Scientists uncover hidden superconductivity in material once thought only magnetic

A new study shows that iron telluride, or FeTe, long thought to be a simple magnetic metal, is actually a superconductor when excess iron atoms are removed. Superconductors carry electricity with no energy loss as heat, enabling technologies such as MRI machines, particle accelerators and potentially quantum computers. Researchers found that extra iron atoms hidden in the material are responsible for its magnetism and suppress superconductivity. Once those atoms are eliminated, electricity flows with zero resistance, and the material’s superconducting properties can be further tuned using layered structures and moiré effects.

The findings are detailed in two papers published back-to-back, both led by Penn State physicist Cui-Zu Chang. The first paper explains how to activate superconductivity in FeTe. The second describes a new type of “quantum dance” in which superconductivity interacts with the material’s atomic structure when a different top layer is added, allowing scientists to adjust its properties.

Mystery Behind FeTe’s Missing Superconductivity

“Unlike the well-known iron-based superconductor iron selenide (FeSe), FeTe has long been considered a magnetic metal without superconductivity, despite having an almost identical crystal structure,” Chang said. “It has remained a mystery why FeTe doesn’t share this important property.”

To investigate this difference, the team created thin films of FeTe using molecular beam epitaxy. This method produces extremely clean, atomically thin materials by slowly depositing source elements onto a suitable surface.

When the researchers examined the samples at the atomic level using scanning tunneling microscopy, they found that the structure was not perfectly uniform. Extra iron atoms were embedded within the crystal lattice of FeTe.

Excess Iron Atoms Disrupt Superconductivity

“These excess iron atoms disrupt the ideal one-to-one ratio of iron and tellurium atoms in FeTe and upset the balance of magnetism and superconductivity,” Chang said, explaining that the researchers theorized that removing the excess atoms to make truly pure FeTe might result in a superconductor.

To test this idea, the researchers developed a way to control the material’s purity by exposing the FeTe films to tellurium vapor. This process offsets the excess iron and pushes the material toward its ideal composition.

“The resulting ideal FeTe exhibits superconductivity with a critical temperature of around 13.5 Kelvin, or about negative 435 degrees Fahrenheit,” Chang said. “The excess iron atoms had disguised its superconductivity, leading to the decades-old view that FeTe was an ordinary magnetic metal. Our findings redefine the phase diagram of this class of iron-containing compounds. Similar phenomena are likely to be present in other correlated materials, where hidden superconducting states or competing magnetic orders remain concealed until disorder is removed or carefully controlled. Understanding the crucial role of disorder will help us to uncover and stabilize such hidden superconducting states in other materials.”

Engineering Superconductivity with Layered Structures

In the second study, after confirming that FeTe is inherently a superconductor, the researchers investigated how its superconducting behavior could be controlled. They built layered structures by placing a thin material with a different crystal lattice on top of FeTe. Because the two materials have different atomic arrangements, they form a larger repeating pattern at their boundary, known as a moiré superlattice.

“The mismatch between the crystal structures at the interface creates what we call a moiré superlattice, which modifies the superconducting properties of FeTe,” Chang said. “In recent years, moiré superlattices in two‑dimensional materials have emerged as an important platform for discovering new quantum states.”

Using scanning tunneling microscopy, which allows imaging at the atomic scale, the team observed that superconductivity appears as a repeating, droplet-like pattern that follows the moiré superlattice, described by the researchers as a “quantum dance.” They also found that this pattern can be tuned by changing the material used in the top layer.

“The role of crystal lattices has often been overlooked in superconductors,” Chang said. “Our findings encourage a renewed focus on the interplay between superconductivity and lattice structure and highlight how moiré interface engineering can serve as a potentially powerful tool for tuning superconductivity and designing next‑generation quantum materials.”

For more information: Nature

Image: A sample of a thin film of the compound iron telluride (FeTe)—the dark region on the clear substrate at the center of the image—created using molecular beam epitaxy. Long thought to be an ordinary magnetic metal, researchers have now shown that exposing the thin film of FeTe to tellurium vapor removes disorder created by excess iron atoms trapped in the crystal structure of the material, revealing that FeTe is a superconductor. Credit: Chang Laboratory/Penn State.

St. Olaf researchers built a computer that doesn’t require electricity

Researchers from St. Olaf College and Syracuse University have built mechanical computers made entirely of common materials that can perform simple calculations without electricity or batteries. Led by St. Olaf College physics professor Joey Paulsen, the team used steel springs and bars to create devices that store and process information by physically responding to motion and force. One machine counts how many times it is moved, another determines whether it has been pushed an odd or even number of times, and a third remembers whether it experienced a medium or large force, demonstrating that everyday materials can both retain memory and perform basic computation.

“We now have a rational way of building these machines that can perform simple computations without a computer chip or a power source,” Paulsen said.

Key findings from the research include:

  • Mechanical computers can perform simple computations without a computer chip or power source.
  • Mechanical computers are able to harvest their power from physical force, rather than electricity.
  • Proof of design that mechanical computers could be a viable alternative to conventional computers in harsh settings—such as extreme temperatures or exposure to corrosive chemicals—when only simple computations are needed.

“Our results are one step towards designing materials that can sense their environment, make a decision, and then respond,” said Paulsen. “Frequently called smart materials, what we learned could help improve people’s lives by having more responsive artificial limbs or tactile rooms.”

Paulsen recommends that future research on mechanical computers focus on understanding their limitations and scalability. Under his leadership, St. Olaf students are currently testing how the state of one rotor affects its interaction with a second rotor –– and potentially a third. This research will continue in the coming months, with opportunities for students to participate through the college’s Collaborative Undergraduate Research and Inquiry (CURI) program.

For more information:  Nature Communications

Image: St. Olaf College students Faten Abu Al Ardat ‘27 and Harry Maakestad ‘26 work on building the mechanical computer.

Graphene just defied a fundamental law of physics

Scientists have observed electrons in graphene flowing like a nearly frictionless liquid, revealing an unusual quantum state that challenges long-standing assumptions in physics and could enable future technologies. Physicists have long sought evidence that electrons can behave collectively as a smooth fluid governed by universal quantum properties, but such behavior is difficult to detect because atomic defects and impurities in real materials typically disrupt these effects. The new observation overcomes those challenges, offering rare experimental insight into electron behavior under exceptionally clean and controlled conditions.

Now, researchers at the Department of Physics, Indian Institute of Science (IISc), working with collaborators from the National Institute for Materials Science in Japan, have finally identified this elusive quantum fluid in graphene. This material consists of a single layer of carbon atoms arranged in a flat sheet. Their findings, reported in Nature Physics, open a new path for studying quantum phenomena and position graphene as a powerful platform for exploring effects that were previously out of reach in laboratory settings.

“It is amazing that there is so much to do on just a single layer of graphene even after 20 years of discovery,” says Arindam Ghosh, Professor at the Department of Physics, IISc, and one of the corresponding authors of the study.

Breaking a Fundamental Law of Physics

To uncover this behavior, the team created exceptionally clean graphene samples and carefully measured how they conduct both electricity and heat. What they found was unexpected. Instead of increasing together, the two properties moved in opposite directions. As electrical conductivity rose, thermal conductivity dropped, and vice versa.

This result directly contradicts the Wiedemann-Franz law, a well-established principle that states heat and electrical conduction in metals should be proportional. The researchers observed deviations from this law by more than 200 times at low temperatures, revealing a striking separation between how charge and heat move through the material.

A Universal Quantum Connection

Despite this unusual split, the behavior is not random. Both types of conduction appear to follow a universal constant that does not depend on the material itself. This constant is tied to the quantum of conductance, a fundamental quantity that describes how electrons move at the smallest scales.

The Dirac Fluid and Liquid-Like Electrons

This remarkable effect occurs at a special condition known as the “Dirac point,” where graphene sits at a boundary between being a metal and an insulator. By adjusting the number of electrons, researchers can reach this precise state.

At this point, electrons stop behaving like individual particles. Instead, they move collectively, flowing like a liquid. This fluid-like motion resembles water but with far lower resistance to flow. “Since this water-like behaviour is found near the Dirac point, it is called a Dirac fluid — an exotic state of matter which mimics the quark-gluon plasma, a soup of highly energetic subatomic particles observed in particle accelerators at CERN,” says Aniket Majumdar, first author and PhD student at the Department of Physics. The team also measured how easily this fluid flows and found that its viscosity is extremely low, making it one of the closest realizations of a perfect fluid ever observed.

A New Window Into Extreme Physics

These results establish graphene as an accessible and cost-effective system for exploring ideas that are usually associated with extreme environments. Scientists can now investigate phenomena linked to high-energy physics and astrophysics, including black-hole thermodynamics and entanglement entropy scaling, within a laboratory setting.

Future Applications in Quantum Technology

Beyond its scientific importance, this discovery could have practical implications. The presence of a Dirac fluid in graphene may enable the development of highly sensitive quantum sensors. Such devices could amplify extremely weak electrical signals and detect faint magnetic fields, opening the door to new technologies in sensing and measurement.

For more information: Nature Physics

Researchers use large language models to discover recipes for novel materials

Researchers at the University of Rochester have developed an artificial intelligence-based method that uses large language models, similar to ChatGPT, to help chemical engineers discover and manufacture new materials, potentially accelerating efforts such as converting carbon dioxide into fuel. The approach allows researchers to describe desired materials in natural language, receive AI-generated recommendations for experimental procedures, and iteratively refine those experiments by feeding results back into the model. By lowering technical barriers to using AI in catalysis research, the method aims to speed experimentation, improve accessibility and advance materials discovery.

“We’re able to leverage the pre-trained knowledge of large language models and well-established statistical methods for materials discovery to help us as researchers navigate large experimental design spaces more efficiently,” says Marc Porosoff, an associate professor in the Department of Chemical and Sustainability Engineering.

Porosoff likens the new AI method to describing a cup of coffee, noting that someone could describe the coffee by its taste, color, and aroma, or by the type of beans, grind size, apparatus, and water temperature used to make the brew. Both representation methods describe the same cup of coffee, but the second approach gives you a recipe to reproduce it that others can easily replicate.

Porosoff and his team are applying the same principle to catalysts for energy applications, using language-based representations to describe materials not just by their properties, but by the steps needed to create them.

To build on their success, the US Department of Energy Advanced Research Projects Agency-Energy (ARPA-E) announced it will provide nearly $3 million in funding to apply the URochester team’s method toward creating catalysts for the production of fuel from abundant materials, specifically methanol and ethanol from carbon dioxide and hydrogen. Porosoff will lead a multi-institution project team that includes URochester, Virginia Polytechnic Institute and State University, Stanford University, Northwestern University, A*STAR Institute of Sustainability for Chemicals, Energy and Environment (ISCE2) in Singapore, and OxEon Energy, a small business based in Salt Lake City.

Leveraging the power of LLMs

Traditional AI methods for materials discovery typically use a strategy called Bayesian optimization to identify and design the best candidates. But the result is complex numerical data about a material’s structure, which requires deep expertise to use effectively. The new LLM method instead produces a set of procedures that researchers can easily understand, execute, and verify to determine if the experiment’s output matches the predicted results.

This can be extremely useful for working with complex materials such as trimetallic catalysts, which are made of three metals.

“Our method reduces the technical barrier associated with using Bayesian optimization, which is a well-established method for efficiently exploring large and complicated parameter spaces,” says Shane Michtavy, a URochester chemical engineering PhD student who helped develop the AI method, synthesize materials, and run the chemical reactions described in the paper. “Using pre-trained LLMs allows users to explore using less data than traditional models, as they are deployed in a frozen state with built-in knowledge of the physical world and catalysis.”

The paper shows how the researchers applied the method to several live experiments, including one to identify catalysts for turning carbon dioxide and hydrogen into carbon monoxide and water using trimetallic catalysts made from low-cost metals. Porosoff says that there are about 360,000 possible experiments that could have been run to find the ideal catalyst, but by using procedures produced by the AI model and providing it with the results from the experiments, they were able to find an ideal candidate in just ten experiments.

The study was supported by funding from the National Science Foundation, the National Institutes of Health, and the US Department of Energy. Additional authors included Mayk Caldas, technical staff at Edison Scientific.

Next steps

Now that they have shown the model works as a proof of concept in the lab, Porosoff aims to take the method further using the funding announced through ARPA-E’s Catalytic Application Testing for Accelerated Learning Chemistries via High-throughput Experimentation and Modeling Efficiently (CATALCHEM-E) program.

“Right now, it takes a decade or longer to go from conceptualizing a new catalyst to testing it in a lab to putting it in a real reactor,” says Porosoff. “The CATALCHEM-E program aims to cut that by an order of magnitude to a single year, and we think using AI with text-based representations will be a big factor in shortening the development cycle.”

Porosoff and his collaborators will first demonstrate their workflow on carbon dioxide-to-methanol and then extend the process to higher alcohols such as ethanol, which is a key additive for gasoline and used in pharmaceuticals, cosmetics, and many other applications. Ultimately, they hope to commercially deploy the model for industries to create catalysts to synthesize alcohols for fuel.

For more information: ACS Central Science

Wisconsin Oven ships draw batch oven to the military

Wisconsin Oven Corp., East Troy, Wisconsin, announced the shipment of an Electrically Heated Standard Draw Batch Oven (SDB Series) to a United States Military Base. The industrial oven will be used for heat treating aerospace components and features combination-style airflow delivering both horizontal and vertical upward flow for optimal heating rates and consistent temperature distribution. Temperature uniformity of ±5°F was verified at set points of 200°F, 700°F, and 1200°F in accordance with a Class 1 Temperature Uniformity Survey per AMS 2750H. The oven incorporates CAN-style construction with a heavy plate exterior and 6 inches of high-temperature industrial insulation, a 15,000 CFM recirculation system with variable frequency drive, and a UL508A-certified control panel.

Read further here

Nanoscale hotspots in OLEDs may shorten their lifespans in phones, TVs

University of Michigan engineers have found that the light in OLED displays comes from nanoscale hotspots—some of which flicker—rather than from a perfectly uniform surface, a behavior that could shorten device lifespans by causing certain areas to carry more current and burn out faster. The studysuggests that these uneven emission patterns may also affect the performance of organic electronics such as solar cells and transistors. As a potential fix, researchers propose using crystalline instead of amorphous structures to improve durability. The work was supported by the U.S. Department of Energy and Universal Display Corp.

 

Charge rivers in hilly OLED energy landscapes

“The calculations that motivated us to look for this are actually pretty old. In the mid-2000s, people were predicting what they called a current channeling phenomenon,” said Chris Giebink, U-M professor of electrical and computer engineering and also senior author of the study. “You could liken it to a hilly landscape. The electrons, or charge carriers, that move through the device tend to want to follow the lowest energy pathways, so they’ll travel along the valleys.”

Charge carriers come in two flavors, electrons and positively charged “holes,” which run in opposite directions through the landscape. Where those rivers cross, light-emitting molecules convert electron-hole pairs into photons, or particles of light.

Because some valleys are deeper than others, they tend to support high-traffic rivers of charge carriers, with densities that are thought to be 10 to 100 times higher than the rest of the material. In contrast, crystalline materials are more uniform. Their landscape is flatter, leading the charge carriers to spread out more evenly, reducing the hotspot effect.

Spotting nanoscale hotspots with superresolution

Theory suggests that the hotspots are just a few tens of nanometers across. They appear as graininess in images from an optical microscope, which is limited to details of a couple hundred nanometers or bigger. 

“An initial concern was whether we were seeing a microscope artifact,” said Joshua Springsteen, a Ph.D. student in electrical and computer engineering and first author of the study. “We examined the same area of the device with our microscope using both photoluminescence and electroluminescence, confirming that it was an electrical phenomenon.”

Because some of the lights flicker and aren’t always in sync, Springsteen could take a video of the device and run it through software that keyed in on the changes in brightness when one hotspot switched off while another stayed on. This technique, called superresolution optical fluctuation imaging, helped the team confirm that the hotspots were smaller than half the wavelength of the green light they emitted.

The researchers believe the blinking is due to charge carriers that are temporarily trapped in dips in the energy landscape. When that happens, they act more like dams, repelling other charge carriers, which seek alternate routes and cause downstream hotspots to go dark. Eventually, the charge carrier absorbs enough heat to pop back out of the dip, and the original hotspot lights up again. Because the flickering hotspots aren’t in sync, the human eye perceives amorphous OLEDs as glowing steadily.

To confirm that they had the mechanism right, the team used those earlier calculations that showed where charge-carrier rivers flow and meet. Springsteen took these theoretical freeze-frames and processed them to mimic the way the hotspots would have been blurred by the superresolution microscopy technique they used. These modified modeling images resembled the experimental images well enough that the team is confident they were seeing the hotspots.

The device was built in the Lurie Nanofabrication Facility and studied at the Michigan Center for Materials Characterization, both of which are operated and maintained with support from indirect cost allocations in federal grants.

For more information:  Michigan Center for Materials Characterization

Image: Hotspots appear in an optical microscope image examining the surface of a green OLED, processed to quantify brightness. These hotspots may limit the lifetimes of amorphous OLEDs. Image: Joshua Springsteen, Optoelectronic Components and Materials Group, University of Michigan.

Tough, reusable adhesive can glue a variety of materials

Researchers at the Department of Energy’s Oak Ridge National Laboratory have developed a reusable adhesive made from waste polymers that is stronger than commercial glues, works in wet and dry conditions, and bonds materials ranging from wood and glass to metal and plastics. Inspired by mussels’ sticking power, the adhesive uses reversible chemical crosslinkers that let it soften, release and be reused, unlike conventional single-use glues. The innovation could simplify manufacturing and repair by replacing the many specialized adhesives used today and may have significant economic impact in the global adhesives and sealants market, valued at about $87 billion and projected to reach nearly $119 billion by 2032.

“Most adhesives are made for one specific application,” said Anisur Rahman, a research and development staff member at ORNL who led a study with former ORNL postdoctoral researcher Mary Danielson, now a research assistant professor with the University of Tennessee-Oak Ridge Innovation Institute. “Our adhesive can be used for diverse applications, including structural or pressure-sensitive uses, and it performs reliably in both wet and dry environments,” he said. “None of the commercial adhesives can be used this way.” 

Beginning with common polymers from beverage bottles, fabric fibers and packaging films, the research team developed a process that saves materials, energy and money. “We took material destined for the landfill and turned it into something valuable,” Danielson said.

The researchers have applied for a patent for their versatile glue. 

How the reversible bonds work

Traditional structural adhesives rely on permanent crosslinks that make removal difficult. “You apply traditional adhesives once; you cannot reuse them,” Rahman said. 

“You basically have to rip an assembly apart to debond it,” Danielson added. “You’ve damaged both the part you’re glueing to and the part you’re glueing from. If you make a mistake when you’re gluing something and you allow it to cure, it’s done.“

In the ORNL adhesive, crosslinkers act like reversible attachments, akin to Velcro. Heating breaks dynamic chemical bonds in the polymer, allowing the adhesive to release without damaging surfaces. As the material cools, the bonds reform.

“If something is damaged or misapplied, you’re able to completely remove it and put it back on with full integrity,” Danielson said.

The team debonded and rebounded the adhesive more than 10 times with no loss in performance. 

“Normally in the marketplace, structural adhesives typically have shear strength — a measure of adhesion — in the 7- to 10-megapascal range,” Rahman said. “Our adhesives also stay well above that range but maintain reusability.”

The researchers can also retrieve the glue chemically, using an excess of amine molecules to break the adhesive into its monomer subunits. “We can recover all chemicals used in this adhesive,” Rahman said.

Mussel-inspired design

A polymer is a long chain or network made of monomers, or chemical subunits of one type. Using no solvents or catalysts, the scientists added amine, a nitrogen-containing chemical group, to the waste polymer and heated it to just below the polymer’s melting temperature. Under these mild conditions, the amine broke the polymer down into monomers that each contained four amine groups.

Next, to design the adhesive, ORNL researchers mimicked mussel foot proteins, which contain both hydrophilic and hydrophobic components that enable strong adhesion even in wet environments.

“We used a crosslinker, or hardener, that has both water-loving (hydrophilic) and water-hating (hydrophobic) components together in the same molecule,” Rahman said. “We mix the hardener and the monomer. It creates an adhesive resin that acts like a mussel foot protein.”

“For any glue that is a cross-linked network of two components, it takes time to complete the reaction between the two components,” Danielson said. “To repair boats, submarines and pipelines, our glue can be applied underwater using hand pressure until it sets.” 

Curing happens when a large four-armed monomer interacts with the crosslinking hardener. The monomer’s amine group reacts with the hardener’s acetoacetate group to produce a resin, or matrix with hydrophilic and hydrophobic characteristics. Whether the protein sticks or releases depends on the balance of those properties. 

“Our glue maintained strong adhesion across different environmental conditions, including seawater, extremely low temperature (100 degrees Celsius below zero), and both acidic and basic conditions,” Rahman said.

National lab capabilities enabled the achievement

Rahman conceived the concept of transforming deconstructed polymer waste into an adhesive. He and Danielson designed and led experiments and drafted the paper. Chuyi Pan, a summer intern from the University of Pennsylvania, assisted in synthesizing the adhesive. Tomonori Saito of ORNL and the University of Tennessee, Knoxville, reviewed and edited the manuscript drafts.

ORNL researchers performed vital characterizations. Bobby Sumpter simulated the energies with which the adhesive bound to different surface materials. Catalin Gainaru used rheology to characterize its stress and relaxation. Honghai Zhang and Vilmos Kertesz performed mass spectrometry to quantify different molecules. Zoriana Demchuk’s lifecycle analysis of the ORNL glue showed it was more energy-efficient to make than commercial adhesives. 

Toward strong and weak bonding applications

The team has also explored using this pioneering chemistry to advance vehicles. ORNL’s glue maintained strong adhesion between dissimilar substrates — a crucial requirement in automotive and aerospace applications, where joining composites to aluminum or steel presents notable challenges.

Next, the scientists aim to tune crosslinking to enable weaker, temporary bonds for removable labels, adhesive bandages, drug-delivery patches and other applications. 

ORNL’s versatile, high-performance glue is poised to make an impact that sticks in situations from the mundane to the extraordinary. Potential uses range from household items that require gentle removal, like press-on nails and price tags, to repairs in remote or extreme environments, including underwater or outer space — settings where specialty glues may be unavailable. 

The DOE Office of Science supported the research. The work used resources of the Center for Nanophase Materials Sciences, a DOE Office of Science user facility at ORNL.

For more information: Science Advances

Image: From left, Mary Danielson and Anisur Rahman, leaders of an ORNL project to invent a versatile reusable glue from polymer waste, examine its bonding performance. Credit: Carlos Jones/ORNL, U.S. Dept. of Energy

Terahertz microscope reveals the motion of superconducting electrons

MIT physicists have used a new imaging technique to observe terahertz-frequency vibrations—described as quantum “jiggles”—inside a superconducting fluid for the first time, revealing behavior that was previously undetectable. By shining terahertz light, which probes matter differently than optical, infrared or X-ray wavelengths, the team captured inherent quantum motions within the material, opening a new window into the fundamental properties of superconductors.

Terahertz light is a form of energy that lies between microwaves and infrared radiation on the electromagnetic spectrum. It oscillates over a trillion times per second — just the right pace to match how atoms and electrons naturally vibrate inside materials. Ideally, this makes terahertz light the perfect tool to probe these motions.

But while the frequency is right, the wavelength — the distance over which the wave repeats in space — is not. Terahertz waves have wavelengths hundreds of microns long. Because the smallest spot that any kind of light can be focused into is limited by its wavelength, terahertz beams cannot be tightly confined. As a result, a focused terahertz beam is physically too large to interact effectively with microscopic samples, simply washing over these tiny structures without revealing fine detail.

The scientists report that they have developed a new terahertz microscope that compresses terahertz light down to microscopic dimensions. This pinpoint of terahertz light can resolve quantum details in materials that were previously inaccessible.

The team used the new microscope to send terahertz light into a sample of bismuth strontium calcium copper oxide, or BSCCO (pronounced “BIS-co”) — a material that superconducts at relatively high temperatures. With the terahertz scope, the team observed a frictionless “superfluid” of superconducting electrons that were collectively jiggling back and forth at terahertz frequencies within the BSCCO material.

“This new microscope now allows us to see a new mode of superconducting electrons that nobody has ever seen before,” says Nuh Gedik, the Donner Professor of Physics at MIT.

By using terahertz light to probe BSCCO and other superconductors, scientists can gain a better understanding of properties that could lead to long-coveted room-temperature superconductors. The new microscope can also help to identify materials that emit and receive terahertz radiation. Such materials could be the foundation of future wireless, terahertz-based communications, that could potentially transmit more data at faster rates compared to today’s microwave-based communications.

“There’s a huge push to take Wi-Fi or telecommunications to the next level, to terahertz frequencies,” says Alexander von Hoegen, a postdoc in MIT’s Materials Research Laboratory and lead author of the study. “If you have a terahertz microscope, you could study how terahertz light interacts with microscopically small devices that could serve as future antennas or receivers.”

In addition to Gedik and von Hoegen, the study’s MIT co-authors include Tommy Tai, Clifford Allington, Matthew Yeung, Jacob Pettine, Alexander Kossak, Byunghun Lee, and Geoffrey Beach, along with collaborators at Harvard University, the Max Planck Institute for the Structure and Dynamics of Matter, the Max Planck Institute for the Physics of Complex Systems and the Brookhaven National Lab.

Hitting a limit

Terahertz light is a promising yet largely untapped imaging tool. It occupies a unique spectral “sweet spot”: Like microwaves, radio waves, and visible light, terahertz radiation is nonionizing and therefore does not carry enough energy to cause harmful radiation effects, making it safe for use in humans and biological tissues. At the same time, much like X-rays, terahertz waves can penetrate a wide range of materials, including fabric, wood, cardboard, plastic, ceramics, and even thin brick walls.

Owing to these distinctive properties, terahertz light is being actively explored for applications in security screening, medical imaging, and wireless communications. In contrast, far less effort has been devoted to applying terahertz radiation to microscopy and the illumination of microscopic phenomena. The primary reason is a fundamental limitation shared by all forms of light: the diffraction limit, which restricts spatial resolution to roughly the wavelength of the radiation used.

With wavelengths on the order of hundreds of microns, terahertz radiation is far larger than atoms, molecules, and many other microscopic structures. As a result, its ability to directly resolve microscale features is fundamentally constrained.

“Our main motivation is this problem that, you might have a 10-micron sample, but your terahertz light has a 100-micron wavelength, so what you would mostly be measuring is air, or the vacuum around your sample,” von Hoegen explains. “You would be missing all these quantum phases that have characteristic fingerprints in the terahertz regime.”

Zooming in

The team found a way around the terahertz diffraction limit by using spintronic emitters — a recent technology that produces sharp pulses of terahertz light. Spintronic emitters are made from multiple ultrathin metallic layers. When a laser illuminates the multilayered structure, the light triggers a cascade of effects in the electrons within each layer, such that the structure ultimately emits a pulse of energy at terahertz frequencies.

By holding a sample close to the emitter, the team trapped the terahertz light before it had a chance to spread, essentially squeezing it into a space much smaller than its wavelength. In this regime, the light can bypass the diffraction limit to resolve features that were previously too small to see.

The MIT team adapted this technology to observe microscopic, quantum-scale phenomena. For their new study, the team developed a terahertz microscope using spintronic emitters interfaced with a Bragg mirror. This multilayered structure of reflective films successively filters out certain, undesired wavelengths of light while letting through others, protecting the sample from the “harmful” laser which triggers the terahertz emission.

As a demonstration, the team used the new microscope to image a small, atomically thin sample of BSCCO. They placed the sample very close to the terahertz source and imaged it at temperatures close to absolute zero — cold enough for the material to become a superconductor. To create the image, they scanned the laser beam, sending terahertz light through the sample and looking for the specific signatures left by the superconducting electrons.

“We see the terahertz field gets dramatically distorted, with little oscillations following the main pulse,” von Hoegen says. “That tells us that something in the sample is emitting terahertz light, after it got kicked by our initial terahertz pulse.”

With further analysis, the team concluded that the terahertz microscope was observing the natural, collective terahertz oscillations of superconducting electrons within the material.

“It’s this superconducting gel that we’re sort of seeing jiggle,” von Hoegen says.

This jiggling superfluid was expected, but never directly visualized until now. The team is now applying the microscope to other two-dimensional materials, where they hope to capture more terahertz phenomena.

“There are a lot of the fundamental excitations, like lattice vibrations and magnetic processes, and all these collective modes that happen at terahertz frequencies,” von Hoegen says. “We can now resonantly zoom in on these interesting physics with our terahertz microscope.”

For more information: Nature

Argonne National Laboratory’s new facility is giving researchers an inside look at irradiated nuclear materials

Argonne National Laboratory has opened its Activated Materials Lab, a radiological facility next to the Advanced Photon Source that lets scientists use an ultrabright X-ray beam to safely examine the inner structure of irradiated metals and nuclear fuels. The new capability allows researchers to see how radiation affects reactor materials in real time, supporting advances in reactor component design and helping utilities better plan maintenance and repair schedules.

Activated Materials Lab

Equipped with fume hoods, glove boxes, shielded containers, and approved sample containments, the new Activated Materials Lab can safely handle samples with higher radioactivity than previously allowed at the facility.

Staffed with a dedicated team that is responsible for receiving radioactive samples and transferring them safely to the Advanced Photon Source x-ray beamlines for measurements, the lab is able to reduce the turnaround time on user experiments while allowing researchers to focus on data collection and interpretation.

The Advanced Photon Source recently underwent a major upgrade project which enhanced its brightness over 100-fold and built nine new x-ray beamlines, including the High-Energy X-ray Microscope beamline at 20-ID which is adjacent to the Activated Materials Lab. With enhanced access to upgraded x-ray capabilities, researchers now have the ability to pursue a wider range of experiments. 

“By safely enabling higher activity samples at the Advanced Photon Source, this new capability allows for clearer views of how materials change during their time in a reactor, speeding progress toward safer, longer-lasting components,” said Brenden J. Heidrich, director of the U.S. Department of Energy’s Nuclear Science User Facilities program. 

Going Granular 

The new facility recently completed its first-ever user experiments examining the origins of stress corrosion cracks in irradiated materials.

Researchers looked at stainless steel parts that after decades of use were removed from a light-water reactor. The parts had developed microcracks due to a combination of irradiation, mechanical stress, and exposure to a corrosive environment.

Using a combination of three-dimensional x-ray techniques, researchers were able to examine the metal at the polycrystal grain level to see if they could find characteristics that correlated to cracking.

Finding the mechanisms that contribute to this behavior in alloys used in nuclear reactors will inform future material designs and long-term performance of existing reactors.

“This first experiment shows that the Activated Materials Laboratory can safely bridge irradiated samples to world-class X-ray tools, lowering barriers for the nuclear materials community,” said Xuan Zhang, Principal Materials Scientist at Argonne and facility lead of the AML. ​“By coordinating shipping, encapsulation, and safety reviews, AML helps users make the most of beam time and investigate questions that were out of reach before.”

“Access to higher-activity samples at beamline 1-ID, under strict controls, lets us probe grain-level behavior in materials that reflect real conditions,” said Jon Almer, Group Leader and Scientist, Materials Physics and Engineering Group, at Argonne. ​“Combining tomography with far-field and near-field high-energy diffraction microscopy gives a fuller picture that can inform models and materials design.”

The experiment was led by the University of Illinois Urbana–Champaign’s Professor James Stubbins, with collaborators at Oak Ridge National Laboratory, The University of Alabama, and Argonne National Laboratory, and was funded by the U.S. Department of Energy’s Nuclear Energy University Program and the Light Water Reactor Sustainability program. 

What’s Next?

The Activated Materials Lab stands ready to receive its next user experiment, adding its capacity to the list of innovative resources and capabilities the U. S. Department of Energy provides to researcher and industry to advance nuclear materials and technologies.

The Activated Materials Laboratory is supported by the Nuclear Science User Facilities (NSUF) program. The Advanced Photon Source is a Department of Energy (DOE) Office of Science user facility operated by Argonne National Laboratory.

For more information: NSUF

Electric field tunes vibrations to ease heat transfer

New research from the Department of Energy’s Oak Ridge National Laboratory found that applying an electric field to a ceramic material removes barriers to phonon transport, conducting heat almost three times more efficiently along the field direction, challenging conventional understanding about controlling heat flow in solid materials.

Continue reading

How rough grinding makes stainless steel prone to corrosion

Stainless steel is widely known for its corrosion resistance. But when it is exposed to environments containing chloride ions, such as seawater, the risk of corrosion increases. Manufacturers typically grind the surface to smooth it. However, this finishing process reduces corrosion resistance even further. Researchers at Tohoku University in Japan have recently shed light on why this occurs.

Continue reading

Stacking up: A new take on diamond electronics

Researchers at the U.S. Department of Energy’s Argonne National Laboratory have overcome a critical barrier in diamond-based electronics and microelectronics by using nanotechnology to integrate two-dimensional materials for efficient n-type doping, a breakthrough in electronic materials designed to operate in high-temperature environments and other harsh environments.

Continue reading

Scientists achieve sub-second 3D printing using rotating light field

Researchers have developed a new sub-second volumetric 3D printing technique that eliminates the need to rotate the printed sample, a long-standing mechanical challenge in the field. The system, called Digital Incoherent Synthesis of Holographic Light Fields, or DISH, instead rotates the illumination using a high-speed periscope, allowing millimeter-scale structures to be printed in 0.6 seconds with about 19-micrometer resolution across a 1-centimeter depth range. The advance addresses a persistent trade-off in volumetric additive manufacturing between resolution, stability and printable volume.

Volumetric 3D printing has long hoped to fabricate entire objects simultaneously – rather than layer by layer. But established approaches, such as computed axial lithography, typically rotate the resin container during exposure.

Fast rotation introduces vibration and alignment errors. Slow rotation, meanwhile, requires highly viscous resins, often thousands of centipoise, to prevent features from drifting before polymerization completes.

When it comes to optics, higher resolution demands higher numerical aperture (NA) objectives. Yet higher NA optics come with a shallow depth of field.

The system used in the study has an intrinsic NA of 0.055 at 405 nm, with a native depth-of-field of roughly 0.4 mm. That’s far smaller than the centimetre-scale volumes desirable for practical manufacturing. 

DISH tackles both precision and scale at once.

In their method, instead of moving the resin container the researchers mounted a rotating periscope on a hollow stage to deliver synchronized angular illumination while keeping the sample stationary.

A 405 nm coherent laser is modulated by a Digital Micromirror Device operating at 17 kHz, projecting optimized binary patterns as the periscope rotates at speeds up to 10 revolutions per second.

The demonstrated sub-second fabrication corresponds to the specific exposure timing used in the reported experiments.

The team abandoned conventional ray-based approximations and implemented a wave-optics model that explicitly incorporates diffraction and refraction at the air–material interface.

A coarse-to-fine iterative optimization algorithm generates projection patterns that maintain intensity modulation well beyond the native focal plane.

An adaptive calibration scheme using two orthogonal cameras corrects single-pixel misalignments in the synthesized 3D light field, improving angular registration and exposure fidelity.

Performance tests show that DISH maintains approximately 19 μm feature fidelity across a 1 cm depth range, far exceeding the objective’s intrinsic 0.4 mm depth of field.

Relief-structure experiments demonstrated approximately 11 μm uniform linewidth across the full centimeter span, while the smallest independently resolved positive feature measured 12 μm.

Comparative tests against conventional back-projection approaches showed sharper edges and improved consistency, particularly in off-center regions where optical blur typically increases.

The single-sided illumination geometry does introduce a missing-cone trade-off that slightly affects axial resolution. The authors note that alternative periscope geometries could mitigate this limitation in future implementations.

One of the more practically significant findings is material compatibility. The system printed successfully in aqueous solutions of polyethylene glycol diacrylate with viscosities as low as 4.7 cP. Because polymerization completes within 0.6 seconds, gravitational drift occurs only after solidification.

By contrast, conventional volumetric systems often require viscosities between 6,000 and 10,000 cP to maintain positional stability during slower exposures.

The researchers also demonstrated printing in higher-viscosity resins and bio-derived hydrogels, including gelatin methacrylate (GelMA) and silk fibroin methacrylate (SilMA).

The single-sided geometry further enables in situ fabrication on fixed substrates and within confined environments such as petri dishes.

Integration with a fluidic channel allowed successive fabrication of multiple structures, pointing toward continuous production workflows.

The authors estimate voxel rates on the order of 1.25 × 108/second, calculated for a defined voxel size and build volume. They suggest that higher-power lasers and faster modulation hardware could further increase build rates.

Surface analysis indicates that inclined projection reduces the prominence of stripe-like speckle artefacts compared with perpendicular illumination systems.

However, the hologram optimization process currently requires substantial offline computation. The authors propose GPU acceleration or neural-network-based approaches as pathways to reduce processing time and enable more automated deployment.

By decoupling angular illumination from sample motion and synthesizing holographic light fields through wave-optics modeling, DISH demonstrates a way to extend effective depth performance without sacrificing resolution.

While industrial deployment remains prospective, the work outlines a credible pathway toward faster, continuous volumetric manufacturing using both acrylate-based systems and selected biomaterials.

Future efforts are likely to focus on accelerating hologram computation, refining optical geometries to address missing-cone effects, and scaling projection hardware.

For more information: Nature

Golden experiment reveals the invisible forces holding the universe together

Scientists in Sweden have developed a simple method to make the universe’s normally invisible binding forces visible by using gold flakes, salt water and light. The Chalmers University of Technology team created a platform that displays these tiny forces—often described as “nature’s invisible glue,” responsible for effects like dust sticking to surfaces or lizards climbing walls—as shifting colors. The technique offers a faster, more accessible way to study how matter organizes itself at the smallest scales.

Inside the lab, doctoral student Michaela Hošková holds up a small glass container filled with millions of micrometer-sized gold flakes suspended in a salt solution. Using a pipette, she places a droplet of the mixture onto a gold-coated glass plate positioned under an optical microscope. The flakes are immediately drawn toward the surface but stop just short of fully attaching, leaving nanometer-sized gaps between the flakes and the gold substrate.

These tiny liquid-filled gaps act like miniature light chambers. Light reflects back and forth inside them, producing visible colors. When the setup is illuminated with a halogen lamp and the reflected light is analyzed with a spectrometer, the different wavelengths become clear. On the connected monitor, flakes shimmer and shift in colors such as red and green against a golden yellow background.

“What we are seeing is how fundamental forces in nature interact with each other. Through these tiny cavities, we can now measure and study the forces we call ‘nature’s glue’ – what binds objects together at the smallest scales. We don’t need to intervene in what is happening, we just observe the natural movements of the flakes,” says Michaela Hošková, a doctoral student at the Department of Physics at Chalmers University of Technology.

By examining the light captured in the cavities, the team can analyze the balance between two competing forces – one that pulls the flakes together and one that keeps them apart. The attractive force, known as the Casimir effect, causes the gold flakes to move toward each other and toward the surface. The opposing electrostatic force develops in the salt solution and prevents the flakes from fully sticking. When these forces reach equilibrium, a process called self-assembly occurs, forming the cavities that make the measurements possible.

“Forces at the nanoscale affect how different materials or structures are assembled, but we still do not fully understand all the principles that govern this complex self-assembly. If we fully understood them, we could learn to control self-assembly at the nanoscale. At the same time, we can gain insights into how the same principles govern nature on much larger scales, even how galaxies form,” says Michaela Hošková.

The new platform builds on several years of research in Professor Timur Shegai’s group at the Department of Physics. Four years ago, the team showed that a pair of gold flakes could form a self assembled resonator. They have now expanded that discovery into a broader method for investigating fundamental forces.

In this system, the gold flakes function as tiny floating sensors. According to the researchers, the approach could be valuable across physics, chemistry, and materials science.

“The method allows us to study the charge of individual particles and the forces acting between them. Other methods for studying these forces often require sophisticated instruments which cannot provide information down to the particle level,” says research leader Timur Shegai.

The platform may also help scientists better understand how particles behave in liquids, including whether they remain stable or tend to clump together. That knowledge could improve how medicines move through the body, support the design of more effective biosensors, and contribute to better water filtration systems. It is also relevant for everyday products such as cosmetics, where preventing unwanted clumping is essential.

“The fact that the platform allows us to study fundamental forces and material properties shows its potential as a truly promising research platform,” says Timur Shegai.

In the laboratory, Hošková opens a small box containing a finished version of the device. Using tweezers, she places it into the microscope. Two thin glass plates enclose everything needed to examine nature’s invisible glue.

“What I find most exciting is that the measurement itself is so beautiful and easy. The method is simple and fast, based only on the movement of gold flakes and the interaction between light and matter,” says Michaela Hošková, zooming in on a gold flake whose colors immediately reveal the forces at work.

Gold flakes about 10 micrometers in size are placed in a salt solution, meaning water that contains free ions. When a droplet is added to a gold-coated glass surface, the flakes are drawn toward it, and nanometer-sized cavities (100-200 nanometers) form. This self assembly results from the balance between two forces: the Casimir force, a measurable quantum effect that pulls objects together, and the electrostatic force that arises between charged surfaces in a salt solution.

A halogen lamp shines light into the cavities, where it becomes trapped and reflected. An optical microscope and spectrometer then separate the light into its component wavelengths so the colors can be identified. By adjusting the salt concentration and observing how the flakes shift relative to the surface, researchers can measure the underlying forces. To prevent evaporation, the droplet containing the gold flakes is sealed and covered with a second glass plate.

The platform was developed at Chalmers’ Nanofabrication Laboratory, Myfab Chalmers, and at the Chalmers Materials Analysis Laboratory (CMAL).

For more information: PNAS

Image: Researchers at Chalmers University of Technology, Sweden, have discovered a quick and easy way to study the hidden forces that bind the smallest objects in the universe together. Using gold, salt water, and light, they have created a platform on which the forces can be seen through colors. Two thin glass plates hold everything needed to study nature’s invisible glue. Credit: Chalmers University of Technology | Mia Halleröd Palmgren

Research team shows nanoparticles adhere to quantum mechanics

Researchers at the University of Vienna have demonstrated quantum interference in sodium nanoparticles containing more than 7,000 atoms—and produced clusters of nearly 10,000—showing that even large metallic particles can exhibit quantum behavior. Lead author Sebastian Pedalino said the results challenge the assumption that such objects must behave classically, confirming that quantum mechanics remains valid at this scale without the need for alternative models. 

The team, led by Markus Arndt and Stefan Gerlich, achieved the breakthrough using sodium clusters exceeding 170,000 atomic mass units, larger than most proteins, marking a significant advance toward the long-sought goal of sustaining quantum superposition in increasingly complex systems.

The test was conducted using the multi-scale cluster interference experiment (MUSCLE). 

To observe interference, the team had to minimize environmental disturbances that would destroy coherence. They produced the sodium clusters under cryogenic conditions (77 K, the temperature of liquid nitrogen) to reduce thermal decoherence, and ran the interferometer in ultra-high vacuum (about 9 × 10⁻⁹ mbar, roughly 10⁻¹¹ of atmospheric pressure) to limit collisions with gas molecules.

The scientists produced sodium clusters containing roughly 5,000 to 10,000 atoms and sent them through a three-grating interferometer made from standing ultraviolet laser light. The first grating spatially confined the particles to build up coherence, and the second grating acted as a beam splitter for the clusters’ matter wave. Scanning the third grating revealed an interference fringe pattern, and the team showed the fringe visibility followed the quantum prediction rather than a classical shadow-pattern model

The study achieved a macroscopicity value of 15.5. This metric quantifies how strictly an experiment tests the limits of quantum theory. This study has a value 10 times higher than any previous experiment. 

This new level of sensitivity presents new opportunities for nanotechnology and materials science, opening new doors for measuring properties of nanoparticles and investigating the transition of matter from individual atoms to bulk metallic solids. 

For more information: University of Vienna

Image: MUSCLE at the University of Vienna, where quantum interference of massive nanoparticles was detected. To isolate vibrations, the experiment is mounted on a table weighing several tons that floats on an air cushion. Credit: S. Pedalino / Uni Wien

Building a sustainable metals infrastructure: NIST report highlights key strategies

NIST has released a report outlining strategies to build a more efficient, sustainable, and resilient U.S. metals processing infrastructure, emphasizing the need for improved standards for recycled content and stronger supply chains for critical materials. Covering the full lifecycle—from mining and alloy design to manufacturing, reuse, and recycling—the report highlights that addressing these challenges is essential for innovation, industrial competitiveness, and national security. The findings stem from a NIST workshop held in July 2024.

“The workshop brought together a diverse group of experts from industry, academia and the policy world to take on some of the biggest challenges in the metals processing space,” said NIST materials research engineer Andrew Iams, a co-author on the report. “Meeting these challenges requires a new approach in how to source, process, use and recycle metals.”

The report covers various topics related to metals manufacturing, from new technologies for extracting and processing bulk materials, like aluminum and steel, to developing new modeling and data tools to design advanced alloys.

The report highlights the importance of critical materials, including minerals containing lithium and cobalt that are key manufacturing elements for smartphones, batteries, semiconductors and medical devices, as well as superalloys used in military hardware and jet engines.

These materials can be challenging to obtain due to limited availability and the risk of supply chain disruptions. Industries can address these issues by diversifying their supply chains with new sources, identifying substitute materials, and improving recycling methods to enable greater recirculation of existing materials.

The report also highlights the need to improve standards for metals reuse and recycling. Better standards can make the separation of metals for recycling more efficient, reducing industry costs. New certification programs can help ensure that products made with recycled content meet performance standards, which could expand the market for recycled materials.

The report highlights five strategies that would help the industry tackle these and other challenges:

  • Advance measurement science for sustainable metals manufacturing, including new separation techniques for recycling.
  • Develop the technical basis to support standards development, including the data needed to create or improve performance-based standards for highly recycled metals, such as aluminum and steel.
  • Enhance data and modeling tools for addressing supply risks and designing products for improved recyclability.
  • Promote workforce development and education by establishing training programs and creating partnerships between universities, labs and industry.
  • Convene stakeholders to establish collaborations that foster knowledge-sharing and innovation.

The NIST workshop brought together manufacturers, technology companies, researchers and other experts from all stages of the metals processing chain. NIST has a long history of convening stakeholders across industrial sectors to solve shared problems through better technology and standards.

“We are always seeking ways to help industrial partners solve tough engineering or scientific problems,” Iams said. “Part of NIST’s mission is to help keep U.S. industry competitive. We can do that by identifying promising technologies and helping to move them out of the lab so they can be implemented on an industrial scale.”

For more information: Material Challenges in Developing a Sustainable Metal Processing Infrastructure – Workshop Report

Lehigh’s Blacksmithing Club to open its first dedicated lab

Lehigh University’s Blacksmithing Club is set to move into its first dedicated lab space in Whitaker Laboratory, marking a major step for its growing community of student metalworkers. The club, supported by professor Laura Moyer and partnerships with Historic Bethlehem Museums & Sites and Lehigh Heavy Forge, offers hands-on experience in both traditional and modern metalworking. Students have also practiced basic techniques at the historic 1750 Smithy, while The Loewy Institute continues to provide education in advanced metal-forming technology.

Moyer said the team had discussed building a lab for years but wasn’t sure how to achieve it financially or where it would be located. 

With student interest rising, Moyer said the idea for the blacksmithing club started to solidify two years ago when a student — now club president Josh Swavely, ‘26 — mentioned his passion for blacksmithing in one of Misiolek’s classes.

Under the guidance of Moyer and Misiolek, the group helped launch a one-credit blacksmithing elective offered each spring. Moyer said this course, now in its second year, is open to students of all majors, as is the club. 

Moyer also said they are hoping to grow the one-credit course that is currently offered for half a semester into a full-semester three-credit elective for students. 

“Within the course we are developing, the idea is to balance time between the laboratory and the lecture hall, between practice and theory,” Misiolek said. “It will be much more of an opportunity to design your own products, analyze the proposed processes, and learn while you are going through the process.”

Moyer said the new laboratory will allow students to complete every stage of the design cycle in one place — from heating and shaping metal at the forge to examining the microstructure of their finished work.

She said students will also be able to fabricate hooks, blades and small hardware entirely on campus supported by new ventilation hoods, anvils and space for larger tools.

Misiolek said the club’s growth has been driven by access to working forges across Bethlehem. Through its partnership with Historic Bethlehem Museums & Sites, students learn centuries-old techniques at the 1750 Smithy.

He also said their work with the Lehigh Heavy Forge in Bethlehem allows students to witness a modern industrial facility that provides advanced forgings to clients.

“We are connecting the historical aspects of forging technology with hands-on experience and current industrial practices,” Misiolek said. 

He also said beyond traditional forging skills, the new lab will support lessons on heat treatments, controlled cooling and other methods that help students compare how different variables affect the strength and performance of their designs.

Nick Rockwell, a researcher at the Loewy Institute, said community partnerships have been essential in turning the club’s ideas into practice. 

“With all the support we’ve received, it’s really helped us connect with Historic Bethlehem,” he said. 

He also said Mike Rex, the shop and laboratory operations supervisor, has been instrumental in building facilities for the lab and giving advice on how to best use the space.

The club also gives students opportunities beyond the classroom. Members have participated in the Forging Industry Educational and Research Foundation’s annual competition and secured funding from the nonprofit organization to support their work.

“We pride ourselves in our hands-on approach,” Moyer said. “There are a lot of materials science and engineering departments across the country, but many don’t allow undergraduate students into labs or use highly specialized equipment.”

For more information: Lehigh University

Image: Josh Swavely, ’26, is pictured forging a point in the soon to be blacksmithing lab. Swavely is the president of the Lehigh Blacksmithing Club. (Max Randall/B&W Staff)

Mapping the future: AI method to transform alloy properties prediction and design

Researchers at The Grainger College of Engineering have integrated their expertise in metals with advanced machine learning to create detailed spatial maps, enabling faster and more precise autonomous material design. Similar to how fingerprint technology captures intricate ridge and valley patterns for biometric identification, their approach leverages spatial mapping of fine details to revolutionize material engineering.

This evolution of recognition technology is mirrored in the field of materials science, where researchers seek new and efficient ways to fully characterize materials, accelerating the discovery of additional new materials. Much like human fingerprints, the performance of metal mixtures called alloys relies on the intricate spatial arrangement of microstructural features. Traditional methods reduce this complexity into a handful of averaged values, causing each alloy to lose its distinctive “fingerprint.”

 In a recent complement of papers from the lab of Jean-Charles Stinville, assistant professor of materials science and engineering, Illinois Grainger engineers have introduced new machine learning approaches for identifying alloy microstructures and predicting their properties rapidly. The Illinois researchers’ method will provide new avenues for faster and more efficient materials design. 

Microstructures are tiny structural features of metals that influence their strength and behavior. Scientists look to the microstructural properties of metals to assess their functionality. Metals used in propulsion devices like rockets and airplanes have special requirements. 

“We are sending these materials into increasingly extreme environments,” Stinville said. “They are exposed to intense environments; for instance, structural materials for space applications must be resistant to mechanical loading under extremely low or high temperatures. Conventional alloys don’t do as well in these conditions because their mechanical properties tend to degrade under these extreme environments. We want to find new ways to accelerate the identification of alloy chemistries and microstructures that can withstand these harsh conditions.”

The complete details of these microstructures, including small-scale influential variances called heterogeneities, cannot be easily captured by existing methods. Instead, Stinville and his colleagues used deep learning to analyze diffraction patterns, or the way electrons interact with metals. By encoding these interactions through a machine learning method onto a spatial latent representation, the researchers captured the full extent of an alloy’s microstructure and its heterogeneity — an approach Stinville calls Material Spatial Intelligence.

“Traditionally, we have used single descriptors or average values to guide data-based alloy design,” he said. “But spatial information from local measurements over a large field of view allows us to capture microstructure heterogeneity of the alloy. Using such spatial information in a data-based model provides significant improvement in prediction accuracy and enables alloy and microstructure design.” 

The initial model is a machine learning approach that successfully identified microstructures and material heterogeneity in unprecedented detail. In a second paper published in Scripta Materialia, Stinville further progressed the model towards the prediction of mechanical properties using the developed approach of material spatial intelligence. This method accelerates alloy property prediction by orders of magnitude and provides a rapid fundamental understanding of structure properties in metals. 

“I started my career as an experimentalist, where I developed tools that allowed us to collect large fields of view with very high resolution,” he said. “Then I went over to the numerical side to develop machine learning tools to actually use all this spatial information. As a metallurgist, I have an understanding that metals are controlled by local properties and their heterogeneities. My unique material scientist background really helped me in developing these novel models.”

By combining high-resolution digital image correlation with alloy microstructure characterization, Stinville examined tiny regions of metal surfaces and how they deformed at a small scale when loaded. Training a new model to recognize these deformation fingerprints allowed him to reliably predict important properties like strength, fatigue life, and ductility (the ability to extend without breaking). The model significantly decreases the time for testing, lessening the time needed to evaluate new alloys. This acceleration brings the field one step closer to intelligent alloy design.

Stinville envisions a future model that works backwards from a user’s desired properties to suggest a chemical composition and microstructure that best suits the given parameters. By integrating these approaches with his group’s advances in automated characterization, Stinville’s lab is setting the stage for fully autonomous alloy design, marking their next frontier.

But even as exciting advancements loom, Stinville still marvels at his field’s early beginnings. 

“This approach unites our field’s fundamental understanding of metals with new and efficient AI database tools,” he said. “We’re not just taking these new tools and leaving behind what we’ve already learned. We’re integrating the present with the past.” 

Mathieu Calvat, Chris Bean and Dhruv Anjaria significantly contributed to this research.

For more information: NPJ Computational Materials

Image: Electron backscatter diffraction (EBSD) maps of the investigated Inconel 718 alloys. Inverse pole figure (IPF) maps along the X direction (horizontal) are presented for a A wrought and fully recrystallized 718 alloy, and a B additively manufactured as-built 718 alloy

Study shows light can reshape atom-thin semiconductors for next-generation optical devices

Rice University researchers have discovered that light can induce a physical shift in the atomic lattice of transition metal dichalcogenides (TMDs), a class of atom-thin semiconductors. This tunable effect, observed in a Janus-type TMD, opens the door to technologies that use light instead of electricity, enabling faster, cooler computer chips, ultrasensitive sensors, and flexible optoelectronic devices.

“In nonlinear optics, light can be reshaped to create new colors, faster pulses or optical switches that turn signals on and off,” said Kunyan Zhang, a Rice doctoral alumna who is a first author on a study documenting the effect. “Two-dimensional materials, which are only a few atoms thick, make it possible to build these optical tools on a very small scale.”

TMDs are layered crystals made of a transition metal such as molybdenum and two layers of a chalcogen element like sulfur or selenium. Their combination of electrical conductivity, light absorption and mechanical flexibility has made them one of the most versatile classes of materials for next-generation electronics and optoelectronics.

Within this family, Janus materials stand out for their asymmetry — an idea reflected in their name. Like their mythological namesake, these materials have two different faces: The top and bottom atoms are made of different chemical species, creating an internal imbalance that gives the crystal a built-in electrical polarity, making it particularly sensitive to light and external forces.

“Our work explores how the structure of Janus materials affects their optical behavior and how light itself can generate a force in the materials,” Zhang said.

Using laser light of different colors, the team studied how a two-layer Janus TMD material — molybdenum sulfur selenide stacked on molybdenum disulfide — converts light through a process called second harmonic generation (SHG), in which the material emits light at twice the frequency of the incoming beam. They found that when the incoming light matched the material’s natural resonances, the doubled-frequency light pattern became distorted, signaling that the atoms inside were being displaced.

“We discovered that shining light on Janus molybdenum sulfur selenide and molybdenum disulfide creates tiny, directional forces inside the material, which show up as changes in its SHG pattern,” Zhang said. “Normally, the SHG signal forms a six-pointed ‘flower’ shape that mirrors the crystal’s symmetry. But when light pushes on the atoms, this symmetry breaks — the petals of the pattern shrink unevenly.”

The team traced the distortion to optostriction, a process in which the electromagnetic field of light itself exerts a mechanical push on atoms. In Janus materials, that push is amplified by strong coupling between the atomic layers, allowing even minute forces to produce measurable strain.

“Janus materials are ideal for this because their uneven composition creates an enhanced coupling between layers, which makes them more sensitive to light’s tiny forces — forces so small that it is difficult to measure directly, but we can detect them through changes in the SHG signal pattern,” Zhang said.

That sensitivity could make these materials useful far beyond the lab. Components that switch or route light using this principle could make optical chips faster and far more energy-efficient, since light-based circuits generate less heat than conventional electronics. The same responsiveness could be harnessed to create precise sensors capable of detecting the smallest vibrations or pressure changes or tunable light sources for advanced displays and imaging tools.

“Such active control could help design next-generation photonic chips, ultrasensitive detectors or quantum light sources ⎯ technologies that use light to carry and process information instead of relying on electricity,” said Shengxi Huang, an associate professor of electrical and computer engineering and materials science and nanoengineering at Rice and a corresponding author on the study. Huang is also a member of the Smalley-Curl Institute, the Rice Advanced Materials Institute and the Ken Kennedy Institute at Rice.

By showing how Janus TMDs’ built-in imbalance opens new ways to steer the flow of light, the study highlights how small structural features can unlock large technological potential.

For more information: ACS Nano

Image: Shengxi Huang is an associate professor of electrical and computer engineering and materials science and nanoengineering at Rice University. (Photo by Jeff Fitlow/Rice University)

Argonne-led Q-NEXT quantum center renewed for five years

The U.S. Department of Energy (DOE) has renewed funding for Q-NEXT, a National Quantum Information Science Research Center led by Argonne National Laboratory with SLAC National Accelerator Laboratory, for another five years. Backed by $125 million—$25 million in fiscal year 2026 and future funding subject to congressional approval—this investment ensures Q-NEXT will continue driving advancements in quantum information science and technology, reinforcing the United States’ leadership in this transformative field.

“Quantum information science is a cornerstone of the nation’s technological future, with the potential to transform industries including computing, healthcare and national security,” said inaugural Q-NEXT Director David Awschalom, who is also a senior scientist at Argonne, the director of quantum engineering at the University of Chicago Pritzker School of Molecular Engineering, and director of the Chicago Quantum Exchange. ​“Through our renewed mission, Q-NEXT will continue to push the boundaries of what is possible in quantum science, delivering the foundational knowledge, tools and technologies needed to ensure national leadership in this critical field.”

Awschalom has now assumed the role of chief science officer.

Q-NEXT’s mission is to unlock the future of quantum information by seamlessly integrating quantum and traditional information systems across optical networks. Building on its achievements over the past five years, the center will focus on demonstrating the potential of distributed quantum entanglement — a phenomenon where qubits, the fundamental unit of quantum information, remain connected even when separated by large distances.

“We’re building on the strong foundation we’ve laid over the past five years to take on our renewed mission — harnessing distributed entanglement to show what’s possible with scalable quantum platforms,” said Q-NEXT Director and Argonne scientist Martin Holt. ​“By uniting quantum technologies across optical networks, we will pave the way for systems capable of revolutionizing how we process, transmit and receive information.”

Q-NEXT brings together a strong network of partners: two DOE national laboratories, 11 leading universities and six tech companies. This cross-sector collaboration ensures that the center’s work is both innovative and practical, bridging the gap between scientific discovery and real-world application. Universities contribute expertise in quantum sensing and communication, while industry partners provide access to state-of-the-art prototypes and manufacturing capabilities. Together, Q-NEXT partners form a vibrant ecosystem that drives progress in quantum science.

“We envision quantum systems that work across chip-to-chip, lab-to-lab and city-to-city scales,” said Q-NEXT Deputy Director Jennifer Dionne, who is a professor of materials science and engineering and, by courtesy, of radiology at Stanford University and SLAC. ​“We’re excited to advance a shared set of hardware, software and protocols to make quantum networks and sensors efficient and practical.”

Q-NEXT’s renewed efforts will focus on three scientific goals:

  1. Communication — developing quantum communication networks that link devices across metropolitan areas. Q-NEXT aims to demonstrate algorithms that run across multiple remote, connected quantum processors.
  2. Sensing — using quantum entanglement to achieve unprecedented precision in sensing applications. Q-NEXT will demonstrate real-world uses, such as in medicine and navigation, as well as foundational scientific discoveries in gravitation and quantum mechanics where quantum entanglement gives a clear advantage in sensing and measurement.
  3. Materials — developing new approaches to integrating materials that can be scaled for industry use. Q-NEXT will tackle the most important challenges involved in merging distinct quantum materials systems with advanced functionality and integrating them into practical quantum devices.

Q-NEXT will pursue its goals through large-scale, team-based projects that combine materials science, device engineering and quantum physics theory. By leveraging world-class science facilities, Q-NEXT aims to deliver breakthroughs that will shape the future of quantum technology. These facilities include the Argonne and SLAC Quantum Foundries, the Stanford Synchrotron Radiation Lightsource and several DOE Office of Science user facilities: Argonne’s Advanced Photon Source and Center for Nanoscale Materials, the Argonne Leadership Computing Facility and SLAC’s Linac Coherent Light Source.

In addition to its scientific mission, Q-NEXT will continue to build the next generation of quantum scientists, engineers, technicians and other professionals. Through educational programs, internships and training opportunities such as the DOE Science Undergraduate Laboratory Internship and the Open Quantum Initiative Undergraduate Fellowship, the center is preparing students, early-career researchers and operations specialists to thrive in the rapidly growing quantum industry.

“By fostering a skilled workforce, Q-NEXT is ensuring that the U.S. remains at the forefront of quantum innovation,” Awschalom said.

Q-NEXT was established in 2020, and in its first five years, it led the establishment of the Argonne and SLAC Quantum Foundries. Together these two national facilities contribute to a robust supply chain of standardized materials and devices.

“We’ve built an active, cross-disciplinary collaboration that’s laid the groundwork for networked quantum information by intertwining computing, sensing and communication,” Holt said. ​“Over the next five years, we’ll strengthen these efforts, strategically coordinating with the other NQISRCs and the national quantum ecosystem to accelerate the arrival of transformative quantum technologies.”

Collaboration with industry has been essential to the center’s mission — turning cutting-edge research into innovations that can transform technology and strengthen the nation’s quantum leadership.

“IBM’s work with the National Quantum Information Science Research Centers, such as Q-NEXT led by Argonne, is vital to our mission to build the future of computing,” said Jay Gambetta, director of IBM Research and IBM Fellow. IBM is a Q-NEXT partner. ​“Together, we are exploring how efficient quantum networks can be created through optical links connected to IBM’s quantum networking units. This could deliver the fundamental technology needed to link the multiple, interconnected fault-tolerant quantum computers of the future over kilometer distances, allowing the convergence of quantum computation and communication within a future quantum computing internet that could revolutionize scientific and industrial discovery.”

Q-NEXT is one of five DOE National Quantum Information Science Research Centers renewed for another five years.

“Quantum technologies are driving innovations across society, and our laboratory is focusing on delivering science breakthroughs to accelerate these innovations,” said Argonne Director Paul Kearns. ​“With a renewed Q-NEXT, we will continue to play an integral role in the national quantum ecosystem through coordinated, complementary efforts with DOE’s other quantum research centers. We are committed to realizing the promise of quantum information to build a more connected world and shape a future of scientific progress that strengthens our nation’s security, prosperity and technological leadership.”

For more information: Argonne National Laboratory

Image: The DOE has renewed the Q-NEXT quantum center for five years with a $125 million investment. One of Q-NEXT’s major first-run accomplishments was the establishment of two national quantum foundries — one at Argonne (pictured) and one at SLAC. (Image by Argonne National Laboratory.)

New ORNL aluminum alloy to strengthen domestic auto supply chain

Over the next decade, large amounts of aluminum auto body scrap will enter salvage systems, but its impurities have traditionally limited reuse in critical automotive parts. Researchers at the Department of Energy’s Oak Ridge National Laboratory (ORNL) have addressed this challenge by developing RidgeAlloy, an innovative aluminum alloy that transforms low-value scrap into high-quality material for structural vehicle components. Produced by remelting and recasting post-consumer aluminum, RidgeAlloy meets strength, ductility, and crashworthiness standards, creating a sustainable domestic supply chain. This breakthrough supports DOE’s critical materials goals, as aluminum is essential for energy technologies that produce, transmit, store, and conserve energy.

“The team advanced from a paper concept to a successful, full-scale part demonstration of a new alloy in only 15 months,” said Allen Haynes, director of ORNL’s Light Metals Core Program. “That’s an unheard-of pace of innovation in developing complex structural alloys.” 

Aluminum-intensive vehicles entered the U.S. market around 2015, with Ford’s F-150 truck series among the first to be mass produced. By the early 2030s, many of these vehicles are projected to reach end-of-life, creating a surge of high-quality aluminum body sheet scrap — up to 350,000 tons annually in North America. Much of this sheet scrap is expected to be downcycled into low-grade castings or exported, which is a missed opportunity to use those resources as a source of high-quality domestic aluminum. 

“You can repurpose post-consumer aluminum into something non-structural like engine blocks,” said Alex Plotkowski, ORNL group leader of Computational Coupled Physics. “But it won’t have the properties needed for higher value, structurally sound body applications.” 

That’s because the vehicle shredding process introduces impurities, such as iron, from various parts, including fasteners like rivets. This makes the scrap chemistry too unpredictable and low performing for commercial automotive structural alloys. As a result, most lightweight parts are made using primary aluminum, which is produced from raw ore in an energy-intensive process. 

While primary aluminum is mostly imported, the U.S. has some of the world’s best infrastructure for vehicle shredding and aluminum scrap recovery. 

“Using remelted scrap instead of primary aluminum is estimated to result in up to 95% reduction in the energy needed for processing a part,” said Amit Shyam, leader of ORNL’s Alloy Behavior and Design Group. 

To make that possible, the team applied world-class scientific tools such as high-throughput computing, which involved more than two million calculations to predict the optimal alloy compositions with targeted properties, as well as materials characterization and neutron diffraction at ORNL’s Spallation Neutron Source, a DOE Office of Science user facility. These tools helped the researchers understand how specific impurities affect alloy behavior. Neutrons are uniquely suited for this kind of research because they can penetrate deep into dense metals without damaging the material, allowing scientists to observe internal structures and atomic-scale changes. 

After pinpointing the desired blend through rapid computational predictions and laboratory trials, the new alloy was tested in a real-world environment. PSW Group’s Trialco Aluminum in Chicago supplied recycled aluminum ingots, metal blocks ready for remelting, cast from mixed auto body sheet scrap and tailored to RidgeAlloy’s specifications. The ingots were shipped to Falcon Lakeside Manufacturing in Michigan, where they were successfully cast into automotive parts using high-pressure die-casting. 

“The part we chose was medium-sized and moderately complex,” Plotkowski said. “The ultimate goal is to eventually cast larger parts, perhaps even automotive giga-castings, but this is the first step.” 

The cast parts confirmed that RidgeAlloy, consisting of aluminum, magnesium, silicon, iron and manganese, had the combination of properties necessary for structural vehicle castings, even when made from recycled blends with higher iron and silicon content. It delivers strength, corrosion resistance and ductility, enabling the production of structural castings of underbodies, frame components and other critical parts from post-consumer aluminum scrap. This breakthrough offers the opportunity to reshape the value equation of how North American auto body sheet scrap is sorted and reused.

“This team figured out how to take full advantage of a national lab’s world-class suite of capabilities to rapidly fill a huge gap in our understanding of lightweight automotive materials,” Haynes said. 

By the early 2030s, RidgeAlloy could enable recycled structural castings at volumes equal to at least half of the annual primary aluminum production in the U.S. This would reduce energy use, cut costs and strengthen domestic supply chains.

“RidgeAlloy offers the first technology capable of recapturing the value of a fast-approaching and historically massive wave of domestic, high-quality recycled automotive aluminum sheet alloys,” Haynes said. “That’s the big picture supply chain impact our team aimed for.”

There is also potential for future applications in industrial machinery, agricultural equipment, aerospace, mobile power generation equipment, off-road vehicles such as snowmobiles, motorcycles, and marine vehicles including jet skis.

For more information: Oak Ridge National Laboratory

Image: This automotive part was manufactured from RidgeAlloy, a new structural alloy developed by researchers at ORNL. It was cast using metals recycled entirely from post-consumer aluminum auto body sheets. Credit: ORNL, U.S. Dept. of Energy

Scientists turn common semiconductor into a superconductor

Researchers have long sought to make semiconductors—essential for computer chips and solar cells—function as superconductors, which can carry electricity without resistance for faster, more efficient performance. This has been difficult because it requires a precise atomic structure that enables free electron movement. Now, an international team has achieved a breakthrough by creating a form of germanium that exhibits superconductivity, allowing electric currents to flow indefinitely without energy loss. This advancement could significantly enhance electronic and quantum devices while reducing power consumption.

“Establishing superconductivity in germanium, which is already widely used in computer chips and fiber optics, can potentially revolutionize scores of consumer products and industrial technologies,” explains Javad Shabani, a physicist at New York University and director of its Center of Quantum Information Physics and Quantum Institute.

Peter Jacobson, a physicist at the University of Queensland, adds that the findings could accelerate progress in building practical quantum systems. “These materials could underpin future quantum circuits, sensors, and low-power cryogenic electronics, all of which need clean interfaces between superconducting and semiconducting regions,” he says. “Germanium is already a workhorse material for advanced semiconductor technologies, so by showing it can also become superconducting under controlled growth conditions there’s now potential for scalable, foundry-ready quantum devices.”

Germanium and silicon, both group IV elements with diamond-like crystal structures, occupy a unique position between metals and insulators. Their versatility and durability make them central to modern manufacturing. To induce superconductivity in such elements, scientists must carefully alter their atomic structure to increase the number of electrons available for conduction. These electrons then pair up and move through the material without resistance — a process that is notoriously difficult to fine-tune on the atomic scale.

In the new study, researchers developed germanium films heavily infused with gallium, a softer element commonly used in electronics. This technique, known as “doping,” has long been used to modify a semiconductor’s electrical behavior. Normally, high levels of gallium destabilize the crystal, preventing superconductivity.

The team overcame this limitation using advanced X-ray methods to guide a refined process that encourages gallium atoms to take the place of germanium atoms in the crystal lattice. Although this substitution slightly distorts the crystal, it preserves its overall stability and allows it to carry current with zero resistance at 3.5 Kelvin (about -453 degrees Fahrenheit), confirming that it had become superconducting.

“Rather than ion implantation, molecular beam epitaxy was used to precisely incorporate gallium atoms into the germanium’s crystal lattice,” says Julian Steele, a physicist at the University of Queensland and a co-author of the study. “Using epitaxy — growing thin crystal layers — means we can finally achieve the structural precision needed to understand and control how superconductivity emerges in these materials.”

As Shabani notes, “This works because group IV elements don’t naturally superconduct under normal conditions, but modifying their crystal structure enables the formation of electron pairings that allow superconductivity.”

For more information: Nature Nanotechnology

Image: Josephson junction structures—quantum devices made of two superconductors and a thin non-superconducting barrier—using different forms of germanium (Ge): super-Ge (in gold), semiconducting Ge (in blue), and super-Ge on wafer-level scale. Millions of Josephson junction pixels (10 micrometer square) can be created with this new material stack on wafer scale. Inset shows crystalline form of Super-Ge on the same matrix of semiconductor Ge, a key for crystalline Josephson junction. Credit: Patrick Strohbeen/NYU

Checking the quality of materials just got easier with a new AI tool

Advancing technologies like batteries, electronics, and pharmaceuticals relies on discovering and verifying new materials, a process traditionally slowed by costly, time-consuming quality checks using specialized instruments. While AI has accelerated material discovery by identifying promising candidates from vast databases, MIT engineers have now developed a new AI tool that streamlines the verification process, potentially reducing delays and costs in materials-driven industries.

In a new study the researchers present “SpectroGen,” a generative AI tool that turbocharges scanning capabilities by serving as a virtual spectrometer. The tool takes in “spectra,” or measurements of a material in one scanning modality, such as infrared, and generates what that material’s spectra would look like if it were scanned in an entirely different modality, such as X-ray. The AI-generated spectral results match, with 99 percent accuracy, the results obtained from physically scanning the material with the new instrument.

Certain spectroscopic modalities reveal specific properties in a material: Infrared reveals a material’s molecular groups, while X-ray diffraction visualizes the material’s crystal structures, and Raman scattering illuminates a material’s molecular vibrations. Each of these properties is essential in gauging a material’s quality and typically requires tedious workflows on multiple expensive and distinct instruments to measure.

With SpectroGen, the researchers envision that a diversity of measurements can be made using a single and cheaper physical scope. For instance, a manufacturing line could carry out quality control of materials by scanning them with a single infrared camera. Those infrared spectra could then be fed into SpectroGen to automatically generate the material’s X-ray spectra, without the factory having to house and operate a separate, often more expensive X-ray-scanning laboratory.

The new AI tool generates spectra in less than one minute, a thousand times faster compared to traditional approaches that can take several hours to days to measure and validate.

“We think that you don’t have to do the physical measurements in all the modalities you need, but perhaps just in a single, simple, and cheap modality,” says study lead Loza Tadesse, assistant professor of mechanical engineering at MIT. “Then you can use SpectroGen to generate the rest. And this could improve productivity, efficiency, and quality of manufacturing.”

The study was led by Tadesse, with former MIT postdoc Yanmin Zhu serving as first author.

Tadesse’s interdisciplinary group at MIT pioneers technologies that advance human and planetary health, developing innovations for applications ranging from rapid disease diagnostics to sustainable agriculture.

“Diagnosing diseases, and material analysis in general, usually involves scanning samples and collecting spectra in different modalities, with different instruments that are bulky and expensive and that you might not all find in one lab,” Tadesse says. “So, we were brainstorming about how to miniaturize all this equipment and how to streamline the experimental pipeline.”

Zhu noted the increasing use of generative AI tools for discovering new materials and drug candidates, and wondered whether AI could also be harnessed to generate spectral data. In other words, could AI act as a virtual spectrometer?

A spectroscope probes a material’s properties by sending light of a certain wavelength into the material. That light causes molecular bonds in the material to vibrate in ways that scatter the light back out to the scope, where the light is recorded as a pattern of waves, or spectra, that can then be read as a signature of the material’s structure.

For AI to generate spectral data, the conventional approach would involve training an algorithm to recognize connections between physical atoms and features in a material, and the spectra they produce. Given the complexity of molecular structures within just one material, Tadesse says such an approach can quickly become intractable.

“Doing this even for just one material is impossible,” she says. “So, we thought, is there another way to interpret spectra?”

The team found an answer with math. They realized that a spectral pattern, which is a sequence of waveforms, can be represented mathematically. For instance, a spectrum that contains a series of bell curves is known as a “Gaussian” distribution, which is associated with a certain mathematical expression, compared to a series of narrower waves, known as a “Lorentzian” distribution, that is described by a separate, distinct algorithm. And as it turns out, for most materials infrared spectra characteristically contain more Lorentzian waveforms, while Raman spectra are more Gaussian, and X-ray spectra is a mix of the two.

Tadesse and Zhu worked this mathematical interpretation of spectral data into an algorithm that they then incorporated into a generative AI model.

“It’s a physics-savvy generative AI that understands what spectra are,” Tadesse says. “And the key novelty is, we interpreted spectra not as how it comes about from chemicals and bonds, but that it is actually math — curves and graphs, which an AI tool can understand and interpret.”

The team demonstrated their SpectroGen AI tool on a large, publicly available dataset of over 6,000 mineral samples. Each sample includes information on the mineral’s properties, such as its elemental composition and crystal structure. Many samples in the dataset also include spectral data in different modalities, such as X-ray, Raman, and infrared. Of these samples, the team fed several hundred to SpectroGen, in a process that trained the AI tool, also known as a neural network, to learn correlations between a mineral’s different spectral modalities. This training enabled SpectroGen to take in spectra of a material in one modality, such as in infrared, and generate what a spectra in a totally different modality, such as X-ray, should look like.

Once they trained the AI tool, the researchers fed SpectroGen spectra from a mineral in the dataset that was not included in the training process. They asked the tool to generate a spectra in a different modality, based on this “new” spectra. The AI-generated spectra, they found, was a close match to the mineral’s real spectra, which was originally recorded by a physical instrument. The researchers carried out similar tests with a number of other minerals and found that the AI tool quickly generated spectra, with 99 percent correlation.

“We can feed spectral data into the network and can get another totally different kind of spectral data, with very high accuracy, in less than a minute,” Zhu says.

The team says that SpectroGen can generate spectra for any type of mineral. In a manufacturing setting, for instance, mineral-based materials that are used to make semiconductors and battery technologies could first be quickly scanned by an infrared laser. The spectra from this infrared scanning could be fed into SpectroGen, which would then generate a spectra in X-ray, which operators or a multiagent AI platform can check to assess the material’s quality.

“I think of it as having an agent or co-pilot, supporting researchers, technicians, pipelines and industry,” Tadesse says. “We plan to customize this for different industries’ needs.”

The team is exploring ways to adapt the AI tool for disease diagnostics, and for agricultural monitoring through an upcoming project funded by Google. Tadesse is also advancing the technology to the field through a new startup and envisions making SpectroGen available for a wide range of sectors, from pharmaceuticals to semiconductors to defense.

For more information: Matter

Image: The circle with the chip symbolizes SpectroGen, with the connecting threads depicting the process of generating a material’s spectrum.