The space ones are expensive partially because weight is an important consideration. For a terrestrial RTG, you can use Strontium or something else that's much cheaper (but much less weight-efficient).
They are used in space because it is much more difficult to use solar beyond Mars. Modern solar has helped in Jovian missions but RTGs are still preferred (even Curiosity uses a RTG for low solar radiance reasons). And beyond Jupiter good luck having a powered device with anything except for a RTG.
yes of course, what I mean is why the space RTGs are so much more expensive than terrestrial RTGs, where weight is not a consideration. For space, it makes sense to use the exotic plutonium isotope if it saves on weight.
Oh yes. Another thing to consider is the weight of the titanium that encases the RTG just so you don't fry the other electronics. Dealing with radiation in space is pretty difficult and radiation shielding in general is still a pretty complex problem. On Earth we pretty much solve it my mass (more mass == more shielding) but we don't have the luxury with space applications. There's a lot of advanced composites there and layered material. It is a really fascinating subject. There's also people trying to harvest some of this energy into usable electricity. I worked on one of these devices (focusing on betavoltaics) and it isn't going to power your house, but you can power things like a heartbeat signal for your craft (and of course use it to trickle charge batteries on long missions).
I don't work in this space anymore. Some I can't talk about but part of what I can talk about is still a pretty big problem, which is finding layer ordering, materials, thicknesses, etc of the shielding. You have problems like that neutrons are absorbed differently than protons, alpha particles, and beta particles (all those are charged). So you want to use thing like hydrocarbons for neutrons (read plastic) and you probably want to dope it. BUT there's a big problem that the energy level matters a lot. Gadolinium is known as having a good neutron cross section, but that is only for thermal neutrons and hot neutrons (as you'd find in space) don't see gadolinium differently from dense materials like titanium and aluminum (good for charged particles). So the problem is to layer, dope, etc. And to do that while accounting for secondary factors like that you can have materials become hot as exposed to radiation and then you also have to consider physical shielding. The solution space is extremely large and you search it by simulation.
As for getting electricity you can probably imagine that if you have two conductive plates that they will get charge levels across them and that's a capacitor. There are other ways to extract energy though and finding ways to do this is very helpful. But there is a theoretical limit to the energy and don't expect to replace solar panels unless you can capture those particles and use a nuclear process instead of an electromagnetic one.
If you're interested in this start searching for betavoltaics[0]. That uses the E&M process whereas an RTG uses a thermal process. There's nothing stopping you from using both though.
That’s still showing a break even around ~30$/gallon fuel costs in 2020 dollars for pure heating applications. It’s hard to reach those kinds of delivery costs. It’s even worse when you want to use RTG’s for electricity as you add complexity and lower efficiency.
You're missing the autonomous part of the equation. An RTG and light will run independently with occasional checkups for decades. A gas generator will not. So now your costs have to account for a permanent local crew, sending supplies for them (food, etc.) and constructing a building they can live in. That's on top of shipping the fuel itself into the desolate location.
They needed to go to these lighthouses multiple times a year even with RTG’s, presumably to change bulbs etc. Truly remote areas without people don’t need lighthouses. In that context operate generator N for X hours then swap to generator N + 1 and repeat as needed.
The more modern solution for even more remote areas is fuel cells which can last significantly longer between inspections.
yes, but if you want to leave a device buried in the middle of nowhere on the ice sheet for 10 years that draws ~30W, an RTG is much simpler (only have to transport it once, no moving parts, no need for exhaust, etc.).
Delivery and deployment costs to the middle of the ice sheet are not cheap.
If you only need 30W for 10 years then Lithium thionyl chloride batteries are a viable option down to -55C. In continuous operation waste heat will give you significant temperature leeway. That’s going to be expensive and very heavy, but there really isn’t a good option for truly remote applications.
30W is kind of the no mans land of remote power. Batteries are becoming seriously impractical, but nothing is the clear winner.
I mean, outside of polar regions, solar + battery hybrids work well enough year-round. Solar + fuel cell + battery hybrids can work in Polar winter if you can transport enough fuel, but the exhaust requirement is a problem (moving parts + can't let your fuel cell + batteries get buried in the snow). The solar panels also need to be raised every once in a while (or mounted high enough in the first place).
Eventually reversible fuel cells might be a good option (use excess solar in the summer to produce methanol or whatever, then consume it in winter).
Lighthouses typically shine a beam up to 20-30 miles, depends on height obviously. I would imagine the loads would be at least 1kW especially with old fashioned incandescent bulbs?
It's not hard when you're delivering fuel to one location 1000km in desolate wilderness with no road access. Your only option is basically ship borne helicopter, and then you have to make this trip semi-regularly.
Plus. nuclear material are only expensive because there are not enough usage - hence the economy of scale is small. They're arguably cheaper in the Soviet Union (even if they followed market terms).
These where used for lighthouses which aren’t exactly useful if people stay 1000km away from them. Now for unmanned Antarctic observation or something then that’s a possibility, but hardly going to feed economies of scale. Outside the Arctic circle solar panels + batteries win hands down.
20 years ago, UChicago Scav Hunt had an item on the list: "Item 240. A breeder reactor built in a shed, and the boy scout badge to prove credit was given where boy scout credit was due. [500 points]"
Well, for the application I was looking into this for (remote deployments in Greenland and Antarctica) it would be almost certainly politically impossible, so did not pursue, but I doubt it's possible to purchase Sr-90 in non-negligible quantities as a civilian (you can purchase it as a calibration source easily enough, I think, but at great expense).
I found some speculation that you might be able to build one out of thorium from smoke detectors and lanterns, as in the David Hahn case, but not much. No way this would be legal though.
See e.g. https://inis.iaea.org/search/search.aspx?orig_q=RN:9398623