Ten ways in which science may change the world - from building new materials atom by atom to robots resembling worms and octopuses
- Electronic displays / Charles K. Choi
- Computing in the cloud / Charles K. Choi
- Microbiome / Catherine Harmon Courage
- Metamaterials / Lee Billings
- Soft robots / Larry Greenmeyer
- Super light materials / Marisa Fessenden
- Carbon storage / Dave Levitan
_________________________________________________________________________________________________________________________________________________________________
To dream in matter / Garbrand Seder and Kristin Persson
Scientists build new materials atom by atom, using supercomputers and quantum equations, before conducting even a single experiment
In 1878, Thomas Edison decided to reinvent electric lighting. To develop small light bulbs that are suitable for home use, he had to find a light-emitting factor that would be long-lasting, emit little heat and consume little electricity. Edison, guided mainly by intuition, began to examine thousands of carbon materials: boxwood, coconut shell, hairs from the beard of his laboratory assistant. 14 months later he registered a patent for a light bulb with a filament made of carbonized cotton wick. The press heralded his invention with the words "a tremendous achievement for a great inventor of electric lighting." But there were better materials for filaments. At the beginning of the 20th century, another American inventor created a filament from the metal tungsten, which we still use today. Edison's cotton thread was history.
Materials science, which deals with "engineering" materials into new and useful forms, has come a long way since the days of Edison. Quantum mechanics gave scientists a deeper understanding of the behavior of matter, and as a result also a better ability to base research on scientific theory instead of hypotheses. Nevertheless, material development is still a lengthy and very expensive process. Companies invest billions in designing new materials, but success stories are few and far between. The researchers think of new ideas based on intuition and experience; But translating these ideas into practical language and testing them involves a long and tedious process of trial and error. Testing a single new substance can take months, and the result is often negative. Thomas Iger, our colleague at the Massachusetts Institute of Technology (MIT), found that it takes an average of 15 to 20 years to move from the stage of laboratory tests to the stage of commercial application, even if it is a successful material. For example, when Sony announced the first commercial lithium-ion battery in 1991, it seemed like an amazing scientific leap forward, but in reality, hundreds or even thousands of battery researchers toiled over two decades of tentative progress and failures until their work was hailed as a success.
However, materials science is on the verge of a revolution. We can harness a hundred years of development in physics and computer science to advance beyond the Addisonian process. The exponential growth in the power of computer processing, combined with the scientific work of Walter Cohn and the late John Poppel, who developed simple but accurate solutions to the quantum equations in the 60s and 70s of the 20th century, allow the design of new materials from scratch using supercomputers And based on the basic laws of physics. This technique is called "designing materials using high-throughput computational methods", and the underlying idea is simple: using supercomputers to examine, virtually, hundreds or even thousands of chemical compounds at the same time, and quickly and efficiently locate the best building blocks to create a new material, whether it's an electrode of a battery or in a metallic alloy and whether it is a new type of semiconductor.
Most materials are made of a large number of chemical compounds. Battery electrodes, which are made of several compounds, are a good example of this. But there are also much simpler materials. Graphene, which is often talked about as the material of the future of electronics, is actually a sheet of carbon that is one atom thick. Regardless of the complexity of the material, one thing is always true: its properties, compressibility, hardness, luster, electrical conductivity, are determined by the quantum properties of the atoms from which it is made. Therefore, the first step in designing materials using high-throughput computational methods is to "grow" new materials virtually through thousands of calculations based on quantum mechanics. A supercomputer organizes virtual atoms into hundreds or thousands of crystal structures. Next, we calculate the properties of these virtual compounds. What do the crystalline structures look like? How stiff are they? Do they absorb light? What happens when you apply force on them? Do they conduct electricity and heat or insulate? We instruct the computer to find compounds with unique desired properties, and in a short time promising compounds are discovered. At the end of the process, the data collected during the research is transferred to a database, from which other researchers can mine information in the future.
Since 2011, we have been leading a collaboration between researchers, called the "Materials Project", whose goal is to accelerate the computerized materials revolution. The project is designed to create freely accessible databases that will include the basic thermodynamic and electronic properties of all known inorganic compounds. To date, we have calculated the basic properties (organization of the crystal structure, electrical conduction, light transmission, etc.) of almost all 35,000 known inorganic substances that exist in nature. In the same way we also reviewed the properties of several thousand other substances that exist only in theory. So far, about 5,000 scientists have signed up for access to the database where this information will be stored, and they are using it to design new materials for solar cells, batteries and other technologies.
Our research team is not alone in adopting this approach. A group of researchers led by Stefano Cortrollo of Duke University calculated the properties of tens of thousands of alloys; The research he conducted can help in the production of car bodies, beams for building skyscrapers, exterior parts for airplanes and more, all of which are lighter and stronger than those currently in use. As part of the project for quantum informatics of materials, conducted by researchers from the American National Laboratory Argonne, from Stanford University and from the Technical University of Denmark, high-throughput computational methods are applied to investigate catalytic processes on metallic surfaces, and this has a valuable use, especially in research in the field of energy.
Already in the very near future, materials scientists will be able to use high-throughput computational methods to design almost anything. We believe that this will lead to the development of technologies that will reshape our world: breakthroughs that will change the face of computing, eliminate environmental pollution, create abundant clean energy and improve our lives in ways that are hard to imagine today.
The genome of matter
The modern world is based on the success of materials science. The creation of transparent conductive glass led to the development of the touch screens of our smartphones. Those devices are able to transmit information around the world at the speed of light thanks to materials scientists, who discovered a way to produce glass without contaminating ions, on which optical fiber communication is based. What's more, these phones are able to work a whole day without recharging thanks to the lithium storage oxides, developed by materials engineers in the 70s and 80s, which serve as the basis for the lithium-ion battery.
Credit: Mark Haydock
It was the research we conducted on the subject of batteries that led us to design materials using high-throughput computational methods. We devoted the best years of our careers to this, but until we discussed the issue, in 2005, with senior executives at Proctor and Gamble, we had no idea what it would be possible to produce with the help of the world's most powerful supercomputers in a sufficiently long computer time. The company's people wanted to find a better material for the cathode of the alkaline batteries produced by Duracell, the company's battery division. They asked us a surprising question: "Would it be possible to scan the All known compounds To find better material?” When we thought about it we realized that the only real obstacles were computer time and money, and they were happy to provide us with both. They funded the project with a million dollars and gave our small team free access to the company's supercomputer center.
During this venture, which we called the "Alkalinity Project", we reviewed 130,000 compounds, both those that exist in nature and hypothetical ones, and produced for the company a list of 200 compounds that met the necessary criteria and were probably much better than the compounds that had been used in batteries until then. At that time, we were already convinced that material design using high-throughput computational methods was the future of the field.
We recruited staff members and additional budgets, and in 2011 we launched the collaboration between the Massachusetts Institute of Technology and the American Lawrence Berkeley National Laboratory, in a project we initially called the "Materials Genome Project". Since then, teams from the University of California at Berkeley, from Duke University, from the University of Wisconsin-Madison, from the University of Kentucky, from the Catholic University of Leuven in Belgium and other institutions have joined the project, and all contribute the information they produce to our central, free and open database at Lawrence Berkeley.
It wasn't long before we dropped the word "genome" from the name of the project, to distinguish it from another project initiated by the Office of Science and Technology Policy of the White House. And to be honest, the properties of chemical compounds are not really "genes", they are not units of information that are inherited as a unique sequence of data. Nevertheless, there is a direct relationship between the function or property of a material and its basic properties. Just as it is possible to associate blue eyes with a certain gene, so it is possible to find, for example, the source of the electrical conductivity of a certain material in the properties and the way of organization of the elements of which it is composed.
Such relationships are the basis of materials science. Here is a simple example: we know that the color of minerals can be "tuned" by creating intentional defects in their crystal structure. Take, for example, the ruby stone. The source of its red hue is the accidental exchange of aluminum ions (Al+3) found in the common mineral corundum (Al2O3) in chromium ions (Cr3+) at a rate of 1%. When the chromium ions are introduced into this environment, the electronic arrangement in them changes, and this changes the way the mineral absorbs and emits light. Once we know the microscopic origin of a property (in this case, the ruby's red color), we can change it using synthetic methods. Computerized tuning of these chemical defects allows us to design new synthetic rubies with customized colors at will.
The quantum equations provide us with the necessary information to make such adjustments: which elements to use and how to organize them. However, the equations are so complicated that only a computer can process and solve them. Suppose we want to test several hundred compounds to find out which of them have certain desired properties. It takes tremendous computing power to process these equations. Until recently, this was impossible, so materials science has so far progressed by trial and error. Now that computing power is in our hands, we can finally take advantage of the full predictive capabilities of quantum mechanics.
Suppose we are studying thermoelectric materials, which produce an electric current when subjected to large temperature changes (the opposite is also true: a thermoelectric material is able to withstand temperature changes when an electric current is passed through it; for example: when it is exposed to accelerated cooling). Modern society wastes a huge amount of heat in combustion processes, industrial processing and cooling. If we had efficient, cheap and stable thermoelectric materials we could capture this heat and reuse it as electricity. Thermoelectric devices will be able to convert excess heat emitted in industrial processes into electricity. Heat from car exhaust pipes could activate the dashboard electronics. Such thermoelectric materials could also perform solid-state cooling on demand: small devices we could weave into our clothes could cool us at the push of a button, without the need for fans or compressors.
One of the best thermoelectric materials we know today is lead tellurium (PbTe), a toxic material too expensive for commercial use. Imagine you are a researcher looking for a better thermoelectric material. If you couldn't use high-throughput computational methods, you would have to start looking for known compounds that, like tellurian lead, have a high Seebeck coefficient (a measure of the amount of electricity obtained from the temperature difference to which the material is exposed), but which, unlike tellurian lead, are not made of rare, toxic or expensive elements. You would have to delve into data tables and compare numbers. And with a little luck, you would discover some materials that could, in theory, suit your purpose. Then, you had to create these compounds in the lab. But the synthesis of materials is an expensive, lengthy and difficult matter. Usually, when you start work, you can't even tell if the new material will be stable. Even if the new substance turns out to be stable, the compound must be synthesized and the process repeated until a sufficiently pure sample is obtained, before its properties can be measured. The work on any such compound can take months.
To date, the researchers have not been able to discover alternative thermoelectric materials, but they have not yet attempted to design materials using high-throughput computational methods. The situation is about to change soon. Starting this year, 2014, we will collaborate with researchers at the California Institute of Technology and five other institutions in the search for new thermoelectric materials. We intend to keep at it until we find the compounds that will allow us to realize these miracle technologies for energy efficient cooling.
The golden age of materials planning
Our ability to automatically access, search, and review material data is still in its infancy. What will happen in the field when it develops? Let's make some assumptions here.
There are quite a few clean energy technologies that are just waiting for the development of applicable advanced materials. Photocatalytic compounds [that catalyze chemical reactions under the influence of light], such as titanium dioxide, could be used to convert sunlight and water into oxygen and hydrogen, and these would be processable into liquid fuels. Other photocatalytic materials can be used for the same purpose in combination with carbon dioxide. The ambition is to create an "artificial leaf" that can convert sunlight and air into liquid fuel similar to methanol, which will be used to power cars and planes [see: Reinventing the leaf, Scientific American Israel, February 2011]. Researchers at the Joint Center for Artificial Photosynthesis, which operates under the auspices of the US Department of Energy, are using high-throughput computational methods to find materials that will make this technology practical.
And what about the search for new metallic alloys, from which these cars and planes will be built? Reducing vehicle weight by 10% can improve fuel economy by 6% to 8%. The US industry invests billions of dollars every year in the research and development of materials and alloys. Designing materials with the help of a computer could multiply the benefit of this investment many times over. Real progress in the field of strong, light and recyclable alloys will have a huge impact on the global economy, thanks to improved energy efficiency in transport and construction.
Another field that needs breakthrough materials is computing. Recently, many serious predictions have been published that we are close to the exhaustion point of Moore's Law, the contract that computing power will double every two years or so. We have known for a long time that silicon is not the best semiconductor, it is simply common and familiar. What could work better? The key lies in finding materials that can quickly switch between conducting and insulating states. A team at the University of California, Los Angeles, created incredibly fast transistors from graphene, while a group at Stanford reported that they were able to operate a magnetite as an electronic switch that transmits an electric current or stops the passage of current in a trillionth of a second - several thousand times faster than the speed of transistors in use today. Designing materials using high-throughput computational methods will allow us to locate the options that suit us best.
The list is still very long. Researchers use computational materials design to develop new superconductors, catalysts, and light-emitting materials. These three could change the face of information technology, the capture of carbon dioxide from the air and the discovery of nuclear materials.
Computer-based material design could also enable breakthroughs beyond our imagination. Perhaps we will succeed in inventing a new liquid fuel based on silicon instead of carbon, which will provide more energy than gasoline and whose by-products will be harmless to the environment, such as sand and water. The idea was raised decades ago, but no one has yet been able to find a workable formula for its implementation. By designing materials with high-throughput computational methods we can, at the very least, know if this is possible, or if we should focus our efforts in other directions.
In light of these developments in the field, we believe that this is the beginning of the golden age of material planning. The tremendous computing power available to us today gives us a greater ability than ever to turn raw material into useful technologies. In view of the challenges facing us, in a world that is getting hotter, that is becoming more and more crowded, the sooner this golden age arrives, the better.
__________________________________________________________________________________________________________________________________________________________________
About the authors
Gerbrand Ceder is a professor of materials science and engineering at the Massachusetts Institute of Technology (MIT). Sider initiated together with Kristin Persson the "Materials Project", which provides the research community with data produced by computational methods regarding the properties of materials.
Kristin Persson is a scientist on the staff of the American Lawrence Berkeley National Laboratory. She received her doctorate in theoretical physics at the Royal Institute of Technology in Stockholm.
in brief
Engineered materials, such as high-purity silicon for the production of chips, and glass for the production of optical fibers, stand at the basis of technological development in the modern world. However, until now, designing new materials has largely involved a frustrating and inefficient process of trial and error.
Simplified versions of quantum equations, combined with supercomputers capable of testing thousands of materials virtually, make most of the deception redundant.
Researchers are currently applying this method, called "high-throughput computational materials design", to develop new batteries, solar cells, fuel cells, computer chips and other technologies.
The basic principles
Why do we need supercomputers?
Broadly speaking, the Addisonian method of planning materials involves testing everything imaginable, and discarding what doesn't work. By designing materials using high-throughput computational methods, however, researchers can quickly locate candidate materials virtually, in supercomputer farms, thus saving time and money and avoiding frustration. After the computer tests hundreds or thousands of materials and produces a list of the top ten candidates, the researchers begin to prepare actual samples of the materials in the laboratory and test them using accepted laboratory methods.
More on the subject
Opportunities and Challenges for First-Principles Materials Design and Applicationsto Li Battery Materials. Gerbrand Cedar in MRS Newsletter, Vol. 35, no. 1, pages 693-701; September 2010.
The Materials Project: A Materials Genome Approach to Accelerating MaterialsInnovation. Anubhav Jain et al. in APL Materials, Vol. 1, no. 1; July 2013.
The article was published with the permission of Scientific American Israel
2 תגובות
A very fascinating article. Thanks.
not trash, trash*