Category Archives: Applications

Spectrometers — devices that distinguish different wavelengths of light and are used to determine the chemical composition of everything from laboratory materials to distant stars — are large devices with six-figure price tags, and tend to be found in large university and industry labs or observatories.

A collection of mini-spectrometer chips are arrayed on a tray after being made through conventional chip-making processes. Credit: Felice Frankel

A new advance by researchers at MIT could make it possible to produce tiny spectrometers that are just as accurate and powerful but could be mass produced using standard chip-making processes. This approach could open up new uses for spectrometry that previously would have been physically and financially impossible.

The invention is described today in the journal Nature Communications, in a paper by MIT associate professor of materials science and engineering Juejun Hu, doctoral student Derek Kita, research assistant Brando Miranda, and five others.

The researchers say this new approach to making spectrometers on a chip could provide major advantages in performance, size, weight, and power consumption, compared to current instruments.

Other groups have tried to make chip-based spectrometers, but there is a built-in challenge: A device’s ability to spread out light based on its wavelength, using any conventional optical system, is highly dependent on the device’s size. “If you make it smaller, the performance degrades,” Hu says.

Another type of spectrometer uses a mathematical approach called a Fourier transform. But these devices are still limited by the same size constraint — long optical paths are essential to attaining high performance. Since high-performance devices require long, tunable optical path lengths, miniaturized spectrometers have traditionally been inferior compared to their benchtop counterparts.

Instead, “we used a different technique,” says Kita. Their system is based on optical switches, which can instantly flip a beam of light between the different optical pathways, which can be of different lengths. These all-electronic optical switches eliminate the need for movable mirrors, which are required in the current versions, and can easily be fabricated using standard chip-making technology.

By eliminating the moving parts, Kita says, “there’s a huge benefit in terms of robustness. You could drop it off the table without causing any damage.”

By using path lengths in power-of-two increments, these lengths can be combined in different ways to replicate an exponential number of discrete lengths, thus leading to a potential spectral resolution that increases exponentially with the number of on-chip optical switches. It’s the same principle that allows a balance scale to accurately measure a broad range of weights by combining just a small number of standard weights.

As a proof of concept, the researchers contracted an industry-standard semiconductor manufacturing service to build a device with six sequential switches, producing 64 spectral channels, with built-in processing capability to control the device and process its output. By expanding to 10 switches, the resolution would jump to 1,024 channels. They designed the device as a plug-and-play unit that could be easily integrated with existing optical networks.

The team also used new machine-learning techniques to reconstruct detailed spectra from a limited number of channels. The method they developed works well to detect both broad and narrow spectral peaks, Kita says. They were able to demonstrate that its performance did indeed match the calculations, and thus opens up a wide range of potential further development for various applications.

The researchers say such spectrometers could find applications in sensing devices, materials analysis systems, optical coherent tomography in medical imaging, and monitoring the performance of optical networks, upon which most of today’s digital networks rely. Already, the team has been contacted by some companies interested in possible uses for such microchip spectrometers, with their promise of huge advantages in size, weight, and power consumption, Kita says. There is also interest in applications for real-time monitoring of industrial processes, Hu adds, as well as for environmental sensing for industries such as oil and gas.

Two leading French and Taiwanese research institutes today announced their new collaboration to facilitate a scientific and technological exchange between France and Taiwan.

Leti, a research institute of CEA Tech in Grenoble, France, and the Taiwanese National Applied Research Laboratories (NARLabs), two key nanotechnology research providers in their respective countries, will explore opportunities for joint research-and-development projects in high-performance computing and networks, photonics, bio-medical nanotechnologies and brain-computer interface. Their scientists will meet in a series of workshops to initiate joint R&D projects. This agreement also includes access to each other’s unique equipment and platforms, and will offer opportunities to researchers with a specific exchange program.

The agreement was signed by CEA-Leti CEO Emmanuel Sabonnadière and NARLabs President Yeong-Her Wang during the recent Leti Day Taiwan in Hsinchu.

“CEA-Leti and NARLabs have the same goals: to create differentiating technologies and transfer them to industry,” Sabonnadière said. “This cooperation agreement will be the starting point for a strategic research cooperation between our organizations that will strengthen R&D and inspire microelectronics innovation in both Taiwan and France.”

“The National Chip Implementation Center (CIC) and the National Nano Device Laboratories (NDL) of National Applied Research Laboratories (NARLabs) have fostered close ties with CEA-Leti since 2017,” said NARLabs Vice President Wu Kuang-Chong. “Around the Leti Day Taiwan, we held seminars together, and our researchers were able to meet and exchange ideas. Topics included silicon photonics, intelligent image sensors, RF technology, 3D IC+ and device fabrication technology, among others. We believe that with this memorandum of understanding, CEA-Leti and NARLabs will continue to collaborate together to complement and to enlighten each other to formulate innovative research projects.”

By Nishita Rao

Nicolas Sauvage, senior director of Ecosystem at TDK InvenSense, will present at the fast-approaching MEMS & Sensors Executive Congress on October 29-30, 2018 in Napa, Calif. SEMI’s Nishita Rao spoke with Sauvage to offer MSEC attendees advance insights on Sauvage’s feature presentation.

SEMI: What is “autonomy value” and why is it important?

Sauvage: How do you increase the perceived value of an electronic device? If it’s an autonomous car, its value is closely tied to the autonomy level — i.e., the independence — that it offers people. Higher autonomy value for a self-driving car, for example, means that even a blind person could use it. It’s been almost two years since Waymo demonstrated this, and here’s the video that shows it.

Countless other sensor-based electronic products have their own “autonomy value.” Imagine the need to get medicine to people during a humanitarian health crisis. Drones could be your best option because they can deliver to inaccessible or remote locations. Unlike older drones, which require active piloting by a person, a drone with higher autonomy value could deliver medicine to Doctors Without Borders without ongoing human intervention.

This drone could navigate objects, such as trees and birds, and would have excellent location-awareness. It could fly through any landscape in bright sunlight or during the night. To increase the drone’s autonomy value, you would need better sensors, including those sensors that can enable sensing in sunny conditions or in pitch-black night, as well as better machine learning.

SEMI: In this example, what types of sensors would the drone manufacturer need?

Sauvage: The manufacturer would need a “surrounding-sensing” solution that includes ultrasonic and pressure sensors as well as image sensors. Start with high-quality image sensors combined with ultrasonic range-finding sensors — high-accuracy devices that function in all lighting conditions and can detect objects of any color. Add motion sensors and a pressure sensor, which would capture the height of the drone to make known the drone’s location in space. The drone would need this combination of sensors, plus smart sensor fusion, because GPS alone cannot avoid obstacles: its signal can be sporadic in certain parts of the world or in certain terrain, making it unreliable.

A key attribute of all these sensors would be low power consumption since the drone would run on battery.

SEMI: To what extent might autonomy value cause manufacturers to consider multi-vendor solutions?

Sauvage: I would like to see it inspire the MEMS and sensors ecosystem to work together, to arrive at multi-vendor solutions that will benefit humanity through greater autonomy value. Whether we’re looking at autonomous cars, drones, robotics or other applications, there are cases where we need to prioritize safety and security over industry competition.

SEMI: Where are we today in terms of achieving true autonomy value – and where are we going?

Sauvage: The sky is the limit, literally. Machine learning and surrounding-sensing solutions applied to cars, drones and robots will increase autonomy value to the point where we can justifiably call it artificial intelligence.

SEMI: What would you like MEMS & Sensors Executive Congress attendees to take away from your presentation?

Sauvage: I hope that attendees will recognize the value of ecosystem solutions in increasing autonomy value. Together we can expand the variety of sensor types that address novel use-cases and jobs-to-be-done. Instead of waiting for customers to ask for ecosystem-level solutions, we need to articulate a complete MEMS and sensors supply-chain ecosystem if we want the Internet of Things (IoT) and Industrial IoT (IIoT) to grow more quickly.

As senior director of Ecosystem, Nicholas Sauvage is responsible for all strategic relationships, including Google and Qualcomm, and other HW/SW/System companies. He is also responsible for strategic and market-driven goal-setting of our SensorStudio developer program, and driving select partnerships with SoC sensor hub platforms. Prior to joining InvenSense, Nicolas was part of NXP Software management team, responsible for worldwide sales, as well as for P&L and product management of their OEM Business Line. Nicolas is an alumnus of Institut supérieur d’électronique et du numérique, London Business School and INSEAD.

Register today to connect with Nicolas Sauvage at the event. You can also connect with him on LinkedIn.

Nishita Rao is a marketing manager at SEMI.

Computer bits are binary, with a value of 0 or 1. By contrast, neurons in the brain can have all kinds of different internal states, depending on the input that they received. This allows the brain to process information in a more energy-efficient manner than a computer. University of Groningen (UG) physicists are working on memristors, resistors with a memory, made from niobium-doped strontium titanate, which mimic how neurons work. Their results were published in the Journal of Applied Physics on 21 October.

The brain is superior to traditional computers in many ways. Brain cells use less energy, process information faster and are more adaptable. The way that brain cells respond to a stimulus depends on the information that they have received, which potentiates or inhibits the neurons. Scientists are working on new types of devices which can mimic this behavior, called memristors.

Memory

UG researcher Anouk Goossens, the first author of the paper, tested memristors made from niobium-doped strontium titanate. The conductivity of the memristors is controlled by an electric field in an analog fashion: ‘We use the system’s ability to switch resistance: by applying voltage pulses, we can control the resistance, and using a low voltage we read out the current in different states. The strength of the pulse determines the resistance in the device. We have shown a resistance ratio of at least 1000 to be realizable. We then measured what happened over time.’ Goossens was especially interested in the time dynamics of the resistance states.

She observed that the duration of the pulse with which the resistance was set determined how long the ‘memory’ lasted. This could be between one to four hours for pulses lasting between a second and two minutes. Furthermore, she found that after 100 switching cycles, the material showed no signs of fatigue.

Forgetting

‘There are different things you could do with this’, says Goossens. ‘By “teaching” the device in different ways, using different pulses, we can change its behavior.’ The fact that the resistance changes over time can also be useful: ‘These systems can forget, just like the brain. It allows me to use time as a variable parameter.’ In addition, the devices that Goossens made combine both memory and processing in one device, which is more efficient than traditional computer architecture in which storage (on magnetic hard discs) and processing (in the CPU) are separated.

Goossens conducted the experiments described in the paper during a research project as part of the Master in Nanoscience degree programme at the University of Groningen. Goossens’ research project took place within the group of students supervised by Dr. Tamalika Banerjee of Spintronics of Functional Materials. She is now a Ph.D. student in the same group.

Questions

Before building brain-like circuits with her device, Goossens plans to conduct experiments to really understand what happens within the material. ‘If we don’t know exactly how it works, we can’t solve any problems that might occur in these circuits. So, we have to understand the physical properties of the material: what does it do, and why?’

Questions that Goossens want to answer include what parameters influence the states that are achieved. ‘And if we manufacture 100 of these devices, do they all work the same? If they don’t, and there is device-to-device variation, that doesn’t have to be a problem. After all, not all elements in the brain are the same.’

STMicroelectronics (NYSE: STM) and Fidesmo, the contactless-services developer and Mastercard Approved Global Vendor, have created a turnkey active solution for secure contactless payments on smart watches and other wearable technology.

The complete payment system-on-chip (SoC) is based on ST’s STPay-Boost IC, which combines a hardware secure element to protect transactions and a contactless controller featuring proprietary active-boost technology that maintains reliable NFC connections even in devices made with metallic materials. Its single-chip footprint fits easily within wearable form factors.

Fidesmo’s MasterCard MDES tokenization platform completes the solution by allowing the user to load the personal data needed for payment transactions. Convenient Over-The-Air (OTA) technology makes personalization a simple step for the user without any special equipment.

STPay-Boost, featuring our secure element and performance-boosting active contactless technology, is a unique single-chip payment solution that fits the design constraints of wearable devices,” said Laurent Degauque, Marketing Director, Secure Microcontroller Division, STMicroelectronics. “Fidesmo’s personalization platform provides the vital ingredient to create a turnkey payment solution that device makers can simply take and use with minimal engineering and certification effort.

Our cooperation with STMicroelectronics opens up a new market for us with support for lightweight boosted secure elements, and broadens choices for our customers,” said Mattias Eld, CEO of Fidesmo. “Together, we have created a unique offering that is sure to impact the hybrid watch market and drive the emergence of innovative new products such as wristbands, bracelets, key fobs, and connected jewelry.

Kronaby, the Malmö, Sweden-based hybrid smart-watch maker, has embedded the STPay-Boost chip in its portfolio of men’s and women’s smart watches that offer differentiated features such as freedom from charging and filtered notifications. The SoC with Fidesmo tokenization enables Kronaby watches to support a variety of services such as payments, access control, transportation, and loyalty rewards.

Jonas Morän, Global Product Manager at Kronaby, said, “We have achieved an intuitive and seamless user experience leveraging the ability to communicate with the Secure Element in the watch using Bluetooth. In addition, the ST/Fidesmo chip’s boosted wireless performance and compact size gives us extra freedom to style our watches to maximize their appeal in our target markets.”

STPay-Boost chips with Fidesmo OTA personalization are sampling now to lead customers and are scheduled to enter production in November 2018, priced from $3.50 for orders of 1000 pieces (excluding Fidesmo license).

By Nishita Rao

ULVAC Technologies’ David Mount is working with The CIA. Is he the Jack Reacher of the MEMS and sensors industry, jetting around the world to secret meetings, you wonder? While David isn’t quite the super-spy that you might have imagined, he is doing some fascinating work on behalf of ULVAC Technologies, the world leader in vacuum technology.

ULVAC has been collaborating with The Culinary Institute of America (CIA) on Menus of Change, “a ground-breaking initiative from The Culinary Institute of America and Harvard T.H. Chan School of Public Health that works to realize a long-term, practical vision integrating optimal nutrition and public health, environmental stewardship and restoration, and social responsibility concerns within the foodservice industry and the culinary profession.”

ULVAC also partners with Menus of Change (MOC) University Research Collaborative, a group of elite universities and food-service executives working together to “accelerate efforts to move Americans toward healthier, more sustainable, plant-forward diets.”

MEMS & Sensors Industry Group’s Nishita Rao caught up with David, a featured speaker at MEMS & Sensors Executive Congress on October 29-30, 2018, in Napa, Calif. to give MSEC attendees a preview of David’s talk.

SEMI: How did ULVAC get involved with The CIA on Menus of Change?

Mount: People in the MEMS & sensors industry may not know that ULVAC started as an equipment supplier to the food industry. In 1952 ULVAC began supplying freeze-drying equipment – which relies on vacuum technology — to food companies tasked with providing long-lasting foods and beverages for the U.S. military under the Marshall Plan. Think instant soup, ramen noodles and Tang. While ULVAC’s technology portfolio is now very broad — spanning deposition equipment for the semiconductor industry, vacuum brazing for automotive, and even vacuum freeze-drying of vaccines that can be shipped dry but combined with distilled water for administration — the company has kept a hand in food technology. ULVAC’s vacuum cooling equipment rapidly and safely cools foods, dramatically increasing shelf life.

The CIA is at the forefront of innovation in food technology, so we worked with them to test a vacuum cooling system that can also be used in the kitchen or in the field. In the Central Valley of California, for example, it can be 104ºF in the fields where lettuce is picked; our vacuum cooling system can cool that lettuce down to 47ºF in minutes.

The CIA is also developing prepared foods for industrial settings such as university cafeterias and airlines. A prepared chicken dish, for example, might be cooked at 350ºF and then cooled to refrigeration temperatures. The potential problem is that bacteria can grow when you cool that food for storage. Some of The CIA test kitchens in California are using ULVAC’s vacuum cooling system to quickly and safely cool prepared foods.

Vacuum-cooling is just one stage in food production, of course. Sensors are also widely used in food production and safety.

SEMI: How do The CIA test kitchens use sensors?

Mount: Nearly all aspects of production, processing and management in agricultural and food systems involve measurement of product and resource attributes. Sensors are a natural fit here as they can provide inspection capabilities that are accurate, fast and consistent. I plan to dive into some specific examples of the ways that The CIA and the MOC Research Collaborative are employing sensors to increase the safety of food and agricultural production.

SEMI: What would you like MSEC attendees to take away from your presentation?

Mount: I love knowing that the work that we do in this industry can benefit humanity. Applying our various technologies to food and agricultural production is just one way to do that. I encourage MSEC attendees to explore those markets that improve human quality of life – as well as the life and health of our planet and its other inhabitants.

ULVAC Technologies senior advisor David Mount is a 35-year veteran of the vacuum and thin film equipment industry. He tried to retire from ULVAC but they would not let him go! David consults with ULVAC on strategic projects such as the company’s collaboration with the CIA.

He will present Sensors in Food and Agriculture on Tuesday, October 30 at the MEMS & Sensors Executive Congress.

Register today to learn more about how sensors are transforming the food industry.

Nishita Rao is a marketing manager at SEMI.

Originally published on the SEMI blog.

Cynthia Wright, a retired military officer with over 25 years of experience in national security and cyber strategy and policy, now Principal Cyber Security Engineer at The MITRE Corporation, will give the opening keynote at the upcoming MEMS & Sensors Executive Congress, October 29-30, 2018 in Napa, Calif. SEMI’s Maria Vetrano interviewed Wright to give MSEC attendees an advance look at Wright’s highly anticipated presentation.

SEMI: MEMS and sensors suppliers provide intelligent sensing and actuation to hundreds of billions of autonomous mobility devices – but historically, our community has not been at the forefront of cybersecurity. Why is now a good time for us to get involved?

Wright: From wearables, smartphones, refrigerators and agriculture to medical devices and military hardware, autonomous mobility devices pervade our lives. At the same time, Internet of Things (IoT) botnet attacks like Mirai — and other demonstrated cyberattacks on home devices, vehicles and infrastructure — highlight the increasingly urgent need to address cybersecurity and privacy in MEMS/sensors-enabled devices.

As building-block players in autonomous devices, MEMS and sensors suppliers have several good reasons to get involved.

The number of IoT cyber security bills before state and federal legislatures suggest that regulation is coming, and it is in everyone’s best interest to prepare. While original equipment manufacturers (OEMs) would generally be held liable in cases of component malfunction or data breach, if insecurity stems from a microelectromechanical component, OEMs would most likely choose component suppliers with secure products.

Beyond legislation and competitive advantage, we must consider that people’s well-being, even lives, could be at stake. Imagine what could happen if someone hacks into an insulin pump, the accelerometer on a train, or the LIDAR of an autonomous car. Intrusions of this sort could prove catastrophic.

SEMI: Where do you perceive the biggest potential threats to consumers, industry, government?

Wright: In good military fashion, I would say that it depends. If a person is a consumer of medical implants, that’s a big threat. On the government side, we could be talking about networked devices involved in military situational awareness. In industry, it could be sensors governing critical manufacturing or safety processes.

I am not saying that every sensor must be secure. In every sector, there are areas of greater or lesser vulnerability, depending on context.

SEMI: What is security or privacy by design?

Wright: Addressing security flaws is cheaper and more easily accomplished at the design stage and not after the vulnerabilities are discovered. At MITRE, we practice systems- and design-oriented thinking as we consult with people doing development. We help them to develop security standards and approaches that are broadly applicable, rather than focusing on a specific product.

For example, MITRE looks at the ways that a person might hack into a car to steal location and life history data — or alter its functions — to facilitate general standards and approaches that will help manufacturers better ensure the privacy and security of autonomous vehicles. Hackers have demonstrated that they can interfere with vehicle transmissions and brakes. Ignition, steering and other critical systems are theoretically accessible through the same types of attacks. To what degree can MEMS/sensors suppliers help automotive manufacturers ensure the privacy and security of autonomous cars, and the safety of their drivers?

SEMI: What would you like MSEC attendees to take away from your presentation?

Wright: MEMS/sensors suppliers are on the leading edge of computing and should take some responsibility for considering cybersecurity and privacy, for the safety of their customers and their own competitive advantage. Recognize which devices should be secure and act accordingly. Get involved at the design stage. The market for secure microelectronics is only going to grow, and this will benefit suppliers who take secure design seriously.

Cynthia Wright will present Cyber Security and Privacy in the Age of Autonomous Sensing on Monday, October 29 at MEMS & Sensors Executive Congress in Napa, Calif.

Register today to connect with her at the event.

Maria Vetrano is a public relations consultant at SEMI.

Researchers from Graduate School of Bio-Applications and Systems Engineering at Tokyo University of Agriculture and Technology (TUAT) have sped up the movement of electrons in organic semiconductor films by two to three orders of magnitude. The speedier electronics could lead to improved solar power and transistor use across the world, according to the scientists.

They published their results in the September issue of Macromolecular Chemistry and Physics, where the paper is featured on the cover.

Led by Kenji Ogino, a professor at Graduate School of Bio-Applications and Systems Engineering at TUAT, Japan, the team found that adding polystyrene, commonly known as Styrofoam in North America, could enhance the semiconducting polymer by allowing electrons to move from plane to plane quickly. The process, called hole mobility, is how electrons move through an electric field consisting of multiple layers. When a molecule is missing an electron, an electron from a different plane can jump or fall and take its place.

Through various imaging techniques, it’s fairly easy to follow the electron trail in the crystal-based structures. In many semiconducting polymers, however, the clean, defined lines of the crystalline skeleton intertwine with a much more difficult-to-define region. It’s actually called the amorphous domain.

“[Electrons] transport in both crystalline and amorphous domains. To improve the total electron mobility, it is necessary to control the nature of the amorphous domain,” Ogino said. “We found that hole mobility extraordinarily improved by the introduction of polystyrene block accompanied by the increase of the ratio of rigid amorphous domain.”

The researchers believe that the way the crystalline domain connects within itself occurs most effectively through the rigid amorphous domain. The addition of polystyrene introduced more amorphous domain, but contained by flexible chains of carbon and hydrogen atoms. Even though the chains are flexible, it provides rigidity, and some degree of control, to the amorphous domain.

Electrons moved two to three times quicker than normal.

“The introduction of a flexible chain in semicrystalline polymers is one of the promising strategies to improve the various functionalities of polymer films by altering the characteristics of the amorphous domain,” Ogino said. “We propose that the rigid amorphous domain plays an important role in the hole transporting process.”

Enhanced hole mobility is a critical factor in developing more efficient solar devices, according to Ogino. Next, Ogino and the researchers plan to examine how the enhanced hole mobility affected other parameters, such as the chemical composition and position of the structures within the polymer film.

Just like their biological counterparts, hardware that mimics the neural circuitry of the brain requires building blocks that can adjust how they synapse, with some connections strengthening at the expense of others. One such approach, called memristors, uses current resistance to store this information. New work looks to overcome reliability issues in these devices by scaling memristors to the atomic level.

A group of researchers demonstrated a new type of compound synapse that can achieve synaptic weight programming and conduct vector-matrix multiplication with significant advances over the current state of the art. Publishing its work in the Journal of Applied Physics, from AIP Publishing, the group’s compound synapse is constructed with atomically thin boron nitride memristors running in parallel to ensure efficiency and accuracy.

Hardware that mimics the neural circuitry of the brain requires building blocks that can adjust how they synapse. One such approach, called memristors, uses current resistance to store this information. New work looks to overcome reliability issues in these devices by scaling memristors to the atomic level. Researchers demonstrated a new type of compound synapse that can achieve synaptic weight programming and conduct vector-matrix multiplication with significant advances over the current state of the art. They discuss their work in this week’s Journal of Applied Physics. This image shows a conceptual schematic of the 3D implementation of compound synapses constructed with boron nitride oxide (BNOx) binary memristors, and the crossbar array with compound BNOx synapses for neuromorphic computing applications. Credit: Ivan Sanchez Esqueda

The article appears in a special topic section of the journal devoted to “New Physics and Materials for Neuromorphic Computation,” which highlights new developments in physical and materials science research that hold promise for developing the very large-scale, integrated “neuromorphic” systems of tomorrow that will carry computation beyond the limitations of current semiconductors today.

“There’s a lot of interest in using new types of materials for memristors,” said Ivan Sanchez Esqueda, an author on the paper. “What we’re showing is that filamentary devices can work well for neuromorphic computing applications, when constructed in new clever ways.”

Current memristor technology suffers from a wide variation in how signals are stored and read across devices, both for different types of memristors as well as different runs of the same memristor. To overcome this, the researchers ran several memristors in parallel. The combined output can achieve accuracies up to five times those of conventional devices, an advantage that compounds as devices become more complex.

The choice to go to the subnanometer level, Sanchez said, was born out of an interest to keep all of these parallel memristors energy-efficient. An array of the group’s memristors were found to be 10,000 times more energy-efficient than memristors currently available.

“It turns out if you start to increase the number of devices in parallel, you can see large benefits in accuracy while still conserving power,” Sanchez said. Sanchez said the team next looks to further showcase the potential of the compound synapses by demonstrating their use completing increasingly complex tasks, such as image and pattern recognition.

As artificial intelligence has become increasingly sophisticated, it has inspired renewed efforts to develop computers whose physical architecture mimics the human brain. One approach, called reservoir computing, allows hardware devices to achieve the higher-dimension calculations required by emerging artificial intelligence. One new device highlights the potential of extremely small mechanical systems to achieve these calculations.

A group of researchers at the Université de Sherbrooke in Québec, Canada, reports the construction of the first reservoir computing device built with a microelectromechanical system (MEMS). Published in the Journal of Applied Physics, from AIP Publishing, the neural network exploits the nonlinear dynamics of a microscale silicon beam to perform its calculations. The group’s work looks to create devices that can act simultaneously as a sensor and a computer using a fraction of the energy a normal computer would use.

A single silicon beam (red), along with its drive (yellow) and readout (green and blue) electrodes, implements a MEMS capable of nontrivial computations. Credit: Guillaume Dion

The article appears in a special topic section of the journal devoted to “New Physics and Materials for Neuromorphic Computation,” which highlights new developments in physical and materials science research that hold promise for developing the very large-scale, integrated “neuromorphic” systems of tomorrow that will carry computation beyond the limitations of current semiconductors today.

“These kinds of calculations are normally only done in software, and computers can be inefficient,” said Guillaume Dion, an author on the paper. “Many of the sensors today are built with MEMS, so devices like ours would be ideal technology to blur the boundary between sensors and computers.”

The device relies on the nonlinear dynamics of how the silicon beam, at widths 20 times thinner than a human hair, oscillates in space. The results from this oscillation are used to construct a virtual neural network that projects the input signal into the higher dimensional space required for neural network computing.

In demonstrations, the system was able to switch between different common benchmark tasks for neural networks with relative ease, Dion said, including classifying spoken sounds and processing binary patterns with accuracies of 78.2 percent and 99.9 percent respectively.

“This tiny beam of silicon can do very different tasks,” said Julien Sylvestre, another author on the paper. “It’s surprisingly easy to adjust it to make it perform well at recognizing words.”

Sylvestre said he and his colleagues are looking to explore increasingly complicated computations using the silicon beam device, with the hopes of developing small and energy-efficient sensors and robot controllers.