Decades of advancements have resulted in the miniaturization of circuits on a computer chip, allowing engineers to pack billions of tiny components onto a single silicon wafer the size of a thumbnail. However, silicon chips are now approaching their physical limitations in terms of size while maintaining reliable performance. The solution lies in two-dimensional (2D) materials, which consist of materials that are only a single layer of atoms thick and can be scaled down even further, offering superior electronic properties.
Until now, the challenge with 2D materials like graphene has been the limited complexity of chips that could be constructed with them, as well as the difficulty in connecting them to traditional processors. In a recent study published in the prestigious journal Nature, Chunsen Liu and his team at Fudan University in Shanghai have successfully overcome these obstacles. They have managed to combine atomically thin 2D memory cells directly onto a standard silicon chip, creating the world’s first two-dimensional silicon-based hybrid architecture chip.
So how did they achieve this remarkable feat? The scientists developed a unique technique, known as ATOM2CHIP, to grow a memory material only a few atoms thick directly onto a standard silicon chip. This innovative approach resolved the significant engineering challenge of effectively connecting a super-thin 2D layer and establishing communication with the thicker silicon circuits beneath. Additionally, they devised a special type of packaging to safeguard delicate 2D materials from stress, heat, and static electricity.
To demonstrate the chip’s functionality, the team conducted comprehensive tests on the entire chip, clocking it running at five megahertz (MHz). They also utilized a test platform called checkerboard programming to verify the reliability of every aspect of the memory system.
A new era of computing?
The chip created by the researchers is not merely a laboratory prototype but a functional chip capable of operating in real-world scenarios. It consumes minimal power, outperforms older silicon memory, and can handle the complex operations required by modern systems and devices. The researchers emphasized in their paper that “these system-level results signify a significant milestone in extending the superiority of 2D electronics to practical applications.”
This breakthrough could usher in a new era of ultra-fast memory for smaller, faster, and more dependable chips. This advancement has the potential to drive innovations in advanced AI computing and fundamentally alter how computers store and process information.
Crafted by our writer Paul Arnold, reviewed by Gaby Clark, and fact-checked by Robert Egan—this article is a product of meticulous human effort. We depend on readers like you to sustain independent science journalism.
If this content resonates with you,
please contemplate making a donation (especially monthly).
As a token of appreciation, you’ll receive an ad-free account.
Further Information:
Chunsen Liu et al, A full-featured 2D flash chip enabled by system integration, Nature (2025). DOI: 10.1038/s41586-025-09621-8
© 2025 Science X Network
