add arrow-down arrow-left arrow-right arrow-up authorcheckmark clipboard combo comment delete discord dots drag-handle dropdown-arrow errorfacebook history inbox instagram issuelink lock markup-bbcode markup-html markup-pcpp markup-cyclingbuilder markup-plain-text markup-reddit menu pin radio-button save search settings share star-empty star-full star-half switch successtag twitch twitter user warningwattage weight youtube

Ultra Small Transistors

Rector-Stingray

1 month ago

After current lithography gets down to 1nm, will there be super small transistors like 346 picometers (one trillionth of a meter) and femtometers (one quadrillionth of a meter)?

Comments

  • 1 month ago
  • 2 points

I doubt we will ever get that small. Silicon atoms are about 0.2 nm wide. The chances of a CPU that uses transistors under 1nm in size would be pretty slim, and even if we get there, it will be far in the future. Femtometers are virtually impossible. (correct me if I'm wrong on any of this)

  • 1 month ago
  • 1 point

While new materials may help I think Quantum Computing will be the next long term development. And maybe there will be a push to be able to take advantage of more cores, via parallel computing.

  • 1 month ago
  • 1 point

Once you shrink the Transistor too much Quantum Tunneling will become a problem. You cannot beat the rules of Physics. Chip designers will either have to find a new material and a new way of making a transistor or go for Quantum Computing.

  • 1 month ago
  • 1 point

No. You might go 3d (flash might already have a third digit on how far the transistors are stacked), but 1nm is going to be really, really, hard and won't be a result of anything resembling normal Moore's law.

Also going 3d makes cooling an even bigger issue (why flash goes there and logic doesn't), so don't expect a whole lot on that front.

Perhaps something other than silicon will be the next big thing. But anything made out of atoms will have the same size limitations (although it might be possible that 3d [many layers of] carbon nanotubes @ 28nm will go a lot faster than silicon @ single digit nm).

I'm not even sure that GaAs is still "the tech of the future" (and always will be...).

Sort

add arrow-down arrow-left arrow-right arrow-up authorcheckmark clipboard combo comment delete discord dots drag-handle dropdown-arrow errorfacebook history inbox instagram issuelink lock markup-bbcode markup-html markup-pcpp markup-cyclingbuilder markup-plain-text markup-reddit menu pin radio-button save search settings share star-empty star-full star-half switch successtag twitch twitter user warningwattage weight youtube