Science & Technology News (1160)
Posted: February 3, 2016 07:51AM
Radio Frequency Identification (RFID) chips are a somewhat common security tool that is used on credit cards, key cards, and on palettes of goods to easily track their use with little overhead as only a special reader is needed. Among their advantages are that the radio frequencies used can penetrate various materials and that the chips themselves do not require a power source. Both of these advantages also make them susceptible to hacks, but researchers at MIT and Texas Instruments has found a solution to the problem.
One way to hack an RFID chip is with side-channel attacks, which analyze the memory access or power usage to gather some information. By repeating this many times, the cryptographic key stored on the chip can be learned. One way to defeat this attack is to have a random number generator on the key that matches that on a server, so after each use the chip can change the key, but this can be beat by a power glitch attack. This attack cuts power to the chip, which is easily done as it requires power from the reader to operate, before the key can be updated, so the key never changes and the side-channel attacks will still work. The solution the researchers developed to both problems is to integrate ferroelectric materials into the chip. These materials can act as nonvolatile memory but also as capacitors, because they work by separating positive and negative charges, creating a voltage. Texas Instruments manufacturers both 1.5 V and 3.3 V ferroelectric cells, and both have been integrated into the new chip. The 3.3 V capacitors are used to power the RFID chip, in the event power is lost so that the key update can complete, while the 1.5 V capacitors serve as nonvolatile memory, protecting the new key from being lost. Once power is restored, the 3.3 V capacitors are recharged before anything else is done.
Texas Instruments has already built multiple prototypes and while they are slower than traditional RFID chips, at 30 readouts a second, they should still be fast enough to work in most scenarios.
Posted: February 2, 2016 07:01AM
Batteries are an integral part of many modern technologies, especially those that use lithium ions for storing energy. Typically these batteries use a graphite anode but for some time researchers have been trying to build silicon anodes, which offer greater energy storage, but they have significant resiliency problems. Now researchers at SLAC National Accelerator Laboratory and Stanford University have found a way to make viable silicon anodes with the help of graphene.
Batteries work by moving ions between the anode and the cathode, and charging involves the ions entering the anode. A silicon anode in a lithium-ion battery could hold ten times the energy of a graphite anode, but silicon swells when it absorbs the ions, and this swelling can cause it to break and the battery to fail. To One way to address this issue is to work with silicon particles so small that the swelling does not matter, but previous attempts tended to be inefficient. This new design though works a bit differently as it places the silicon particles inside of graphene cages. Graphene is an atom-thick sheet of carbon that is very strong, flexible, and conductive, and it actually improves the conductivity of the silicon particles. To create the cages, the researchers start with the silicon particles and coat them with a layer of nickel. The nickel then acts as a catalyst to grow the graphene on top of, and the nickel can then be etched away to leave the particle plenty of room to swell without damage.
Just as important as the viability of a silicon anode is that the silicon particles can be micrometers in width, which are much easier and cheaper to produce than smaller nanostructures. In fact, silicon particles microns in width are a waste product to the milling of silicon for computer chips, like sawdust is for cutting wood.
Posted: February 1, 2016 07:15AM
On all scales of Nature, it is possible to find vortices in some form, and some can be more useful than others. Researchers at Berkeley Lab have discovered polar vortices in a ferroelectric material for the first time, after being predicted over a decade ago. These vortices could be similar to so-called skyrmions and could be used to create some advanced technologies.
Skyrmions are vortices that can be created within ferromagnetic materials when properly exposed to an external magnetic field. These vortices can then be moved around across macroscopic distances, making them quite interesting for spintronic applications. By applying an external electric field to a specially prepared ferroelectric material, the researchers were able to create what look like electrical skyrmions. The material was comprised of lead titanate and strontium titanate with a superlattice that had elastic, electrostatic, and gradient energies competing with each other. This competition is what allows the polar vortices to form.
If it turns out the vortices are effectively electrical skyrmions then we could see them used for ultracompact data storage and processing. They could also lead to discovering new states of matter, which this actually already is, and many other applications depending on how our understanding of the phenomenon evolves.
Source: Berkeley Lab
Posted: January 29, 2016 06:32AM
Modern fabrication methods for computer chips rely on stacking layers of different materials and etching them as needed. These materials have to be compatible with each other though, because their crystalline structures must align for this to work. Researchers at MIT have developed a new fabrication method without that requirement as it allows very different materials to be combined in a single layer.
To demonstrate their method, the researchers worked with molybdenum disulfide and graphene, but the technique can be applied on many other materials provided they are combinations of elements from group six and sixteen on the periodic table. Graphene is an atom-thick sheet of carbon atoms with very useful properties, including great strength and the highest recorded electron mobility. The researchers started by depositing a layer of graphene onto a silicon substrate, and then etched away the areas they wanted the molybdenum disulfide to go. Next a bar of a material known as PTAS was attached to one side of the substrate and heated. A gas flow then moves across the PTAS and substrate, carrying PTAS molecules that attach to the silicon but not the graphene. These molecules then catalyze a reaction with a different gas, resulting in the deposition of a molybdenum disulfide layer.
The resulting device consisted of layers just one to three atoms thick, which is far smaller than the layers in modern computer chips. At that size, this fabrication method could be used to build tunneling transistors, which use quantum mechanics to operate at very low power, producing little heat.
Posted: January 27, 2016 01:43PM
Later this year we are going to see significant shifts made by component manufacturers as they switch to using smaller process nodes. This switch will allow for faster components that also use less power to run. Miniaturization is far from easy as we approach the limitations of current manufacturing methods, which is why many are working on completely new methods. Researchers at MIT have recently found how to have block co-polymers stack themselves into nanogrids, which could allow for further miniaturization.
Currently our various computer chips are made by a process known as photolithography, which uses light to etch circuitry onto the appropriate materials. While this method has worked well since the 1960s, we are approaching circuitry sizes that are smaller than the wavelengths this method works with, requiring either clever modifications to the technique or a completely new method. The MIT researchers are opting for the latter and have found a way to produce mesh structures with block co-polymers, which are capable of self-assembly because they are comprised of two different polymers that repel each other. One of the polymers was carbon based and the other was silicon based, and the researchers found that the silicon, to escape the carbon, would fold themselves into cylinders. After laying down one layer, it could be exposed to an oxygen plasma to burn away the carbon-based polymer and oxidize the silicon into glass-like cylinders. If a second layer is applied, the cylinders will align themselves perpendicularly to those of the first layer, forming the nanomesh, provided the chemical interactions between the layers are weak enough. This is something that can be predicted by simulations, which could also predict other polymers that could be used for this process.
If the first layer is placed without preparation, the cylinders will form haphazard patterns, but if the substrate the first layer is put on is properly treated, they will align themselves into parallel rows. While the resulting glass-like wires are not immediately useful for electronics, seeding them with other molecules could make them electronically active. Also the square gaps between the cylinders are of interest in electronic circuitry, as a place to run the back-end wiring of processors.
Posted: January 26, 2016 07:13AM
Believe it or not, but knots are a fairly interesting topic in different scientific fields. They are not just some means of connecting lengths of rope or string, but also mathematical systems in topology, which can then be related to various other systems. Now researchers at Aalto University of Amherst College have succeeded in knotting a quantum mechanical system, which could have implications in cosmology, nuclear fusion, and quantum computers.
For a long time it has been believed that it should be possible to tie knots in quantum fields, but this is the first time it has been achieved. To do it, the researchers started with a superfluid of rubidium atoms. Superfluids, also known as Bose-Einstein condensates are fluids that share a single quantum function across all of the atoms, which results in some interesting properties. By exposing this superfluid to a magnetic field and then changing it to leave a null point in the center, so no magnetic field was present there, the knots were tied in less than a millisecond. Each ring in the knot has its own field direction and are linked with each other in a way that makes it topologically stable. What that means is it is impossible to untie the knot without cutting it.
After first tying the knot, the researchers have gotten quite good at tying it again, and have done so several hundred times. From this they hope to develop more complex knots and eventually create a knot stable enough for more in depth study. As we delve deeper into quantum mechanics for quantum computers and nuclear fusion, this research will likely become quite important.
Source: Aalto University
Posted: January 25, 2016 06:21AM
Not too long we saw an interesting trend develop of using GPUs to do more than render graphics but also perform various computations. Now you can find supercomputers leveraging the power of GPUs for very complex computations and various pieces of software to accelerate tasks, like simulations and video encoding. To help advance GPU designs and improve their capabilities to perform these tasks, researchers at Binghamton University have created the first open-source GPU for research.
The name of this new GPU is Nyami and was created to allow researchers and even enthusiasts to see how modifications can affect performance. While it is possible to simulate it for experiments, the researchers intentionally designed it to be synthesizable, so no corners were cut in the process. This makes it much more reliable to work with, as any results can be experimentally verified.
The hope for Nyami is to allow people to approach issues with it as research problems, resulting in performance and efficiency improvements. It could also be used for research that is not necessarily GPU-specific, such as studies focused on energy efficiency and reliability.
Source: Binghamton University
Posted: January 22, 2016 09:32AM
When it comes to memory technologies, speed and capacity are probably the two most important factors for most people, and typically they also tend to be mutually exclusive. This could significantly change though, thanks to researchers at the University of Nottingham who have discovered how to turn antiferromagnets into memory devices. These materials have the potential to be a thousand times faster than current technologies and have significantly greater data densities, while also being more resilient.
The magnetic properties of a material are determined by the alignment of the spins of the various molecules within that material. For a ferromagnetic material, the spins all align and in a nonmagnetic material, the spins point in all directions, cancelling each other out. In an antiferromagnetic material though, the spins will be parallel to each other, but will also be in the opposite direction of their neighbors. This means an antiferromagnet does not produce a magnetic field. A memory device built using antiferromagnetic materials then can have a very high density, as neighboring bits will not interfere with each other. Also the device will be resilient to external magnetic fields and radiation.
What the researchers discovered is that a specific crystal structure of copper, manganese, and arsenic can have its magnetic moments controlled via electrical impulses by tilting the individual spins by 90 degrees. This means it should be possible to create an electrically controlled memory device from antiferromagnetic materials, and because it will be all electrically controlled, it should be exceptionally fast. To test this, the researchers intend to build prototype USB memory devices using this research.
Source: University of Nottingham
Posted: January 21, 2016 08:13AM
There are several examples in technology of integrating components to improve performance and capability, and now researchers at Northwestern University have successfully integrated a mid-infrared tunable laser with an on-chip amplifier. This combination improves the significantly improves the operation of both components over their predecessors, and allows multiple components to all be part of one package.
The mid-infrared frequency range of this laser gives it potential applications in detecting chemicals, including potentially hazardous and explosive ones. Such a sensor could also work at a distance, so the user would not need to be physically near a suspicious object. Compared to other mid-infrared laser packages, this design has ten times the power of others, and the range of frequencies it can be tuned to is more than twice as large. These improvements will also help it be used as a chemical sensor. Beyond that, this technology could be applied for free-space optical communications and aircraft protection.
Source: Northwestern University
Posted: January 20, 2016 11:03AM
I remember being taught in school that the Solar System has nine planets, with Pluto being the most distant, but not too long ago, Pluto got demoted to the newly created dwarf planet status, bringing the planet count to eight. Now it looks like we may be back at nine planets thanks to researchers at the California Institute of Technology, including Mike Brown, one of the researchers that led to Pluto's demotion.
Back in 2014 a paper was published by one of the Brown's former students, in which 13 distant Kuiper Belt objects were described because of their very unusual orbits. To explain these orbits, the paper suggested a small planet might exist. This explanation seemed unlikely to Brown, but it got him thinking, so he went down the hall and began collaborating with a theorist. The theorist constructed a computer model to test the mechanics involved, which constantly tested and was tested against Brown's observations. Quickly Brown realized there was an odd alignment between six of the 13 objects that simply should not happen randomly.
Initially the planet explanation seemed to be the only answer, but it did not quite fit right, until, almost by accident, they tried simulations where the planet of these objects had anti-aligned orbits. This means the part of their orbits that are closest to the Sun, called the perihelion, are 180º opposite each other. While this would explain it, the mechanics involved are so rare that it still was hard to accept, but then they found this solution answers other questions. Other Kuiper Belt objects have been discovered with unusual orbits that cannot normally be explained, including some that orbit perpendicularly to the plane the planets orbit on (known as the plane of the ecliptic). A ninth planet perfectly explains these orbits.
Currently the researchers are calling it Planet Nine and the math indicates it should have about 10 times the mass of Earth, and its average distance from the Sun is about 600 times that of the Earth, or 20 times that of Neptune. At this distance it should take between 10,000 and 20,000 years to complete one orbit. The speculation for its origin is that instead of four planetary cores existing in the early Solar System, which eventually became the gas giants, a fifth core could have also existed, but was ejected to its current eccentric orbit when it got too close to Jupiter or Saturn. Now we just need to find it, which could be very difficult depending on where it is in its orbit, and how much Sunlight may be hitting it. If it is near its farther point from the Sun, only the world's most powerful telescopes might spot it, but if it is closer other telescopes have a chance. In fact, it may have already been captured in previous surveys.
Posted: January 20, 2016 08:05AM
Defects can be very important in the semiconductor industry as these imperfections can translate into missed release dates, broken products, and lost revenue. Naturally then, many methods are used to reduce the number of defects, but as the circuitry gets smaller it becomes much harder to maintain a standard of less than one defect per 100 cm2. Luckily researchers at Argonne National Laboratory have made a discovery that should help, even as circuitry shrinks.
Current methods for creating computer chips involve photoresist polymers, which allow one to control where a material is etched. The problem is that there is a limit to how small these polymers can be. The new solution is to turn to block copolymer molecules, because they can self-assemble into desired shapes with great density and precision. The trick has been making sure the copolymers are in a stable state, so that the pattern will not change, as opposed to metastable states. Copolymers can exist in a metastable state for long periods of time though, so the researchers used supercomputers to figure out the energy barriers between these and lower-energy stable states. This allows the researchers to find the path the molecules can take from a metastable state to a stable state.
Armed with this knowledge, manufacturers should be able to take advantage of block copolymers to significantly reduce the defects in the products they produce. Now the researchers will continue their work by looking into more materials, building more complex patterns, and potentially creating 3D structures for more advanced technologies.
Source: Argonne National Laboratory
Posted: January 19, 2016 06:37AM
Geckos are very interesting animals in part because of their ability to climb on smooth surfaces at any angle. For a long time now, researchers have been trying to replicate the adhesive quality of gecko toes, but it appears this is not a means to allow humans to climb walls. Researchers at the University of Cambridge examined various species that also evolved sticky foot pads and found there is a size limit.
When it comes to successfully sticking to a surface, an animal's volume and surface area must be considered. As animals become larger, their volume and surface area both increase, but volume, which directly relates to mass, increases much faster than surface area. As it is the surface area that relates to the adhesion, more and more of an animal's body would have to become adhesive to successfully hold it. While a gecko only needs about 2% of its surface area to be sticky pads, a human would need about 40% to be sticky, or 80% on just one side.
Obviously such large pads are infeasible for humans, but we should not abandon hope just yet. There are some animals larger than geckos that developed separate means of sticking to surfaces by evolving stickier pads, but the exact mechanisms involved are not well understood. Also adhesives that replicate gecko feet could still find other applications than scaling buildings.
Source: University of Cambridge
Posted: January 15, 2016 07:21AM
There are a lot of people that want to see graphene used in electronics, because it has great electric properties, including very high electron mobility. It is still a relatively new material though, and so its properties and interactions are still unknown. Researchers at Tohoku University though have recently made a discovery about how graphene nanoribbons connect, which could help see them made into electronics.
Graphene itself is an atom-thick sheet of carbon atoms, and a nanoribbon of it would just be a long but narrow piece. The molecular structure of graphene is comprised of hexagons, like chicken wire, which means it can have multiple edge geometries, and these greatly affect its properties. Some of these geometries provide interconnection points, while others do not and it is important to know how connecting nanoribbons at these points will affect their properties. The researchers found that when connecting nanoribbons with zigzag edges, forming an elbow structure, the electronic properties are not affected. Electrons will travel smoothly across the connection and not suffer increased resistance.
This is very important to know because if graphene nanoribbons are to enter electronics, we need to know how to connect them and how their properties will be affected. With these answers, we are a step closer to using them to create high performance and efficiency electronics based on graphene.
Source: Tohoku University
Posted: January 14, 2016 07:02AM
There are a number of ways we can connect materials and objects today, and each method has its own pros and cons. Glues are relatively simple to apply but are not typically conductive or strong compared to soldering or welding, which also both require high heat. Researchers at Northeastern University however, have developed a metallic glue that could combine the best of both worlds.
This glue utilizes nanotechnology to adhere two objects together, and does so at room temperature with little pressure. It is comprised of metallic nanorods with one side coated in indium and the other in gallium. These nanorods are placed on substrates like teeth on a comb, but at an angle. When the substrates are pushed together and the nanorods interlace, the indium and gallium touch and form a liquid. The metal core of the nanorods though, turn that liquid into a solid, adhering the substrates together. The resulting bond is strong, resistant to air and gas leaks, and a good conductor of heat and electricity. These are qualities more typically seen with soldering or welding, but the temperatures they required can damage the objects involved and are relatively expensive processes.
Many of the applications for this glue can be found in the electronics industry as a potential replacement for solders and a replacement for thermal grease. Only time will tell, but it definitely has some good potential.
Source: Northeastern University
Posted: January 13, 2016 07:35AM
For years now there has been a broad effort to get rid of incandescent bulbs, citing their low efficiency at converting electricity to visible light. Compact fluorescents (CFLs) and LEDs are both more efficient technologies, but come with issues of their own such as color rendering and cost. Researchers at MIT and Purdue University though have found a means to significantly increase the efficiency of incandescent lights, which could potentially help them make a comeback.
Modern incandescent bulbs work by heating a tungsten filament to temperatures high enough that it emits a broad spectrum of light. The problem is that about 95% of what radiates from the bulb is heat and not the desired visible light. What the MIT and Purdue researchers have done is sandwich the filament between special photonic crystals that are made of common elements and with conventional methods. These crystals allow visible light to pass through them, but reflect the heat back to the filament. Because it is the heating of the filament that produces the light, this reflected heat causes more visible light to be emitted, effectively recycling it. In theory this approach could raise the luminous efficiency of incandescent lights to 40%, which beats the 7-15% of CFLs and 5-15% of LEDs. The proof-of-concept bulbs though only reached 6.6%, so there is still a ways to go, but that value is still triple that of traditional incandescent bulbs.
Obviously this research could potentially be used to bring back the warm glow of incandescent bulbs, without the waste heat few situations call for, but it has more applications than that. Thermo-photovoltaics could also benefit, as they convert heat into light that is then converted to electricity by a photovoltaic absorber.
Posted: January 12, 2016 06:15AM
Conservation of energy tells us that energy cannot can neither be created nor destroyed, but can be converted to different forms. Because the Sun provides us with so much energy, many are trying to find efficient ways to convert sunlight into something more useful, like electricity or heat, but storage becomes a problem. The Sun does not shine the entire day, so that energy is not always available, but researchers at MIT may have an interesting and very effective solution.
Currently many methods for storing the energy of Sunlight involve converting it to electricity, which is not always efficient. Instead the MIT researchers have examined how to store solar heat by a chemical reaction, which can easily be undone to release the energy on demand. This work builds on previous work, where the material involved was a liquid, but now it is a solid state polymer. When Sunlight strikes the molecules in the polymer, they are kicked into a new configuration with greater energy, and can stay there for long periods of time. By hitting it with a specific stimulus, like a jolt of heat, light, or electricity, the molecules will fall back to a configuration with a lower energy state, releasing the difference as heat.
This polymer material is based on azobenzenes and can be made very easily and cheaply with a two-step process. It can also be made into transparent thin films, which is important as those both open up possibilities. One potential application is as a film on car windshields that can store up solar heat and then release it to suddenly melt ice that may have built up on them. The researchers are currently working to remove a slight yellow tint to the material, making it more transparent, and increase its ability to release bursts of heat. Currently the bursts can be about 10 ºC over the ambient temperature, but they think they can reach 20 ºC.
Posted: January 11, 2016 08:18AM
Detecting single photons may not seem like something most people will have to worry about, but as we come closer to developing quantum communication networks, it is becoming more important. To that end, researchers at NIST have developed a new single-photon detector that cuts in half the jitter of a previous design. Reducing jitter allows for a higher bit rate.
The new detector uses nanowires of the superconductor molybdenum silicide. Even though a single photon does not carry much energy, it will still transition the nanowire back to being a regular conductor, which can be detected. What makes this new detector better is that molybdenum silicide is a superconductor at a higher temperature than that in NISTs previous design. This simplifies the necessary refrigeration and allows a higher electrical current to be run through it. This higher current cuts the uncertainty of when a photon is detected, or the jitter, from 150 picoseconds down to 76 ps. It also has a high detector efficiency of 87% at telecommunication wavelengths.
Beyond its uses for quantum communications, this new detector will likely also see use in tests of different physics theories that require examining the properties of billions of photons and entangled photons.
Posted: January 6, 2016 09:30AM
Most, if not all people know not to use their phone when driving, because the distraction can lead to accidents. That view may be changing slightly, thanks to researchers at MIT and the spinoff Cambridge Mobile Telematics (CMT).
Using MIT research, CMT developed the DriveWell app which monitors and scores your driving from 1 to 100. This score is based on data collected on road types, driving smoothness, speeding, phone distraction, and more, and all of the data is collected in the background. Some insurance companies offer devices that similarly track driver safety, but DriveWell takes it a bit further with its feedback and scoring. Instead of just offering insurance discounts, the scores can be collected into leaderboards, creating social games with badges and other prizes.
In South Africa, which has one of the highest traffic accident fatality rates (31.9 per 100,000 inhabitants), a competition was run with some 65,000 participants. While the competition ran for four months, after just two weeks the riskiest participants improved significantly with 40-50% less speeding, less hard braking, corning, and less phone usage, and in some cases the improvements were noticeable after just two days. The next step for CMT is, naturally, to bring DriveWell and its other apps and products to as many drivers as possible.
Posted: January 5, 2016 07:22AM
If there is one thing scientists have to bear in mind at all times, it is to always take a second look at bad results. Researchers at NIST and IBM were disappointed when their attempts to create nanowires failed due to a contaminant, but then they put the results under the scanning electron microscope. To their surprise, there were long straight channels etched into the semiconductor they were working with, which could have some interesting applications.
The researchers were working with gold nanoparticles on indium phosphide and were expecting nanowires to be created, but the presence of water completely changed that. Initially a gold layer was deposited on the semiconductor, and when heated this film broke apart to form droplets, and some of the indium phosphide was absorbed by the droplets to form a gold alloy. With water vapor present, these nanoparticles are surrounded by water molecules that will etch into the semiconductor by oxidizing it. This results in indium oxide and phosphorus, which evaporate away. At temperatures below 300 ºC, pits are formed, but at 440 ºC and above, long V-shaped channels are created. These channels are the size of the nanoparticle, which can be controlled.
While this is not the nanowires the researchers were aiming for, this ability to create precise and straight nanochannels could be used to bring lasers, sensors, wave guides and more to lab-on-a-chip devices. So far the researchers have found this works with indium phosphide, gallium phosphide, and indium arsenide, which are used to create LEDs and high-speed electronics, but we could see it adapted for patterning channels into silicon and other materials in the future.
Posted: January 4, 2016 07:33AM
Carbon is an amazing element as its various forms can have wildly different and useful properties, with graphene exhibiting great strength and conductivity. This would make it ideal for use in electronics, except that it lacks a natural band gap and thus requires special doping to act as a semiconductor. Now researchers at Berkeley Lab have found a new dopant that is very effective at turning graphene into a p-type semiconductor.
Graphene is a sheet of carbon just one atom thick, making it effectively two dimensional, and it is able to carry electrons at speeds near that of light. Without a means of switching off that conductivity though, its usefulness is limited. By doping it with other atoms and molecules though, its properties can be altered and the Berkeley researchers found that F4TCNQ does a very good of this, when the graphene is on a boron nitride substrate. This pairing causes the F4TCNQ molecules to self-assemble islands on the graphene that pull in electrons. This causes the graphene to take on a positive charge, making it a p-type semiconductor, and also reduces its energy, which leads to island cohesion.
On its own this discovery could help graphene enter electronics, but the mechanics behind the island forming should occur with other materials as well. This could open the door to actually tuning the properties of graphene for use in devices.
Source: Berkeley Lab
Posted: December 24, 2015 11:19AM
When we think about how fast the various components in our computers perform, we likely forget about important distance matters. The greater the distance between two components, the less data can be efficiently transported without bumping up the power to possibly unsustainable levels. One way around this would be to use light to transmit data between components, and now researchers at MIT, the University of California, Berkeley, and the University of Colorado have successfully built a working optoelectronic microprocessor.
Successfully designing an optoelectronic microprocessor is far from easy because the materials involved can have conflicting properties. For example, the free electrons in conductors tend to absorb light, which interferes with optical transmissions, so the researchers had to figure out a way to use the positive charge carriers left behind to protect the transmissions. To convert the optical signals into electricity the researchers patterned metal onto the optical ring resonators within the chip. On its own the metal will not interfere, but when a voltage is applied it can alter the optical properties of the resonator or register changes in the optical signal. It turns out the energy cost of these detectors can be as low as a picojoule, which is a tenth of what traditional computer chips require.
Built on a 45 nm process at GlobalFoundries and with 850 optical components and 70 million transistors, this microprocessor is not about to outperform modern processors, but it does not have to. It has proven that it is possible to build optoelectronic microprocessors with standard semiconductor fabrication methods. There is still plenty of work to do, but this is a significant step towards bringing optics into computer components, and the benefits that come with them.
Posted: December 23, 2015 07:15AM
Generally, liquids and electronics do not mix well and can lead to the electronics' failure, which can be a problem if applying a material to a device requires a liquid. This has been the case with metal organic frameworks (MOFs) as they are grown and applied with a liquid solvent, but because MOFs could increase the performance of electronic devices, researchers have been searching for a solution. Now those at CSIRO, the University of Leuven, and the National University of Singapore have discovered how to grow these crystals from a vapor.
The two points that make MOFs so interesting are that we can design their size and shape and that they are very porous. A gram of MOF crystals can have a surface area equal to that of a football field. By trapping different molecules within these crystals, their properties can be altered, and that could translate to fitting more transistors into a microchip. Beyond that, these MOFs could also be used to create advanced chemical sensing devices.
Source: CSIRO Australia
Posted: December 22, 2015 07:28AM
Indium tin oxide, or ITO, is a transparent conductor that is used in a great many displays, but because indium is a rare and expensive material, researchers have been searching for a replacement for years now. Many alternatives have been created, but none have quite matched the optical and electrical properties needed to replace ITO. Now researchers at Penn State have discovered that a class of materials that can compete with ITO, while being significantly cheaper.
These materials are known as correlated metals, which differ from conventional metals because of how their electrons flow. In normal metals the electrons move like particles in a gas, but in a correlated metal they can interact with each other, like particles in a liquid. This actually grants the materials high optical transparency and metal-like conductivity, and when light shines on it, it becomes even more transparent. The two materials the researchers specifically worked with were strontium vanadate and calcium vanadate, but the work could lead to other correlated metals being discovered.
Right now a kilogram of indium costs about $750, and as we use the limited supply, its price will increase even more, while vanadium, which is far more common, costs just $25 per Kg, and strontium is even cheaper. Combined with the potential of being able to just replace ITO with strontium vanadate in the manufacturing process and it becomes an even more enticing alternative.
Source: Penn State
Posted: December 21, 2015 05:48AM
If you peruse the Guinness World Records you are certainly going to find some unusual entries, and one of the new ones is the world's smallest inkjet-printed color image. The image of clown fish around sea anemones is just 80 micrometers by 115 micrometers, which is small enough to require a special microscope to see. It was printed using 3D NanoDrip technology developed at ETH Zurich, and has now been spun off to the company Scrona.
To create the 24 bit color image, the quantum dots were used, which are nanoparticles that can be designed to emit specific colors. The colors were put down in layers of red, green, and blue dots with their thicknesses controlled with sub-nanometer precision. The distance between each pixel is 500 nm, so the image itself is 25,000 DPI.
Currently quantum dots are finding uses in displays, thanks to their intense colors, and with this level of precision having been achieved we could see some new and interesting applications for them in electronics and optics. Right now though, the precision is being used to print images and you can have one made for you by backing Scrona's Kickstarter, which ends January 9.
Source: ETH Zurich
Posted: December 17, 2015 02:20PM
For any number of reasons, chances are we have all gotten upset by something while we are at our computers. Thanks to researchers at Brigham Young University, computers can now determine your emotional state by following your mouse movements.
The researchers discovered that there is a link between our mouse movements and our emotional states, with negative emotions being associated with slower movements. This includes when you are frustrated. Being upset of confused, however, makes our movements more jagged and sudden.
While there can definitely be many applications for this work, the researchers see it potentially being used to identify when a website experience turns bad. With this information, a web designer knows where to rework a site to make it easier to use. Potentially the research could also be applied to mobile devices that rely on touchscreens for interaction.
Source: Brigham Young University
Posted: December 15, 2015 05:51AM
In a great many fields, heat transfer is of significant importance, and computer chips are no exception as circuitry reaches ever smaller sizes. On the macroscale we live on, the rules for how heat flows are well understood, but at the nanoscale the game can change and for decades a debate has continued over how to describe it. At last we appear to have an answer, thanks to researchers at the University of Michigan.
Radiative heat is the light that objects emit from the movement of particles, and over a century ago Max Planck wrote the equations that explained the process. However, as the gap between two objects shrinks, the equations start to fall apart. A new theory was developed in the middle of last century by Sergei Rytov, but some experiments reported significant differences from it. Now the Michigan researchers with a very advanced lab have confirmed Rytov's theory experimentally. The experiment involved placing a heated surface of silica, silicon nitride, or gold beneath a scanning thermal microscopy probe of the same material, and moving the probe closer to the surface. The researchers discovered that at very small distances between the surfaces and the probes, the heat flow would jump to 10,000 times that we experience on the macroscale. The reason is because the surface and evanescent waves of the two objects start to overlap at these distances.
The significance of this research is with how it will impact nanotechnologies that require controlling heat flow. It was for this reason that the researchers worked with silica, silicon nitride, and gold as all three materials are used in nanotechnologies.
Source: University of Michigan
Posted: December 14, 2015 02:27PM
For a number of reasons, there are millions of people around the planet using systems like Tor to hide their online activities and communications. While Tor is relatively popular, it does have some flaws that could allow an adversary to compromise it, especially if the adversary is particularly powerful. A team of researchers at MIT decided to design a system that could thwart such powerful attackers and have appropriately named it Vuvuzela, after the noisemakers from the 2010 World Cup.
Vuvuzela is a dead-drop system, so when someone sends a message, it is sent to a specific memory address in a server that the receiver checks later. To protect against someone monitoring traffic patterns to identify users, messages are sent by every user at regular intervals, even if the messages contain no information. Also the messages are wrapped in layers of encryption, which are peeled away by the servers when they arrive, but the order the messages arrive at the servers is a random permutation. Additionally the servers generate multiple dummy messages that are also encrypted and meant for other locations. If three servers are being used by Vuvuzela, an attacker could compromise two of them and still the messages would be safe because there would be too much noise to know what messages are real, where they came from, or where they are going.
As it is now, Vuvuzela is not ready to be deployed, and it has some limitations that may prevent it from being rolled out. However, it could inspire new systems with its intriguing use of differential privacy to ensure secure communication.
Posted: December 14, 2015 07:17AM
While real artificial intelligence has not reached the heights predicted by science fiction yet, the fact is that it has existed for decades. Continually researchers have been working to improve upon machine learning methods, with the ultimate goal of matching the capabilities of the human mind. Now researchers at New York University have successfully taught a computer to recognize and draw symbols similarly to the way humans do.
From a young age, humans are able to learn simple visual concepts, like characters to a language, with just a few examples, but it can take thousands of examples before a computer gets it right. To cut down on that number, the researchers developed a Bayesian Program Learning (BPL) framework. It works by having the program recreate the symbol using computer code, and does so in a probabilistic way, so each time the code is run it produces a slightly different version of the symbol. This generative approach allows it to learn the symbol more quickly, and how a symbol can vary just because of who draws it.
To test the effectiveness of this approach, the researchers set up a visual Turing test with various characters. Less than a quarter of the judges did better than chance in the test, proving this approach's power.
Source: New York University
Posted: December 11, 2015 02:26PM
The plan for quantum computers is to take advantage of exotic phenomena like entanglement and superposition to perform computations classical computers typically struggle to. Still some problems would be difficult for a quantum computer to solve, but about ten years ago an idea was suggested to make it possible. That idea was to send information back in time and now researchers at the Center of Quantum Technology (CQT) at the National University of Singapore, and other universities, have determined that this can be done without risking causality paradoxes.
The original idea was to send information back in time on closed timelike curves, with the time-traveling particles being entangled with a system in a laboratory. These would be complete loops in spacetime, so the particles would be able to interact with their past selves. This creates the risks of paradoxes, like the grandfather paradox where an entity destroys its predecessor, which means it should never have existed to destroy its predecessor. The solution the researchers have come up with is to use open timelike curves, which avoid the causality issues by preventing the particles from interacting with themselves. The computing benefits remain however, so we can solve otherwise unsolvable problems without risking paradoxes.
Of course we still lack any means of creating these timelike curves, even though General Relativity technically allows them to exist.
Posted: December 11, 2015 07:32AM
There are a variety of display technologies available now, and each has their advantages and disadvantages. Many more technologies are also being developed now to satisfy different needs, such as efficiency, brightness, and color accuracy. Among these technologies are displays that utilize nanoparticles to produce colors, and researchers at Rice University have discovered the first reversible method for changing the color between metal nanoparticles.
Originally the researchers were working on dimers made from gold nanoparticles, because theory predicted that changes in charge density would result in color changes. While this work was successful, the changes were small and differed from the predictions. They investigated and found the difference was because the silver also used in the experiments was oxidizing and forming bridges between the gold nanoparticles. Instead of searching for a way to remove these bridges, the researchers decided to work with these bridges and found they could control the distance between the nanoparticles via these bridges. It just took changing the voltage applied to the nanoparticles to alter the bridge.
By changing the distance between the gold nanoparticles, the plasmon coupling between them changes, and that affects what color of light scatters from them. Potentially this could be used to create displays and many other plasmonic devices, such as switches and modulators that would have many other applications.
Source: Rice University