the onion

[vansig]

vansig wrote:
a 15 kJ pulse would then raise 1.2 J on each cell

i’m starting to think that, to keep PV cells from saturating, it will be necessary to bring the onion’s inner radius to ~4.4m

back-of-the-envelope calculation follows.

radius of 50 cm and 12,500 layers leads to 1.2 joule per layer.

Assuming saturation has something to do with capacitance, and breakdown voltage,
1.2 joule = .5 C V^2; and the cell is ~ 4.4 volts, therefore C needs to be 124 millifarad.

to meet this at spacing of .01 mm, (breakdown ~30 V), a single pair of plates needs overlap of 378 x 378 m (in air).

instead, bringing the onion’s radius out to 4.4 m, and using glass as the insulator, only 1/77 th of the energy needs to be absorbed by each plate, and C becomes 1610 uF per pair, yielding overlap of 15.5 x 15.5 m, which is about the surface area of a sphere at 4.4m radius.

Someone please prove me wrong

[Rezwan]

Fascinating posts. Can someone do us a favor and summarize this (and other posts) – extracting the information and explaining to the broader readership what this is about and how it connects with the project?

[Tulse]

which is about the surface area of a sphere at 4.4m radius

Does the onion really need to be a sphere? Creating a multi-layer sphere seems like a fairly daunting engineering task compared to creating flat panels. Is it necessary for the x-rays to hit the surface at a normal angle? If not, I would think that simply six panels, put together as a cube, would be a much simpler design — the panels themselves could be mass-produced far more easily. Presumably one wouldn’t even need to join the panels, as long as there was no space for the x-rays to leak through externally — just ensure that the panels overlap sufficiently that there are no through gaps.

[vansig]

some kind of geodesic should enable flat panels, yes.

intensity drops with the cosine of the angle from the normal, leading to large currents along layers, if they receive much different amounts of light. but if panels are kept to with ~ ±10% it should not cause huge problems.

[vansig]

Rezwan wrote: Fascinating posts. Can someone do us a favor and summarize this (and other posts) – extracting the information and explaining to the broader readership what this is about and how it connects with the project?

The onion was proposed some years ago, as a way to capture otherwise wasted x-ray radiation (Bremsstrahlung radiation) at high efficiency (perhaps 80%). And it really captured my imagination, when i saw drawings of it for the first time, shortly after Eric’s Google tech talks video was published.

Such a system is as yet undeveloped, as no existing photo-voltaic power systems consider x-ray energy as a source, probably because Earth’s atmosphere blocks such light of cosmic origin.

However, in principle, x-ray photo-voltaics work much the same way as regular ones. it’s a matter of choosing the right semi-conductor combinations to catch the photon energy.

Some x-ray photodetectors are using wide band-gap semiconductors, but they saturate under the kind of intense pulses that FF-1 will produce, which would reduce efficiency and could harm the detector. But if the photo-voltaic system consists of thousands of layers of nearly x-ray transparent cells, (eg: 0.008% absorption), then we can limit the saturation in individual cells.

As this is fresh science, the fabrication methods and cost, and ultimate efficiency, of this is unknown presently. Our discussion, here, should try to pin down the constraints.

[vansig]

vansig wrote: But if the photo-voltaic system consists of thousands of layers of nearly x-ray transparent cells, (eg: 0.008% absorption), then we can limit the saturation in individual cells.

if we want to get really fancy, then each layer could also be tuned to receive different wavelengths, from longer to shorter as you move outward.

any idea about the fabrication costs?

[MTd2]

Don’t you think this onion would waste a lot of energy by thermal loses? Why not just heating a liquid?

[vansig]

vansig wrote:
Someone please prove me wrong

Doh! it turns out i did make a silly calculation error, after all. can you tell where it is?

MTd2 wrote: Don’t you think this onion would waste a lot of energy by thermal loses? Why not just heating a liquid?

Just as a point of comparison, in fact the only point to doing it this way is IF it can achieve better efficiency than a heat engine.

“The maximum theoretical efficiency of a heat engine (which no engine ever obtains) is equal to the temperature difference between the hot and cold ends divided by the temperature at the hot end, all expressed in absolute temperature or kelvins.”

this would be, about:
(900 – 300)/900 = 66.6% if operating at 627 °C, cooling to 27 °C; or,
(1273 – 300)/1273=76.4% if operating at 1000 °C, cooling to 27 °C.
both of these are challenging because they exceed the creep limit of several candidate structural materials at the hot side.

so if the 80% claim for the onion carries any weight, it’s well-worth investigating.

[Aeronaut]

Glad to see a potential backup plan. Thanx for the numbers, Vansig.

[vansig]

vansig wrote:
tightening up the interlayer spacing helps a lot, but reduces breakdown voltage. it’s doubtful that spacing could be less than .003mm; yielding 214 m³ volume. still kinda large.

it turns out that dielectric strength is easier to control on thin films than it is on thicker areas, (go figure?)
and i read some claims that thin films of silicon dioxide end up with a dielectric strength near 1000 kV/mm. These end up being used in chip fabrication. If true, this could enable shrinking the cells to a tolerable size, (eg: 25,000 layers, total thickness ~38 cm, 20m x 20m surface area, or a sphere 5.6 m radius).

[Tulse]

vansig wrote: the only point to doing it this way is IF it can achieve better efficiency than a heat engine.

Is there a maximum theoretical efficiency for photo-voltaic solutions? As I understand it, current commercial photo-voltaic solar panels are rather inefficient (~15-20%). Can we reasonably expect to get significantly better out of the onion?

[vansig]

probably not, unless we can force a situation where secondary electrons have exactly the right energy to be captured most-efficiently; eg: if the PV cell lases,
…which might be the case if, for a 215 nm band gap, for example, the semiconductor is, say, 215 nm thick, or forced to occupy a channel 215 nm wide.
perhaps something similar to this photonic crystal defect cavity : http://www.dtic.mil/cgi-bin/GetTRDoc?Location=U2&doc=GetTRDoc.pdf&AD=ADA471262

[Tulse]

vansig wrote: probably not, unless we can force a situation where secondary electrons have exactly the right energy

If not, then would a thermal approach be better than photo-voltaic, especially since the latter is unproven?

[vansig]

Well, x-ray photo-voltaics is not-only unproven, it’s a completely new regime. There is much imagination and science to do, and it wont be ready by next year.

So if you want to use parts that were available to Nikola Tesla, then we’ll start another thread for Heat Engine

[vansig]

Tulse wrote:
As I understand it, current commercial photo-voltaic solar panels are rather inefficient (~15-20%). Can we reasonably expect to get significantly better out of the onion?

that figure is for single-layer monolithic silicon. as of 2009, 45% has been achieved in laboratory, using triple-layer thin films, tuned to different band, R-G-B. (very similar to what the onion is considering, but with visible light, rather than x-ray).

by the way, the principle of up-conversion could make it possible to relax this constraint of whether secondary electrons can be given exactly the right energy. — http://www.rp-photonics.com/upconversion.html

[Author]

Posts