Will one-terahertz MRAM memory revolutionize memory access?

A thousand billion changes per second: this is the mind-boggling speed of MRAM. Will this real Arlesian of technologies, which has been in development since the 1980s, overthrow our computers?

Will a paper published by the University of Tokyo change the face of our electronic devices? It’s hard to tell unless you have a time machine and look back ten years from now. But one thing’s for sure: this publication has what it takes to fuel the dreams of those who have been waiting for the ultrafast memory that has been in development for four decades.

Also read: Intel and Samsung have been developing the memory of the future for decades! (December 2018)

MRAM or Magnetic random access memory in a broad sense, it is a memory tower on paper. Judge for yourself: its access speed is one millisecond, it consumes no energy when not in use, and it retains data after the device is turned off. Basically, it combines the strengths of RAM (random access memory) and ROM. It’s beautiful, it’s perfect, it can solve a lot of problems, especially on supercomputers, and… it’s not yet mass industrialized.

Advantages of MRAM:

  • It is non-volatile (like our hard drives or SSDs)
  • Up to 1000 times faster than DRAM or Flash,
  • You don’t need to delete previously saved data before you can write another one
  • It consumes little energy
  • It is theoretically indestructible

As the first letter “M” in its name suggests, MRAM is based on magnetic principles. Kim instead says “magnetic field”, fields that cause problems for researchers and engineers. In particular, the impossibility of randomly placing magnetic cells in space. This forces alignment of the cells (complicated and expensive), which in turn creates a magnetic field that slows down the reading speed. The researchers working on it have many tricks, but they mostly consist of playing with temperatures close to absolute zero… Basically: for the moment, MRAM works well, but it is not yet ready for mass production of our electronic devices. Well, that was before.

Reduce resistance and work at room temperature

The main cell of an already existing MRAM module. © Avalanche Technology

Researchers from the University of Tokyo claim what they achieved in a paper they published in the January 18 issue. Nature under the very sexy heading ” Eight-pole controlled magnetoresistance in an antiferromagnetic tunnel junction », appears to have overcome many of the scientific hurdles in the development of more efficient and easier-to-use MRAM. Digested and made more readable by our colleagues at ScienceDaily (thanks to them: popular science is hard!), this article highlights two major successes of these researchers.

First, they developed a completely new antiferromagnetic component. Unlike ferromagnetic magnets, which generate a field according to the magnetic order prevailing at room temperature, this magnet does not generate this field. You will understand it here, its absence prevents data writing and reading from slowing down. This allows researchers to predict the rate of change of the cell state on the order of terahertz. That is, 10-12 times per second. Yes, it can go very quickly!

Second, this antiferromagnetic component works at room temperature. Materials do not need to be supercooled to discover their quantum properties. Which brings MRAM a little closer to our real world.

Other obstacles including industrialization

Samsung Foundry is also working on mass production of MRAM.
Samsung Foundry is also working on mass production of MRAM.

Do you think your time jump in 2033 is said to show you machines equipped with MRAM? Curb your enthusiasm, even the researchers working on it aren’t there yet. This is the success of a laboratory researcher at a certain point in science. The sum of the developments necessary for the mass production of a very dense module of MRAM operating on this principle is still considerable.

In an interview with ScienceDaily, Professor Satoru Nakatsuji from the University of Tokyo’s Department of Physics explains that the way antiferromagnetic magnets are formed is not insignificant: We grow crystals in vacuum, in incredibly thin layers, using two processes called molecular beam epitaxy and magnetron sputtering. […] This is an extremely difficult procedure, and if we improve it, it will make our lives easier and also allow us to produce more efficient devices. As you can see, this is science straight out of the cutting edge lab right now.

Also read: Memory giant Samsung relies on ultra-fast MRAM to power our connected devices (Feb 2021)

Assuming the process improves, even if researchers eventually achieve sophisticated memory modules from MRAM, the technology faces a major challenge: industrialization. For the rest of the research, getting a technology from the lab to small-scale production for space for defense requires a big leap. But it often takes a giant leap to reach the scale of mass production. The only lever that allows technology to really take off. An obvious parallel are the back-illuminated image sensors. Although many sensors have been produced in laboratories, especially for space sensors for imaging satellites, these ultra-sensitive sensors began to appear in low light until Sony designed the Exmor R industrial manufacturing process. First in compacts, then in smartphones.

IBM has also been working on MRAM for a long time and has already considered various applications for different MRAM variants.  © IBM
IBM has also been working on MRAM for a long time and has already considered various applications for different MRAM variants. © IBM

What if researchers could benefit from the discovery of this Japanese section, a large brick or even a foundation? – perhaps it was asked today. And the whole industry is listening to the results of this research: the Korean giant Samsung is also betting on it. It must be said that the promises of MRAM memory are huge. And to today’s memory problems. Especially in supercomputers, where the speed of processors (CPU, GPU, etc.) is much higher than the speed of memories.



Leave a Reply

Your email address will not be published. Required fields are marked *