Materials science

Xinpeng Zhao, 32, invented a glass-based coating that can cool the roof of a building by around 3.5 °C below ambient air temperature, which could reduce air-conditioning energy use by nearly 10% for a mid-rise building.

Today, air conditioning and other cooling systems emit around 7% of global greenhouse gases, a number that is projected to triple by mid-century. Passive cooling materials are an appealing alternative, because they cool themselves without using any energy.

While such cooling materials have been available for years, most are based on polymers and plastics and can’t withstand decades of exposure to the elements. Those materials lose up to 30% of their performance in just a few days on the roof of a building as they turn yellow, says Zhao, and, “after a few months they will totally lose their cooling efficiency.”

Zhao solved the durability problem by working with finely ground glass and aluminum oxide instead of plastics. His new coating can be sprayed or painted on buildings, vehicles, or roads. It’s composed of an aluminum oxide particle of around 500 nanometers, which is perfectly tuned to reflect up to 99% of the sun’s light to prevent heat absorption in the first place.

To take his innovation to market, Zhao founded a startup called Ceracool, and he is working to expand the range of coatings for different applications.

Julia Carpenter, 33, invented a simple, cost-effective way to make a foam-like material out of metal that has a thousand times more surface area than any of its predecessors. The new material could serve as a heat sink and reduce the amount of energy required to cool microprocessors and other electronics. 

Right now, around 40% of the energy used by data centers—such as servers for AI, cloud computing, and crypto currency—goes toward cooling. Similar cooling methods are used in industrial manufacturing and the batteries in electric vehicles. 

For decades, electronics have been cooled by systems that circulate liquid or air through heat sinks built from sheets of solid metal, or by using refrigerants and compressors. Both types of systems are costly to build, maintain, and run. They’re also prone to failure, and their creation emits significant greenhouse gases. 

Metal foams have existed for nearly a century, but until now they were produced using an expensive multistep process that involved making a plastic template and replacing it with metal. As a result, their commercial use has been limited. 

“We changed the process fundamentally,” says Carpenter.

Carpenter’s process starts with a slurry of metal particles infused with bubbles as small as 1 micron—she’s used iron, nickel, and stainless steel, among others—that she essentially pipes into a lattice-like structure with various sized holes. “It’s like making meringue in the kitchen,” she says.

Once air-dried, the material is heated to between 600 °C and 1,700 °C to solidify the structure into something so highly porous that it can float.

To get her technology into the market, Carpenter started Apheros, in her native Switzerland—and industrial clients immediately lined up. 

Brain-computer interfaces have returned some aspects of vision to blind people and provided critical monitoring of conditions like epilepsy, but usually only for a short time before having to be uninstalled.

That’s because brain-computer interfaces tend to rely on silicon chips that degrade in the body or are rejected by the immune system. They are rigid, easily dislodged, and can cause injuries. And powering them requires bulky battery systems or hardwired connections that are impractical for devices that must be implanted inside the skull.

Claudia Cea, 33, circumvented these problems by replacing silicon chips with polymers. She developed the first flexible neural recording device based on an ion-gated organic electrochemical transistor. Such transistors depend on the movement of charged atoms known as ions, instead of pure electrical signals. They existed before Cea built hers, but were not fast enough to record brain signals in real time.

“The reason organic, ion-based transistors are slower than silicon is that they rely on ions to switch on and off—the ions have to migrate from the body into the back of the transistor,” says Cea. “I created ion reservoirs in the transistor itself, so that the distance the ions have to move is much smaller.”

Next Cea designed the data processing, transmission, and power modules to complete the system, all using ion-embedded polymers. One advantage of this approach is that ion-based signals propagate through human tissue. That means her implant communicates wirelessly from the inside of the skull to gold terminals placed along the inner and outer edges of the skull, which also send power back to the device.

Inho Kim, 34, built artificial muscle fibers that are light, flexible, and strong. He envisions them being used in wearable devices to assist people with cerebral palsy. But because his new material produces around six times more power than human muscle fibers, and can lift up to 5,000 times its own weight, it is also one of the best candidates for suits to augment the strength of people such as soldiers, construction workers, or the elderly. The fibers might also be used to build robots that move more like humans or animals. 

Exoskeleton suits already exist, but they are heavy and bulky, and they can cost up to $200,000, putting them out of reach for most people and applications. 

Kim’s solution was to replace actuator motors and rigid frames with artificial muscle fibers inspired by the real thing—made from a flexible substrate and infused with a second material to enhance movement and power. While others have attempted similar approaches, artificial muscles have so far failed to match their natural counterparts in speed, power, weight, and feedback sensing. 

Kim’s new material, dubbed Hercules fibers, is cheap to produce and scalable, because the fibers, on average around 200 microns in diameter, can be bundled by the thousands like real muscle fiber. And thanks to the use of graphene, which conducts electricity, the fibers provide real-time feedback during a contraction.

“My goal is to make reliable, commercialized, wearable robotics,” says Kim—ideally, for those most in need, like babies with cerebral palsy who are learning how to walk.  

The world is losing about 10 million hectares of forests every year to both natural factors and human activities. Spraying seeds from planes or drones is one of the most efficient ways to plant trees in remote areas and help the environment recover. But there’s little that people can do to make sure those seeds actually germinate. Animals, rain, and wind can all prevent seeds from penetrating the soil. 

Danli Luo, 32, a materials scientist and PhD candidate at the University of Washington, looked to nature for inspiration. She found Erodium, a flowering plant whose coil-shaped seeds drill into the soil by winding and unwinding as the humidity changes. 

Based on those mechanics, she led a team that designed “E-seed,” a tiny seed carrier. It has coiled tails made of wood veneer that can help a variety of seeds (measuring up to 10 millimeters in diameter) bury themselves in soil. It can also carry fertilizers or other substances to aid seed germination. 

The key was in “programming” the wood material. Much like software developers who program an application, Luo tinkered with the wood veneer by removing a component called lignin so it’s more pliable; she also simulated how different shapes of coils would act once deployed, and ended up designing a carrier with three curved tails (while the natural Erodium seed has one) to keep the seed upright and boost its chances of germination.

The result is a simple, versatile, naturally sourced, and completely biodegradable solution to deforestation. Her technology has already been licensed by a Fortune 500 company, she says. Now Luo is working to find manufacturers or 3D printing services that can scale up production.

Amil Merchant, 27, built a machine-learning model that unearthed atomic recipes for a vast trove of new materials, some of which could lead to breakthroughs in fields such as supercomputing and renewable energy.

As technical lead at Graph Networks for Materials Exploration, or GNoME, a program run by Google’s AI subsidiary DeepMind, Merchant faced a tricky proposition. While models like those behind today’s AI chatbots are good at making predictions based on information that they’ve seen, they’re less adept at making new discoveries.

Merchant and his colleagues believed materials science was due for a shake-up: Efforts to use machine learning to find new molecular structures had struggled to pinpoint ones that would be unlikely to decay or combust.

With Merchant leading the charge, GNoME set out to find these stable structures. The project team trained a machine-learning model on an open-access database of known materials to learn the patterns that set the stable ones apart and then let the model loose on the entire periodic table. The model identified millions of potentially stable combinations. Next, the team used existing techniques from quantum mechanics to appraise those possibilities and plugged the vetted data back into the model, a process repeated several times with increasing levels of accuracy. In the end, they reported finding 380,000 new stable structures—a nearly tenfold increase from the 40,000 previously known to humanity.

For now, most of these new materials exist only on paper. But the team’s findings are available to outside researchers, and they could eventually help build more powerful solar panels, batteries, or semiconductors.