Article which actually has some numbers.[3] 100μW for 5,000 years. Maybe power a dumb watch. 10 to 15 of those might power a hearing aid.
This might have potential for tracking devices where cost is not a big issue. Trickle-charge a capacitor until there's enough energy for a burst transmission.
Tile tracker tags use a 3V 130mAh CR1632 battery, which lasts for a year... I wish I remembered the math to figure out what trickle wattage that works out to, for comparison.
3*0.130 = 0.39 Wh
Over 5000 years it would amount to about 8.9 nW (nanowatt), of course assuming the chemistry to be able to sustain such time span (which is not).
In that case, one of these diamond batteries would be able to substitute for the CR1632 and give a sealed tracker an effectively infinite lifetime (at least compared to, well, human civilization). I imagine it would be pricey, though.
It also wouldn't be able to do drive a speaker, another key purpose of a Tile. Its actual tracking is primarily passive, so it's hard to locate with just a phone. (You might be able to triangulate from multiple phones... I recall an HN article with somebody doing that a few weeks ago). The main use case is to get close enough to trigger it, then follow your ears. That's thousands of milliwatts.
I suppose it might be possible to add a capacitor that's being continually charged.
Stupid title, but interesting article. They try to use radioactive Carbon-14 to make diamonds to produce heat in a RTG power source.
Obviously it doesn't have near infinite power. The amount of power will be very small, but it may be useful for some applications. Also, the half-life of C14 is 5750 years, so after a few milenia it will need a replacement.
Half-life means that after 5750 years half of it will have decayed, then half of that in another 5750 years, half of what's left in another 5750 years, and so on. That's about 20000 years for still 10% of the original power output, and it won't go totally dead for some million years more.
Yes, but it's the operational energy that he's talking to and it's even shorter than that.
If the power output is expected to be 1000W, usually you design the system to utilize all of it. So as the power starts to degrade, even when you're really far from the half life, it may not be enough to power the systems.
Specifically, the Voyager's RTGs have degraded a lot since Pu-238's half life is much shorter[0], but most of the machine's systems can't run at ~67% power. We're basically down to the very last few experiments running intermittently.
So for useful energy, parent comment is largely right. You've got <5750 yrs of usable work, which is absolutely fantastic.
usually you design the system to utilize all of it
So, don't do that then. Adding capacitance is a problem, dumping heat is not. Seems a small issue (at low power) for a doubling of operational lifespan.
You'd have to design a capacitor, or the equivalent, that could last thousands of years for stuff that you can't service (like a space probe). That is a pretty big ask for many materials in general.
Right out of the gate aluminum electrolytic capacitors are out as getting 10-20 years out of consumer grade ones is pretty good. I have a single monitor I've replaced caps in 3 times now (and they weren't even a known bad batch) in the past 10 years. Electrolytic caps will go bad even faster if they aren't regularly receiving electricity, many have a 2 year shelf life. If you've ever plugged in a vintage electronic (say an Atari/Commodore computer) and you've had smoke-like stuff come out, that's usually the electrolytic capacitors going. If you open up an old device that has had one or more fail, you'll find this brownish-black stuff all over the insides after they go.
Most super-capacitors would die from use long before useful power output, they're usually good for a million cycles or less so even at one cycle a day you're only looking at 2700 years and change.
Polymer electrolytic capacitors are good in use for 10-15 years.
Pretty much every capacitor out there right now has a 5-15 year lifespan.
A data point from the Apollo Guidance Computer restoration: we found all the capacitors worked perfectly after 50 years. This included a bunch of electrolytic and tantalum capacitors.
They weren't in use that entire time though and I'm guessing they were kept somewhat climate controlled too which helps a ton unlike other vintage electronics that get shoved in attics/garages and exposed to wide temperature ranges year after year. First thing I do when I buy a new vintage computer or computer accessory (I'm an 8-bit Atari guy) is change the caps (period) and then remove anything socketed and clean the leads.
You actually worked on restoring them? Do public images of the innards exist anywhere from that? I'd love to see that, the innards of old electronics and machines are just so beautiful in their simplicity. You can look at them and actually understand what is going on.
Edit: found your blog, getting lost in the images.
Climate controlled? No, the AGC was in a junkyard in Houston and then years in a barn. The capacitors worked because they were aerospace quality and thoroughly tested.
Already subscribed and a few posts deep now, really cool stuff! If you see a spike in visitors I shared it on the AtariAge Facebook group as a lot of them will probably also dig your content.
Voltage is part of the problem for sure, which a variable transformer can definitely help keep voltage at a steady operation (you may not even need that since it's a Thermoelectric generator... I'm not familiar enough).
But it's primarily wattage - at some point the overall power dips too low to maintain voltage at that amperage to keep the system alive.
Power is a measure of energy per time, so to get nitpicky about it that's the equivalent of saying "near infinite speed" about something that goes a mile per hour for a few thousand years.
It'll cover some distance, yeah, but it's not what I'd call speedy.
It could be interesting coupled with an ultra-capacitor, if you have something that has sporadic high energy demands, but not always. Think a remote sensor, or pace maker, or such devices that need a short spike of power, but on average the energy requirements are not huge.
I guess you would not get much power given the long half life (as the stuff decaing is what gives you power and this thing doesn not decay a whole lot), but it will give you that power level for quite a while.
There's another option I don't see used often for things like this and that's infinite standby.
Batteries self-discharge. If you have a power source that is greater than the self-discharge rate of a battery, then standby time can exceed the useful life of the device. That's a different problem, but it's also a useful one, and has a much bigger power budget than devices powered solely by the new power source.
I could see wide applications for emergency equipment, for instance.
One of those, "If cars can fly we wouldn't need to build any more roads!" concepts.
I got to learn a bit about betavoltaic[1,2] devices which seem pretty neat, but I'm not going to hold my breath on seeing production versions any time soon :-).
This actually seems like it has potential application if they could be safe enough. As we get the last drops of power efficiency out of Moore's Law, we might end up being able to get a useful amount of compute out of a tiny amount of energy. This could enable things like ambient sensors that never need replaced. Shower a country with a million microSD-sized soil sensors that run for decades and you have a real-time agricultural monitoring system.
Probably an incredibly stupid question but I've always wondered why we can't take small amounts of nuclear waste - say, size of a AA battery, put it into a metal package and use it in laptops / cars / whatever. Short of the risk of leakage, the amount of power that can be generated from it is surely enough to power a laptop or put them into a battery pack for a car, no ?
That's the main risk. A major cause of civilian radiation accidents is radioactive material accidentally mixed with unlabeled scrap [1], and if you put any noticeable amount of the stuff into products meant for untrained civilians, the rate of that happening would skyrocket.
I'm not an expert but I know of two different technologies to convert heat to power: thermocouples and heat engines.
Thermocouples are great because they're small, can fit in enclosed spaces, have no moving parts and last nearly forever. However, they're not a very efficient way to convert heat to energy.
Heat engines + generators can be very efficient, but they're bulky and have many moving parts (you're not going to attach a diesel generator to your cell phone).
"With the majority of the UK’s nuclear power plants set to go offline in the next 10-15 years this presents a huge opportunity to recycle a large amount of material to generate power for so many great uses."
1. Nitrogen-14 has 0.99998807418 of the mass of carbon-14.
2. After 1 year, 0.99987945993 of the original carbon-14 will remain.
3. 1 carat of pure carbon-14 diamond will therefore lose 287.5 picograms of mass in 1 year.
4. 287.5 picograms of mass equates to 25,839.2 joules of energy.
This amounts to 0.818 milliwatts / carat for the first year (and a very slow decline after that). Note that this is the maximum possible output for a pure carbon-14 diamond. Usable energy will be less, and I'm assuming that the diamonds would not start out as pure carbon-14. Practical outputs are probably in the small-number-of-microwatts / carat range.
So 1 gram of perfectly pure carbon-14 would produce theoretically 4mW essentially continuously, or 35Wh per year.
I think that’s pretty neat. There are bursty low energy communication devices which could perhaps work with that power budget?
The problem is, nothing in the world actually needs to stay on for 5,000 years. So the alternative of using a 100x cheaper battery that just lasts 6 months or something is going to be hard to beat.
For more human-scale applications, you're probably looking for something that has a half-life of ~650 years -- that gives you > 90% original power for 100 years.
I wonder, if natural disasters start increasing in impact and frequency, at what point people would start reconsidering economics in order to accelerate environmental protection.
It doesn't seem like they've made much progress since this snopes article came out in 2017[0]. I'm also skeptical that they could use these batteries to practically power a cellphone. So a cellphone requires about an average power of about 200 mw assuming we deplete the battery once every 24 hours. 1 gram of Carbon-14 experiences about 156*10^9 decays per second, which release at most 156 KeV in the form of beta particles, so carbon 14 has a power density of about 4 mW/g. By itself this is decent, but conversion of these high energy electrons to electricity won't be very efficient. 9.8% efficiency has been attained for diamond films under electron irradiation[1], although because the beta particles emitted have a much larger energy than the bandgap of all the semiconductors we know, the efficiency is likely to be less. So assuming 10% conversion efficiency, we need 500 grams of just C-14. This about twice the mass of the iphone 11 pro max. It is probably not very practical to haul a large brick of radioisotopes around to power a phone.
However, in theory it is possible to make a nuclear battery capable of providing this amount of power taking up only a fraction a phone's mass. [3] shows that it is theoretically possible to make nuclear batteries with power densities of 1-50 mW/g. The way they propose to accomplish this is to use the alpha particle emissions from americium to generate power. Alpha particles have on the order of MeV of energy rather than KeV beta particles typically have. Even if the conversion efficiency is low, because there's so much energy available we still attain high power densities. The issue is that because alpha particles have so much energy and are massive they do quite a bit of damage to semiconductors. This means that alphavoltaic batteries do not last very long before they degrade, and it is not expected that alphavoltaic devices can made with current semiconductors that last much longer than month. I've also heard that some alphavoltaic devices have not lasted much longer than an hour. Although, self healing liquid semiconductors appear to be a promising option with one test demonstrating minimal degradation over a 57 day period.
Oh, okay. I was just going on the article, and "absorbs any radiation and makes them safe" is a really bad way to explain that. It makes it sound like pure containment, not that it's also converting it to electricity.
Press release from University of Bristol.[2]
Article which actually has some numbers.[3] 100μW for 5,000 years. Maybe power a dumb watch. 10 to 15 of those might power a hearing aid.
This might have potential for tracking devices where cost is not a big issue. Trickle-charge a capacitor until there's enough energy for a burst transmission.
"Near-infinite power", no.
[1] https://en.wikipedia.org/wiki/Diamond_battery
[2] https://www.bristol.ac.uk/news/2020/january/recycling-nuclea...
[3] https://www.electronicsweekly.com/news/research-news/diamond...