4 ms·

Researchers discover ferroelectricity at the atomic scale

89 points by dnetesn 7y ago

guerrilla 7y ago
I was wonderimg if this was expected and the answer is yes. It was predicted 80 years ago. Here's another article with a simple explanation of it and its dual phenomenon: https://phys.org/news/2018-10-validate-year-old-ferroelectric-theory.html https://phys.org/news/2018-10-validate-year-old-ferroelectri...
- jarfil 7y ago
  If I get that right, a 1nm layer should behave more like an ideal material, rather that the average of several multi-nm hysterons.
  - SiempreViernes 7y ago
    It’s just 1nm in one direction, so you merely force the hysterons to lie along the plane. They claim ferroelectricity, not a new form of ferroelectricity, presumably this means the old theory applies.
peter_retief 7y ago
Is this not the problem graphene was going to solve?
- aspenmayer 7y ago
  Graphene, always the bridesmaid and never the bride.
yummypaint 7y ago
Pardon my ignorance on this topic. Anyone have a sense of what the response time is like? Is the idea to use the hysteresis for storage? How do cycle limitations compare with other storage tech?
wallacoloo 7y ago
Edit: oops, I confused ferroelectricity with ferromagnetism. This comment is about the latter... not strictly related to the article, although there’s at least some similarity between the effects. I’ve been wondering if the ferromagnetic effect could be used to make mid-scale integrated logic devices at home. CMOS devices (and really just anything which depends on PN junctions) are notoriously difficult to manufacture outside of a controlled setting and often require dangerous chemicals. On the other hand, the ferromagnetic effect works on a macro scale (read: hobbyist-friendly) with fewer process steps and friendlier materials for the most basic implementations. You can make memory devices out of it from just two materials — a toroid-shaped ferromagnetic “core” (e.g. nickel) and a wire running through the core (which can be used to polarize the core, or to read its polarization). That’s core memory, used extensively after the era of vacuum tubes but before the era of flash or silicon-based memories. To my knowledge it was never used much for logic, but the row + column select of a core memory circuit perform an AND operation, and if you search around you can find at least a few papers proposing combinatorial logic based on those ferromagnetic cores. It’s two materials (a ferromagnetic material and a conductor), it’s something that was actually used for a decade in industry (core memory), and you can get to 1 MHz even with macro-scale devices. Ferromagnetic devices seem like a hobbyist’s dream! The biggest missing link seems to be a way to automate their manufacturing, since threading wire through a thousand cores to make a multiplier sounds too tedious for me.
phendrenad2 7y ago
So maybe FRAM will finally have it's day.

89 points by dnetesn 7y ago

I was wonderimg if this was expected and the answer is yes. It was predicted 80 years ago. Here's another article with a simple explanation of it and its dual phenomenon: https://phys.org/news/2018-10-validate-year-old-ferroelectric-theory.html https://phys.org/news/2018-10-validate-year-old-ferroelectri...

jarfil 7y ago
If I get that right, a 1nm layer should behave more like an ideal material, rather that the average of several multi-nm hysterons.
- SiempreViernes 7y ago
  It’s just 1nm in one direction, so you merely force the hysterons to lie along the plane. They claim ferroelectricity, not a new form of ferroelectricity, presumably this means the old theory applies.

peter_retief 7y ago

Is this not the problem graphene was going to solve?

aspenmayer 7y ago
Graphene, always the bridesmaid and never the bride.

yummypaint 7y ago

Pardon my ignorance on this topic. Anyone have a sense of what the response time is like? Is the idea to use the hysteresis for storage? How do cycle limitations compare with other storage tech?

wallacoloo 7y ago

Edit: oops, I confused ferroelectricity with ferromagnetism. This comment is about the latter... not strictly related to the article, although there’s at least some similarity between the effects. I’ve been wondering if the ferromagnetic effect could be used to make mid-scale integrated logic devices at home. CMOS devices (and really just anything which depends on PN junctions) are notoriously difficult to manufacture outside of a controlled setting and often require dangerous chemicals. On the other hand, the ferromagnetic effect works on a macro scale (read: hobbyist-friendly) with fewer process steps and friendlier materials for the most basic implementations. You can make memory devices out of it from just two materials — a toroid-shaped ferromagnetic “core” (e.g. nickel) and a wire running through the core (which can be used to polarize the core, or to read its polarization). That’s core memory, used extensively after the era of vacuum tubes but before the era of flash or silicon-based memories. To my knowledge it was never used much for logic, but the row + column select of a core memory circuit perform an AND operation, and if you search around you can find at least a few papers proposing combinatorial logic based on those ferromagnetic cores. It’s two materials (a ferromagnetic material and a conductor), it’s something that was actually used for a decade in industry (core memory), and you can get to 1 MHz even with macro-scale devices. Ferromagnetic devices seem like a hobbyist’s dream! The biggest missing link seems to be a way to automate their manufacturing, since threading wire through a thousand cores to make a multiplier sounds too tedious for me.

phendrenad2 7y ago

So maybe FRAM will finally have it's day.