Max Planck Institute for Intelligent Systems
Max Planck Institute for Intelligent Systems
News Article | April 17, 2017
We are difficult for computers to understand. Our actions are sufficiently unpredictable that computer vision systems, such as those used in driverless cars, can’t readily make sense of what we’re doing and predict our next moves. Now fake people are helping them to understand real human behaviour. The idea is that videos and images of computer-generated bodies walking, dancing and doing cartwheels could help them learn what to look for. “Recognising what’s going on in images is natural for humans. Getting computers to do the same requires a lot more effort,” says Javier Romero at the Max Planck Institute for Intelligent Systems in Tübingen, Germany. This, he says, is one of the biggest things holding back progress with driverless cars. Using synthetic images to train computers could give them more meaningful information about the human world. At the moment, the best computer vision algorithms are trained using hundreds or thousands of images that have been painstakingly labelled to highlight important features. This is how they learn to distinguish an eye from an arm, for example, or a table from a chair. But there is a limit to how much data can be realistically labelled this way. Ideally, every pixel in every frame of a video would be labelled. “But this would mean instead of creating thousands of annotations, people would have to label millions of things, and that’s just not possible,” says Gül Varol at École Normale Supérieure in Paris, France. So Varol, Romero and their colleagues have generated thousands of videos of “synthetic humans” with realistic body shapes and movement. They walk, they run, they crouch, they dance. They can also move in less expected ways, but they’re always recognisably human – and because the videos are computer-generated, every frame is automatically labelled with all the important information. The team created their synthetic humans using the 3D rendering software Blender, basing their work on existing human figure templates and motion data collected from real people to keep the results realistic. The team then generated animations by randomly selecting a body shape and clothing, and setting the figure in different poses. The background, lighting and viewpoint were also randomly selected. In total, they generated more than 65,000 clips and 6.5 million frames. With all this information, computer systems could learn to recognise patterns in how pixels change from one frame to the next, indicating how people are likely to move. This could help a driverless car tell if a person is walking close by or about to step into the road. As the animations are in 3D, they could also teach systems to recognise depth – which could help a robot learn how to smoothly hand someone an object without accidentally punching them in the stomach. The work will be presented at the Conference on Computer Vision and Pattern Recognition in July. “With synthetic images you can create more unusual body shapes and actions, and you don’t have to label the data, so it’s very appealing,” says Mykhaylo Andriluka at Max Planck Institute for Informatics in Saarbrücken, Germany. He points out that other groups are using graphics from video games like Grand Theft Auto to improve computer vision systems, as these can offer a relatively lifelike simulation of the real world. “There’s been huge advances in the realism of virtual images. We can use this to teach computers to see things,” says Romero.
News Article | April 24, 2017
Plasmonic nanoparticles exhibit properties based on their geometries and relative positions. Researchers have now developed an easy way to manipulate the optical properties of plasmonic nanostructures that strongly depend on their spatial arrangement. The plasmonic nanoparticles can form clusters, plasmonic metamolecules, and then interact with each other. Changing the geometry of the nanoparticles can be used to control the properties of the metamolecules. “The challenge is to make the structures change their geometry in a controlled way in response to external stimuli. In this study, structures were programmed to modify their shape by altering the pH,” tells Assistant Professor Anton Kuzyk from Aalto University. Utilization of Programmable DNA Locks In this study plasmonic metamolecules were functionalized with pH-sensitive DNA locks. DNA locks can be easily programmed to operate at a specific pH range. Metamolecules can be either in a “locked” state at low pH or in relaxed state at high pH. Both states have very distinct optical responses. This in fact allows creating assemblies of several types of plasmonic metamolecules, with each type designed to switch at different a pH value. The ability to program nanostructures to perform a specific function only within a certain pH window could have applications in the field of nanomachines and smart nanomaterials with tailored optical functionalities. This active control of plasmonic metamolecules is promising for the development of sensors, optical switches, transducers and phase shifters at different wavelengths. In the future, pH-responsive nanostructures could also be useful in the development of controlled drug delivery. The study was carried out by Anton Kuzyk from Aalto University, Maximilian Urban and Na Liu from Max Planck Institute for Intelligent Systems and the Heidelberg University, and Andrea Idili and Francesco Ricci from the University of Rome Tor Vergata.
News Article | May 15, 2017
If you want a robot to pick up a coffee cup, a cherry tomato, or a bag of packaged food, you’ll have to deal with a lot of programming and quite possibly some cleanup in aisle 3. A child can handle all kinds of shapes and materials, and apply the right combination of force and delicacy when lifting them, but getting machines to reproduce these seemingly simple skills hasn’t been easy. Now, a team of researchers has developed a new way to pick up objects that solves many of these problems. Modeled on geckos’ toes, it could also help robots climb irregularly shaped walls. Engineers have long sought inspiration from nature, and one of nature’s most inspiring specimens is the gecko, that fearlessly climbing lizard with the power to stick to walls and ceilings. These feats rely on Van der Waals forces. At any moment, atoms tend to be slightly positively charged on one side and negatively charged on the other, as a result of the random behavior of their electrons. When two atoms come close to each other, Van der Waals forces can generate an attraction. A gecko’s toe pads are covered with tiny, hairlike fibers that maximize its contact with surfaces, amplifying the Van der Waals effect. Scientists have created materials that exploit synthetic arrays of microfibers to replicate geckos’ stickiness, but there’s a trade-off. Making geckolike materials stick well takes pressure, which requires a rigid backing like a paddle. But such a backing in turn prevents these materials from adhering to curved surfaces. The new research avoids this tradeoff, offering both flexibility and grip, by arraying microfibers on a thin, stretchy membrane to create what the researchers call fibrillar adhesives on a membrane (FAM), and providing a new type of backing for it. In the soft robotic gripper, the FAM covers the wide opening of a shallow funnel made of soft rubber, 18 millimeters across, with the narrow opening connected to an air pump. After the FAM touches a flat or curved object, air is pumped out of the funnel, flattening the shallow cone against the object, whatever its contours. In tests, researchers found that with a contact area of only 2.5 square centimeters (about the size of a dime), their gripper could lift objects weighing more than 300 grams (close to the weight of a can of soda), they report today in the Proceedings of the National Academy of Sciences . It could also pick up a coffee cup from the convex outside, the concave inside, or a point on the handle; lift a cherry tomato without harming it; or grab a plastic bag of packaged food. Inflating the gripper releases the objects. “This is a very nice contribution to a field which in its early days was just focused on the nanoscale fibrils rather than the backing,” says Kimberly Turner, a mechanical engineer at the University of California, Santa Barbara (UCSB), who was not involved in the research. “The design of the backing is key to making these adhesives function properly for most applications, and this is a very exciting development.” The technology has several potential applications, says Metin Sitti, an author of the study and a mechanical engineer at the Max Planck Institute for Intelligent Systems in Stuttgart, Germany, and Carnegie Mellon University in Pittsburgh, Pennsylvania. In factories, such grippers could help assemble delicate electronics or move around objects with complex shapes such as custom car parts. They could also be used in biomedicine for picking up organs. And they would enable robots to climb everything from airplanes to nuclear plants, for inspection and maintenance. Still, the gripper must meet several requirements before it can be used widely, says Elliot Hawkes, a mechanical engineer at UCSB and Stanford University in Palo Alto, California, who has also worked on geckolike grippers. It must be durable (able to lift and release things hundreds of thousands of times), scalable (able to lift things heavier than a kilogram), and worth the cost relative to other types of grippers such as clamps or suction cups. Sitti sees no obstacle to making bigger grippers, tens of centimeters wide, and using many of them in concert to grab and lift heavy objects. The team still needs to test the gripper’s durability, however. When a climbing robot fails, cleanup means more than mopping up tomato juice; it means sweeping up the remains of an expensive robot.
News Article | May 18, 2017
A soft gripping system capable of strong and reversible adhesion to non-flat surfaces is particularly critical for applications in transfer printing, robotics, and precision manufacturing. In these areas, adhesion-controlled grasping of complex three-dimensional surfaces is very challenging because the adhesive must be soft enough to enable intimate contact under light pressure but stiff enough to support high load and fracture strength. "Until now, there has been a fundamental trade-off between soft, mechanically compliant interfaces and high adhesion strength," said Metin Sitti, a director in Max Planck Institute for Intelligent Systems. "With this soft gripping system, we've demonstrated that you can have both." Sitti, also an adjunct professor of mechanical engineering at Carnegie Mellon University, has long focused his research on the unique features of the gecko lizard. The gripping system contains a soft adhesive membrane with microscopic pillars that are inspired by the sticky foot hairs on a gecko's foot. The membrane is supported by a deformable gripper body that controls the gripping strength of the adhesive through changes in internal air pressure. It exhibits enhanced and robust adhesion on various sizes of curved and deformable surfaces, outperforming other adhesion methods. Another area of application for this soft gripping system is in personal robotics. "One day, we might have robots that act as assistants and care-givers—for example, a robot that can help unload a dishwater or manage medication. The ability for the robot to manipulate a broad range of objects is crucial, " said Carmel Majidi, an associate professor of mechanical engineering at Carnegie Mellon. Majidi and Sitti acknowledge Sukho Song and Dirk-Michael Drotlef in performing this multidisciplinary research and the Max Planck Society for the financial support. Majidi also acknowledges support from the Office of Naval Research. Explore further: Flexible soft gripper mimics gecko to pick up objects with curved surfaces More information: "Controllable load sharing for soft adhesive interfaces on three-dimensional surfaces," Proceedings of the National Academy of Sciences, DOI: 10.1073/pnas.1620344114
News Article | December 5, 2016
Whether building organs or maintaining healthy adult tissues, cells use biochemical and mechanical cues from their environment to make important decisions, such as becoming a neuron, a skin cell or a heart cell. Scientists at UC Santa Barbara have developed a powerful new technique that reveals for the first time the mechanical environment that cells perceive in living tissues -- their natural, unaltered three-dimensional habitat. "Knowing how cells respond to mechanical cues in the living embryo and how they physically sculpt tissues and organs in the 3D space will transform the way we think about developmental processes," said Otger Campàs, a professor in the Department of Mechanical Engineering at UCSB and senior author on the paper that reports this novel technique in Nature Methods. "Importantly, this knowledge will help us better understand healthy tissue homeostasis and the wide range of diseases that involve abnormal tissue mechanics, especially cancer." The growth and development of a living organism is a choreography of cellular movements and behaviors that follow internal genetic guidelines and specific biochemical and mechanical signals. All these events conspire over time to create a variety of complex forms and textures that make our tissues and organs functional. Scientists have focused for decades on the role of biochemical cues in embryonic development, Campàs said, because no techniques existed to measure the mechanical cues that cells are exposed to during the formation of tissues and organs. "We know that the mechanical environment of cells is important," explained Campàs, who holds the UCSB Mellichamp Endowed Chair in Systems Biology and Bioengineering. "Growing stem cells on synthetic surfaces with different levels of compliance showed that stem cells would become a different cell type depending solely on the mechanical environment they perceive. If you put embryonic stem cells on a substrate like Jell-O -- mechanically similar to brain tissue -- they turn into neurons. But if you put them on something harder, similar to embryonic bone, they turn into bone-like cells." Until now, scientists did not have a method for studying the mechanical characteristics of native cellular environments -- that is, cells surrounded by other cells and matrix scaffolds within living tissues. As a consequence, it was not possible to know how cells respond to the mechanical cues they perceive as they build tissues and organs. "The technique we developed allows the measurement of the mechanical properties such as stiffness and viscosity within living tissues," said author Friedhelm Serwane, who is currently at the Max Planck Institute for Intelligent Systems in Stuttgart, Germany. "This is exciting because important cell functions are controlled by those mechanical properties. If we can measure the mechanical properties within living organisms now we might be able to understand better how this relationship between mechanics and biology works." Key to this method are tiny magnetically responsive droplets inserted between cells in the developing embryo. When exposed to a magnetic field, these magnetic droplets deform, pushing on nearby cells. By carefully controlling the composition of the droplets and the strength of the magnetic field, the forces applied by the droplet can be controlled, and the response of the surrounding tissue reveals its mechanical characteristics as well as the cues that cells are exposed to as the tissue grows. This technique is complementary to a previous methodology developed by Campàs and colleagues that revealed the forces that cells apply to each other in growing tissues. The scientists applied their new technique to study how the vertebrate body axis is mechanically built. Using embryos of zebrafish, which was selected for its rapid development and optical transparency, they could show that the mechanical properties of the tissue change along the body axis, facilitating the extension of the body at its posterior end. Inserting magnetic droplets at different locations in the tissue, and generating forces by applying a magnetic field to the droplets, the researchers showed that the tissue behaves like a fluid while growing, with similar mechanical characteristics as thick honey. The data showed that the tissue is more fluid at the posterior end where it was growing, and becomes less fluid far from the growing region. "It is similar to glass-blowing," said Campàs. "The tissue is more fluid in growing regions and 'fixes' its shape by becoming less fluid where it does not need to expand." The scientists' findings have wide implications in the effort to understand how organs are sculpted into their shapes and how cells respond to their native mechanical environment both in healthy tissues and during disease. The Campàs lab is studying several of these questions, including how limbs are built and how mechanical changes in tumors affect the behavior of malignant cells and the growth of the tumor. Research on this paper was conducted also by Alessandro Mongera, Payam Rowghanian, David A. Kealhofer, Adam A. Lucio and Zachary M. Hockenbery. The project received support from the National Institutes of Health and the National Science Foundation.
News Article | October 5, 2016
Microbots have made great strides in recent years as scientists and engineers work on creating cell-sized robots that can swim through the bloodstream and act as tiny medical commandos. However, such tiny automata are tricky to steer and control, so researchers led by Clemens Bechinger of the Max Planck Institute for Intelligent Systems are taking a page from nature and developing simple microswimmers that can mimic the light-seeking behavior of some bacteria. Phototaxis is a common behavior in the animal and plant kingdoms. When we talk about being drawn like a moth to a flame, that's an example of phototaxis. But it can also act in the opposite direction, as turning on the lights in a cheap hotel room and watching the cockroaches scatter can show. What makes phototaxis so attractive to roboticists is that it is a very simple behavior based on light intensity. Some creatures are drawn to light, some are repelled by it, and some seek out an area of just the right intensity. What the Max Planck team wanted to do was to mimic this behavior in a way that could be applied to robots the size of bacteria as a way to not only steer, but to move them at the same time. According to the team, simplicity was the key to creating the microswimmers. Building phototactic robots on a human scale is very easy and the first robotic "tortoises" were built by pioneering cyberneticist William Grey Walter back in the late 1940s. But for all their being "simple" life forms, bacteria use highly complex mechanisms to sense and respond to light, which the team had no hope of duplicating in a microbot, so they settled on a remarkably minimalist design. The microswimmers don't look much like robots or bacteria. In fact, they are nothing but glass microbeads with a width of a few thousandths of a millimeter with one hemisphere covered in carbon black. This led to their being called "Janus particles" after the two-faced Roman god of beginnings and endings. These are suspended in a solution of water and a soluble organic chemical, which under heat separates from the water. When a light is shone uniformly on the Janus particles, the black side heats up more than the other and the solution next to it breaks up, causing a lower concentration. Then, like adding a drop of fresh water to a salt solution, the higher concentrate rushes in to reestablish the balance. This acts like oars in the particle as the solution flows past, pushing the particle in the direction of its transparent side. But the Janus particles were still a long way from being controllable. According to Bechinger, when an even light was set over the microswimmers, they'd go off in any direction. However, if the light was changed to produce a gradient of light to dark, the particles would swim toward the light. More importantly, they would steer toward it because the light would cause any particle not pointing at it to heat unevenly, which would make the solution flow faster around one side than the other and cause it to rotate. Unfortunately, Bechinger's team found that in ordinary light, the effect was only effective over a tenth of a millimeter before the particles started to veer off in random directions. Researcher Celia Lozano found a way to navigate over longer distances through the use of a system of lasers, lenses, and mirrors that produced a saw-toothed light field made up of areas of increasing and decreasing brightness. Any particles in the areas of decreasing brightness moved farther into the darkness, but the ones in the light areas went straight for the light and maintained their course even when passing through the areas of decreaasing brightness because the area was too narrow for them to have time to reverse course. The result was controllable, stable movement. The Janus particles are still in the laboratory phase, but their simple, easy to manufacture design holds great promise. According to the team, the swimmers can not only be steered by light, but also chemically, which means they could one day be tailored to seek out tumours in the body to deliver precise doses of chemotherapy.
News Article | February 15, 2017
An enzyme-propelled nanorobot: urease-coated nanotubes turn into a propulsion system in a urea-containing liquid because the enzyme breaks down the urea into gaseous products. Since the tubes always have small asymmetries, the reaction products generate a current in the fluid which propels them out of the tube like a jet. Credit: MPI for Intelligent Systems Nanorobots and other mini-vehicles might be able to perform important services in medicine one day – for example, by conducting remotely-controlled operations or transporting pharmaceutical agents to a desired location in the body. However, to date it has been hard to steer such micro- and nanoswimmers accurately through biological fluids such as blood, synovial fluid or the inside of the eyeball. Researchers at the Max Planck Institute for Intelligent Systems in Stuttgart are now presenting two new approaches for constructing propulsion systems for tiny floating bodies. In the case of one motor, the propulsion is generated by bubbles which are caused to oscillate by ultrasound. With the other, a current caused by the product of an enzymatic reaction propels a nanoswimmer. Jet aircraft have led the way. They burn fuel, eject the combustion products in one direction and as a result move in the opposite direction. Researchers at the Max Planck Institute for Intelligent Systems in Stuttgart do it in a very similar way - albeit on a much smaller scale. Their underwater-nanorobot is a single-walled nanotube made of silicon dioxide, a mere 220 nanometres (billionths of a metre) in diameter. A particle of that nature would not normally be able to propel itself in fluids. The scientists therefore coated either only the inner or the inner as well as the outer surface or of the nanotube with the enzyme urease which breaks down urea into ammonia and carbon dioxide. If a nanotube prepared in this way is introduced into a fluid containing urea, this urea is broken down at the urease-coated internal wall. The reaction products generate a current in the fluid which propels them out of the tube like a jet. As such a nanoswimmer either is thinner at one end than at the other or the the urea is not distributed homogeniously over its surface, this results in a thrust, so that the micro-swimmer experiences propulsion in the opposite direction – as in a jet aeroplane. The nanojets reached speeds of 10 micrometres per second, i.e. almost four centimetres per hour. The smallest jet engine in the world Admittedly, coating a nanorobot to achieve a chemical drive is by no means new. However, the tube now presented, with its 220 nanometre opening, represents the smallest jet propulsion system so far constructed in the world. "Our previous record, which is still in the Guinness Book of Records, was around three-times bigger", explains Samual Sanchez who leads the Smart NanoBioDevices Group at the Max Planck Institute for Intelligent Systems in Stuttgart and at the same time holds a professorship at the Institute for Bioengineering of Catalonia in Barcelona. And there is another new aspect of the nanojet which scientists from the Harbin Institute of Technology in Shenzhen in China also helped to develop: for the first time, all the materials and reaction partners used are fully biocompatible. "Previous chemical drives of this kind were usually based on a metallic catalyst at the surface of which hydrogen peroxide was broken down into hydrogen and oxygen molecules", says Sanchez. Oxygen bubbles are created in the process, which creates a thrust in the opposite direction. Both the hydrogen peroxide and the gas bubbles would have disadvantages if used in the human body. But this is not the case with the urease-coated version with its water-soluble – and therefore bubble-free – reaction products. "Urease occurs anyway in the human organism", Sanchez explains. The researchers now want to test the biocompatibility more precisely – and in the process examine whether they can succeed in implanting such micro-tubes into individual cells. "That would be necessary, of course, in order to bring drug molecules to their destination, for example", says Sanchez. While gas bubbles were still unwanted in the approach specified, they form the very centrepiece of a entirely new principle of propulsion for minirobos, which colleagues at the Institute in the Micro, Nano and Molecular Systems Group led by Peer Fischer propose. However, here the gas bubbles are not bubbling freely through the fluid and therefore cannot damage the organism. Rather, the researchers enclose the micro-bubbles in small cylindrical chambers along a plastic strip. To provide the drive, therefore, the gas bubbles expand and contract cyclically because ultrasound causes them to oscillate. As the pulsating bubbles are in chambers open on one side, they only expand through this opening. In the process, they exert a force on the opposite wall of the chamber which propels the plastic strip. In order to achieve propulsion worth mentioning, the researchers arranged several chambers with air bubbles in parallel on their polymer strip. A notable aspect: the sound wave frequency required to cause them to oscillate depends on the size of the tiny bubbles. The bigger the bubbles, the smaller the corresponding resonant frequency. The researchers used this connection to cause their swimmer to rotate alternately clockwise and anti-clockwise. To do so, they placed bubbles of different sizes on the two halves of the four, long cuboid faces divided lengthwise. Two different sound frequencies were then used in a liquid to each cause all the bubbles of one size to oscillate. In this way, the scientists generated thrusts exclusively on one-half of the cuboid face which caused it to rotate on its own axis. This small acoustically driven rotation motor with longitudinal areas each five square millimetres in size achieved up to a thousand rotations per minute in the process. "The variation in the size of the bubbles thereby enables a mini-swimmer to deliberately steer in different directions", says Tian Qiu, who also conducts research at the Max Planck Institute in Stuttgart and played an appreciable role in the study. According to Qiu, a further benefit of the new principle of propulsion is that even swimmers with a complicated geometric structure can be coated with the wafer-thin strips together with chambers for the bubbles. He goes on to explain that the use of ultrasound is also suited to optically impenetrable media such as blood. Light waves, which are also a potential control instrument for micro-drives, can achieve nothing in this case. The researchers now want to use tests in real biological media to check whether the new drive principle is also able to make the most of its advantages in practice. More information: Xing Ma et al. Bubble-Free Propulsion of Ultrasmall Tubular Nanojets Powered by Biocatalytic Reactions, Journal of the American Chemical Society (2016). DOI: 10.1021/jacs.6b06857
News Article | February 28, 2017
Reservoir for heavy hydrogen: Molecules of the heavy hydrogen isotopes deuterium and tritium preferentially bind to copper atoms in a metal-organic framework compound. The metal atoms are therefore symbolically represented as shells in this image. Credit: University Leipzig / Thomas Häse Deuterium and tritium are substances with a future - but they are rare. The heavy isotopes of hydrogen not only have numerous applications in science but could also contribute to the energy mix of tomorrow as fuels for nuclear fusion. Deuterium is also contained in some drugs that are currently undergoing regulatory approval in the US. However, the process of filtering deuterium out of the natural isotopic mixture of hydrogen is at present both difficult and expensive. Scientists from the Max Planck Institute for Intelligent Systems, the Max Planck Institute for Solid State Research, the University of Leipzig, Jacobs University Bremen, the University of Augsburg, and Oak Ridge National Laboratory (USA) may be able to remedy this problem. They have presented a metal-organic framework compound that can be used to separate the two isotopes from normal hydrogen more efficiently than previous methods. In drugs, deuterium has a life-prolonging effect – albeit initially only for the active substance itself. The human metabolic system breaks down molecules carrying the deuterium isotope, which is twice as heavy as hydrogen, more slowly than the same substance incorporating normal hydrogen. Drugs containing deuterium can therefore be given in smaller doses, which means that their side effects are also reduced. Deuterium, like the even heavier radioactive hydrogen isotope tritium, also plays a role in nuclear fusion. This process, which makes stars shine, may some day fuel power plants in which atomic nuclei are fused together, releasing large amounts of energy in the process. Whereas deuterium has only been used in pharmaceuticals for a short time and its potential use in power plants still lies in the future, it has long been used in science, for example to track the path of nutrients through the metabolic system. "Deuterium and, to a certain extent, tritium are useful in some applications," says Michael Hirscher, who, as Leader of a Research Group at the Max Planck Institute for Intelligent Systems, has played a key role in the research. "To date, however, it has been very difficult to separate deuterium from light hydrogen," he says. Deuterium is obtained from heavy water, which occurs in natural water at a concentration of just 15 parts per thousand. The heavy water is first isolated by a combination of chemical and physical methods, such as distillation, to obtain deuterium gas. The whole process is so intricate and energy-intensive that a gramme of deuterium with a purity of 99.8 percent costs around 100 euros, making hydrogen's heavy brother around three times more precious than gold, although deuterium is more than 300 times more abundant in the oceans and Earth's crust than gold. "Our metal-organic framework compound should make it easier and less energy-intensive to isolate deuterium from the naturally occurring mixture of hydrogen isotopes," says Dirk Volkmer, whose colleagues in the Department of Solid-State Chemistry at the University of Augsburg synthesized the material. In a metal-organic framework, or MOF for short, metal ions are linked by organic molecules to form a crystal with relatively large pores. Such substances are able to absorb large quantities of gas in relation to their weight. In the compound that the research team proposes for use as a deuterium and tritium filter, zinc and copper ions form the metallic nodes. As early as 2012 the scientists presented a metal-organic framework compound containing only zinc as the metallic component. It was able to filter out deuterium – but only at a temperature of minus 223 degrees Celsius. With copper instead of zinc, the filter can be cooled with liquid nitrogen The Augsburg-based chemists therefore replaced some of the zinc atoms with copper atoms, whose electron shells more selectively filter out deuterium and does so at higher temperatures. Michael Hirscher and his staff at the Max Planck Institute for Intelligent Systems and researchers at the Oak Ridge National Laboratory confirmed this property in various tests. Among other things, they determined the quantities of deuterium and normal hydrogen that the material absorbs from a mixture of equal parts of the two isotopes at various temperatures. They found that at minus 173 degrees Celsius it stores twelve times more deuterium. "At this temperature the separating process can be cooled with liquid nitrogen, which makes it more cost-effective than methods that only work at minus 200 degrees," says Michael Hirscher. The team of theoretical chemists headed by Thomas Heine, who has recently assumed a chair at the University of Leipzig after previous teaching at Jacobs University in Bremen, helped interpret the collected data. "Our calculations fitted the various parts of the experimental puzzle together into a coherent picture," the scientist says. The metal-organic framework has to absorb even more gas The data for deuterium and normal hydrogen showed that the predictions of the calculations agreed very well with the experimental results. The theoreticians are therefore confident that those calculations, which cannot easily be tested experimentally, are just as valid. "Our calculations for tritium would probably be right too. But this can only be experimentally confirmed under stringent safety procedures," Thomas Heine says. The material also absorbs the radioactive hydrogen isotope very efficiently from a mixture of isotopes. That could be a useful property in a particular application in which the aim is not to obtain the isotope but to get rid of it. Water from power plants – including the water that flooded the Fukushima reactors in the 2011 disaster – contains tritium. The new metal-organic framework compound may provide a way to dispose of this radioactive waste, although the radioactively contaminated water first has to undergo electrolysis to convert the tritium-containing water molecules into tritium-containing hydrogen gas. However, before tritium and deuterium can be filtered out of the isotope mixture using large-pore crystals in practice, the technique first has to be refined – not least, so that it absorbs more gas. Neutrons are ideal to study the adsorption of molecular hydrogen Neutron scattering is a very sensitive tool to study the motion of hydrogen, the neutron also distinguishes the signals coming from different isotopes like hydrogen and deuterium. "In the metal organic framework, hydrogen molecules adsorbe on different sites, by tracking the relative populations of hydrogen and deuterium in each site, neutrons clearly elucidated the mechanism of isotopic separation." Timmy Ramirez-Cuesta, from the Spallation neutron scource at Oak Ridge National Laboratory says. The research made use of ORNL's VISION spectrometer, the world most powerful chemical neutron spectrometer. More information: Capture of heavy hydrogen isotopes in a metal-organic framework with active Cu(I) sites. Nature Communications, 28 February 2017; DOI: 10.1038/ncomms14496
News Article | December 15, 2016
« Global Bioenergies reports first production of green isobutene at demo plant | Main | BMW Group and IBM collaborate on research on future driver assistance systems; IBM Watson IoT » Daimler AG is reinforcing its involvement in research into artificial intelligence by participating in the Cyber Valley research initiative in Germany as a partner. The research alliance comprises the Max Planck Institute for Intelligent Systems with its two facilities in Tübingen and Stuttgart, the University of Tübingen, the University of Stuttgart and six partners from industry. Partly supported through the State of Baden-Württemberg, the Cyber Valley partners will establish new research groups and professorships in the fields of machine learning, robotics, and computer vision in a new research center in the Stuttgart-Tübingen area in Germany. A key element of the project will be the training of up to 100 doctoral students. One of the aims of the scientific and business initiative is to create the same kind of constructive conditions for successful start-ups in the field of automated learning as Stanford University did in Silicon Valley in the field of digital technology. That is why Daimler is funding an endowed professorship for “Entrepreneurship in the Digital Transformation”. With Cyber Valley, we are reinforcing the knowledge landscape in Baden-Württemberg and promoting a key future topic. Artificial intelligence has the potential to take digitization in the automotive industry to a new level, having left the realms of science fiction a long time ago. Progress with respect to autonomous driving and the many potential applications for development, production, sales or even entirely new mobility services prove this impressively. This will create one of Europe’s biggest research partnerships in the field of artificial intelligence, in which the state of Baden-Württemberg will invest more than €50 million in the years ahead.
News Article | March 1, 2017
Glass artisans in medieval times exploited the effect long before it was even known. They coloured the magnificent windows of gothic cathedrals with nanoparticles of gold, which glowed red in the light. It was not until the middle of the 20th century that the underlying physical phenomenon was given a name: plasmons. These collective oscillations of free electrons are stimulated by the absorption of incident electromagnetic radiation. The smaller the metallic particles, the shorter the wavelength of the absorbed radiation. In some cases, the resonance frequency, i.e., the absorption maximum, falls within the visible light spectrum. The unabsorbed part of the spectrum is then scattered or reflected, creating an impression of colour. The metallic particles, which usually appear silvery, copper-coloured or golden, then take on entirely new colours. Researchers are also taking advantage of the effect to develop plasmonic printing, in which tailor-made square metal particles are arranged in specific patterns on a substrate. The edge length of the particles is in the order of less than 100 nanometres (100 billionths of a metre). This allows a resolution of 100,000 dots per inch – several times greater than what today's printers and displays can achieve. For metallic particles measuring several 100 nanometres across, the resonance frequency of the plasmons lies within the visible light spectrum. When white light falls on such particles, they appear in a specific colour, for example red or blue. The colour of the metal in question is determined by the size of the particles and their distance from each other. These adjustment parameters therefore serve the same purpose in plasmonic printing as the palette of colours in painting. The trick with the chemical reaction The Smart Nanoplasmonics Research Group at the Max Planck Institute for Intelligent Systems in Stuttgart also makes use of this colour variability. They are currently working on making dynamic plasmonic printing. They have now presented an approach that allows them to alter the colours of the pixels predictably – even after an image has been printed. "The trick is to use magnesium. It can undergo a reversible chemical reaction in which the metallic character of the element is lost," explains Laura Na Liu, who leads the Stuttgart research group. "Magnesium can absorb up to 7.6% of hydrogen by weight to form magnesium hydride, or MgH2", Liu continues. The researchers coat the magnesium with palladium, which acts as a catalyst in the reaction. During the continuous transition of metallic magnesium into non-metallic MgH2, the colour of some of the pixels changes several times. The colour change and the speed of the rate at which it proceeds follow a clear pattern. This is determined both by the size of and the distance between the individual magnesium particles as well as by the amount of hydrogen present. In the case of total hydrogen saturation, the colour disappears completely, and the pixels reflect all the white light that falls on them. This is because the magnesium is no longer present in metallic form but only as MgH2. Hence, there are also no free metal electrons that can be made to oscillate. The scientists demonstrated the effect of such dynamic colour behaviour on a plasmonic print of Minerva, the Roman goddess of wisdom, which also bore the logo of the Max Planck Society. They chose the size of their magnesium particles so that Minerva's hair first appeared reddish, the head covering yellow, the feather crest red and the laurel wreath and outline of her face blue. They then washed the micro-print with hydrogen. A time-lapse film shows how the individual colours change. Yellow turns red, red turns blue, and blue turns white. After a few minutes all the colours disappear, revealing a white surface instead of Minerva. The scientists also showed that this process is reversible by replacing the hydrogen stream with a stream of oxygen. The oxygen reacts with the hydrogen in the magnesium hydride to form water, so that the magnesium particles become metallic again. The pixels then change back in reverse order, and in the end Minerva appears in her original colours. In a similar manner the researchers first made the micro image of a famous Van Gogh painting disappear and then reappear. They also produced complex animations that give the impression of fireworks. The principle of a new encryption technique Laura Na Liu can imagine using this principle in a new encryption technology. To demonstrate this, the group formed various letters with magnesium pixels. The addition of hydrogen then caused some letters to disappear over time, like the image of Minerva. "As for the rest of the letters, a thin oxide layer formed on the magnesium particles after exposing the sample in air for a short time before palladium deposition," Liu explains. This layer is impermeable to hydrogen. The magnesium lying under the oxide layer therefore remains metallic − and visible − because light is able to excite the plasmons in the magnesium. In this way it is possible to conceal a message, for example by mixing real and nonsensical information. Only the intended recipient is able to make the nonsensical information disappear and filter out the real message. For example, after decoding the message "Hartford" with hydrogen, only the words "art or" would remain visible. To make it more difficult to crack such encrypted messages, the group is currently working on a process that would require a precisely adjusted hydrogen concentration for deciphering. Liu believes that the technology could also be used some day in the fight against counterfeiting. "For example, plasmonic security features could be printed on banknotes or pharmaceutical packs, which could later be checked or read only under specific conditions unknown to counterfeiters." It doesn't necessarily have to be hydrogen Laura Na Liu knows that the use of hydrogen makes some applications difficult and impractical for everyday use such as in mobile displays. "We see our work as a starting shot for a new principle: the use of chemical reactions for dynamic printing," the Stuttgart physicist says. It is certainly conceivable that the research will soon lead to the discovery of chemical reactions for colour changes other than the phase transition between magnesium and magnesium dihydride, for example, reactions that require no gaseous reactants. Explore further: Rotate an image, another one appears (w/ Video)