Chinese Researchers Develop Flash Memory Chip That Stores Data With a Single Electron

Every time you send a text message, take a photo or post on social media, billions of electronic components process and store your data. Now, researchers in China have developed a two-dimensional flash memory chip capable of storing information with just one electron.

The device, named “Guiyi”—a Chinese Buddhist term meaning “return to oneness”—can trap individual electrons at room temperature. The breakthrough could reduce energy consumption and improve the speed of data transfer in future memory and artificial intelligence systems.

In a study published July 16 in the journal Science, the researchers described how their single-electron flash memory technology generates a measurable electrical signal while using significantly less energy than conventional memory devices.

How single-electron flash memory works

A flash memory cell typically stores data by moving electrons into or out of a floating gate. The presence or absence of electrical charge represents digital information, such as a 1 or a 0. In Guiyi, researchers reduced the process to the movement of a single electron—the theoretical minimum required to store one bit of data.

Scientists have explored single-electron data storage since the late 1990s. However, earlier devices produced signals that were too weak to read reliably. Detecting the signal was comparable to trying to observe the tiny ripple made by a raindrop falling into a large reservoir.

To solve this problem, the researchers placed a layer of graphene in front of the chip’s floating gate. Graphene consists of a single layer of carbon atoms arranged in a hexagonal lattice. Electrons can travel through the material quickly, with low electrical resistance and minimal energy loss.

The graphene layer helps accelerate electrons before they enter the floating gate, where they remain trapped even after the power is switched off. This gives the device nonvolatile memory capabilities while improving the strength of the electrical signal.

A stronger signal from one electron

According to the researchers, Guiyi produces a signal of approximately 0.5 volts when a single electron is captured. That signal is about 10 times stronger than those produced by previous single-electron memory experiments, making the stored information easier to detect.

“Our demands for storage speed, capacity, energy efficiency and stability are reaching new levels in the AI era,” study co-author Liu Chunsen, an engineer at Fudan University, told China Daily.

“Being able to store a single bit of information by changing the state of a single electron would significantly reduce power consumption and pave the way for greater storage capacity,” Liu said.

The researchers also believe the technology could help data move more efficiently between processors and memory. Faster transfers would reduce delays, improve computing performance and potentially support the expansion of AI applications across multiple industries.

Why AI is increasing demand for better memory

Artificial intelligence workloads and large language models require enormous amounts of data to move between processors and memory. In many systems, the processor operates faster than the memory can supply data, creating a performance bottleneck known as the memory wall.

At the same time, the semiconductor industry is increasing its use of high-bandwidth memory for advanced computing systems. Major manufacturers—including Samsung, SK Hynix and Micron—continue to dominate the NAND flash memory market, while shifting resources toward memory technologies designed for AI and data-center applications.

These changes have contributed to pressure on NAND flash supplies and prices. A memory architecture such as Guiyi, which aims to combine low energy use with rapid data transfer, could eventually help address some of these challenges.

The challenge of scaling up the technology

Although Guiyi is a promising proof of concept, turning a laboratory device into a commercial memory product could take years. The chip must be manufactured reliably, integrated with existing semiconductor processes and produced at a cost that is competitive with conventional flash memory.

Andrew Humphrys, a professor of nanoimaging at the University of Bristol and founder of semiconductor metrology company Infinitesima, told Live Science that scientific advances only transform industries when they can be produced “reliably, repeatedly and economically.”

He pointed to extreme ultraviolet, or EUV, lithography as an example. EUV technology is now used by companies such as Intel and Samsung to manufacture advanced 3- and 5-nanometer chips, but it took ASML more than two decades to move the technology from the laboratory to commercial production, according to the company.

“The companies that will form the next generation of AI chips will not just be the ones with the best technology, but the ones that can manufacture them at scale at a reasonable cost,” Humphrys said.

Even so, the team led by Zhou Peng, a microelectronics professor at Fudan University, has ambitious plans. According to the South China Morning Post, Zhou plans to establish a company and pursue commercialization within three to five years.

If the researchers can overcome the manufacturing and reliability hurdles, the Guiyi single-electron flash memory chip could become an important new approach to reducing energy consumption and easing memory bottlenecks in future AI hardware.