It's impossible to read much into articles that are so incredibly buzz word inundated. It sounds like a transistor with ultra low leakage power which is becoming an increasing problem. You'd still have dynamic power consumption.
There's something about running entirely on leakage power, never powering on into the normal region? I can't figure out what they're talking about with scavenging power at all; if it could scavenge why would it only last 'months or years'.
Scavenge is just used as a buzzword. Rather than having a battery, the IoT device could be low enough power to use just piezo capture, EM capture, etc.
It's not much better. That graph looks like nonsense and it doesn't really explain things. The abstract is fine but only explains a small fraction of the article by itself.
"Wearable devices and environmental sensors ideally should consume very little power to avoid the need for batteries that would have to be replaced. Lee and Nathan developed a thin-film transistor (TFT) from In-Ga-Zn-O thin films. To make the material less conductive, the films were fabricated to avoid oxygen vacancies. The TFT operated at ultralow power (less than 1 nW) and at switching voltages of less than 1 V with very high intrinsic gain. The devices work by changing the height of the so-called Schottky barrier formed between the semiconductor gate material and the metal drain contact."
How credible is Futurism.com as a source? Futurism, LLC, a Delaware limited liability company, was formed less than a year and a half ago [1]. In my experience, their content tends to be breathlessly hyperbolic.
[1] State of Delaware Department of State: Division of Corporations
It's rather badly explained, but it seems they've developed the Schottky transistor, using a TFT manufacturing process. Conventionally TFTs are slow and leaky compared to a regular CMOS FET in silicon. It sounds like they've greatly improved leakage, although I can't figure out where they say this in the paper. This makes printable low-power circuits more plausible.
It is a "Schottky-barrier
indium-gallium-zinc-oxide thin-film transistor
operating in the deep subthreshold regime." According to the original article in Science Mag. the currents involved are significantly lower then of the other possible technologies. It's not only that the OFF current is apparently an order of magnitude lower (as a property of the use of that material, IGZO) it's that the transistor does the useful work in the "almost off" state (as they operate it in that regime). The low currents typically mean also "slow" but for the applications intended it could be exactly what's enough, and the advantages huge.
I don't see that they give any exact estimate of some real life application, it can be that the practical uses technologically introduce some limitations that aren't discussed. At least it looks promising.
Could this type of transistor be used in stacked silicon chips? [1] It would seem that such a low power transistor would produce much less heat, solving one of the largest problems with the technology.
"Operation at such low voltages is problematic due to susceptibility to noise, high sensitivity to temperature, and variability due to process drift. In developing the SPOTTM platform, Ambiq Micro has addressed all of these challenges by redesigning every analog and digital circuit in the chip to adapt for the temperature, voltage, and manufacturing variations that are problematic at these voltage levels."
Yeah, they are using the transistors near their highest gain point. It's impressive that they got this to work- If you try this with through-hole transistors, you will definitely be picking up your local AM station.
24 comments
[ 2.7 ms ] story [ 54.7 ms ] threadPaper abstract: http://science.sciencemag.org/content/354/6310/302
[1] State of Delaware Department of State: Division of Corporations
I don't see that they give any exact estimate of some real life application, it can be that the practical uses technologically introduce some limitations that aren't discussed. At least it looks promising.
[1]https://en.wikipedia.org/wiki/Three-dimensional_integrated_c...
http://ambiqmicro.com/low-power-microcontroller
http://ambiqmicro.com/spot-platform
"Operation at such low voltages is problematic due to susceptibility to noise, high sensitivity to temperature, and variability due to process drift. In developing the SPOTTM platform, Ambiq Micro has addressed all of these challenges by redesigning every analog and digital circuit in the chip to adapt for the temperature, voltage, and manufacturing variations that are problematic at these voltage levels."
https://www.fujitsu.com/uk/Images/Ambiq_whitepaper.pdf
Also always fun to look at their patents, but who knows which ideas are really used in the chip:
https://www.google.com/patents/US7009265?dq=ambiq&hl=en&sa=X...
This thesis seems to have a good overview:
http://www.eng.auburn.edu/~agrawvd/THESIS/KIM_S/S.Kim_Honors...