or RESET operations as well as a change in memory type. Hence, RS in photonic memristors, regarding the operation modes, can be sorted into photoassisted and photoinduced processes (transitions), both of which are strongly connected with ionic dynamics. Accordingly, the choice of active materials and junctions at interfaces for RS events is of critical importance since they work as switching media to be modulated not only by electrical signals but also by incident light, which is considered to be an additional dimension for specific control of RS behavior. To do that, light, as an effec- tive means, is supposed to achieve deep engagement in ionic reconfiguration across the dielectric active layer and the inter- facial region or endure photoinduced chemical reactions where new products are formed with distinct electrical characteristics. Beyond that, another important implementation for effective control of RS behavior is believed to be the photogating effect by means of electrostatic force (potential), which is rarely reported relating to RS behavior.[45] Slightly different from that of the electrically induced RS memory, the physical mechanisms of photonic memristors with respect to the effect of light are classified into four categories, depicted in Figure 1, including the photovoltaic effect-mediated Schottky barrier, photovoltaic effect-induced formation/annihilation of conductive filaments, photogating effect, and photoinduced chemical reaction/confor- mation changes, as discussed in detail in the following.
2.1. Photovoltaic Effect-Mediated Schottky Barrier
The photovoltaic effect typically involves the separation of photogenerated electron–hole pairs, the creation of free car- riers, and the production of an electrical voltage or current from incident photons.[46] Photovoltaic materials are supposed to possess traits, such as efficient light harvesting and large absorption coefficients, which allow full absorption of photons with energy above their bandgap and impedes the undesired charge recombination, a sufficient exciton diffusion length to ensure the efficient charge and voltage generation. The
electron–hole separation is highly related to the internal electric field induced by the heterojunction interface (heterojunction system) or by the Schottky barrier between the semiconductor and metal,[47] which leads to a movement of holes toward the positive electrode and electrons toward the negative electrode, followed by the extraction of charges to external circuits and the generation of the open-circuit voltage. Therefore, the devices based on photovoltaic effects were applied in field-effect transis- tors, where the extended light detection range, high gain, and bandwidth have been achieved, which can then be extended to the memristor application.[48] The photovoltaic effect induced excitons can be separated at the interfaces between electrodes and active materials (single component) or at the interfaces between different components (hybrid system), contributing to ion migration or carrier trapping to tune the interfacial barrier critical for RS behavior.
In the migration of ions, such as positively charged oxygen vacancies, diffusion in oxides has been demonstrated with the capability of modulating the energy barrier in the depletion region, which may induce persistent photoconductivity under light illumination. The interfacial barrier between the electrode (ITO) and active materials can be lowered by the photovoltaic effect. Li et al. demonstrated an optically programmable and electrically erasable memristor with a simple structure of ITO/ CeO2-x/AlOy/Al, which can be labeled as the photoinduced SET process.[11,23] CeO2-x was carefully chosen to act as a photoac- tive material with a wide range of photoresponses attributed to the defect energy level embedded in the bandgap as well as an active material for RS events due to the nonstoichiometric nature and was beneficial for the easy formation of a native AlOy layer as a thin Schottky junction to induce an insulating barrier, which suppresses the leakage current while the stable interfacial trapping sites can be created. During positive voltage sweeps (electrical mode), additional positively charged oxygen vacancies were attracted toward AlOy, leaving electrons driven toward the ITO electrode under the electric field. The accumu- lation of oxygen vacancies at the CeO2-x/AlOy interface results in a thinner Schottky barrier; thus, the resistance changes from HRS to a low resistance state. Upon the introduction of light, the optical programming was achieved, which stems from the narrower junction barrier and reduced junction resistance that subsequently enhanced charge transport at the interface by the diffusion of photoexcited electrons away from the interfacial region with the accumulation of more oxygen vacancies in the opposite direction, as shown in Figure 2a.[8,11] In addition, the multilevel nonvolatile state was realized by the modulation of light intensity and wavelength, implying a possible applica- tion in future high-density optoelectronic memory.[18] The ITO/Nb:SrTiO3 heterojunction also shows an analogous phe- nomenon that was reported by the same group in 2019.[49] Similarly, the high photosensitivity of the Nb:SrTiO3 and ZnO nanorods heterojunction ensures the versatile reconfiguration of ionized oxygen vacancies and the alteration of the effective energy barrier in the interfacial region under the optical stimuli (Figure 2b).[44,50] Multiple oxide materials have been employed as the active materials for memristors under electrical and optical stimuli, exhibiting multilevel data storage.[22,51] The generation of oxygen vacancies and the tuning of the Schottky junction at the interfaces (SrTiO3) can improve the RS behaviors
Adv. Optical Mater. 2019, 7, 1900766 1900766 (4 of 15) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advopticalmat.de
Figure 2. Photovoltaic effect-induced interfacial Schottky barrier and formation and annihilation of conductive filaments. a) Energy band diagram depicting the operation mechanism of electrical erasing and optical programing. Reproduced with permission.[11] Copyright 2017, American Chemical Society. b) Energy band diagram of the reversely biased heterojunction showing sustained photocurrent induced by charge separation process after the illumination of UV light. Reproduced with permission.[50] Copyright 2013, Wiley-VCH. c) Switchable photovoltaic model and mechanism study. Sketch diagrams of positively and negatively charged ion drift to the opposite direction in perovskite under positive and negative bias, resulting in the p–i–n junction (energy diagram). A section of Au electrode was removed for further KPFM potential measurements. Reproduced with permission.[47] Copyright 2014, Springer Nature. d) Light modulation on LRS and HRS with different intensities. Reproduced with permission.[54] Copyright 2017, Wiley-VCH. e) Sketch of the conductive filament consisting of iodine vacancies in the planar configuration device. f) RESET process under broadband visible light with different light intensities. Reproduced with permission.[34] Copyright 2017, Wiley-VCH.
by introducing multiple resistance states and lowering opera- tion voltage under the application of light.[23,50,52]
Perovskite materials have attracted enormous attention due to their extraordinary broadband light absorption and high bandgap tenability. In addition to their potential application in next- generation solar cells and light-emitting diodes, the dynamic ion migration and ferroelectric behavior related to current– voltage hysteresis have been considered the basic form for data storage, although they are not preferred in photovoltaic applica- tions.[47] For perovskite-based memristors, the interfacial bar- rier can be readily modulated by photovoltaic effect-induced ion
migration.[53] Huang et al. discovered the switchable photovoltaic effect in perovskite optoelectronic devices. The perovskite device with symmetric structure (Au electrodes) transformed from a neutral state to a p–i–n junction with obvious variation in open- circuit voltage (VOC) and photocurrent.[47] Importantly, the mem- ristive hysteresis was mainly attributed to the charge accumula- tion induced by negatively charged Pb and MA vacancies and positively charged I vacancy migration, resulting in p- and n-type doping at perovskite/PEDOT:PSS and perovskite/Au interfaces, respectively (Figure 2c). Conversely, the perovskite memris- tors can be read out in two modes by both electricity (current
Adv. Optical Mater. 2019, 7, 1900766 1900766 (5 of 15) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
www.advopticalmat.de
on/off ratio of ˜103) and light (different VOC). Wu et al. system- atically studied the switching mechanism of hybrid perovskites CH3NH3PbBr3 (MAPbBr3).[54] By tuning the maximum posi- tive/negative voltage during sweeping, various degrees of hys- teresis were addressed, which originated from the modulation of the perovskite/ITO Schottky junction through the migration of charged ions/defects (MA vacancies). Since perovskite mate- rials exhibit impressive photoresponse, light signals were also employed for the involvement of tuning the perovskite-based memristors, where UV illumination gave rise to multilevel data storage and the photocurrent increased at both HRS and LRS (Figure 2d). More obvious current variation was observed for HRS, indicating a thicker depletion region in this scenario. The overall enhancement in current was due to the neutralization of charged ions with additional photogenerated charge carriers at the interfacial region, while interface-based switching (tuning perovskite/ITO Schottky barrier height) was responsible for RS operation when the possibility of metallic filament formation was excluded.
2.2. Photovoltaic Effect-Induced Formation/Annihilation of Conductive Filaments
For cation memristors, active metal electrodes, such as Cu or Ag, are often chosen as anodes for the purpose of redox reac- tions (ECM), while in anion memristors, the formation and annihilation of anion conductive filaments are the main mecha- nism for RS.[55] Upon the application of external electric fields, conductive filaments are formed via the migration of either charged ions or vacancies (electrical SET process), whereas light has not been reported to accelerate this process. In contrast, light-assisted diffusion of charged ions was proved to facilitate the RESET process. As discussed above, ion migration inside lead halide perovskite results in the variation of the interfacial barrier that further influences the RS behavior. In addition, filamentary RS involving either Ag or halide ion migration was also proposed as the switching mechanism of recently reported halide perovskite-based memristors, depending on the selec- tion of electrodes. Lu and co-workers reported the RS effect based on MAPbI3 with tunable SET/RESET voltages,[56] owing to the formation and rupture of conductive channels consisting of iodine vacancies (Figure 2e). In MAPbI3, both I- and MA+ ions have been confirmed capable of migrating while the diffusion of I- is preferable to induce RS due to the lower diffu- sion energy barrier.[57] However, with visible light illumination, the filaments composed of iodine vacancies became unsteady and tended to rupture, arising from the spontaneous diffusion and recombination of iodine vacancies with iodine ions, which leads to a photoinduced RESET process. Figure 2f depicts the light intensity effect on the RESET voltages of the MAPbI3 device with a planar structure. An obvious increase in the SET voltage and decrease in the RESET voltage can be observed with increased light intensities. The coupling of electrical and optical effects resulted in the implementation of an electrical- SET and photoinduced RESET memory cell with an ultrahigh on/off ratio and the ability to remotely measure the device con- ductance. Later, a memristor with the analogous function had been realized in a doped inorganic perovskite material system
(Ti-doped BiFeO3) based on ferroelectric photovoltaic effects,[50] in which the redistribution of electrons and atomic partial densities of states in the BiFeO3 film caused by Ti doping can improve the photovoltaic effect and induce the filament forma- tion. The ferroelectric photovoltaic VOC was properly modulated by RS behavior in Ti-doped BFO film. Thus, the electric-optical memory can be operated in the form of electrically induced resistance switching between HRS and LRS and optical reading of the VOC at different resistance states.
Similarly, Han and co-workers reported a near-infrared (NIR) controlled memristor that can induce unusual RESET operations under 790 nm NIR illumination.[58] After the SET process, in which Ag conductive filaments were formed to bridge the top and bottom electrodes, photogenerated elec- trons were arrested by the MoSe2/Bi2Se3 hybrid layer, leaving untrapped holes that rendered the oxidation reaction of Ag clusters to Ag+ cations. Gradual oxidation of Ag atoms within conductive filaments results in photovoltaic effect induced annihilation of the Ag conductive filaments operation in the form of a current decline of the RS device.
2.3. Photogating Effect
Photogating effect, literally defined as resistance modulation through a light-induced electric field, is widely used in pho- todetectors or transistor-based flash memories. When photo- generated charge carriers (e.g., holes) are trapped in localized defects or trapping layers, the introduction of a trapped hole-induced electric field gives rise to more electrons in the conductive channel, thus inducing the increase in channel conductance.[59] The electrostatic force induced by trapped charge carriers contributes to the modulation of the charge transport along the conductive channel by means of either an electrical or a light signal. Thus, it will be intriguing as well as challenging to develop optical memory devices on the basis of the photogating effect. 2D van der Waals (vdWs) mate- rials, especially the heterostructures, have aroused remark- able interest in photonic memories and synaptic devices for neuromorphic computing due to their intriguing optoelec- tronic and mechanical properties.[37,60–63] In recent years, flash memories based on the 2D vdWs heterostructures exhibit highly competitive performance with a large memory window, high on–off current ratio and long retention time.[64,65] How- ever, the complexity of the three-terminal structure hinders the variability, expansibility, and integration density of the memory device for further development. Lee and co-workers presented a two-terminal tunneling random access memory (TRAM) device based on a 2D graphene/h-BN/MoS2 hetero- junction, analogous to the device structure of flash memory, in which the major discrepancy yields the number of termi- nals to modulate the device performance.[45] Graphene with a proper thickness of h-BN and MoS2 is used as the floating gate, tunneling insulating layer and semiconductor material. TRAM, which displays a simplified device structure that acts as a memory device that converts from a three-terminal flash memory structure to a two-terminal configuration, exhibited a reduced complexity with an ultralow off-state current of 10-14 A and ultrahigh write/erase current ratio of 109. In terms of
Adv. Optical Mater. 2019, 7, 1900766 1900766 (6 of 15) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advopticalmat.de
Figure 3. Photogating effect and photo-induced conformation change. a) I–V characteristics of the MoS2/h-BN/graphene heterostructure optical flash memory. b) Reproducible programing and erasing cycles by electrical and optical pulses, respectively. c) Working mechanism of the optical memory. (1) Electrical programming by applying a -10 V bias to drain electrode. (2) Reading of ON state under dark condition by applying a 0.5 V bias. (3) Optical erasing for 1 s with a 458 nm laser. (4) Reading of OFF state under dark condition by applying a 0.5 V bias. Reproduced with permission.[45] Copyright 2018, Wiley-VCH. d) Different I–V characteristics and threshold voltages under dark and UV light illumination of donor-bridge-acceptor compound (DBA) memory device. Reproduced with permission.[81] Copyright 2012, American Chemical Society. e) Scheme of the BMThCE-based device as well as the chemical structures of the light-induced transition between o-BMThCE and c-BMThCE. f) I–V curves of the memory devices based on o-BMThCE and c-BMThCE. Reproduced with permission.[85] Copyright 2017, Wiley-VCH.
the operation mechanism of the TRAM, the charge carriers can tunnel through the insulating h-BN layer and maintain the highly conductive graphene layer under the large elec- tric potential between the drain electrode and the floating gate, which cannot tunnel through the h-BN layer away from the floating gate to the source electrode due to the limited potential difference, thereby achieving the storage of charge carriers. TRAM also presented excellent durability over 105 cycles and superb stretchability with no obvious device perfor- mance degradation changes up to 19% strain, resulting from the superior mechanical stability of the 2D materials assembly. Recently, on the basis of previous work, they developed a multilevel nonvolatile optical memory based on the MoS2/ h-BN/graphene heterostructure, where the MoS2 monolayer
acts as both photoactive and semiconductor material.[66] The 2D material heterostructure-based optical TRAM device also exhibited an ultralow off-state current of 10-14 A and a signifi- cant current increase was induced by laser light illumination, leading to nonvolatile memory behavior shown in Figure 3a. Reproducible electrical programming and light erasing opera- tions were carried out (Figure 3b). Under light illumination, the photogenerated holes in MoS2 can easily tunnel through h-BN to the graphene layer, neutralizing the stored electrons and enabling the optical erasing operation (Figure 3c). The optical-tunable TRAM revealed robust memory performance, such as an ultrahigh current on/off ratio over 106, stable endurance over 104 cycles, a long retention time over 104 s and multilevel storage (four-bit). Therefore, the optical TRAM
Adv. Optical Mater. 2019, 7, 1900766 1900766 (7 of 15) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
www.advopticalmat.de
opens up a promising strategy for high-density integration of novel multifunctional optoelectronic devices with 2D vdWs heterostructures.
Recently, Han et al. carried out the investigations on the effect of incident light with different wavelengths and inten- sities on RS characteristics and the underlying mechanism of light-tunable memory devices based on quantum dots (QDs), which are widely used and studied for their quantum confinement effect, including CsPbBr3 QDs and carbon QDs– silk protein bicomponent blend.[67,68] In both cases, with the application of UV light, SET and RESET voltages were effec- tively reduced, facilitating the development of energy-efficient memory. After light absorption, photogenerated electron–hole pairs at the interfaces were separated into electrons and holes that were driven to the anode and cathode, respectively. Along with the externally injected charge carriers, photogenerated charge carriers were also trapped inside the QDs, leading to the photogating effect, which further accelerated the formation of the conduction path as well as the decreased voltage required for the resistance transition. To confirm the charge trapping behavior within the QDs, in situ Kelvin probe force micros- copy (KPFM) measurements were carried out. The increase in potential difference with QDs and further enhancement under UV light illumination after carrier injections from the conduc- tive tip confirmed the feasibility of light-tunable memory char- acteristics and was accounted for in the modulation of charge trapping ability within QDs (photogating effect). These works emphasize the importance of the development of light-tunable memory devices based on QDs and other materials.[69]
Most light-tunable memristors require the application of high energy UV or visible photons, whereas NIR photonic memristors have rarely been reported.[62,70] Upconversion materials have excellent photoluminescent properties that can absorb long-wavelength light and radiate light with higher energy.[70] Heterostructures combining upconversion materials with 2D materials can provide a novel way of light modulation in the RS process, contributing to enhanced photo energy utilization and photosensitivity. Han et al. integrated directly grown MoS2-NaYF4:Yb3+, Er3+ upconversion nanoparticle (UCNP) nanocomposites into memristors, in which MoS2 nanosheets could absorb the visible light emitted from the UCNPs when excited by NIR illumination.[62] In material char- acterizations, ideal matches were found between the absorption peaks of MoS2 and emission peaks of UCNPs under 980 nm NIR irradiation, resulting in efficient nonradiative energy transfer from UCNPs to MoS2. An evident decreasing trend was found for SET and RESET voltages of light-controllable MoS2–UCNP nanocomposite memory devices, while current at LRS increased drastically under NIR irradiation, which can be extended to diverse resistance levels (at different NIR intensi- ties). Moreover, conductive atomic force microscopy (C-AFM) mapping measurements as well as the well-fitted space charge limited current (SCLC) model were exploited to explain the underlying mechanism of tunable charge trapping endowed by MoS2. Upon the application of NIR irradiation, excitons were generated in UCNPs followed by the rapid charge carrier sepa- ration at the MoS2–UCNPs interface that further led to charge trapped in MoS2 and photogating accelerated SCLC. This study paves the way towards the modulation of NIR memory behavior
and provides insight into further light-controlled encrypted data storage using invisible light illumination.
2.4. Photoinduced Chemical Reaction/Conformation Change
Photoinduced chemical reactions have attracted increasing attention in many promising applications such as energy con- version and environmental safety due to their atom-economy (only photons are needed) and low energy consumption (depending weakly on temperature).[71] High selectivity of a photoinduced chemical reaction is achieved originating from the adjustable photon energies and fast electron transfer. Photoinduced chemical reactions involve the absorption of photons to render the molecules excited and induce the chem- ical changes such as ionization and isomerization.[72] Addi- tionally, the electrical properties, such as the energy bandgap and dipole moment of small molecules, can be precisely manipulated by the photoinduced ionization and isomeriza- tion, resulting in the optimization of RS behavior in optical memristors. In RS memory, light-induced transformation between switching behaviors is strongly related to conforma- tion change within the photoactive materials, which may lead to the change in chemical bonding and energy band.
Reduced graphene oxide (rGO) has attracted remarkable interests due to its advantages of tunable bandgap, proper electrical properties, and large-scale synthesis, as well as its application in nonvolatile RS.[36] Han and co-workers dem- onstrated a phototunable memristor based on ITO/phospho- tungstic acid (PW)/graphene oxide (GO)/Al structure.[73] PW can act as a multifunctional catalyzer to reduce GO under UV irradiation and induce the transition of memory types from WORM to bipolar RS. The irreversible transformation in RS behavior is due to the discrepancy in the energy level between GO and rGO, as well as the decreased interface energy barrier in the PW/rGO junction. Zhao et al. proposed a photocata- lytic reduction method via the introduction of TiO2 into GO, which provided the desired control of rGO domains.[74] A sig- nificantly improved performance of Al/GO-TiO2/ITO memory was observed after photocatalytic reduction of GO, presenting a reduced SET voltage (2.1–0.52 V) without the electrical forming process. Increased photoreduction time as well as the concentration of TiO2 can lead to the generation of local rGO domains through controlled photoreduction. In addition, this method allows low temperature and solution process of high- performance memristors on flexible substrates.
Organic molecules with functional donor and acceptor groups are artificially tailored to achieve intramolecular charge transfer between these groups.[75,76] By means of deliberate control of the functional groups (mostly acceptors), multi- level resistance states can be clearly realized arising from the charge transfer from a donor to multiple acceptors.[77–80] Ye et al. reported a meta-conjugated donor-bridge acceptor (DBA) molecule as the RS media to achieve ternary states storage with the help of UV stimuli, which originates from the charge transfer from triphenylamine (donor) to 2-dicyano- methylen-3-cyano-4,5,5-trimethyl-2,5-dihydrofuran (acceptor) (Figure 3d).[81] Additionally, the UV–vis spectra provided crucial information on the UV-assisted step-by-step oxidation based on
Adv. Optical Mater. 2019, 7, 1900766 1900766 (8 of 15) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
www.advopticalmat.de
the extent of bathochromic-shift of the absorption peak at long wavelengths. The cooperative effect of both redox systems of the organic small molecules and the assistance of UV ensures the effective modulation on the optoelectrical memory device with multilevel storage.[82]
As a well-established class of photochromic molecules, the isomerization between two structural conformations in diary- lethene (DAE) molecules can be induced via optical irradiation, and the material properties of the isomers can be changed with the external optical stimulus.[47,83,84] The significant variations in the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) energies as well as the dipole moment make them promising candidates as the active material in optoelectronic devices. Ling et al. demonstrated a reversible RS characteristic in a DAE-type molecule, named BMThCE, upon UV or visible light irradiation.[85] The device demonstrated WORM characteristics when the BMThCE mole- cule was in a ring-open state (o-BMThCE) under visible light, which could then change to the c-BMThCE (ring-close state) under UV irradiation (Figure 3e), resulting in an erasable non- volatile bipolar RS characteristic (Figure 3f). In consequence, the discrepancies in RS characteristics originated from the tun- able HOMO/LUMO of BMThCE under different wavelengths of light irradiation. The HOMO of o-BMThCE is lower than that of c-BMThCE. Therefore, the deeper hole trapping in the o-BMThCE compared to c-BMThCE occurred. It can be briefly concluded that light-induced transformation between confor- mational structures with a change in energy level indeed has a pronounced impact on the RS types, which greatly modulates the device performance in an accurate and energy-efficient way.
3. Photonic Memristor for Future Photonic Computing
3.1. Light-Involved Logic Operations
Memory devices such as conventional transistor-based flash have their intrinsic and fixed function to store information. By designing memristors capable of storing while processing information based on incident signals, logic operations can then be performed in a highly parallel and energy-efficient way. Novel light-modulated two-terminal memristor devices operating as data storage with the ability to conduct logic operations with simple device configuration are drawing great research attention since logic plays a fundamental role in the binary system and the introduction of light brings an ingenious approach to merge electric signals with the additional dimen- sion of light.[86] Apart from the tunable memory characteristics of the photonic memristors we discussed above, novel memory integrated with other functions, including arithmetic and demodulation, is a fundamental basis for future in-memory computing using electrical and optical signals as inputs. Two light sources, blue and green lights, along with wavelength and light intensity were capable of realizing four memory states, enabling the information storage and transmission of 8-bit codes in ITO/CeO2-x/AlOy/Al devices in which the physical mechanism was discussed earlier.[23] Upon the basis of the linear dependency of photocurrent on the number of applied
optical pulses, arithmetic operations including “Counter” and “Adder” were hence implemented. The simple integration of memory and logic functions offers a feasible route to reduce system complexity in the integrated circuits.[87,88] In addition to basic nonvolatile memory and arithmetic function, as well as the light-induced RS effect, memlogic devices with arith- metic functions were also proposed by Li et al. The device can operate as a light-controlled logic gate, which implements fundamental Boolean logic including the “AND,” “OR,” and “NOT” operations by means of controlling the charge trap- ping at the CeO2-x/AlOy/Al interface.[11] Both electrical and light pulses are termed as logic input while device current is considered an output of the logic gate. When the output cur- rent exceeds a given threshold value, such as in the case where the light pulse is applied along with the voltage pulse, logic “1” is achieved. If no or either stimulus is applied, the output is logic “0,” realizing the “AND” function. In addition, “OR” and “NOT” functions can be implemented by preprograming the device to an intermediate resistance state and adding an addi- tional RESET electrical pulse. Figure 4a demonstrates the input of two patterns to the memlogic array as an image processor and are recognized together with memory characteristics that were validated to implement SAME FINDER and ALL FINDER functions (Figure 4b,c). Furthermore, the parallel operation of data recording and processing inside a single memory cell as well as the reconfigurable logic operations indicate a promising potential in future in-memory computing.
3.2. System-Level Integration of a Photonic Memristor
The human visual system allows for the collection, transpor- tation, and processing of visual information from ambient objects. To simplify these processes, photodetectors can be regarded as retinas to sense the light signal and then trans- form the light signal into an electrical one to be processed and recorded in memory devices.[89,90] Chen et al. reported a novel combination of In2O3-based image sensors and Al2O3- based memristors, which constructed artificial visual memory motivated by UV light for the application of image sensing and storage.[91] The visual memory system could be operated by excellent detection of UV light while the information was recorded in the memristors. Figure 4d demonstrates the config- uration of visual memory, where the image sensor and memory unit are connected in series with a conjunct electrode. When the UV light is applied, the resistance decrease of image sen- sors leads to the increase in partial voltage on the memristors, thus switching the memory device from HRS to LRS. Once the light illumination is removed, the recorded image shows a long retention time of more than a week to be recognized, and the stored information can be RESET via a reverse voltage (Figure 4e). The obtained results signify the potential applica- tion of visual memory systems in the development of artificial electronic eyes and bioinspired systems.[92] In 2016, Kim et al. presented a photodestructible memristor based on modified UCNPs.[93] The broadened absorption range for the modified UCNPs to a visible range by introducing multiple fluorescent dyes extends the choice of light source, enabling the erasure of specific information that is locally stored inside the memory
Adv. Optical Mater. 2019, 7, 1900766 1900766 (9 of 15) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advopticalmat.de
Figure 4. Memristor-involved photonic applications. a) Patterns “X” and “Y” fed into the memlogic devices; output sketch of b) the same regions and c) all the parts of both input patterns. Reproduced with permission.[11] Copyright 2017, American Chemical Society. d) Schematic illustrations of imaging and memory system corresponding to human visual memory. e) Schematic illustration of the detected and memorized image of the artificial visual memory. Reproduced with permission.[91] Copyright 2018, Wiley-VCH. f) Graphs depict the selective destruction of data under 980 and 800 nm laser illumination. The left side corresponds to the UCNPs with etched Mg layer while the one on the right side indicates the UCNPs with sensitizer III. g) Magnified image of the integrated device before (left) and after (right) 800 nm laser illumination. Reproduced with permission.[93] Copyright 2017, Wiley-VCH.
device. Figure 4f,g depicts the destruction of selective regions by light illumination at a wavelength of 800 nm only, which provides insight into data security applications as an invisible light signal is a fascinating candidate for information manipu- lation.[94] Memristors with integrated functions, including light detection, data storage, and in-memory computing, are primary steps toward complex artificial neural networks. The abovemen- tioned photonic memristors are beneficial to the construction of optoelectronic synaptic devices, which come with advantages in novel optical communications.
4. Photonic Memristor for Optogenetics-Tunable Neuromorphic Computing
In contrast to conventional digital computers based on the von Neumann architecture, where the memory is separated from the central processing unit, the human brain has tens of bil- lions of neurons connected to each other in a vast network with the synapses.[35,95] In this system, a great deal of information is transmitted all the time, constituting our memory, logical reasoning, linguistic competence, and visual-spatial cognition. In a single neuron, information is transmitted by electrical sig- nals. The electric potentials inside and outside the cell mem- brane are reversed with signal conduction and followed by a large amount of calcium ion influx. The traditional method for neuroscientists to activate a nerve cell is with electrical stimu- lation (electrode placed near a neuron to change its electrical
potential) or the chemical approach (apply small molecules that act as neurotransmitters to the cell receptors). However, electrical stimulation cannot accurately localize the cell while chemicals cannot provide accurate time control due to their slow release. Therefore, the development of unique and reliable techniques that are able to control neuronal activity with pre- cise temporal and spatial precision is highly desirable.
In recent years, a new technology, termed “optogenetics,” which uses external light stimulation to control the behavior and function of neuron cells with high specificity, has been developed, and its temporal resolution has been exten- sively utilized in neuroscience.[96,97] In optogenetics, light is employed to activate/deactivate the light-sensitive ion chan- nels and change the membrane potential on both sides of the cell membrane. Thus, the excitation or inhibition manipula- tion of the neuron cell can be achieved. The basic principle of optogenetics is shown in Figure 5a.[29,30] Photosensitive mem- brane channel proteins, such as channelrhodopsin (ChR2) and halorhodopsin (NpHR), are transferred into neuron cells for ion channel expression. In the case of ChR2, when a neuron cell is exposed to a blue laser at 470 nm, a cationic channel of neurons opens, allowing a large number of cations (such as Na+ and Ca2+) to flow into the cell. The action will depo- larize the neuronal membrane and positively modulate the action potentials, thus placing the neuron in an excited state. For NpHR, when a neuron cell is exposed to a yellow laser at 580 nm, the anion (Cl-) is allowed to enter the cell through the ion pump, inducing the membrane hyperpolarization and
Adv. Optical Mater. 2019, 7, 1900766 1900766 (10 of 15) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advopticalmat.de
Figure 5. Light-tunable synaptic plasticity. a) Diagrams of ion conduction under light illumination in optogenetics. Reproduced with permission.[30] Copyright 2011, Elsevier Inc. b) Schematic of the connection between adjacent neurons (biological synapse). Reproduced with permission.[103] Copy- right 2018, Wiley-VCH. c) Current response under the application of single light pulse with different incident light intensities (0.4, 2.2, 2.8, and 4.0 mW cm-2); d) current response under the application a sequence of light pulses. The inset demonstrates the amplitude change of peak current after the application of two consecutive pulses. Reproduced with permission.[108] Copyright 2018, American Chemical Society. e) A single-layer artificial neural network is composed of 784 input and 10 output neurons, jointed by independent synapse with different synaptic weight. Each input neurons correspond to a pixel of the input pattern. f) Pattern recognition rate against the number of learning phases. g) Sketch images of eventual synaptic weights correlated to the input patterns for all the three training modes. Reproduced with permission.[121] Copyright 2018, Wiley-VCH. h) Biological and artificial synapse under optical simulation and the corresponding muscle response. Reproduced with permission.[125] Copyright 2018, AAAS.
negatively modulating the action potentials, thereby inhibiting the neuron.
In biological systems, calcium ions (Ca2+) play an impor- tant role in many physiological processes, such as intracellular information transmission, gene expression, neurotransmitter release and so on. Benefitting from the optogenetics tech- nology, it was recently demonstrated that light can activate/ deactivate the Ca2+ channels to modulate the synaptic plasticity of mice.[98] It is noted that the memristors resemble typical biological synapses in both structure and kinetic processes. The current change in the memristor stimulated by voltage pulses resembles the Ca2+ dynamics in biological counterparts during the modulation of synaptic plasticity.[99,100] Thus, the strategy of using light as an external stimulus to modulate synaptic plasticity in memristors is a feasible method to emulate optoge- netics-mediated synaptic functions.
Memristors, as artificial synapses, are termed RS devices, of which the present resistance state is closely related to the voltage stimuli that were previously applied and can be well tuned with analog switching characteristics extensively studied to emulate biological synapses functions with only electrical input.[101,102] Photonic memristors open up novel access for the modulation of synaptic weight by light for further transmis- sion and processing of stimulus information (Figure 5b).[103]
Very recently, light acting as a noninvasive synaptic modu- lator that is physically separated from the conductive channel can also effectively and energy-efficiently tune the synaptic functions of photonic memristors based on photoactive mate- rials.[104] On-chip photonic computing has been proposed as an alternative to replace the electronic computers to eliminate the von Neumann bottleneck with high speed and low power consumption in data transmission.[105] Phase-change mate- rials (PCMs) have emerged as viable materials for nonvolatile photonic memories due to their distinct difference in optical and electrical properties between their amorphous and crystal- line states.[106] In another aspect, the promising fast processing ability of PCMs make it a potential candidate for photonic neu- romorphic systems.
Recent works by Cheng et al. demonstrated an on-chip photonic synapse with PCMs incorporated into a tapered waveguide structure.[15] Based on the rationally designed structure, the photonic synapse shows an ultrahomogeneous electric field distribution. Therefore, the synaptic weight of the photonic synapse can be simply and effectively controlled by varying the pulse numbers sent down the waveguide while retaining the pulse duration and energy. Synaptic plasticity, including the spike-time-dependent plasticity, can be mim- icked via an all-optical method. This work provided a simple
Adv. Optical Mater. 2019, 7, 1900766 1900766 (11 of 15) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
www.advopticalmat.de
but powerful approach to realize a low power consumption artificial photonic synapse, which may establish a prototype for future large-scale photonic neuromorphic networks. Purely photonic triggered memory or artificial synapse can be imple- mented via waveguide structures, while for electrical-controlled synaptic devices, the light signal offers a novel approach to nondestructively modulate the synaptic weight of the conduc- tive channel by a spatially separated stimulus.[17] In addition, more superiorities, such as separated learning and readout operations, broadband response to light illumination and sig- nificantly lowered energy consumption, open up encouraging prospects for future neuromorphic computing. Two-terminal oxide heterojunction-based memristors including ZnO1-x/AlOy, In2O3/ZnO and ITO/Nb:SrTiO3 were also reported to emu- late multiple fundamental synaptic functions, including STP, LTP along with phototunable excitatory postsynaptic current (EPSC) and paired-pulse facilitation (PPF), and were faithfully emulated by optical stimuli (Figure 5c,d).[49,107,108] Similarly, current–voltage characteristics were attributed to the intrinsic charge trapping capability inside the oxide heterojunction and led to a nonvolatile persistent photocurrent, which is a vital element to realize light-induced LTP behavior. Moreover, other photoactive materials, such as perovskite and 2D semiconduc- tors, also possess the ability to truly emulate the synaptic behav- iors of biological synapses using light as the input to modulate synaptic weight.[16,109–112] Different from two-terminal memris- tors, three-terminal synaptic transistors, especially light-gated synaptic transistors, demonstrate tunable synaptic behaviors through controllable charge trapping and detrapping at inter- faces or trap sites; inherent photoconductivity (PPC)[113–116] also emerged as a promising optoelectronic device for modulating synaptic weight in an energy-efficient way.[103,117] Recently, light-stimulated synaptic transistors based on inorganic perov- skite quantum dots (CSPbBr3 QDs) and organic semiconduc- tors exhibited excellent photoresponsivity and delayed the decay process of the photocurrent and synaptic tunability due to the efficient charge separation of photogenerated excitons.[118,119] Therefore, neuromorphic potentiation and depression were effectively conducted by photonic and electrical stimuli. These works implemented the basic synaptic functions by applying an external light signal and provided a potentially solid strategy for the development of novel low-power neuromorphic computing systems and artificial neural networks.[16,120]
As we mentioned in the previous section, MAPbI3-based memristors exhibited light-tunable SET/RESET voltages attributed to the formation and annihilation of the conductive channel composed of iodine vacancies.[34] Recently, inspired by the manipulation of neuronal activities of optogenetics beyond conventional means, Lu et al. exploited a light-modulated MAPbI3-based two-terminal memristor device as an artificial synapse for the purpose of simulating fundamental synaptic behaviors in biological neuronal systems.[34] Enlightened by the tuning of the Ca2+ dynamics, light can be considered an additional source to modulate the connection strength of the conductive channel. Light illumination applied to the planar channel induces the suppression of iodine vacancies and facili- tates the relaxation process of device conductance, authentically emulating the synaptic behavior of biological synapses. Fur- thermore, the strengthening and weakening of the conductive
channel induced by the adjustment of different incident light parameters display significant resemblance to the memory and forgetting processes.
Inspired by light-excited secretion of dopamine (neurotrans- mitter) and the promotion of fast learning and adaption of neural activities, computing systems capable of rapid learning and recognition can be designed at an extremely low power- consuming level. The light-involved tuning of synaptic weight enables a significant reduction in energy consumption as well as processing time. In 2018, a two-terminal light-assisted artifi- cial synapse with improved synaptic plasticity was proposed by Ham et al.[121] Light illumination facilitated the iodine vacan- cies movement in the perovskite layer, leading to easier realiza- tion of long-term potentiation when combined with electrical input signals, which resembles the phenomenon of dopamine- assisted behaviors in biological synapses. A pattern “3” was input into a single-layer neural network composed of 784 input and 10 output neurons between which are individual synapses with different synaptic weight for recognition (Figure 5e). Each input neuron corresponds to a pixel of the input pattern. After updating the weight, the fully trained artificial network is able to execute the recognition process of input patterns. It is noteworthy that the perovskite-based synaptic memristor dem- onstrated good learning capability and acquired a high pattern recognition rate of 81.8% after limited learning phases with light illumination while consuming much less energy than that of only electrical input (Figure 5f,g). A similar training process and resulting conclusion can also be drawn for photosynaptic transistor devices based on charge trapping/detrapping.[103] To summarize, light involved operations demonstrate improved data storage and synaptic behaviors with the potential for building multidimensional neural networks.[21,103,122] The massless and uncharged nature of photons can avoid the charge-based wiring issue and allow it to perform efficient communication performance.[123] It is highly promising to use light instead of electrical signals for communication in artificial neural networks due to the parallel computing capability and power efficiency of optical systems.[124] Recently, Shen et al. demonstrated a fully optical on-chip implementation of neural networks where the computational speed and power efficiency can be enhanced by the coherent and linear-optics matrix multiplication.[14] This work lays a solid foundation for the real- ization of a truly integrated all-photonic neuromorphic system based on all-photonic artificial synapses to perform ultrafast neuromorphic computations.
Bioinspired nervous systems with sensory, integrative, and motor functions have turned into a core technology for next- generation electronic neurorobotics in the booming artificial intelligence industry.[125–127] This requires that the sensory receptor has hypersensitive sensing of the changes in the external environment, the sensing system has a strong ana- lyzing and interpreting ability of the sensory information to allow appropriate responses and proper decision making, and the motor unit has an accurate response to the sensory infor- mation. Recently, Lee et al. demonstrated the first optoelec- tronic nervous system based on an organic artificial sensory synapse and a neuromuscular system (Figure 5h).[125] The ion gel-gated stretchable organic nanowire synaptic transistor (s-ONWST) can mimic typical short-term synaptic plasticity,
Adv. Optical Mater. 2019, 7, 1900766 1900766 (12 of 15) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
www.advopticalmat.de
including EPSC, PPF, spike-dependent plasticity (the verified parameters including voltage, number, duration and frequency) and high-pass filtering, at both 0% and 100% strains. Addi- tionally, by combining with an organic photovoltaic device, the s-ONWST can convert the light signals into electrical presyn- aptic spikes to cause postsynaptic potentiation of the artificial synapse. More importantly, this organic optoelectronic synapse can actuate an artificial neuromuscle to display the analogous biological muscular contraction mechanism, which is difficult to realize in conventional artificial muscle actuators. This work would suggest a promising strategy for the development of movement systems in neurologically inspired robotics.
5. Summary and Outlook
acknowledge the grants from Natural Science Foundation of China (Grant Nos. 61601305 and 61604097), Guangdong Province Special Support Plan for High-Level Talents (Grant No. 2017TQ04X082), Guangdong Provincial Department of Science and Technology (Grant No. 2018B030306028), the Science and Technology Innovation Commission of Shenzhen (Grant Nos. JCYJ20180507182042530, JCYJ20180507182000722, JCYJ20170818143618288 and JCYJ20180305125314948), Shenzhen Peacock Technological Innovation Project (Grant Nos. KQJSCX20170727100433270 and KQJSCX20170327150812967), NTUT-SZU Joint Research Program (Grant No. 2019006) and the Natural Science Foundation of SZU.
Conflict of Interest
The authors declare no conflict of interest.
Keywords
artificial synapses, neuromorphic computing, photonic computing, photonic memristors, resistive switching
Received: May 7, 2019 Revised: July 31, 2019 Published online: August 26, 2019
[1] D. B. Strukov, G. S. Snider, D. R. Stewart, R. S. Williams, Nature 2008, 453, 80.
[2] J. J. Yang, D. B. Strukov, D. R. Stewart, Nat. Nanotechnol. 2013, 8, 13. [3] Y. Yang, R. Huang, Nat. Electron. 2018, 1, 274.
[4] D. P. Sahu, S. N. Jammalamadaka, Sci. Rep. 2017, 7, 17224.
[5] Y. Yang, P. Gao, S. Gaba, T. Chang, X. Pan, W. Lu, Nat. Commun.
2012, 3, 732.
[6] S. H. Jo, T. Chang, I. Ebong, B. B. Bhadviya, P. Mazumder, W. Lu,
Nano Lett. 2010, 10, 1297.
[7] L. Qi, L. Shibing, L. Hangbing, W. Wei, N. Jiebin, H. Zongliang,
C. Junning, L. Ming, ACS Nano 2010, 4, 6162.
[8] C. Li, B. Gao, Y. Yao, X. Guan, X. Shen, Y. Wang, P. Huang, L. Liu,
X. Liu, J. Li, C. Gu, J. Kang, R. Yu, Adv. Mater. 2017, 29, 1602976. [9] B. G. Chae, J. B. Seol, J. H. Song, K. Baek, S. H. Oh, H. Hwang,
C. G. Park, Adv. Mater. 2017, 29, 1701752.
[10] J. Zheng, J. Zhang, Z. Wang, L. Zhong, Y. Sun, Z. Liang, Y. Li,
L. Jiang, X. Chen, L. Chi, Adv. Mater. 2018, 30, 1802731.
[11] H. Tan, G. Liu, H. Yang, X. Yi, L. Pan, J. Shang, S. Long, M. Liu,
Y. Wu, R.-W. Li, ACS Nano 2017, 11, 11298.
[12] K. Pei, X. Ren, Z. Zhou, Z. Zhang, X. Ji, P. K. L. Chan, Adv. Mater.
2018, 30, 1706647.
[13] J. Gorecki, V. Apostolopoulos, J. Y. Ou, S. Mailis, N. Papasimakis,
ACS Nano 2018, 12, 5940.
[14] Y. Shen, N. C. Harris, S. Skirlo, M. Prabhu, T. Baehr-Jones,
M. Hochberg, X. Sun, S. Zhao, H. Larochelle, D. Englund,
M. Soljac?ic ´, Nat. Photonics 2017, 11, 441.
[15] Z. Cheng, C. Ríos, W. H. P. Pernice, C. D. Wright, H. Bhaskaran,
Sci. Adv. 2017, 3, e1700160.
[16] W. Xu, H. Cho, Y. H. Kim, Y. T. Kim, C. Wolf, C. G. Park, T. W. Lee,
Adv. Mater. 2016, 28, 5916.
[17] B. Gholipour, P. Bastock, C. Craig, K. Khan, D. Hewak, C. Soci,
Adv. Opt. Mater. 2015, 3, 635.
[18] D. Xiang, T. Liu, J. Xu, J. Y. Tan, Z. Hu, B. Lei, Y. Zheng, J. Wu,
A. H. C. Neto, L. Liu, W. Chen, Nat. Commun. 2018, 9, 2966.
[19] Y. Zhai, J.-Q. Yang, Y. Zhou, J. Mao, Y. Ren, V. A. L. Roy, S.-T. Han,
Mater. Horiz. 2018, 5, 641.
Adv. Optical Mater. 2019, 7, 1900766 1900766 (13 of 15) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
In summary, inspired by the improved data communication speed of fiber optics and the high parallelism of optogenetics- tunable biological synapses, photonic memristors with on-chip integration offer a promising approach to develop future com- puting systems past the Moore’s law and von Neumann era. Photonic memristors and their applications in photonic memory and computing functions were discussed in detail with regard to the physical mechanism. Emulation on optogenetics-modu- lated artificial synapses and further artificial neural networks with light-tunable and enhanced synaptic behaviors were pre- sented. Moreover, light-involved logic operations, system-level integration, and light-controlled artificial synaptic memristors along with improved learning tasks performance were pre- sented. However, compared with traditional CMOS technology, photonic memristor applications are very straightforward, though many challenges remain. The most important challenge is the lack of comprehensive understanding of the microscopic mechanism of the RS under light illumination, since the opera- tion principle is more complicated by considering photovoltaic and photogating effects. In addition, there remains a vacancy in RS behavior in view of both the increase and decrease in resistance of optoelectronic memristors by means of com- plete light control. Moreover, requirements on the broadband response or specific narrow wavelength response can vary sig- nificantly depending on the applications. A general constraint in the photonic memristors lies in their inferior IR sensitivity as opposed to their UV–vis sensitivity, making the broadband absorption and discrimination between the signals of different wavelength difficult. Thus, the energy band optimization and the effects limiting the optical-electrical conversion need to be elucidated. Finally, technique issues, such as employing trans- parent electrodes instead of opaque metals and integrating light sources with photonic memristors, should be addressed. Crea- tive and multidiscipline research, including chemistry, physics, electrical engineering, computer science, and neuroscience, is needed to eventually apply photonic memristors in novel photonic and neuromorphic computing systems.
Acknowledgements
J.Y.M. and L.Z. contributed equally to this work. The authors thank Mr. Junhong Li for his kind help in the manuscript writing. The authors
www.advancedsciencenews.com
www.advopticalmat.de
[20] B. Gao, Y. Bi, H.-Y. Chen, R. Liu, P. Huang, B. Chen, L. Liu, X. Liu, S. Yu, H. S. P. Wong, J. Kang, ACS Nano 2014, 8, 6998.
[21] C. Wu, T. W. Kim, H. Y. Choi, D. B. Strukov, J. J. Yang, Nat. Commun. 2017, 8, 752.
[22] P. Russo, M. Xiao, R. Liang, N. Y. Zhou, Adv. Electron. Mater. 2018, 28, 1706230.
[23] H. Tan, G. Liu, X. Zhu, H. Yang, B. Chen, X. Chen, J. Shang, W. D. Lu, Y. Wu, R. W. Li, Adv. Mater. 2015, 27, 2797.
[24] K. A. Paller, A. D. Wagner, Trends Cognit. Sci. 2002, 6, 93.
[25] L. Q. Zhu, C. J. Wan, L. Q. Guo, Y. Shi, Q. Wan, Nat. Commun.
2014, 5, 3158.
[26] Z. Wang, S. Joshi, S. E. Savel’ev, H. Jiang, R. Midya, P. Lin, M. Hu,
N. Ge, J. P. Strachan, Z. Li, Q. Wu, M. Barnell, G. L. Li, H. L. Xin,
R. S. Williams, Q. Xia, J. J. Yang, Nat. Mater. 2017, 16, 101.
[27] H. Tian, W. Mi, H. Zhao, M. A. Mohammad, Y. Yang, P. W. Chiu,
T. L. Ren, Nanoscale 2017, 9, 9275.
[28] C. Du, W. Ma, T. Chang, P. Sheridan, W. D. Lu, Adv. Funct. Mater.
2015, 25, 4290.
[29] L. Fenno, O. Yizhar, K. Deisseroth, Annu. Rev. Neurosci. 2011, 34,
389.
[30] O. Yizhar, L. E. Fenno, T. J. Davidson, M. Mogri, K. Deisseroth,
Neuron 2011, 71, 9.
[31] B. L. Sinnen, A. B. Bowen, J. S. Forte, B. G. Hiester, K. C. Crosby,
E. S. Gibson, M. L. Dell’Acqua, M. J. Kennedy, Neuron 2017, 93,
646.
[32] X. Liu, S. Ramirez, P. T. Pang, C. B. Puryear, A. Govindarajan,
K. Deisseroth, S. Tonegawa, Nature 2012, 484, 381.
[33] K. Deisseroth, Nat. Methods 2011, 8, 26.
[34] X. Zhu, W. D. Lu, ACS Nano 2018, 12, 1242.
[35] Z. H. Tan, R. Yang, K. Terabe, X. B. Yin, X. D. Zhang, X. Guo, Adv.
Mater. 2016, 28, 377.
[36] A. Rani, D. B. Velusamy, F. Marques Mota, Y. H. Jang, R. H. Kim,
C. Park, D. H. Kim, Adv. Funct. Mater. 2017, 27, 1604604.
[37] Y. Shi, X. Liang, B. Yuan, V. Chen, H. Li, F. Hui, Z. Yu, F. Yuan,
E. Pop, H. S. P. Wong, M. Lanza, Nat. Electron. 2018, 1, 458.
[38] J. Lee, W. D. Lu, Adv. Mater. 2018, 30, 1702770.
[39] L. Zhou, J. Mao, Y. Ren, S. T. Han, V. A. L. Roy, Y. Zhou, Small
2018, 14, 1703126.
[40] R. Waser, R. Dittmann, G. Staikov, K. Szot, Adv. Mater. 2009, 21,
2632.
[41] W. Lee, Y. Kim, Y. Song, K. Cho, D. Yoo, H. Ahn, K. Kang, T. Lee,
Adv. Funct. Mater. 2018, 28, 1801162.
[42] J. H. Yoon, J. Zhang, X. Ren, Z. Wang, H. Wu, Z. Li, M. Barnell,
Q. Wu, L. J. Lauhon, Q. Xia, J. J. Yang, Adv. Funct. Mater. 2017, 27,
1702010.
[43] J. H. Yoon, Z. Wang, K. M. Kim, H. Wu, V. Ravichandran, Q. Xia,
C. S. Hwang, J. J. Yang, Nat. Commun. 2018, 9, 417.
[44] H. Ling, M. Yi, M. Nagai, L. Xie, L. Wang, B. Hu, W. Huang, Adv.
Mater. 2017, 29, 1701333.
[45] M. D. Tran, H. Kim, J. S. Kim, M. H. Doan, T. K. Chau, Q. A. Vu,
J.-H. Kim, Y. H. Lee, Adv. Mater. 2019, 31, 1807075.
[46] K. A. Mazzio, C. K. Luscombe, Chem. Soc. Rev. 2015, 44, 78.
[47] Z. Xiao, Y. Yuan, Y. Shao, Q. Wang, Q. Dong, C. Bi, P. Sharma,
A. Gruverman, J. Huang, Nat. Mater. 2015, 14, 193.
[48] V. Adinolfi, E. H. Sargent, Nature 2017, 542, 324.
[49] S. Gao, G. Liu, H. Yang, C. Hu, Q. Chen, G. Gong, W. Xue, X. Yi,
[54] X. Guan, W. Hu, M. A. Haque, N. Wei, Z. Liu, A. Chen, T. Wu, Adv. Funct. Mater. 2018, 28, 1704665.
[55] L. Zhou, J. Y. Mao, Y. Ren, J. Q. Yang, S. R. Zhang, Y. Zhou, Q. Liao, Y. J. Zeng, H. Shan, Z. Xu, J. Fu, Y. Wang, X. Chen, Z. Lv, S. T. Han, V. A. L. Roy, Small 2018, 14, 1800288.
[56] X. Zhu, J. Lee, W. D. Lu, Adv. Mater. 2017, 29, 1700527.
[57] J. Choi, S. Park, J. Lee, K. Hong, D. H. Kim, C. W. Moon, G. D. Park, J. Suh, J. Hwang, S. Y. Kim, H. S. Jung, N. G. Park,
S. Han, K. T. Nam, H. W. Jang, Adv. Mater. 2016, 28, 6562.
[58] Y. Wang, J. Yang, Z. Wang, J. Chen, Q. Yang, Z. Lv, Y. Zhou, Y. Zhai,
Z. Li, S. T. Han, Small 2019, 15, 1805431.
[59] H. Fang, W. Hu, Adv. Sci. 2017, 4, 1700323.
[60] M. Wang, S. Cai, C. Pan, C. Wang, X. Lian, Y. Zhuo, K. Xu, T. Cao,
X. Pan, B. Wang, S.-J. Liang, J. J. Yang, P. Wang, F. Miao, Nat.
Electron. 2018, 1, 130.
[61] H. Zhang, ACS Nano 2015, 9, 9451.
[62] Y. Zhai, X. Yang, F. Wang, Z. Li, G. Ding, Z. Qiu, Y. Wang, Y. Zhou,
S. T. Han, Adv. Mater. 2018, 30, 1803563.
[63] X. Zhou, Q. Zhang, L. Gan, H. Li, T. Zhai, Adv. Funct. Mater. 2016,
26, 4405.
[64] S. Bertolazzi, D. Krasnozhon, A. Kis, ACS Nano 2013, 7, 3246.
[65] Q. A. Vu, Y. S. Shin, Y. R. Kim, V. L. Nguyen, W. T. Kang, H. Kim,
D. H. Luong, I. M. Lee, K. Lee, D. S. Ko, J. Heo, S. Park, Y. H. Lee,
W. J. Yu, Nat. Commun. 2016, 7, 12725.
[66] J. Lee, S. Pak, Y. W. Lee, Y. Cho, J. Hong, P. Giraud, H. S. Shin,
S. M. Morris, J. I. Sohn, S. Cha, J. M. Kim, Nat. Commun. 2017,
8, 14734.
[67] Z. Lv, Y. Wang, Z. Chen, L. Sun, J. Wang, M. Chen, Z. Xu, Q. Liao,
L. Zhou, X. Chen, J. Li, K. Zhou, Y. Zhou, Y.-J. Zeng, S.-T. Han,
V. A. L. Roy, Adv. Sci. 2018, 5, 1800714.
[68] Y. Wang, Z. Lv, Q. Liao, H. Shan, J. Chen, Y. Zhou, L. Zhou,
X. Chen, V. A. L. Roy, Z. Wang, Z. Xu, Y. J. Zeng, S. T. Han, Adv.
Mater. 2018, 30, 1800327.
[69] J. Xu, X. Zhao, Z. Wang, H. Xu, J. Hu, J. Ma, Y. Liu, Small 2019, 15,
1803970.
[70] Y. Zhou, S.-T. Han, X. Chen, F. Wang, Y.-B. Tang, V. A. L. Roy, Nat.
Commun. 2014, 5, 4720.
[71] Q. Yan, Z. Luo, K. Cai, Y. Ma, D. Zhao, Chem. Soc. Rev. 2014, 43,
4199.
[72] M. A. Tasdelen, Y. Yagci, Angew. Chem., Int. Ed. 2013, 52,
5930.
[73] X. Chen, X. Zhu, S.-R. Zhang, J. Pan, P. Huang, C. Zhang, G. Ding,
Y. Zhou, K. Zhou, V. A. L. Roy, S.-T. Han, Adv. Mater. Technol. 2019,
4, 1800551.
[74] X. Zhao, Z. Wang, Y. Xie, H. Xu, J. Zhu, X. Zhang, W. Liu, G. Yang,
J. Ma, Y. Liu, Small 2018, 14, 1801325.
[75] Q. Zhang, J. He, H. Zhuang, H. Li, N. Li, Q. Xu, D. Chen, J. Lu,
Adv. Funct. Mater. 2016, 26, 146.
[76] Q.-D. Ling, Y. Song, S.-L. Lim, E. Y.-H. Teo, Y.-P. Tan, C. Zhu,
D. S. H. Chan, D.-L. Kwong, E.-T. Kang, K.-G. Neoh, Angew. Chem.
2006, 118, 3013.
[77] S. Miao, H. Li, Q. Xu, Y. Li, S. Ji, N. Li, L. Wang, J. Zheng, J. Lu,
Adv. Mater. 2012, 24, 6210.
[78] P.-Y. Gu, F. Zhou, J. Gao, G. Li, C. Wang, Q.-F. Xu, Q. Zhang,
J.-M. Lu, J. Am. Chem. Soc. 2013, 135, 14086.
[79] B. Sun, M. Tang, J. Gao, C. M. Li, ChemElectro Chem 2016, 3,
J. Shang, R.-W. Li, ACS Nano 2019, 13, 2634. 896.
[50] A. Bera, H. Peng, J. Lourembam, Y. Shen, X. W. Sun, T. Wu, Adv. Funct. Mater. 2013, 23, 4977.
[51] M. Ungureanu, R. Zazpe, F. Golmar, P. Stoliar, R. Llopis, F. Casanova, L. E. Hueso, Adv. Mater. 2012, 24, 2496.
[52] D. S. Shang, J. R. Sun, L. Shi, B. G. Shen, Appl. Phys. Lett. 2008, 93, 102106.
[80] C.-L. Liu, W.-C. Chen, Polym. Chem. 2011, 2, 2169.
[81] C. Ye, Q. Peng, M. Li, J. Luo, Z. Tang, J. Pei, J. Chen, Z. Shuai,
L. Jiang, Y. Song, J. Am. Chem. Soc. 2012, 134, 20053.
[82] Y.Li,H.Li,J.He,Q.Xu,N.Li,D.Chen,J.Lu,Chem.-AsianJ.
2016, 11, 2078.
[83] M. Irie, Chem. Rev. 2000, 100, 1685.
[84] M. Irie, T. Fukaminato, T. Sasaki, N. Tamai, T. Kawai, Nature 2002,
420, 759.
Adv. Optical Mater. 2019, 7, 1900766 1900766 (14 of 15) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
[53] O. S. Game, G. J. Buchsbaum, Y. Zhou, N. P. Padture, A. I. Kingon, Adv. Funct. Mater. 2017, 27, 1606584.
www.advancedsciencenews.com
www.advopticalmat.de
[85] H. Ling, K. Tan, Q. Fang, X. Xu, H. Chen, W. Li, Y. Liu, L. Wang, M. Yi, R. Huang, Y. Qian, L. Xie, W. Huang, Adv. Electron. Mater. 2017, 3, 1600416.
[86] Y. L. Kim, H. Y. Jung, S. Park, B. Li, F. Liu, J. Hao, Y.-K. Kwon, Y. J. Jung, S. Kar, Nat. Photonics 2014, 8, 239.
[87] Y. Wu, Y. Wei, Y. Huang, F. Cao, D. Yu, X. Li, H. Zeng, Nano Res. 2017, 10, 1584.
[88] Z. Sun, E. Ambrosi, A. Bricalli, D. Ielmini, Adv. Mater. 2018, 30, 1802554.
[89] K. Mathieson, J. Loudin, G. Goetz, P. Huie, L. Wang, T. I. Kamins, L. Galambos, R. Smith, J. S. Harris, A. Sher, D. Palanker, Nat. Photonics 2012, 6, 391.
[90] C. Choi, M. K. Choi, S. Liu, M. S. Kim, O. K. Park, C. Im, J. Kim, X. Qin, G. J. Lee, K. W. Cho, M. Kim, E. Joh, J. Lee, D. Son, S.-H. Kwon, N. L. Jeon, Y. M. Song, N. Lu, D.-H. Kim, Nat. Commun. 2017, 8, 1664.
[91] S. Chen, Z. Lou, D. Chen, G. Shen, Adv. Mater. 2018, 30, 1705400.
[92] W. Wu, X. Wang, X. Han, Z. Yang, G. Gao, Y. Zhang, J. Hu, Y. Tan,
A. Pan, C. Pan, Adv. Mater. 2019, 31, 1805913.
[93] J. Lee, B. Yoo, H. Lee, G. D. Cha, H. S. Lee, Y. Cho, S. Y. Kim,
H. Seo, W. Lee, D. Son, M. Kang, H. M. Kim, Y. I. Park, T. Hyeon,
D. H. Kim, Adv. Mater. 2017, 29, 1603169.
[94] H. Wang, H. Liu, Q. Zhao, Z. Ni, Y. Zou, J. Yang, L. Wang, Y. Sun,
Y. Guo, W. Hu, Y. Liu, Adv. Mater. 2017, 29, 1701772.
[95] L. Hu, S. Fu, Y. Chen, H. Cao, L. Liang, H. Zhang, J. Gao, J. Wang,
F. Zhuge, Adv. Mater. 2017, 29, 1606927.
[96] S. Chen, A. Z. Weitemier, X. Zeng, L. He, X. Wang, Y. Tao,
A. J. Y. Huang, Y. Hashimotodani, M. Kano, H. Iwasaki, L. K. Parajuli, S. Okabe, D. B. L. Teh, A. H. All, I. Tsutsui-Kimura, K. F. Tanaka, X. Liu, T. J. McHugh, Science 2018, 359, 679.
[97] Z. Wang, M. Hu, X. Ai, Z. Zhang, B. Xing, Adv. Biosyst. 2019, 3, 1800233.
[98] T. Kyung, S. Lee, J. E. Kim, T. Cho, H. Park, Y. M. Jeong, D. Kim, A. Shin, S. Kim, J. Baek, J. Kim, N. Y. Kim, D. Woo, S. Chae, C. H. Kim, H. S. Shin, Y. M. Han, D. Kim, W. D. Heo, Nat. Biotechnol. 2015, 33, 1092.
[99] R. S. Zucker, W. G. Regehr, Annu. Rev. Physiol. 2002, 64, 355.
[100] D. E. Clapham, Cell 2007, 131, 1047.
[101] M. K. Kim, J. S. Lee, ACS Nano 2018, 12, 1680.
[102] J.-Y. Mao, L. Zhou, Y. Ren, J.-Q. Yang, C.-L. Chang, H.-C. Lin,
H.-H. Chou, S.-R. Zhang, Y. Zhou, S.-T. Han, J. Mater. Chem. C
2019, 7, 1491.
[103] J. Sun, S. Oh, Y. Choi, S. Seo, M. J. Oh, M. Lee, W. B. Lee, P. J. Yoo,
J. H. Cho, J.-H. Park, Adv. Funct. Mater. 2018, 28, 1804397.
[104] J.-Y. Mao, L. Hu, S.-R. Zhang, Y. Ren, J.-Q. Yang, L. Zhou,
Y.-J. Zeng, Y. Zhou, S.-T. Han, J. Mater. Chem. C 2019, 7, 48.
[105] T. Tuma, A. Pantazi, M. Le Gallo, A. Sebastian, E. Eleftheriou, Nat.
Nanotechnol. 2016, 11, 693.
[106] [107] [108] [109]
[110]
[111] [112]
[113] [114] [115] [116] [117] [118] [119] [120]
[121] [122] [123] [124] [125]
[126] [127]
H. S. P. Wong, S. Raoux, S. Kim, J. Liang, J. P. Reifenberg, B. Rajendran, M. Asheghi, K. E. Goodson, Proc. IEEE 2010, 98, 2201. D. C. Hu, R. Yang, L. Jiang, X. Guo, ACS Appl. Mater. Interfaces 2018, 10, 6463.
M. Kumar, S. Abbas, J. Kim, ACS Appl. Mater. Interfaces 2018, 10, 34370.
S. Wang, C. Chen, Z. Yu, Y. He, X. Chen, Q. Wan, Y. Shi, D. W. Zhang, H. Zhou, X. Wang, P. Zhou, Adv. Mater. 2019, 31, 1806227.
L. Qian, Y. Sun, M. Wu, C. Li, D. Xie, L. Ding, G. Shi, Nanoscale 2018, 10, 6837.
Z. Xiao, J. Huang, Adv. Electron. Mater. 2016, 2, 1600100.
H. Tian, Q. Guo, Y. Xie, H. Zhao, C. Li, J. J. Cha, F. Xia, H. Wang, Adv. Mater. 2016, 28, 4991.
M. Lee, W. Lee, S. Choi, J. W. Jo, J. Kim, S. K. Park, Y. H. Kim, Adv. Mater. 2017, 29, 1700951.
S. Qin, F. Wang, Y. Liu, Q. Wan, X. Wang, Y. B. Xu, S. Yi, X. Wang, Z. Rong, 2D Mater. 2017, 4, 035022.
S. Dai, X. Wu, D. Liu, Y. Chu, K. Wang, B. Yang, J. Huang, ACS Appl. Mater. Interfaces 2018, 10, 21472.
X. Wu, Y. Chu, R. Liu, H. E. Katz, J. Huang, Adv. Sci. 2017, 4, 1700442.
R. A. John, F. Liu, N. A. Chien, M. R. Kulkarni, C. Zhu, Q. Fu, A. Basu, Z. Liu, N. Mathews, Adv. Mater. 2018, 30, 1800220.
K. Wang, S. Dai, Y. Zhao, Y. Wang, C. Liu, J. Huang, Small 2019, 15, 1900010.
Y. Chen, Y. Chu, X. Wu, W. Ou-Yang, J. Huang, Adv. Mater. 2017, 29, 1704062.
X. Yan, Y. Pei, H. Chen, J. Zhao, Z. Zhou, H. Wang, L. Zhang, J. Wang, X. Li, C. Qin, G. Wang, Z. Xiao, Q. Zhao, K. Wang, H. Li, D. Ren, Q. Liu, H. Zhou, J. Chen, P. Zhou, Adv. Mater. 2019, 31, 1805284.
S. Ham, S. Choi, H. Cho, S.-I. Na, G. Wang, Adv. Funct. Mater. 2019, 29, 1806646.
G. Agnus, W. Zhao, V. Derycke, A. Filoramo, Y. Lhuillier, S. Lenfant, D. Vuillaume, C. Gamrat, J. P. Bourgoin, Adv. Mater. 2010, 22, 702. X. Yang, Y. Fang, Z. Yu, Z. Wang, T. Zhang, M. Yin, M. Lin, Y. Yang, Y. Cai, R. Huang, Nanoscale 2016, 8, 18897.
M. Prezioso, F. Merrikh-Bayat, B. D. Hoskins, G. C. Adam, K. K. Likharev, D. B. Strukov, Nature 2015, 521, 61.
Y. Lee, J. Y. Oh, W. Xu, O. Kim, T. R. Kim, J. Kang, Y. Kim, D. Son, J. B.-H. Tok, M. J. Park, Z. Bao, T.-W. Lee, Sci. Adv. 2018, 4, eaat7387.
H. Wang, Q. Zhao, Z. Ni, Q. Li, H. Liu, Y. Yang, L. Wang, Y. Ran, Y. Guo, W. Hu, Y. Liu, Adv. Mater. 2018, 30, 1803961.
H. E. Lee, J. H. Park, T. J. Kim, D. Im, J. H. Shin, D. H. Kim, B. Mohammad, I.-S. Kang, K. J. Lee, Adv. Funct. Mater. 2018, 28, 1801690.
Adv. Optical Mater. 2019, 7, 1900766 1900766 (15 of 15) © 2
Quantum effects in memristive devices
by INRIM - Istituto Nazionale di Ricerca Metrologica
Quantum effects in memristive devices - The MEMQuD Project
Quantum conductance phenomena in semiconductors and metals. a) Schematic representation of a semiconductor-based device showing conductance quantization, where a 2DEG is formed at the interface of a heterojunction. The quantum point contact is realized by applying a negative voltage to the gate electrodes while measuring transport properties through contacts to the 2DEG at either side of the constriction. The constriction width (W) can be varied by means of the applied gate voltage. b) Schematic representation of a metallic-based device where conductance quantization can be observed when the metallic contact is of atomic dimensions. Credit: Gianluca Milano et al, Advanced Materials (2022). DOI: 10.1002/adma.202201248
At the nanoscale, the laws of classical physics suddenly become inadequate to explain the behavior of matter. It is precisely at this juncture that quantum theory comes into play, effectively describing the physical phenomena characteristic of the atomic and subatomic world. Thanks to the different behavior of matter on these length and energy scales, it is possible to develop new materials, devices and technologies based on quantum effects, which could yield a real quantum revolution that promises to innovate areas such as cryptography, telecommunications and computation.
The physics of very small objects, already at the basis of many technologies that we use today, is intrinsically linked to the world of nanotechnologies, the branch of applied science dealing with the control of matter at the nanometer scale (a nanometer is one billionth of a meter). This control of matter at the nanoscale is at the basis of the development of new electronic devices.
Among these, memristors are considered promising devices for the realization of new computational architectures emulating functions of our brain, allowing the creation of increasingly efficient computation systems suitable for the development of the entire artificial intelligence sector, as recently shown by Istituto Nazionale di Ricerca Metrologica (INRiM) researchers in collaboration with several international universities and research institutes.
In this context, the EMPIR MEMQuD project, coordinated by INRiM, aims to study the quantum effects in such devices, in which the electronic conduction properties can be manipulated to allow the observation of quantized conductivity phenomena at room temperature. In addition to analyzing the fundamentals and recent developments, the review work "Quantum Conductance in Memristive Devices: Fundamentals, Developments, and Applications" recently published in the journal Advanced Materials, analyzes how these effects can be used for a wide range of applications, from metrology to the development of next-generation memories and artificial intelligence.
Doc. Ya think that is because Ploinks is only approved for use in the U S A ? Or?
O T mahalo Doc. Also accepted at U Cinti honors college and U Conn honors college. Zoom meetings so far with Yale , Duke and U Penn. Lots more coming.
We need some news from our BoD. All this silence is for what? I don’t get it. Makes no sense to me. Why would it have any bearing on the TSC decision? Any c e o worth his salt would never treat their shareholders like this. It’s totally uncalled for. Am I missing something?
O T Doc. Lehua just awarded $251,000.00 super free ride at Miami University Honors college in Oxford Oh. Several other scholarships etc. 13 to go out of 18 apps. On way.back from DC and USSYP scholarship she won. Met Joe Jill at WH Antony Blinken Butigeggeeeeggeeeee lol Senators Ambassadors Chairman of joint chief of staff etc at Pentagon .Capitol Hill Unreal. Ivy leagues coming in. Accepted at U Conn $ U Va. Ethos program , others. Interviewed by Yale Duke etc etc unreal kid. She’ll have her pick 4.333
I’d be willing to bet news comes out of some kind of deal/lic/ jv with Broadcom /sbv. We’ll see pretty soon imo T S C should rule within 30 days if not sooner
Wade was caught lying again our attorney filed another brief proving it in black and white with court records. The court asked for the records. Wade has 15 days after the 18th to reply to us at the latest maybe sooner? I doubt his attorney will. It’s hard to dispute facts in black and white on paper records. He’s done. The T SC will dismiss this no later than first week in April imo. 🎉😎🤙
Sounds like we’re in there! etl , sql, arbitrary object framework etc all working together to make AI and ML work . VCSY is The Glue. All the biggies saw it 15 years ago. That’s why they /mirror etc fought so hard to destroy us. We prevailed in our Markman Hearing. After that they all had to go to plan B With Wade gone it’s plan C . The 18th can not come too soon. T S C will dismiss Wade’s case imo and then it’s everyone get in line for a license.