Â鶹AV

Researchers selected for JST Strategic Basic Research Programs (PRESTO) in the FY2022

World¡¯s First! A Single Molecule that Records Information at Room Temperature: Ideal for Big Data Applications with 1000¡Á More Recording Density

My research is centered on developing molecules with various functions. I am currently focused on molecules that can serve as storage-class memory, reading and writing data at high speeds without data loss even in absence of a power supply.

The Internet of Things, big data, and cloud computing have ushered in an age of increasing demand for high-capacity storage, and our increasingly computerized society requires that devices be made more energy-efficient. Storage-class memory, which obviates the need for hard disks to hold data, is seen as a solution and development is progressing in this area, but already we are nearing the physical limits¡ªi.e., miniaturization and recording density¡ªof memory that can retain data at room temperature without a power supply.

That makes our success in developing a single molecule that exhibits ferroelectricity even at room temperature¡ªa world¡¯s first¡ªall the more significant. Ferroelectricity is a characteristic in which a material takes on a positive or negative polarity spontaneously in the absence of an electric field, and the polarity can be reversed by the application of an external electric field. By assigning positive polarity to 1 and negative to 0, the molecules can be used to record digital information.

Recording devices incorporating ferroelectrics are already in practical use; they are called FeRAM. However, ferroelectric substances have a weakness: they become susceptible to thermal fluctuation below a certain size and struggle to maintain polarization. Our aim was to devise a method of miniaturizing ferroelectric memory that used metal oxide molecules, whose polarity can be switched with the use of an electric field, to break through those physical limitations.

The metal oxide molecule we used is ¡°Preyssler-type polyoxometalate,¡± an inorganic molecule shaped somewhat like a basket and composed of 30 tungsten, 110 oxygen, and 5 phosphorus atoms. Inside, it has a tube-shaped cavity, which contains a terbium ion (Tb3+); the molecule¡¯s polarity is determined by where inside the cavity the terbium ion becomes stable. By assigning positive polarity to 1 and negative to 0, it is possible to express one bit of information using only a single molecule. Thus, through dedicated R&D work, we succeeded in doing what the impossible: making a single molecule manifest ferroelectricity at or above room temperature.

It could be said that this molecule is part of a new group of substances not bound by the physical limits that conventional logic placed on recording density. If these molecules can be mounted into memory devices, they could outstrip current recording densities by a factor of 1,000 or more and reduce the power consumption of computers by 90%. Thus, our work has the potential to shape our computerized society.

The Preyssler-type polyoxometalate was created in 1970, but has never been explored in-depth. Its basket-like shape was ideal for what we had in mind, so we made the molecules ourselves and set about measuring them. Our research began in 2011 with a team of then-undergraduates. After six years¡¯ work, our success in demonstrating the potential for use in memory applications spurred us to apply for the Japan Science and Technology Agency¡¯s 2017 PRESTO Strategic Basic Research Program. Our application was accepted.

Knowing that any memory device that can¡¯t retain its data at or above room temperature is useless, we raised the temperature ceiling at which recording becomes impossible by increasing the size of the ions in the molecule, thus making it harder for them to move. Based on this, we built an actual memory device in a semiconductor plant, and this development saw our project accepted for the PRESTO program again in 2022. Furthermore, we launched a company in June 2023 to pursue R&D toward implementation of this technology as a storage-class memory. And, in a full-circle twist, the head of the company was one of my students from my earliest days at Â鶹AV a decade or so ago¡ªhow blessed I am to have had such excellent students.

Currently, the smallest memory we have made is one micrometer in size, but we are determined to go ever smaller¡ªa tenth of that size, a hundredth, and smaller again. This is the first time that basic research has progressed to actual application of the technology, so it is exciting.

My underlying desire has always been to reproduce biological functions with molecules, so my ideal for creating the ultimate memory is to reproduce the human brain. The brain, however, is much more energy-efficient than digital memory: whereas digital memory records data as zeroes and ones, but brain is ¡°analog,¡± and uses ambiguous values to record and calculate. For instance, Google¡¯s AI once beat the world champion in Go, but doing so took an enormous amount of power¡ªroughly the same as running a nuclear power generator at full capacity for an hour or two. Humans, though, can do the same thing and expend only the amount of energy it takes to eat a meal. In terms of what current technology can and cannot achieve, the human brain represents the impossible, but I am determined to close that gap.