Stories
Slash Boxes
Comments

SoylentNews is people

SoylentNews is powered by your submissions, so send in your scoop. Only 18 submissions in the queue.
posted by Fnord666 on Friday November 29 2019, @04:16AM   Printer-friendly
from the 1,0,1 dept.

Arthur T Knackerbracket has found the following story:

Scientists at the Tokyo Institute of Technology (Tokyo Tech) and the University of Tokyo (UTokyo) have developed a new three-valued memory device inspired by solid lithium-ion batteries. The proposed device, which has an extremely low energy consumption, may be key for the development of more energy-efficient and faster random-access memory (RAM) components, which are ubiquitous in modern computers.

[...] A research team from Tokyo Tech led by Prof. Taro Hitosugi and student Yuki Watanabe recently reached a new milestone in this area. These researchers had previously developed a novel memory device inspired by the design of solid lithium-ion batteries. It consisted of a stack of three solid layers made of lithium, lithium phosphate and gold. This stack is essentially a miniature low-capacity battery that functions as a memory cell; it can be quickly switched between charged and discharged states that represent the two possible values of a bit. However, gold combines with lithium to form a thick alloy layer, which increases the amount of energy required to switch from one state to the other.

In their latest study, the researchers created a similar three-layer memory cell using nickel instead of gold. They expected better results using nickel because it does not easily form alloys with lithium, which would lead to lower energy consumption when switching. The memory device they produced was much better than the previous one; it could actually hold three voltage states instead of two, meaning that it is a three-valued memory device. "This system can be viewed as an extremely low-capacity thin-film lithium battery with three charged states," explains Prof. Hitosugi. This is a very interesting feature with potential advantages for three-valued memory implementations, which may be more area efficient.

More information: Yuki Watanabe et al, Low-Energy-Consumption Three-Valued Memory Device Inspired by Solid-State Batteries, ACS Applied Materials & Interfaces (2019). DOI: 10.1021/acsami.9b15366


Original Submission

This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
(1)
  • (Score: 0) by Anonymous Coward on Friday November 29 2019, @04:20AM (3 children)

    by Anonymous Coward on Friday November 29 2019, @04:20AM (#925922)

    "... developed ... proposed ..."

    So did they actually build a prototype, or is it merely "proposed"?

    • (Score: 2) by takyon on Friday November 29 2019, @05:06AM (2 children)

      by takyon (881) <takyonNO@SPAMsoylentnews.org> on Friday November 29 2019, @05:06AM (#925932) Journal

      https://pubs.acs.org/doi/10.1021/acsami.9b15366 [acs.org]

      It seems that they did make it. There is a single photo in the paper.

      Looks like this would be non-volatile, if nothing else.

      --
      [SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
      • (Score: 2) by c0lo on Friday November 29 2019, @05:56AM (1 child)

        by c0lo (156) Subscriber Badge on Friday November 29 2019, @05:56AM (#925939) Journal

        There is a single photo in the paper.

        Looks like this would be non-volatile, if nothing else.

        Confirmed, the photo is non-volatile.
        I mean, I looked at the photo for about 10 mins at Australian-summer room temperatures and it didn't evaporate. (grin)

        ---

        Otherwise, the memory is volatile for some time/temperature values

        When all these factors are taken together, the operation of VolRAM at room temperature is as follows. On a timescale of hours, the device voltage remains reliably within whichever of the three voltage ranges is set despite some relaxation (Figure S4). On a timescale of days, relaxation proceeds enough that the HVS switches to the IVS by itself. On a timescale of months, perhaps, the IVS switches to zero voltage by itself via LVS.

        --
        https://www.youtube.com/@ProfSteveKeen https://soylentnews.org/~MichaelDavidCrawford
  • (Score: 2) by c0lo on Friday November 29 2019, @05:04AM (1 child)

    by c0lo (156) Subscriber Badge on Friday November 29 2019, @05:04AM (#925930) Journal

    Switch time is not yet measured, but TFA estimates it in the (lousy) 5Mhz range.

    The smaller net charge required for switching should also lead to faster switching. The speed of voltage switching largely depends on the resistance of the solid electrolyte. If we employ an electrolyte with an ionic conductivity of 10–2 S/cm (a typical value for Li10GeP2S12),(31) the speed of voltage switching is estimated to be 200 ns (see Section S8 for the calculations).

    For comparison, SDRAM has read access in the 5ns range [electronics-notes.com] with memory clock freq in the 133-200MHz range [electronics-notes.com]

    --
    https://www.youtube.com/@ProfSteveKeen https://soylentnews.org/~MichaelDavidCrawford
    • (Score: 2) by shortscreen on Friday November 29 2019, @02:00PM

      by shortscreen (2252) on Friday November 29 2019, @02:00PM (#926001) Journal

      5ns would be the minimum time for bus transfers (ie. 200MHz) from an on-chip buffer, not the random access time which is more like 30ns.

(1)