Advances in machine learning and computational modeling are transforming materials design by allowing researchers to predict and discover new materials more efficiently than traditional methods. At the forefront of this innovation is Professor Toshiaki Taniike of the Japan Advanced Institute of Science and Technology (JAIST), who leads the Laboratory on Materials Informatics. His team integrates high-throughput experiments, data science, and simulations to accelerate the development of materials like catalysts, polymers, nanocomposites, and nanomaterials such as MOFs and graphene. Driven by a lifelong passion for science, Taniike is focused on solving real-world problems through smarter, data-driven approaches to materials discovery.

“I studied chemical engineering to learn the basics of process design. But I realized I needed a deeper understanding of how chemical reactions actually work. That led me to quantum physics and simulations. However, simulations cannot fully capture the complexity of real reactions, so I returned to experiments, focusing on catalysis, because catalysts drive about 80% of chemical processes,” says Prof. Taniike, sharing how he voyaged through this field. Over time, he learned that discovery often comes through trial and error, so now, he combines high-throughput experiments with data science and machine learning to make this process faster and smarter.

Traditionally, discovering new materials or chemical reactions was mainly pursued as a trial-and-error experiment. Researchers would use what they already knew to take a calculated guess about which features or descriptors matter most when designing a new material. For example, they might think that surface area or crystal structure will affect catalyst performance and then test that idea. This worked well for improving things we already understand, but its utility is limited for discovering truly new reactions or materials because you cannot guess what you do not know.

What Prof. Taniike’s lab does differently is to combine high-throughput experimentation with machine learning techniques like automatic feature engineering. Instead of relying on human intuition to choose descriptors, they let the machine automatically generate and test thousands or even millions of possible descriptors to find the ones that really matter. They then run experiments in parallel to test these ideas, which makes the process 10 to 1000 times faster than doing it manually. Moving beyond the traditional trial-and-error approach, this lets them discover reactions or materials that were impossible to find before.

One such example is the oxidative coupling of methane (OCM), which is sometimes called a “dream reaction.” It aims to convert methane — the most abundant hydrocarbon feedstock — directly into ethylene, which is extremely valuable for producing plastics and chemicals. This is very challenging because methane is such a stable molecule. Usually, when you try to activate it, it just burns completely to CO2. The idea behind OCM is to carefully control this process so that instead of complete combustion, you stop the reaction at the ethylene stage. This could help reduce CO2 emissions from the chemical sector. Published in ACS Catalysis in 2020, Prof. Taniike’s study presents a high-throughput system for the OCM, generating a large, consistent dataset that enables automated performance evaluation, insightful data visualization, and accurate C₂ yield prediction through nonlinear machine learning.

Another study published in Communications Chemistry presents a combination of automatic feature engineering with the above high-throughput system, to streamline the discovery of high-performing catalysts for the OCM through automated descriptor design.

The team is also moving toward discovering entirely new kinds of reactions that could someday transform fields like energy, carbon recycling, or sustainable chemical manufacturing.

For more information: Japan Advanced Institute of Science and Technology

Image: Professor TANIIKE Toshiaki from JAIST.

Pure metals like aluminum and titanium often lack the ideal combination of properties needed for advanced applications, prompting researchers to develop custom alloys by blending various elements. To accelerate this traditionally slow and complex process, scientists at Lawrence Livermore National Laboratory, in collaboration with Cornell University, are creating the APEX platform—an automated system for 3D printing, processing, and analyzing alloy samples—to dramatically reduce alloy development time from years to months.

Funded by LLNL’s Laboratory Directed Research and Development program, this platform will combine industry-matured robotics and automation technologies with cutting-edge machine learning capabilities, turning a traditionally tedious and repetitive research process into a fully automated cycle.

“Our end goal is to make APEX the first self-driving laboratory for alloy discovery at LLNL, capable of working around the clock to collect experimental data and autonomously design, build and test novel alloys,” said Mason Sage, APEX principal investigator and LLNL robotics engineer.

Researching new alloys is complex for several reasons, but primarily because the design space expands exponentially with each added processing or compositional variable.

“Imagine you’re baking cookies and want to create the perfect recipe, so you decide to experiment with three ingredients: sugar, butter and flour,” Sage said. “If you test each ingredient with 10 different amounts, you’d need to bake 1,000 cookies to try every possible combination (10 sugar variations × 10 butter variations × 10 flour variations). Now, imagine you have 20 or 30 ingredients to test, suddenly, the number of cookies you’d need to bake becomes astronomically high — somewhere around 100 quintillion combinations, or 100 with eighteen zeros following it.

“This is the challenge before us. In theory, there hasn’t been enough time since the universe began to test every possible combination — even if we conducted an experiment every second.”

To navigate this enormous and complex parameter space, APEX leverages robotics and machine learning to accelerate sample fabrication, preparation and inspection by intelligently exploring different combinations.

“Our goal is to use machine learning to analyze the information we have collected from all our processes, learn from past experiments and design new experiments,” Sage said.

After selecting an alloy to explore, the APEX platform builds the metal sample layer-by-layer using an additive manufacturing method called directed-energy deposition, a type of 3D printing that uses a high-power laser to melt and fuse metal powders onto a substrate.

Then, to prepare a printed sample to be microstructurally and mechanically characterized, its surface must undergo grinding and polishing. Traditionally, these steps can take two to three days to complete by hand, but APEX is expected to simultaneously process multiple samples at a time, producing dozens of samples a day.

Once a sample is prepped, APEX can evaluate its properties through a series of characterization tests, including examining its surface under a microscope, indenting it with a diamond to measure hardness and crushing it to assess its response to compression.

Finding ways to automate these labor-intensive steps has been a central challenge for the APEX team, one that is critical to the project’s success.

“Trying to not only determine what is really important to our research process but also how to turn that into something you can control a robot to do has been an interesting undertaking,” said LLNL research scientist Michael Juhasz.

For example, when a researcher grinds a sample, they do so by feel, tuning into the vibrational patterns of the machine and sample to ensure they are applying the right amount of pressure. This left the team wondering: How do you relay this highly developed skill to a robotic system? One creative solution the team came up with was to strap a vibration sensor to the machine to capture these “feelings” numerically.

“The sensor produces data in real time, allowing us to put numbers and parameters to the different vibrations, which can guide APEX when it’s grinding a sample,” said LLNL research scientist Alex Baker. “This project is really trying to understand where the automation complexity meets the material science complexity and then translating between the two.”

From sample fabrication to grinding and polishing to characterization, APEX works to collect data at every phase, generating a full history for each sample that passes through the platform. In the future, APEX aims to integrate this data with the Materials Acceleration Platform (MAP), combining physics-based models with machine learning to help design and improve the next round of APEX-produced samples.

The MAP is a framework for integrating computer models that uses algorithms to design optimal materials under different constraints (limitations related to how material properties interact and impact one another) in an effort to find an ideal composition that embodies all desired properties. This is particularly important because optimizing materials often involves trade-offs, where improving one property can sometimes lead to compromises in another. MAP helps strike a balance to achieve the best possible combination of properties.

While one model might excel at predicting a material’s hardness, another might be better suited for predicting its ductility.

“MAP connects all of these separate, siloed models and weaves them together to help us predict new materials that we haven’t tested yet,” said LLNL computational scientist Brandon Bocklund. “The vision of the project is to really close the loop of materials development and acceleration — letting the system learn and update the models autonomously.”

Currently in the early stages of its development, APEX is being trained to work with stainless steel alloys as the research team proves its feasibility.

“We chose stainless steel because it is very a well characterized alloy,” Juhasz said. “This makes it the ideal control, allowing us to ground what APEX is doing and perfect the system before diving deeper into our materials development research.”

The plan is to make APEX as extensible as possible by its anticipated completion in 2027, allowing researchers to adapt the platform to accommodate a variety of unique scientific challenges and incorporate new characterization or diagnostic techniques as needed.

This one-of-a-kind project exemplifies LLNL’s multidisciplinary approach to pushing the boundaries of scientific discovery, demonstrating that when experts from different fields come together, no problem is too “unsolvable.”

 “It’s been great seeing the synergies between deep scientific knowledge of the materials experts meshing with the technical expertise of the engineers,” Sage said.

Baker echoed this sentiment, “There have been many instances where we each could see the subject matter expertise the other side of the project was bringing.”

For more information: Lawrence Livermore National Laboratory

Microsoft’s newly unveiled prototype chip, Majorana 1, has emerged as a strong contender in the race to build practical quantum computers. Unlike traditional computers, quantum machines promise to tackle complex problems—such as simulating vast molecular interactions for drug discovery—at unprecedented speeds. What sets Microsoft’s approach apart is its use of a rare state of matter, making it one of the most unconventional and potentially more stable methods among current quantum technologies, which often rely on manipulating atoms with lasers or magnetic fields.

The chip is the most advanced attempt yet to harness the properties of a strange type of particle that was first predicted to exist in the 1930s by Italian theoretical physicist Ettore Majorana, for whom Microsoft’s chip is named. Majorana’s particle, called a Majorana fermion, remained entirely theoretical for more than seventy years. That is, until researchers (many of them supported by the U.S. National Science Foundation) realized that thin-layered arrangements of certain solid materials might tease out the behavior of the elusive particles and perhaps even put them to use. One of those researchers was theoretical physicist Chetan Nayak.  

In 2000, Nayak was at UCLA, where he received an NSF Faculty Early Career Development (NSF CAREER) grant to explore the physical attributes and theoretical potential of systems that might be made with newly discovered materials now known as topological materials (more on those later). Nayak joined Microsoft in 2005, where he now leads the company’s effort to develop a functional quantum computer by physically implementing some of his NSF-funded theories.  

“NSF was really important in supporting my career in its early stages, without a doubt,” says Nayak. “NSF was there at the start.” 

Microsoft is one of many companies racing to produce a useful quantum computer. A powerful quantum computer could potentially help scientists find new superconducting materials, new medicines and other valuable substances hidden within the practically infinite variety of nature.  

“Genuinely transformative new technologies require entirely new concepts,” says Daryl Hess, program director in the NSF Division of Materials Research. Hess oversaw Nayak’s NSF CAREER grant from 2000 to 2006. “We can’t go down that road without bold new ideas thoughtfully woven together through theory and experiment.”

“We’re basically doing stuff at the one-qubit level,” says Nayak of the currently modest ability of the Majorana 1 to potentially control qubits, the basic unit of information in a quantum computer.  

“That’s pretty basic, but it’s a whole new type of qubit. It’s a whole new way of controlling qubits and doing computations.” 

Quantum mechanics was a young field when Ettore Majorana proposed the existence of a new particle in a 1937 paper. It details his mathematical exploration of equations created by British physicist Paul Dirac, which describe the properties of subatomic particles. From Majorana’s mathematical manipulations, out popped a strange theoretical particle that could be coursing through your body and you’d never know it: the now-eponymous Majorana fermion. 

A fermion is a category of elementary particle (the smallest, most basic particles of matter) that includes electrons, protons and neutrons. Unlike those more familiar fermions, Majorana’s equations predicted a particle that would not interact much with anything. That evasiveness would make Majorana fermions largely invisible to particle colliders and other instruments scientists use to detect and understand elementary particles.  

And so, Majorana’s particle remained entirely theoretical for over 70 years. It wasn’t until the 2010s that physicists came up with new experiments that might reveal the particles by using custom-made nanowires made of a newly found type of matter.

When cooled to near absolute zero and with certain voltages applied, researchers hypothesized that Majorana fermions might be detected at the tips of the nanowires, not as individual particles, but in the collective behavior of many electrons flowing on the surface of the wire.

Scientists call such collective particle behavior a “quasiparticle,” and it’s akin to the wave created by thousands of fans at a stadium standing and sitting in coordination. The key to evoking that stadium-like “wave” of quasiparticle behavior was in the stuff the nanowires were made of: a topological material.  

To the naked eye, a topological material might look indistinguishable from, say, a ceramic tile. But topological materials are very different from your bathroom tiles, and in a peculiar way.

The outer surface of a topological material has properties unlike its interior. And yet, the entire material (both surface and interior) is made of a single, uniform substance. Topological materials get their name from topology, a branch of mathematics that describes properties of shapes that do not change even when deformed in certain ways.  

Scientists have discovered a number of topological materials with seemingly contradictory, yet measurably different and persistent, surface properties — that distinction makes topological materials unique compared to all other known states of matter.

One type of topological material is called a topological insulator because electrical current flows only on its surface, while its interior remains electrically insulating. And, that surface-level flow of electrons is unimpeded by cutting, scratching or otherwise changing the material. Other topological materials allow the unrestricted flow of light or even sound along their surface. 

Microsoft’s Majorana 1 chip — a complex device that uses nanowires made of topological materials and operates at near absolute zero — is based on the idea that Majorana fermions can be used to create qubits. While the state, or value, of a regular computer bit can be either zero or one, a qubit can be in multiple states simultaneously. That uniquely quantum phenomenon is called superposition, and it’s key to how quantum computers can theoretically solve problems far too complex for even the fastest supercomputers in use today.  

While there are several different ways to create qubits, their superposition is generally fragile and easily spoiled by slight environmental changes, like temperature or light. Creating qubits using materials with topologically persistent properties is one way to potentially reduce that fragility.  

Topological computers have greater stability, explains Nayak. “Certain features are really robust and stable.”

Whether using topological materials or other techniques, it will take considerably more research and development before quantum computers are capable of overtaking classical computers.

“Quantum computers would help us understand, predict and discover yet more new states of matter and related phenomena,” says Hess. “Those currently undiscovered states may provide the foundations of as-yet unimagined quantum-based technologies.” 

From fundamental scientific discovery to application, realizing any transformative technology requires time, investment and sustained effort.

Some experts are skeptical about whether the Majorana 1 actually demonstrates functional qubits made from the elusive fermions first predicted by Ettore Majorana nearly 90 years ago. Nonetheless, Nayak and team at Microsoft are continuing to push ahead to validate and scale up the technology with the goal of eventually creating a practically useful quantum computer that is fast, controllable and stable.

“The process of discovery is long and complex, and it took years of hard work from many just to get to this point,” says Alex Klironomos, senior advisor in the NSF Division of Materials Research. “If you want to enable major technological advances, you have to take calculated scientific risks.”  

For more information: Nature

Image: A close-up of some of the intricate features on Microsoft’s Majorana 1 chip.(Credit: John Brecher for Microsoft)