Tag Archives: letter-wafer-tech

Researchers at Queen Mary University of London, University of Cambridge and Max Planck Institute for Solid State Research have discovered how a pinch of salt can be used to drastically improve the performance of batteries.

They found that adding salt to the inside of a supermolecular sponge and then baking it at a high temperature transformed the sponge into a carbon-based structure.

Surprisingly, the salt reacted with the sponge in special ways and turned it from a homogeneous mass to an intricate structure with fibres, struts, pillars and webs. This kind of 3D hierarchically organised carbon structure has proven very difficult to grow in a laboratory but is crucial in providing unimpeded ion transport to active sites in a battery.

In the study, published in JACS (Journal of the American Chemical Society), the researchers demonstrate that the use of these materials in Lithium-ion batteries not only enables the batteries to be charged-up rapidly, but also at one of the highest capacities.

Due to their intricate architecture the researchers have termed these structures ‘nano-diatoms’, and believe they could also be used in energy storage and conversion, for example as electrocatalysts for hydrogen production.

Lead author and project leader Dr Stoyan Smoukov, from Queen Mary’s School of Engineering and Materials Science, said: “This metamorphosis only happens when we heat the compounds to 800 degrees centigrade and was as unexpected as hatching fire-born dragons instead of getting baked eggs in the Game of Thrones. It is very satisfying that after the initial surprise, we have also discovered how to control the transformations with chemical composition.”

Carbon, including graphene and carbon nanotubes, is a family of the most versatile materials in nature, used in catalysis and electronics because of its conductivity and chemical and thermal stability.

3D carbon-based nanostructures with multiple levels of hierarchy not only can retain useful physical properties like good electronic conductivity but also can have unique properties. These 3D carbon-based materials can exhibit improved wettability (to facilitate ion infiltration), high strength per unit weight, and directional pathways for fluid transport.

It is, however, very challenging to make carbon-based multilevel hierarchical structures, particularly via simple chemical routes, yet these structures would be useful if such materials are to be made in large quantities for industry.

The supermolecular sponge used in the study is also known as a metal organic framework (MOF) material. These MOFs are attractive, molecularly designed porous materials with many promising applications such as gas storage and separation. The retention of high surface area after carbonisation – or baking at a high temperature – makes them interesting as electrode materials for batteries. However, so far carbonising MOFs has preserved the structure of the initial particles like that of a dense carbon foam. By adding salts to these MOF sponges and carbonising them, the researchers discovered a series of carbon-based materials with multiple levels of hierarchy.

Dr R. Vasant Kumar, a collaborator on the study from University of Cambridge, commented: “This work pushes the use of the MOFs to a new level. The strategy for structuring carbon materials could be important not only in energy storage but also in energy conversion, and sensing.”

Lead author, Tiesheng Wang, from University of Cambridge, said: “Potentially, we could design nano-diatoms with desired structures and active sites incorporated in the carbon as there are thousands of MOFs and salts for us to select.”

University of Waterloo chemists have found a much faster and more efficient way to store and process information by expanding the limitations of how the flow of electricity can be used and managed.

In a recently released study, the chemists discovered that light can induce magnetization in certain semiconductors – the standard class of materials at the heart of all computing devices today.

“These results could allow for a fundamentally new way to process, transfer, and store information by electronic devices, that is much faster and more efficient than conventional electronics.”

For decades, computer chips have been shrinking thanks to a steady stream of technological improvements in processing density. Experts have, however, been warning that we’ll soon reach the end of the trend known as Moore’s Law, in which the number of transistors per square inch on integrated circuits double every year.

“Simply put, there’s a physical limit to the performance of conventional semiconductors as well as how dense you can build a chip,” said Pavle Radovanovic, a professor of chemistry and a member of the Waterloo Institute for Nanotechnology. “In order to continue improving chip performance, you would either need to change the material transistors are made of – from silicon, say to carbon nanotubes or graphene – or change how our current materials store and process information.”

Radovanovic’s finding is made possible by magnetism and a field called spintronics, which proposes to store binary information within an electron’s spin direction, in addition to its charge and plasmonics, which studies collective oscillations of elements in a material.

“We’ve basically magnetized individual semiconducting nanocrystals (tiny particles nearly 10,000 times smaller than the width of a human hair) with light at room temperature,” said Radovanovic. “It’s the first time someone’s been able to use collective motion of electrons, known as plasmon, to induce a stable magnetization within such a non-magnetic semiconductor material.”

In manipulating plasmon in doped indium oxide nanocrystals Radovanovic’s findings proves that the magnetic and semiconducting properties can indeed be coupled, all without needing ultra-low temperatures (cryogens) to operate a device.

He anticipates the findings could initially lead to highly sensitive magneto-optical sensors for thermal imaging and chemical sensing. In the future, he hopes to extend this approach to quantum sensing, data storage, and quantum information processing.

SILTECTRA GmbH, a developer of advanced wafering technology solutions and services, today announced that it has fortified its market position by adding three new patents to its global portfolio of intellectual property (IP). The first patent covers new technical capabilities relating to the company’s COLD SPLIT laser process and extends the approach to non-polymer applications. The second patent secures COLD SPLIT for all substrate materials.

The third patent covers an extension of the company’s silicon carbide (SiC) process capability to split materials with sub-100-micron material loss, regardless of vendor-specific SiC crystal-growing processes. SILTECTRA’s relentless effort to drive down SiC material loss aims to help accelerate adoption of the superior substrate for power devices and other ICs. Up to now, high cost has inhibited fast adoption. Substantial cost reductions enabled by SILTECTRA’s technology could speed deployment of SiC for a broader range of applications, such as electric vehicles (EVs) and 5G technology.

SILTECTRA’s IP portfolio now consists of 70 patent families with 200 patents. Collectively, the patents cover every innovation associated with the company’s breakthrough laser-based wafer-thinning process.

The growth of SILTECTRA’s IP portfolio reflects the company’s steady march toward commercializing its solution. COLD SPLIT demonstrated early differentiation by thinning wafers to 100 microns and below in minutes with extreme precision and virtually no material loss. These enabling advantages drew high interest from integrated device manufacturers (IDMs) who had previously relied on grinding to thin their wafers. Grinding is a slower, less precise process that generates material loss and reduces overall yield. In contrast, COLD SPLIT is a much faster laser-based thinning approach with higher yield and strong cost-of-ownership benefits.

In a development announced earlier this year, SILTECTRA reported a breakthrough new capability for COLD SPLIT that vastly increased the value of the technology for cost-sensitive IDMs. Thanks to a novel adaptation known as “twinning”, the company demonstrated that COLD SPLIT can reclaim substrate material generated (and previously wasted) during backside grinding and create a second fully optimizable bonus wafer in the process. SILTECTRA validated the breakthrough by producing a gallium nitride (GaN) on SiC high electron-mobility power transistor (HEMT) device on a split-off (or “twinned”) wafer at its new facility in Dresden. The HEMT showed results that were superior to a non- COLD-SPLIT-enabled HEMT when measured for CMP characterization, as well as GaN EPI, metal layer and gate layer outcomes.

The developments drew keen interest from IDMs, as well as substrate manufacturers, and even providers of certain process technologies.

SILTECTRA’s CEO, Dr. Harald Binder, pledged to maintain the rapid pace of innovation at the company to enable IDMs with superior wafering solutions. He noted: “Like all technology companies, SILTECTRA’s leadership and future growth depend on continually innovating to extend our capabilities and further enrich the value of our solution. Naturally, therefore, it’s a strategic priority to protect the innovations along the way so that our competitive differentiation and enabling advantages remain strong in all regions where customers are located. Our robust IP portfolio reflects this priority.”

Dr. Jan Richter, SILTECTRA’s CTO, stated: “Our R&D team is relentlessly pushing the limits of our COLD SPLIT technology to fulfill its enormous potential. The additional patents further strengthen our market position, while enabling us to drive COLD SPLIT’s material loss far below 50 microns.”

In the wake of its recent discovery of a flat form of gallium, an international team led by scientists from Rice University has created another two-dimensional material that the researchers said could be a game changer for solar fuel generation. Rice materials scientist Pulickel Ajayan and colleagues extracted 3-atom-thick hematene from common iron ore. The research was introduced in a paper today in Nature Nanotechnology.

Hematene may be an efficient photocatalyst, especially for splitting water into hydrogen and oxygen, and could also serve as an ultrathin magnetic material for spintronic-based devices, the researchers said.

“2D magnetism is becoming a very exciting field with recent advances in synthesizing such materials, but the synthesis techniques are complex and the materials’ stability is limited,” Ajayan said. “Here, we have a simple, scalable method, and the hematene structure should be environmentally stable.”

Ajayan’s lab worked with researchers at the University of Houston and in India, Brazil, Germany and elsewhere to exfoliate the material from naturally occurring hematite using a combination of sonication, centrifugation and vacuum-assisted filtration.

Hematite was already known to have photocatalytic properties, but they are not good enough to be useful, the researchers said.

“For a material to be an efficient photocatalyst, it should absorb the visible part of sunlight, generate electrical charges and transport them to the surface of the material to carry out the desired reaction,” said Oomman Varghese, a co-author and associate professor of physics at the University of Houston.

“Hematite absorbs sunlight from ultraviolet to the yellow-orange region, but the charges produced are very short-lived. As a result, they become extinct before they reach the surface,” he said.

Hematene photocatalysis is more efficient because photons generate negative and positive charges within a few atoms of the surface, the researchers said. By pairing the new material with titanium dioxide nanotube arrays, which provide an easy pathway for electrons to leave the hematene, the scientists found they could allow more visible light to be absorbed.

The researchers also discovered that hematene’s magnetic properties differ from those of hematite. While native hematite is antiferromagnetic, tests showed that hematene is ferromagnetic, like a common magnet. In ferromagnets, atoms’ magnetic moments point in the same direction. In antiferromagnets, the moments in adjacent atoms alternate.

Unlike carbon and its 2D form, graphene, hematite is a non-van der Waals material, meaning it’s held together by 3D bonding networks rather than non-chemical and comparatively weaker atomic van der Waals interactions.

“Most 2D materials to date have been derived from bulk counterparts that are layered in nature and generally known as van der Waals solids,” said co-author Professor Anantharaman Malie Madom Ramaswamy Iyer of the Cochin University of Science and Technology, India. “2D materials from bulk precursors having (non-van der Waals) 3D bonding networks are rare, and in this context hematene assumes great significance.”

According to co-author Chandra Sekhar Tiwary, a former postdoctoral researcher at Rice and now an assistant professor at the Indian Institute of Technology, Gandhinagar, the collaborators are exploring other non-van der Waals materials for their 2D potential.

Researchers at Duke University and North Carolina State University have demonstrated the first custom semiconductor microparticles that can be steered into various configurations repeatedly while suspended in water.

With an initial six custom particles that predictably interact with one another in the presence of alternating current (AC) electric fields of varying frequencies, the study presents the first steps toward realizing advanced applications such as artificial muscles and reconfigurable computer systems.

The study appears online on May 3 in the journal Nature Communications.

“We’ve engineered and encoded multiple dynamic responses in different microparticles to create a reconfigurable silicon toolbox,” said Ugonna Ohiri, a recently graduated electrical engineering doctoral student from Duke and first author of the paper. “By providing a means of controllably assembling and disassembling these particles, we’re bringing a new tool to the field of active matter.”

While previous researchers have worked to define self-assembling systems, few have worked with semiconductor particles, and none have explored the wide range of custom shapes, sizes and coatings that are available to the micro- and nanofabrication industry. Engineering particles from silicon presents the opportunity to physically realize electronic devices that can self-assemble and disassemble on demand. Customizing their shapes and sizes presents opportunities to explore a wide-ranging design space of new motile behaviors.

“Most previous work performed using self-assembling particles has been done with shapes such as spheres and other off-the-shelf materials,” said Nan Jokerst, the J. A. Jones Professor of Electrical and Computer Engineering at Duke. “Now that we can customize whatever arbitrary shapes, electrical characteristics and patterned coatings we want with silicon, a whole new world is opening up.”

In the study, Jokerst and Ohiri fabricated silicon particles of various shapes, sizes and electrical properties. In collaboration with Orlin Velev, the INVISTA Professor of Chemical and Biomolecular Engineering at NC State, they characterized how these particles responded to different magnitudes and frequencies of electric fields while submerged in water.

Based on these observations, the researchers then fabricated new batches of customized particles that were likely to exhibit the behaviors they were looking for, resulting in six different engineered silicon microparticle compositions that could move through water, synchronize their motions, and reversibly assemble and disassemble on demand.

The thin film particles are 10-micron by 20-micron rectangles that are 3.5 microns thick. They’re fabricated using Silicon-on-Insulator (SOI) technology. Since they can be made using the same fabrication technology that produces integrated circuits, millions of identical particles could be produced at a time.

“The idea is that eventually we’re going to be able to make silicon computational systems that assemble, disassemble and then reassemble in a different format,” said Jokerst. “That’s a long way off in the future, but this work provides a sense of the capabilities that are out there and is the first demonstration of how we might achieve those sorts of devices.”

That is, however, only the tip of the proverbial iceberg. Some of the particles were fabricated with both p-type and n-type regions to create p-n junctions — common electrical components that allow electricity to pass in only one direction. Tiny metal patterns were also placed on the particles’ surfaces to create p-n junction diodes with contacts. In the future, researchers could even engineer particles with patterns using other electrically conductive or insulating materials, complex integrated circuits, or microprocessors on or within the silicon.

“This work is just a small snapshot of the tools we have to control particle dynamics,” said Ohiri. “We haven’t even scratched the surface of all of the behaviors that we can engineer, but we hope that this multidisciplinary study can pioneer future studies to design artificial active materials.”

A simple method that uses hydrogen chloride can better control the crystal structure of a common semiconductor and shows promise for novel high-powered electronic applications.

The electronic components used in computers and mobile devices operate at relatively lower power. But high-power applications, such as controlling electrical power grids, require alternative materials that can cope with much higher voltages. For example, an insulating material begins to conduct electricity when the field is high enough, an effect known as electrical breakdown. For this reason, power electronics often use nitride-based semiconductors, such as gallium nitride, which have a very high breakdown field and can be epitaxially grown to create multilayered semiconductors.

However, ever-increasing energy demands and the desire to make electricity distribution more efficient requires even more electrically robust materials. Gallium oxide (Ga2O3) has a theoretical breakdown field more than twice that of gallium-nitride alloys and so has emerged as an exciting candidate for this function. The latest challenge however is a simple way to deposit high-quality gallium oxide on the substrates commonly used for power electronics, such as sapphire.

Haiding Sun, Xiaohang Li, and co-workers from KAUST worked with industry partners Structured Materials Industries, Inc. in the U.S. to demonstrate a relatively simple method to control the crystal structure of gallium oxides on a sapphire substrate using a technology known as metalorganic chemical vapor deposition (MOCVD). “We were able to control the growth by changing just one parameter: the flow rate of hydrogen chloride in the chamber,” explains Sun. “This is the first time that hydrogen chloride has been used during oxide growth in an MOCVD reactor.”

Working in a clean suit in the lab, Dr. Sun holds up a gallium-oxide template. Credit: © 2018 KAUST

Working in a clean suit in the lab, Dr. Sun holds up a gallium-oxide template. Credit: © 2018 KAUST

The atoms in gallium oxide can be arranged in a number of different forms known as polymorphs. β­­­?Ga2O3 is the most stable polymorph but is difficult to grow on substrates of other materials. ε?Ga2O3 has been grown on sapphire but its growth rate has been difficult to control.

Different polymorphs of gallium oxide can be grown in a MOCVD chamber by controlling the flow of hydrogen chloride.

Different polymorphs of gallium oxide can be grown in a MOCVD chamber by controlling the flow of hydrogen chloride.

Engineers at the University of California, Riverside, have demonstrated prototype devices made of an exotic material that can conduct a current density 50 times greater than conventional copper interconnect technology.

Current density is the amount of electrical current per cross-sectional area at a given point. As transistors in integrated circuits become smaller and smaller, they need higher and higher current densities to perform at the desired level. Most conventional electrical conductors, such as copper, tend to break due to overheating or other factors at high current densities, presenting a barrier to creating increasingly small components.

Microscopy image of an electronic device made with 1D ZrTe3 nanoribbons. The nanoribbon channel is indicated in green color. The metal contacts are shown in yellow color. Note than owing to the nanometer scale thickness the yellow metal contacts appear to be under the green channel while in reality they are on top. Credit: Balandin lab, UC Riverside

Microscopy image of an electronic device made with 1D ZrTe3 nanoribbons. The nanoribbon channel is indicated in green color. The metal contacts are shown in yellow color. Note than owing to the nanometer scale thickness the yellow metal contacts appear to be under the green channel while in reality they are on top. Credit: Balandin lab, UC Riverside

The electronics industry needs alternatives to silicon and copper that can sustain extremely high current densities at sizes of just a few nanometers.

The advent of graphene resulted in a massive, worldwide effort directed at investigation of other two-dimensional, or 2D, layered materials that would meet the need for nanoscale electronic components that can sustain a high current density. While 2D materials consist of a single layer of atoms, 1D materials consist of individual chains of atoms weakly bound to one another, but their potential for electronics has not been as widely studied.

One can think of 2D materials as thin slices of bread while 1D materials are like spaghetti. Compared to 1D materials, 2D materials seem huge.

A group of researchers led by Alexander A. Balandin, a distinguished professor of electrical and computer engineering in the Marlan and Rosemary Bourns College of Engineering at UC Riverside, discovered that zirconium tritelluride, or ZrTe3, nanoribbons have an exceptionally high current density that far exceeds that of any conventional metals like copper.

The new strategy undertaken by the UC Riverside team pushes research from two-dimensional to one-dimensional materials­­– an important advance for the future generation of electronics.

“Conventional metals are polycrystalline. They have grain boundaries and surface roughness, which scatter electrons,” Balandin said. “Quasi-one-dimensional materials such as ZrTe3consist of single-crystal atomic chains in one direction. They do not have grain boundaries and often have atomically smooth surfaces after exfoliation. We attributed the exceptionally high current density in ZrTe3 to the single-crystal nature of quasi-1D materials.”

In principle, such quasi-1D materials could be grown directly into nanowires with a cross-section that corresponds to an individual atomic thread, or chain. In the present study the level of the current sustained by the ZrTe3 quantum wires was higher than reported for any metals or other 1D materials. It almost reaches the current density in carbon nanotubes and graphene.

Electronic devices depend on special wiring to carry information between different parts of a circuit or system. As developers miniaturize devices, their internal parts also must become smaller, and the interconnects that carry information between parts must become smallest of all. Depending on how they are configured, the ZrTe3 nanoribbons could be made into either nanometer-scale local interconnects or device channels for components of the tiniest devices.

The UC Riverside group’s experiments were conducted with nanoribbons that had been sliced from a pre-made sheet of material. Industrial applications need to grow nanoribbon directly on the wafer. This manufacturing process is already under development, and Balandin believes 1D nanomaterials hold possibilities for applications in future electronics.

“The most exciting thing about the quasi-1D materials is that they can be truly synthesized into the channels or interconnects with the ultimately small cross-section of one atomic thread– approximately one nanometer by one nanometer,” Balandin said.

Research appearing today in Nature Communications finds useful new information-handling potential in samples of tin(II) sulfide (SnS), a candidate “valleytronics” transistor material that might one day enable chipmakers to pack more computing power onto microchips.

Valleytronics utilizes different local energy extrema (valleys) with selection rules to store 0s and 1s. In SnS, these extrema have different shapes and responses to different polarizations of light, allowing the 0s and 1s to be directly recognized. This schematic illustrates the variation of electron energy in different states, represented by curved surfaces in space. The two valleys of the curved surface are shown. Credit: Berkeley Lab

Valleytronics utilizes different local energy extrema (valleys) with selection rules to store 0s and 1s. In SnS, these extrema have different shapes and responses to different polarizations of light, allowing the 0s and 1s to be directly recognized. This schematic illustrates the variation of electron energy in different states, represented by curved surfaces in space. The two valleys of the curved surface are shown. Credit: Berkeley Lab

The research was led by Jie Yao of the Department of Energy’s Lawrence Berkeley National Laboratory (Berkeley Lab) and Shuren Lin of UC Berkeley’s Department of Materials Science and Engineering and included scientists from Singapore and China. Berkeley Lab’s Molecular Foundry, a DOE Office of Science user facility, contributed to the work.

For several decades, improvements in conventional transistor materials have been sufficient to sustain Moore’s Law – the historical pattern of microchip manufacturers packing more transistors (and thus more information storage and handling capacity) into a given volume of silicon. Today, however, chipmakers are concerned that they might soon reach the fundamental limits of conventional materials. If they can’t continue to pack more transistors into smaller spaces, they worry that Moore’s Law would break down, preventing future circuits from becoming smaller and more powerful than their predecessors.

That’s why researchers worldwide are on the hunt for new materials that can compute in smaller spaces, primarily by taking advantage of the additional degrees of freedom that the materials offer – in other words, using a material’s unique properties to compute more 0s and 1s in the same space. Spintronics, for example, is a concept for transistors that harnesses the up and down spins of electrons in materials as the on/off transistor states.

Valleytronics, another emerging approach, utilizes the highly selective response of candidate crystalline materials under specific illumination conditions to denote their on/off states – that is, using the materials’ band structures so that the information of 0s and 1s is stored in separate energy valleys of electrons, which are dependent on the crystal structures of the materials.

In this new study, the research team has shown that tin(II) sulfide (SnS) is able to absorb different polarizations of light and then selectively reemit light of different colors at different polarizations. This is useful for concurrently accessing both the usual electronic – and the material’s valleytronic – degrees of freedom, which would substantially increase the computing power and data storage density of circuits made with the material.

“We show a new material with distinctive energy valleys that can be directly identified and separately controlled,” said Yao. “This is important because it provides us a platform to understand how valley signatures are carried by electrons and how information can be easily stored and processed between the valleys, which are of both scientific and engineering significance.”

Lin, the first author of the paper, said the material is different from previously investigated candidate valleytronics materials because it possesses such selectivity at room temperature without additional biases apart from the excitation light source, which alleviates the previously stringent requirements in controlling the valleys. Compared to its predecessor materials, SnS is also much easier to process.

With this finding, researchers will be able to develop operational valleytronic devices, which may one day be integrated into electronic circuits. The unique coupling between light and valleys in this new material may also pave the way toward future hybrid electronic/photonic chips.

Berkeley Lab’s “Beyond Moore’s Law” initiative leverages the basic science capabilities and unique user facilities of Berkeley Lab and UC Berkeley to evaluate promising candidates for next-generation electronics and computing technologies. Its objective is to build close partnerships with industry to accelerate the time it typically takes to move from the discovery of a technology to its scale-up and commercialization.

Spin Transfer Technologies, Inc., the developer of advanced STT-MRAM for embedded SRAM and stand-alone DRAM applications, today announced results of its unique Precessional Spin Current (PSC™) structure. The results from advanced testing of the PSC structure confirm that it will increase the spin-torque efficiency of any MRAM device by 40-70 percent — enabling dramatically higher data retention while consuming less power. This gain translates to retention times lengthening by a factor of over 10,000 (e.g., 1 hour retention becomes more than 1 year retention) while reducing write current. Improved efficiency is critical for enabling MRAM to replace SRAM and DRAM in mobile, datacenter and AI applications, as well as for improving retention and performance in high-temperature automotive applications. The company reported these results at the prestigious Intermag 2018 Conference.

Spin-torque efficiency is one of the core performance metrics of the pMTJ (perpendicular magnetic tunnel junction — the “bit” that stores the memory state in an MRAM memory) and is defined by the ratio between the thermal retention barrier, measuring how long data can be reliably stored in the memory, and the switching current necessary to change the value of the bit. In previous MRAM implementations, increasing the energy barrier to increase retention would require a proportional increase in write current — leading to higher power consumption and much faster wear-out of the pMTJ devices (lower endurance). The PSC structure is a breakthrough because it effectively decouples the static energy barrier that determines retention from the dynamic switching processes that govern the switching current. As a result, when the PSC structure is added to any pMTJ, benefits include:

  • A higher energy barrier when the pMTJ does not have current flowing through it, which is ideal for retaining data for long periods
  • An increased spin polarization when current is flowing and the device is writing a new state, which is ideal for minimizing switching current and extending the life of the device by many orders of magnitude

The PSC structure was designed from the outset to be modular and fabricated with any pMTJ — either the company’s own pMTJs, or a pMTJ from other sources. The PSC structure is fabricated during the pMTJ deposition process and adds approximately 4nm to the height of the pMTJ stack. The structure is compatible with a wide range of standard MRAM manufacturing processes, materials and tool sets — enabling any foundry to readily incorporate the PSC structure into existing pMTJ stacks without adding significant complexity or manufacturing costs.

“MRAM is attracting a lot of attention as an embedded memory for ASICs and MCUs, but issues of write current and data retention have caused concern,” said Jim Handy, general director of Objective Analysis. “Spin Transfer Technologies’ new PSC structure shows a lot of promise to solve a number of those issues and pave the path for MRAM to take a significant share of the embedded memory market.”

Spin Transfer Technologies’ testing of the PSC structure involved comparing the performance of the same pMTJ devices with and without PSC for a large number of devices within CMOS test chip arrays at various temperatures and device diameters. The tests exhibited a robust performance advantage due to the PSC structure, both during writing of the low-resistance (“0”) and the high-resistance (“1”) memory states. Some specific examples of the advantages that the data have shown are as follows:

  • Increase of the spin-torque efficiency by up to 70 percent
  • Demonstration of the efficiency gain across a range of sizes (40-60nm) and temperatures (30°C to 125°C)
  • Increase of the thermal energy barriers by 50 percent corresponding to an increase in data retention time of greater than four orders of magnitude while reducing the switching current
  • Reduction of read disturb error rate up to five orders of magnitude

These advantages have come without degradation to other performance parameters. The data for the PSC structure indicate significant potential for enabling high-speed applications as well as high-temperature automotive and other applications. Furthermore, since the data shows that the PSC structure’s efficiency gains actually increase as the pMTJ get smaller, the PSC structure opens new pathways to achieving embedded SRAMs in the latest 7nm and 5nm generations.

“There is a huge demand for a memory with the endurance of SRAM, but with higher density, lower operating power and with non-volatility. We believe the improvements the PSC structure brings to STT-MRAM technology will make it a highly attractive alternative to SRAM for these reasons,” said Mustafa Pinarbasi, CTO and SVP of Magnetics Technology at Spin Transfer Technologies. “We are excited to enable the next generation of STT-MRAM and to shake up the status quo of the memory industry through our innovation.”

In even the most fuel-efficient cars, about 60 percent of the total energy of gasoline is lost through heat in the exhaust pipe and radiator. To combat this, researchers are developing new thermoelectic materials that can convert heat into electricity. These semiconducting materials could recirculate electricity back into the vehicle and improve fuel efficiency by up to 5 percent.

The challenge is, current thermoelectric materials for waste heat recovery are very expensive and time consuming to develop. One of the state of the art materials, made from a combination of hafnium and zirconium (elements most commonly used in nuclear reactors), took 15 years from its initial discovery to optimized performance.

Now, researchers from the Harvard John A. Paulson School of Engineering and Applied Sciences (SEAS) have developed an algorithm that can discover and optimize these materials in a matter of months, relying on solving quantum mechanical equations, without any experimental input.

“These thermoelectric systems are very complicated,” said Boris Kozinsky, a recently appointed Associate Professor of Computational Materials Science at SEAS and senior author of the paper. “Semiconducting materials need to have very specific properties to work in this system, including high electrical conductivity, high thermopower, and low thermal conductivity, so that all that heat gets converted into electricity. Our goal was to find a new material that satisfies all the important properties for thermoelectric conversion while at the same time being stable and cheap.”

Kozinsky co-authored the research with Georgy Samsonidze, a research engineer at the Robert Bosch Research and Technology Center in Cambridge, MA, where both authors conducted most of the research.

In order to find such a material, the team developed an algorithm that can predict electronic transport properties of a material based only on the chemical elements used in the crystalline crystal. The key was to simplify the computational approach for electron-phonon scattering and to speed it up by about 10,000 times, compared to existing algorithms.

The new method and computational screening results are published in Advanced Energy Materials.

Using the improved algorithm, the researchers screened many possible crystal structures, including structures that had never been synthesized before. From those, Kozinsky and Samsonidze whittled the list down to several interesting candidates. Of those candidates, the researchers did further computational optimization and sent the top performers to the experimental team.

In an earlier effort experimentalists synthesized the top candidates suggested by these computations and found a material that was as efficient and as stable as previous thermoelectric materials but 10 times cheaper. The total time from initial screening to working devices: 15 months.

“We did in 15 months of computation and experimentation what took 15 years for previous materials to be optimized,” said Kozinsky. “What’s really exciting is that we’re probably not fully understanding the extent of the simplification yet. We could potentially make this method even faster and cheaper.”

Kozinsky said he hopes to improve the new methodology and use it to explore electronic transport in a wider class of new exotic materials such as topological insulators.