Not every computer can make use of a disk drive when it needs to store persistent data. Embedded systems especially have pushed the development of a series of erasable programmable read-only memories (EPROMs) because of their need for speed and reliability. But erasing memory and writing it over again, whether it’s an EPROM, an EEPROM, an FPGA, or some other type of configurable solid-state memory is just scratching the surface of what it might be possible to get integrated circuits and their transistors to do. This team has created a transistor that itself is programmable.
Rather than doping the semiconductor material with impurities to create the electrical characteristics needed for the transistor, the team from TU Wien in Vienna has developed a way to “electrostatically dope” the semiconductor, using electric fields instead of physical impurities to achieve the performance needed in the material. A second gate, called the program gate, can be used to reconfigure the electric fields within the transistor, changing its properties on the fly. This still requires some electrical control, though, so the team doesn’t expect their new invention to outright replace all transistors in the future, and they also note that it’s unlikely that these could be made as small as existing transistors due to the extra complexity.
While the article from IEEE lists some potential applications for this technology in the broad sense, we’d like to see what these transistors are actually capable of doing on a more specific level. It seems like these types of circuits could improve efficiency, as fewer transistors might be needed for a wider variety of tasks, and that there are certainly some enhanced security features these could provide as well. For a refresher on the operation of an everyday transistor, though, take a look at this guide to the field-effect transistor.
Fixing security issues.
… or introducing new ones.
… by introducing serious threats to reliability.
Last time I looked,most circuits get pretty unhappy if you suddenly change the type or pinout of a given transistor anyway… your voltage follower is now releasing the magic smoke…
just a few new failure points in VLSI design in addition to the existing thousands that Cadence will have to cover with a few more design rules and maybe two or three more tricks. ;P
cool applications for analogue computers, neural network ICs, adjusting a transistors activation characteristics dynamically as a model is trained.
Does that mean if you leave it unpowered long enough the charge dissipates, the “transistors” forget what they are and it all becomes an inert lump of silicon no longer capable of functioning even when power is re-applied?
If so I don’t think I like that.
I assume it’s much like nand flash that stores data with charge in the gate layer, it’ll retain configuration for some time before electron tunneling does its thing. For flash retention time can be ~10 years without power at room temp for slc cells. Otherwise having volatile configuration may be desirable for specific security applications.
New DRM tech?
I would assume that the functionality would be similar to SSD’s. If at room temperature (powered off) after about 10 years they loose enough data that the parity is no longer able to correct. If you program them while hot, the data lasts longer and if you store them in cold (powered off) the data lasts longer.
Just like regular fpga’s? They just use bitstream stored in external flash for configuration at startup.
maybe they just mix a few memristors into the mix so it will never forget. ;)
As someone who enjoys playing with old electronic devices obtained from hamfests, flea markets, thrift shops, garage sales, etc… 10 years doesn’t sound like that long of a time.
Great, more uncertainty in hw.
Sounds like an EMP will hit these with 10000% destruction (I mean I am 100% sure it will destroy them 100%).
Pretty sure an EMP won’t be kind to any semiconductors and it’s unlikely these parts or any others would be used in anything intended to survive without having had some heavy duty testing and qualification to prove they can be protected so it’s a bit academic really.
This will make excellent noise generators
If these are field programmable then I can imagine some really interesting applications for calibration of instrumentation, audio, RF and I’m sure many more fields but like the other commenters, I can also see problems with obsolescence when manufacturers stop support or just plain refuse to release data
Is it just me or is that picture with the three guys in hats is hysterically funny!
I could be wrong but it certainly sounds like they just obsoleted current FPGA technology. If it’s what I suspect then this will definitely be used to make neural network chips that run much faster and using a LOT less power. This is a Nobel-Prize-level development.
Do not confuse a nice positive news paper article with an easy way forward to industrialization at scale. But maybe now the time is right for this to take off – keeping my fingers crossed for the group.
But it is right, that FPGA topologies can benefit from RFETs in a few ways: https://www.itiv.kit.edu/english/7162_8101.php
Really cool to see the branches of my little niche science work on HAD over a decade later. I am getting old… ;D
There are many more topologies for reconfigurable FETs.
Here is a quite good summary of the recent state of research for those interessted: https://www.sciencedirect.com/science/article/abs/pii/S0038110122001538