One of the most critical issues that has remained unresolved in the semiconductor world for many years is **extreme temperatures**. For engineers aiming to develop systems capable of operating in harsh environments, from the depths of the Earth's crust to the surface of Venus, the threshold of approximately 200 degrees Celsius was considered a significant barrier. However, new research from the University of Southern California (USC) in the USA suggests that this **limit can now be surpassed**.
Researchers have developed an electronic memory device that can operate far beyond previously achieved thermal limits. Experiments showed that this new-generation component continued to function stably at **700 degrees Celsius**, temperatures akin to lava, without showing any signs of degradation.
Professor Joshua Yang, the senior author of the study, described this development as a "revolution," emphasizing that the developed system is "the best high-temperature memory technology demonstrated to date."
Next-Generation Memristor-Based Design
At the heart of this success lies a nanoscale component called a **memristor (RRAM)**, capable of both data storage and processing. The research team designed the device as a three-layered structure. The design features **tungsten** on top, **hafnium oxide** in the middle, and a single-atom-thick layer of **graphene** at the bottom.
Each of these materials plays a critical role in the system. Tungsten, with its high melting point, provides structural durability, while ceramic-like hafnium oxide acts as a stable insulating layer. Graphene, the most remarkable component of the system, prevents short circuits at high temperatures due to its chemical stability and durability.
According to the data, the developed device operated at 700°C for **over 50 hours** without data loss. Furthermore, it achieved this performance without needing any refresh cycles. The system also demonstrated durability for **over 1 billion switching cycles** and operated with only **1.5 volts**. Researchers noted that they could not reach the device's actual failure point, and the limitations of their test equipment restricted the experiments.
Discovered by Chance
Another striking aspect of the study is that this outcome was not the product of a planned objective. The research team discovered this effect while working on a different graphene-based system. Subsequent advanced microscopy, spectroscopy, and quantum simulations revealed why the system was so durable.
The findings indicate that the absence of chemical bonds between graphene and tungsten prevents metal atoms from moving and disrupting the structure at high temperatures. This ensures the device remains stable even under extreme conditions.
The potential applications of this technology are vast. Space agencies have long been trying to develop hardware capable of operating in environments above **500 degrees Celsius**. Considering that the surface of **Venus** is at approximately these temperatures, the newly developed system could pave the way for planetary probes, geothermal drilling equipment, and computers that can perform in-situ processing in nuclear reactors.
Step Taken Towards Commercial Product
Moreover, **memristor technology** offers significant advantages for artificial intelligence. These components can perform **matrix multiplication operations**, which form the basis of AI algorithms, directly through physical processes. This allows for much lower energy consumption and much higher processing speeds compared to classic chips. According to Yang, over 92% of the computations in modern AI systems consist of such operations.
Furthermore, a startup called **TetraMem** has been established to commercialize the technology. The company aims to introduce memristor-based AI hardware, particularly for operation at room temperature. However, researchers acknowledge that there is still a long way to go before practical applications are realized. Another piece of good news in the commercialization process is that a significant portion of the materials used are already employed in semiconductor manufacturing.
0 Comments: