In an era where artificial intelligence (AI) is becoming increasingly intelligent, domestic researchers have found a shortcut to make this technology faster and more efficient.
Professor Kim Se-young from Pohang University of Science and Technology (POSTECH) and Dr. Kwak Hyun-jung's research team, in collaboration with Oki Gunawan, a researcher at IBM Watson Research Center in the U.S., have identified the operating principle of electrochemical memory devices (ECRAM), which is gaining attention as a core technology for next-generation AI, for the first time in the world. This research was published in the international journal Nature Communications in March.
As AI advances, the volume of data processing is increasing exponentially. However, current computers have separate 'memory' for storing data and 'processors' for performing calculations, which requires significant time and energy for data transfer between the two devices. The concept introduced to solve this issue is 'in-memory computing.'
'In-memory computing' technology allows calculations to be performed within the memory, enabling fast and efficient operations without data movement. ECRAM is one of the key technologies to implement this. ECRAM can store and process information through the movement of ions, allowing it to store continuous values similarly to an analog method. However, its complex structure and high-resistance oxide materials made it difficult to clearly understand the operating principle, which remained a significant obstacle to commercialization.
The research team used tungsten oxide to fabricate this ECRAM in a 'multi-terminal structure' and applied a technology that can observe the movement of electrons internally at various temperatures from cryogenic to room temperature.
As a result, the research team observed for the first time that oxygen defects within the ECRAM form a shallow donor level of about 0.1 eV (electron volts) and create a kind of 'shortcut' that allows electrons to move easily. When ECRAM stores and transmits information, not only does the quantity of electrons increase, but an environment is also formed that allows electrons to move freely. The fact that this mechanism is maintained even at cryogenic temperatures is an important finding that proves the stability and durability of ECRAM.
Professor Kim Se-young noted, 'This research is significant in that it experimentally elucidates the operating principle of ECRAM at various temperatures,' and 'If this technology is commercialized, we can expect AI to run faster on devices such as smartphones, tablets, and laptops, with longer battery life as well.'
References
Nature Communications (2025), DOI: https://doi.org/10.1038/s41467-025-58004-0