[Image above] The development of non-silicon electronics will help enable the widespread advancement and deployment of high-temperature aerospace and energy technologies. Credit: robertindiana(opens in new tab) / Shutterstock
The upcoming Labor Day weekend in the United States signals the shift from warm-weather vacations to autumn routines. It also marks the launch of the 2026 NFL regular season next Wednesday, with the Seattle Seahawks facing off against the New England Patriots on home turf.
This season is shaping up to be exceptionally compelling due to a massive offseason reshuffling of players and coaches(opens in new tab) and a historic slate of international games(opens in new tab). Yet it may also be one of the safest seasons for players from a health standpoint thanks to a new suspension rule(opens in new tab) and improved protective gear(opens in new tab).
American football is notorious for having a higher severity of injury than rugby, its direct ancestor(opens in new tab), because of (ironically) its use of hard helmets and thick pads for protection. Although this gear reduces the occurrence and frequency of superficial abrasions, it can give football players a “Goliath complex(opens in new tab).” This feeling of invincibility can embolden players to launch into high-speed, less-controlled collisions, which cause their soft brains to bounce around inside their skulls and experience severe trauma, even though the player’s external features are protected.
The NFL started to prioritize(opens in new tab) protection of a player’s visceral as well as somatic structures in the late 2000s. Numerous football players have spoken out about the importance of this shift in safety measures, as the brain cannot be fixed(opens in new tab) like a torn ACL or a broken hand.
This shift in focus from external to internal injuries within American football is an analogy for the development of next-generation electronics. We have published numerous stories about the increasing thermal demands being placed on advanced aerospace (hypersonics(opens in new tab)) and energy (nuclear(opens in new tab)) systems. The development of high-temperature coatings and components have helped keep the overall structure safe, but scientists have been slower to address vulnerabilities regarding the system’s electronic parts (the “brain” behind these next-generation technologies).
Current electronics rely primarily on silicon-based semiconductors, which break down at high temperatures(opens in new tab). Various passive and active thermal management technologies(opens in new tab) are used to keep these electronics from overheating, similar to how protective gear shields football players’ brains from high-stress external forces. But unlike brains, which cannot be swapped with other materials to improve their impact resilience, different types of semiconductors(opens in new tab) can be used to improve the high-temperature resilience of advanced electronics.
Currently, certain electronic components such as transistors(opens in new tab) have been developed that can operate at temperatures up to 800°C(opens in new tab). However, developing nonvolatile memories that operate reliably above 300°C remains elusive.
In March 2026, researchers led by the University of Southern California announced that they successfully developed a memristor capable of operating above 700°C. This development makes it “the best high-temperature memory ever demonstrated,” says Joshua Yang(opens in new tab), the Arthur B. Freeman Chair Professor of the Ming Hsieh Department of Electrical and Computer Engineering, in a press release(opens in new tab).
The new memristor is based on resistive random-access memory(opens in new tab), which means it stores data by creating and removing tiny defects, called oxygen vacancies, inside a thin insulating oxide layer. Hafnium oxide is used as the thin insulating layer in this memristor, and it is sandwiched on either side by tungsten (top) and graphene (bottom) electrodes.
Testing revealed that the memristor can hold data for more than 50 hours at 700°C without needing to be refreshed. It can also survive more than one billion switching cycles at that temperature, and it can run on just 1.5 volts with an operation speed of tens of nanoseconds.
The 700°C reported in this study is by no means the limit for this memristor. It is simply the highest temperature that the team’s testing equipment could reach, and Yang says they believe the memristor can likely operate efficiently at even higher temperatures.
Yang and three of the paper’s co-authors (Qiangfei Xia, Miao Hu, and Ning Ge) already have a startup called TetraMem(opens in new tab) that is commercializing room-temperature memristor chips for artificial intelligence computing. They expect that someday the new high-temperature memristor could allow their chips to be used in more demanding applications, but first they will need to develop and integrate high-temperature logic circuits alongside the memristor.
The paper, published in Science, is “High-temperature memristors enabled by interfacial engineering(opens in new tab)” (DOI: 10.1126/science.aeb9934).
Interested in reading about more materials with performance beyond the usual expectations? Check out the September 2026 Bulletin(opens in new tab), which published online yesterday with the theme “Ceramics with contrary properties.”
