Slashdot Mirror
Solar Cells Integrated In Microchips
cylonlover writes "In a new, more efficient approach to solar powered microelectronics, researchers have produced a microchip which directly integrates photovoltaic cells. While harnessing sunlight to power microelectronics isn't new, conventional set-ups use a separate solar cell and battery. What sets this device apart is that high-efficiency solar cells are placed straight onto the electronics, producing self-sufficient, low-power devices which are highly suitable for industrial serial production and can even operate indoors."
...now $0.50? maybe go roll-to-roll?
So "high-efficiency solar cells are placed straight onto the electronics" and of course it has to "see" light (even indoors)--then how is this gonna be packaged for protection from high electrostatic discharges?
Chip below 1 MilliWatt power, hmm so Intel and AMD are ruled out
---
In the brave new world, learn to want more
Chip provides its own power
What one fool can do, another can. (Ancient Simian Proverb)
Considering the power drain must be "well below 1 milliwatt", what could such a chip be used for?
Since the solar cell piggybacks on the chip, wouldn't the chip be exposed to the elements in order to generate electricity?
All the article states is that it could be manufactured not what it could be used for..
Hmm... provided you could build some capacitance into the die, it would be trivial to manufacture, en-masse, an array of incredibly inexpensive devices that could respond to a light-signal stimulus, much the same way that RFID tags respond to a suitable RF signal.
I can't think of an immediate application, but the key technical difference would be that you can transmit focussed, directional light in a laser, which is a much more difficult and wasteful feat with RF. You could therefore elicit a response from one of these chips at pinpoint accuracy and great range, which is not currently possible with RFID. I'm sure some more innovative people than I could conceive of a novel application.
Meta will eat itself
Using the solar cell for light-sensing is okay, it is the same as using a photodiode. But with some extra circuit, and band-gap reference you can make the sensor more accurate. and if you are very good, you might be able to store the measurement in a flash memory. (of course you can drive other type of sensors. Like ChemFET base sensors, but that's another story)
The field where this can be applied is rather narrow. -> suboptimal solar-cell + suboptimal integrated circuits.
These low power chips (few mW) are usually very small, and area really matters in chip price. So I see a solar cell which has a small chip manufactured right into it.
But I still don't think it is a very marketable idea.
what could such a chip be used for?
The important question to ask is, "what could such a chip be economically used for?"
Unless one is developing something for military or other national security-type purposes, where cost is typically significantly less important than attaining the ultimate in performance (however "performance" is defined in the application), the question typically is, "what's the cheapest way to do X?"
If, as is frequently the case, X is defined as "power the chip," one has an interesting economic quandary: In terms of money/area, conventional solar cells, especially amorphous solar cells, are about the cheapest form of silicon known to man. Using this new technology, though, these solar cells would be replaced by area on an integrated circuit, which is about the most expensive form of silicon known to man. Worse, the power consumed by the electronics must be minimized ("below 1 milliwatt"), so one is pressured into using very fine-lithography silicon, which is the most expensive form of silicon known to man.
The only way I can see that one wins on cost with this technology is if one has electronics that are so low-powered that they can be powered by an amorphous solar cell with an area equal to that of the circuitry itself. If you need a point of reference on the practicality of this requirement, I point you to your average solar-powered calculator, which has a solar cell area of several cm^2, and an active circuit area of probably less than 5 mm^2.
If, however, X is defined as "power the chip with a monolithic structure," perhaps for acceleration, board area, or other system-level requirements, then using external solar cells is prohibited (by the terms of the game), and then this technology begins to look more appealing. Even then, however, I would wonder if it wouldn't be easier to use chip stacking or similar technology to put a solar cell on a chip, since then each die could be processed in a manner optimum for its purpose (solar cell or integrated circuit). And, of course, you're still left with the problem of getting sufficient power from such a small area.
The machines were using the sun for energy!