NASA boss calls for nuclear reactor on the Moon NASA's Acting Administrator, Sean Duffy, has directed the US space agency to come up with a plan to deploy a nuclear reactor on the Moon. Duffy's directive [PDF] warns about China and Russia's own intentions to put a reactor on the Moon by the mid-2030s. According to the directive, "The first country to do so could potentially …
> Now please give us back the smart people that NASA had at that time.
Give us back the budget that NASA had at that time. Fixed it for you.
(Of course somebody will say "they need to work smarter, not harder" and that kind of nonsense, but fact is, you won't even put a washing machine on the Moon's surface with the current budget. First we need a transport which actually works, and there are a couple million $ between "should eventually be able" and "is reliably able to do so today".)
That part is actually pretty easy.
The trouble with Lunar Night is a lack of heat and power, neither of which are an issue for a nuclear power plant.
Lunar Day on the other hand could be a significant issue, as radiative cooling doesn't work so well in direct sunlight or when covered in sharp, electrostaticly charged dust.
... radiative cooling doesn't work so well in direct sunlight...
I wonder what the temperature of the regolith is down a metre or so or however deep you need to go to get a constant day/night temperature? If it's cooler than the reactor waste heat that would be usable for cooling, sort of reverse geothermal.
If there is a region below the surface where the temperature is about half way between day and night temperatures then Stirling engines might be a better bet, you'd just need to reverse which is the hot and cold sides on each day/night transition.
So, they just need to safely land a nuclear reactor - and digging equipment to safely make a hole, place the reactor, cover it up again - uncover it, plug in the cables this time, cover it up...
Doddle.
Stirling engines. Hmm. That could make for a good film plot: the hero has to choose between topping up the working fluid after a leak to vacuum, to keep the power running and save everyone else, or breathing.
You don't dig a pit, you drill a hole and stick pipes in it and backfill with the spoil. Possibly lot's of holes and pipes.
You'd want the reactor to be on the surface for servicing access, nobody would bury one.
NASA have prototype Stirling engines for space applications. They are the most reliable engines known, as long as the bearings are kept free from wear and contamination they should run for years.
"NASA have prototype Stirling engines for space applications."
A Stirling engine used as the power conversion for an RTG is much more efficient than thermocouples. Since they are mechanical, they will eventually need maintenance, so using them on the moon where they could be serviced is a good fit.
"Just the bearings and possibly the piston rings. NASA has had mechanical gyros last for over 40 years in deep space."
Deep space has the advantage of being much more dust-free. The moon dust is very bad for mechanical bits. If a Stirling engine is designed to be easy to replace bearings and pistons, they could still last a very long time between servicing rounds.
Most inertia management units (IMU) are solid state these days.
Or, since the current trend seems to be aiming for a polar landing where there are permanent light and permanent dark places (hopefully with water ice), just stick the nuke in a permanently shady area where it's own waste heat can keep it at operating temps and never have to worry about the heat from the sun.
"Stick the power plant in there and you avoid the problem of lunar regolith and the issue of solar heat in one go."
We need data on what it looks like below the surface. The pulverized material on the surface could do down quite a way. I'm a big proponent of prioritizing the exploration and exploitation of lunar lava tunnels. I believe that they'd be a big head start to building a base in the moon. Domes on the surface are too risky and a poor investment. Heat from a nuclear source is an obvious benefit to habitable spaces.
"I wonder what the temperature of the regolith is down a metre or so or however deep you need to go to get a constant day/night temperature? If it's cooler than the reactor waste heat that would be usable for cooling, sort of reverse geothermal."
I had a nice long conversation with an engineer from Teledyne about an RTG designed for the lunar environment. It was something the company had discussed and, of course, the engineers were all for giving it a go. RTG's can be much more efficient in places such as the arctic where conduction can be used to keep the delta-T as great as possible. For something on the moon, the waste heat also becomes a highly useful product. During the lunar day, solar panels can be used to provide power and an RTG can trickle charge batteries for night time as well as provide heat. An RTG is also the easiest sort of unit to start with if the Pu is being made again.
This was years ago when the open house at JPL had lots of outside vendors with exhibits and one didn't need to be lucky enough to book tickets in the first 5 minutes of availability. All of the space centers are super popular when they have an open house. The wonders they show are amazing. It's also great to talk with people that work on the projects. I certainly remember Dr Thaller from the Spitzer Space Telescope and Julie that was one of the MER rover "drivers". With the sort of work they were doing, people had to be made to go home.
The reason for the "nuke on the Moon" race is because there is limited area on the south pole where you can find spots that are permanently in shadow and frozen water is thought to be found in quantity. That's probably where you want to put your reactor.
Even if you put it on a crater rim that gets sun 24x7 that wouldn't be a problem. On the pole the sun would always be very low on the horizon so a simple short wall around your reactor would keep it permanently in shade. Radiative cooling on the Moon is easy since the clear sky temperature is 2.73K because unlike Earth there is no atmosphere and no clouds, which is the best you're going to do anywhere in the universe for radiative cooling.
Land area is also not at a premium so (within the limits of how much you can launch/deploy) you can use as much space as you need for your radiator surface lying on the ground.
Solar cells don't let you declare an exclusion zone. This "race" has nothing to do with what is the best alternative for power on the Moon that may (we aren't 100% sure yet) have continuous sunlight, but what lets you "claim" an area of the Moon for your country using a loophole in the 60s era treaty.
Radiative cooling is absolutely a problem, there is a limit to the radiation any body can put out. That's why we use airflow to cool e.g. CPUs. Power plants use evaporative cooling which is even more effective. Those are not an option on the moon.
The surface area of the cooler needs to be pretty big for such a power plant to work. Physics does impose limits.
It is proportional to the temperature. If you had a radiator at 2000C you can radiate 150kw/m^2. That's not likely to be practical, but IIRC 800C (which is likely to be achievable from a nuclear reactor and using normal materials for the radiator) would get you an order of magnitude less - 15kw/m^2. The total radiator surface would be pretty small for a 100kw reactor.
Setting up a reactor is probably quite easy. No atmosphere to transport anything that escapes, easy to put a 1km security border around it. But then how to create electricity? Steam requires water, and the Rankine cycle, like the Carnot cycle, requires cooling....
Meteor impact will burst any pipes close to the surface.
Colour me sceptical for another 50 years or so...
ALL very good points. I thought this one out long ago...
* - Peltier devices require a sink but you can obtain that by drilling down. The moon's average temp is about 50C below Earth's as I recall, because no air. So the sink would be "reverse geo-thermal". Would require excavation and coolant, probably
* HUGE underground ice patch near the south pole offers water AND the opportunity to make O2 and H2 fuel, plus electricity for whatever. Also coolant.
If a large enough bank of high temperature Peltier devices were used for DC electricity, you would avoid the steam plant completely. For efficiency, not so good, but cost would more than likely make up for it. You could also use thermal energy from nuke waste in a similar manner (Space 1999 anyone?)
"If a large enough bank of high temperature Peltier devices were used for DC electricity, you would avoid the steam plant completely."
The efficiency is extremely low. For the Voyager spacecraft, a trade was made so a power supply would last for decades. The same goes for the Curiosity and Perseverance rovers on Mars where solar panels, while more efficient, were the Achilles heel of Spirit and Opportunity. The above mentioned Stirling engine generators would be more efficient and with people about, can be serviced as needed.
> and with people about,
Hey, hey, who said there will be any "people about"? They only talk about a reactor. Do you really think they will build facilities and send up crews to service a totally useless reactor probably just feeding a big "No Trespassing" neon sign? Who will pay for that? NASA's ever-shrinking budget?
It's like they plan to build a high end sewer system in middle of the Sahara desert, just more pointless and way more expensive. Fortunately at this point technology won't allow what common sense should had forbidden. Jeez.
A sealed system much like nuclear submarines use. The excess heat could be dissipated on the shady side of the reactor (or artificial shade). Once we have people actually using the power from the reactor some or a lot of the waste heat could be used to keep the living quarters warm during the lunar night, to heat and distill water, and even for cooking (such as a steam kettle).
> Once we have people actually using the power from the reactor
Do you seriously believe they are about to budget a permanent settlement on the Moon? When the vastly cheaper International Space Station is going to be retired because "too expensive, not useful enough"?
I mean we can dream, I have grown up with visions of space colonization too, but in 2025 you can't deny the harsh truth anymore, those were just dreams. Humanity is not about to expand to the stars, on the contrary it is shrinking into petty clan mentality and aggressive territoriality.
(Didn't downvote you though.)
"Which is ironic given that when Musk set it up no one from Nasa wanted to work there."
There's a big turnover as well. When one is fresh out of university and all glassy-eyed, SpaceX looks like a great job. After a year of 60 hour weeks, no chance at relationships and only earning what others make for 40 hour work weeks, the shiny wears off.
"Now please give us back the smart people that NASA had at that time."
I know I'll get viciously downvoted for saying this, but with the smart people you mean the Nazi war criminals that Uncle Sam paperclipped out of Germany in 1945 instead of putting them on trial as they should have been?
Don't get me wrong, I am in awe of NASA's accomplishments since the 1950s. But let's not forget that that was not in the first place thanks to American skill.
"I know I'll get viciously downvoted for saying this, but with the smart people you mean the Nazi war criminals that Uncle Sam paperclipped out of Germany in 1945 instead of putting them on trial as they should have been?"
If that sort of thing were a thing, there'd be plenty on the other side guilty as well. It's the politicians and planners that are the responsible parties, not the people doing the work. Even engineers have no say in policy matters. They could leave late one night and take their chances crossing over to the other side, but remember that news would have been tightly controlled and the propaganda those people would have seen would show the "enemy" in a very poor light. There was no internet, very little TV and radio/newspapers would have embedded censors and political officers as editors-in-chief.
They’re talking about putting a nuclear reactor on the top of a honkingly big rocket and landing it on the moon, a celestial object we have about a 50% success rate of landing on (crewed and uncrewed).
As every school sports day fan knows, the egg and spoon race isn’t won by someone sprinting for the finish line, so I’m more than fine with them not having another race!
And the water? There are small pockets of ice (at least we think it's ice) in a few crater shadows, but you'd have to mine it, transport it and deposit it into the reactors – which themselves have to be placed in an advantageous spot for settlement, not necessarily near the ice – without ever letting it be exposed to sunlight.
You are right abour 'permanently shaded areas' (PSR's) but there's this:
Conservative estimates suggest the south pole could harbor 100,000 to 1,000,000 metric tons of water ice across all PSRs, though some optimistic models propose up to 10 billion metric tons if deeper subsurface ice exists.
"Conservative estimates suggest the south pole could harbor 100,000 to 1,000,000 metric tons of water ice across all PSRs, though some optimistic models propose up to 10 billion metric tons if deeper subsurface ice exists."
Concentration and purity are still be unknowns. It will also be all sorts of fun to sort out how to retrieve it and keep it from sublimating away as it's exposed.
In the near term, power is more of an issue and locations for missions should be chosen for habitability and usefulness for proposed activities. The south pole is an aggressive target at a time when there's scant real-world experience living off-Earth. Mining water isn't the end-all application for going to the moon. There needs to be things that can't be or shouldn't be done on Earth that are the primary drivers of going and staying.
Nuclear reactors boil water to turn a turbine. It's extremely unlikely they will be launching the entire reactor (SMRs many times larger than the Starship) as a self-contained unit, especially with fluid aboard. It would be far too large and heavy, especially to land.
If it allows them to claim an exclusion zone around a nuke teapot/kettle (likely inactive) then it's simply a "Land Grab".
And if it is active, what's the cooling mechanism going to be - even Radioisotope thermoelectric generators work on heat conversion to electricity ?
There's not a lot of running water up there to do much cooling - unless it's supposed to go critical first to melt the permafrost !!!
*explosion icon of course*
reverse geo-thermal - average temp on moon is about 50C below Earth's average temp because no air. SO, you drill down and/or excavate a geo-thermal COOLING grid. Might need liquid coolant for it though, but I expect conductive heat transfer might also work...
Moon has a lot of titanium, which I believe is a lot like aluminum in its heat transfer capability.
I can't tell if people are joking when the make statements like this. With exactly what were you planning on excavating the moon with. Did I miss something, does Catepillar have a moon office? And with what energy source will the mining equipment tap to dig these big holes? I read these ideas for lunar/martian sites and am amazed at how all the materials to do plumbing, mining, construction are just magically there. Did I miss an article about star trek transporters being available? That would make things so much simpler if we could just beam stuff we needed around. There was that one movie where they transported a whole whale w/ocean so it must be possible to transport a small by comparison cat excavator.
" With exactly what were you planning on excavating the moon with."
There's been work done on that but it's a long way from even good prototypes. There needs to be a solid list of requirements since the equipment will need to be specialized rather than general purpose to keep mass and volume envelope down. The regolith excavation challenge had teams compete in scooping up loose surface material (JSC-1A simulant) that could be carried off for H3 extraction. What seems more interesting and of more immediate need is a look at what appears to be caves on the moon that could be tidied up, sealed and used as habitats. Ships or domes on the surface to house people are highly vulnerable to impacts, solar and galactic radiation and decompression. There needs to be a solid construction shack or 12 to base from and bootstrap into lunar activities as they evolve.
Re the tunnels, while a good idea on the surface (pun not intended!), we have no idea how porous the rock is so it will probably need the surfaces all coated with a sealant. Here on Earth we'd probably used an expanding foam or something for both it's sealing an insulating properties. I'm not sure what could be used in a vacuum at such low temperatures.
It's quite probably still the best option, but I suspect it's a more complex than just bodging in an airlock on the handily shaped and sized entrance we see on SF shows :-)
It could probably work too, so long as the chemical mix can produce enough gas when it reacts. I'm no expert, and my only experience with expanding foam is a packaging machine we used to have. The chemicals are not nice and as health and safety rules tightened, it was easier to just get rid of it than to try to comply with all the hazmat stuff. And being toxic, I'd imagine any later outgassing over time might not be so good for people living with it.
> I bet it goes REALLY big!
Most likely just explodes into a cloud of dust since the internal pressure will be way too high. Using a can of Earth-grade foam would probably just result in a cloud of frozen dust particles. :-D
But I admit they probably can create a vacuum-certified formula with just the minimum of propellant, and a foam which doesn't immediately freezes solid but takes the time to first enter any cracks you wanted to fill. Won't be neither easy nor cheap.
"I suspect it's a more complex than just bodging in an airlock on the handily shaped and sized entrance we see on SF shows :-)"
Yeah, that would too ideal for our universe to be handing out. It does mean that experiments will need to be run in-situ. Before an airlock is installed, any major tunnel work needs to be done as it's easier to haul out spoil if a tight spot needs to be widened. Searching for the best initial candidate might take time.
After it's sealed, any need to take material out is another cycle on the air locks and stirring up very abrasive dust and fines. Somewhere I have some JSC-1A lunar stimulant and some photos from the microscope. It's not sand, it's pulverized glass and sharp. Consolidating the material inside any habitat so it doesn't wind up floating in the air is important. It's electrostatic as hell too and clings everywhere.
> Furthermore, it is not clear who would use the power source.
That's the first thing which hit me when I heard about it. A nuclear reactor on the Moon? What for? Besides they are strangely specific about its size: "a minimum output of 100kW". Does some government contractor have some dead stock >100 kW reactors to get rid of?
I'm maybe too logical about this, but you usually deploy a power plant when you need power, and right now our power requirements on the Moon are nil. Only once (and only if one day) we have a permanent station up there, we could start considering the need for a nuclear reactor, depending on the station's actual power needs. Given a Moon station is a imminent as that on Mars and a reactor has a finite life expectancy, what is this project about except wasting money?
OK, but if you want a deliverable lunar reactor for when you need it to appear, some design and testing will have to have been done in advance ... such as in this project. And if you expect to need a reactor to power your moonbase, you might even want to have it in place and running reliably for a while *before* it becomes a critical part of your infrastructure.
And in the meantime, some physicists will -- most likely -- be along shortly with innumerable projects for a wide variety of sensors, telescopes, and the like, all of which might benefit from a convenient power source. Might be tricky to get your robot to plug in the necessary powerboard/extension cable though. :-)
> if you want a deliverable lunar reactor for when you need it to appear, some design and testing will have to have been done in advance
True, but the question is, do you want a deliverable lunar reactor? I still fail to see the point.
I mean, what about some gantry cranes, for when cargo containers start getting shipped to the Moon? Should we send up some gantry cranes to moon-proof them? This is a slippery slope...
As for any telescopes, the blocking point isn't lack of power (solar panels would do just fine), it's the budget of building one and sending it up there.
"A nuclear reactor on the Moon? What for? Besides they are strangely specific about its size: "a minimum output of 100kW". Does some government contractor have some dead stock >100 kW reactors to get rid of?"
Power is a big chicken/egg problem. There's no point in building out a large power supply if there's nobody lined up to need it and nobody is going to move forward on a project that might need a load of power if it isn't in place. It can be very hard to do a "build it and they will come" approach if there isn't much insight as to needs.
If we don't know what this is going to be used for, and we're just doing it to create a "keep out" zone - then the answer is simple. Heavy looking metal box. Radiation symbols - scatter of blinking lights. Job's a good'un. Oh we also need the acronym. How could I forget that most important bit of space science?!
He's setting up NASA to fail because there's no way they can make that deadline.
It takes a decade or more to construct a nuclear reactor from a known design on Earth. How long do you think it will take to design a nuclear reactor from scratch capable or running in an airless environment on the Moon? Hint: it's take more than the 4.5 years Duffy's asking for it to be constructed in.
We're not talking about lots of megawatts if we assume the lunar outpost is adequately insulated, so presumably almost any of the smaller factory-built package reactors, due to be available Real Soon Now (or so we're told) could equally well be used on the moon and shipped up there as needed. Assuming a set of solar cells with an equivalent output to the reactor are also installed on the moon, the reactor only needs to be run for two weeks at a time during the lunar night because solar cells will provide power during the lunar day, so refuelling or replacing it shouldn't be an issue.
There's another benefit too: since it seems that moon-dust is likely to be a good insulator, that should make it easy to insulate the outpost well enough to keep the crew alive and warm(ish) enough to survive a two week reactor fail by shipping up a replacement unit from Earth or activating a backup reactor that's already on the Moon.
Lastly: if moondust is light and a good enough insulator, we won't need nukes on the Moon: just solar panels and good insulation for the moon base.
"the reactor only needs to be run for two weeks at a time during the lunar night"
Mostly, you don't just turn a nuclear reactor on and off. You might vary the output somewhat, but that usually means mechanical assemblies that could break. To minimise complexity, you just run it at normal rates all the time and shunt off any "waste" energy via a sink of some sort (or into batteries if there may be times you need more than normal output can supply.)
That's quite an assumption considering that it's about 3 times the mass of anything that has been landed on the moon in a controlled manner in the last 60 years.
Soviet Lunas were up to 5.8 tonnes.
A fully laden Apollo descent module is approximately 15 tonnes, but about 10 tonnes of that is fuel, a lot of which was burned on the way down.
Apollo S-IVBs (3rd stages of Saturn V) were dropped on the moon and were approximately 14 tonnes, but they were just crashed into it - probably not the best way to land a nuclear reactor.
"A fully laden Apollo descent module is approximately 15 tonnes, but about 10 tonnes of that is fuel, a lot of which was burned on the way down."
There's a bunch of life support in that mass that won't be needed for cargo flights. Something cheap and cheerful with no ascent capability would also have more down mass to the lunar surface. If the construction of the lander was "Mechano-like", the lander can be taken to bits and used for building materials or the refined metal.
" If the construction of the lander was "Mechano-like", the lander can be taken to bits and used for building materials or the refined metal."
Eventually, yes. Unless bits can be used "as-is", I think it might be a wile before the future colonists get around to setting up a machine shop to be able to cut up and repurpose struts and panels, let alone treat it as raw material for smelting :-)
'Unless bits can be used "as-is""
That's my thinking. A machine shop with a mill and lathe are absolutely needed, but I won't count on them for lots of work. There will be all sorts of needs for gantries and A-frames to service things such as rovers. If there are very many pre-fabricated shafts (as opposed to an adit or incline) to explore, some of them might need a head-frame to get in and out of them.
I'm sure I built a Mechano crane at some point since it was in the "book". Growing up I had two things on my gift list: More Lego and Mechano (and batteries).
"let alone treat it as raw material for smelting :-)"
I was thinking of things like the metal used in the rocket motor. One can't make one out of standard shape "Mechano" parts so the only reuse would likely need to start with melting the metal down and casting it into a new shape. That doesn't need to be a big priority other than a quick thought about making sure the alloys are documented and if it isn't left in place, there's a defined bin/pile to stack it in for later with an inventory being kept.
Might want to tone down the hyperbole. Exactly how much nuclear material do you expect will be there? The moon is huge (compared to a reactor), and still orbiting just fine, still tugging our tides just fine, despite billions of years getting repreatedly whacked by meteors that would make an exploding nuclear reactor seem like a firecracker.
[Icon for what won't happen. And even if it did, it won't cause the problem you feverishly imagine.]
"years getting repreatedly whacked by meteors that would make an exploding nuclear reactor seem like a firecracker."
Reactors don't explode. Not a nuclear explosion, anyway. Fukushima going bang was Hydrogen. Chernobyl was steam, etc. The tend to melt and spew really bad radiation. Not a good thing, but there is the benefit on the moon of debris not being carried by wind to occupied areas.
Nuclear reactors cannot explode, as those are not fueled with sufficiently enriched fuel to result in a nuclear explosion.
A meltdown can result if the cooling system fails and the rector is poorly designed in that convention cannot exhaust the residual heat post-shutdown - which still does not lead to an explosion.
It's only if molten fuel comes into contact with water and produces hydrogen (which then is thoroughly mixed with air and ignited), that there can be a hydrogen explosion (containment buildings are designed to contain explosions of that magnitude).
A moon reactor likely wouldn't use water as a coolant and regardless there would be no substantial amount of oxygen available on the moon, thus there would be no chance of an explosion.
Even if a 100MT nuclear bomb was detonated on the moon, only a smallish crater would be produced (the main issue would be the radiation burst, which would travel long distances unimpeded by any atmosphere, but maybe that wouldn't impact artificial satellites much more than the radiation from the sun due to spread out over the large distance?)
It is one thing to declare a 'keep out zone', quite another to enforce one.
Aside from that, the lunar surface is quite hostile. Solar radiation without the protection of the Earth's magnetosphere or atmosphere, and lots of micro-meteorites, plus huge temperature variations
"The temperature of the Moon can vary from highs of 250° Fahrenheit (120° Celsius) at its equator during lunar daytime and plunge to -208 degrees F (-130° C) at night."
from: https://www.skyatnightmagazine.com/space-science/temperature-moon
The keep-out signs for Earth reactors have those scary "you will get irradiated" signs on them. The Lunar reactor - or its shadow, at least - will probably be a place to stand to avoid being irradiated.
But not to worry, the keep-out zone will be patrolled by US Space Force squaddies on the lookout for intruders.
This is actually what NASA excels at, these days.
They can do excellent powerpoint projects.
This won't happen, any more than https://en.wikipedia.org/wiki/Demonstration_Rocket_for_Agile_Cislunar_Operations happened.
"However, by January 2025, the mission's planned 2027 launch was placed on indefinite hold due to technical and regulatory challenges such as the complex safety and testing requirements for ground-based nuclear reactor validation and the unresolved final design of the propulsion system. The program's status was further impacted by the May 2, 2025 release of the FY2026 federal budget, which proposed a $531 million cut to NASA’s Space Technology Mission Directorate. The budget documentation cited reductions in unspecified advanced space propulsion projects. Some analysts interpreted this as effectively ending nuclear propulsion research, noting similarities to NASA’s earlier cancellation of Project Prometheus."
Those are not actual reactors, they're RTGs (https://en.wikipedia.org/wiki/Radioisotope_thermoelectric_generator), i.e. a blob of radioactive material and a thermocouple.
They are extremely simple, have no moving parts and are thus very robust, but they also have very low outputs (<100W).
"They are extremely simple, have no moving parts and are thus very robust, but they also have very low outputs (<100W)."
That's down to Carnot efficiencies. The energy of heat is the ∆T along with the conversion method. In space, radiating heat is not that easy and thermocouples do a poor job of turning ∆T into electricity. On Luna, being able to conduct heat away from the cold side and using a more efficient conversion method can make an RTG more attractive. The Pu is glowing hot which gives a good indication of its temperature. All that's needed is a good connection to use the moon as a heat sink and Bjorn Stronginthearm is your uncle.
Now I need to find my books on RTG's and run some numbers. It's been two moves since I last looked at it and I still haven't spent the time to organize the book collection. Some say I have a book habit.....
> On Luna, being able to conduct heat away from the cold side and using a more efficient conversion method can make an RTG more attractive.
I don't know why RTGs would be more attractive on the Moon than elsewhere in space. Rocks aren't very good at conducting heat, so you would be forced to mostly use radiative cooling just like in space.
(Besides, the plutonium oxide cylinder in the Wikipedia article is glowing because they covered it with a graphite blanket for several minutes before taking the picture (read picture caption). Left alone it is much cooler and doesn't glow.)
"(Besides, the plutonium oxide cylinder in the Wikipedia article is glowing because they covered it with a graphite blanket for several minutes before taking the picture (read picture caption). Left alone it is much cooler and doesn't glow.)"
Then leave the graphite blanket on it so it stays as hot as possible! BTW, anything above 0K does indeed glow.
Rocks don't have great thermal conductivity, but if the heat is being radiated into a pressurized habitat (underground), there's loads of surface area and a lot of rock. There will be water reservoirs that will need to be kept warm, etc.
Calculating power output will have a lot to do with the thermal difference between the hot and cold sides of the generation system. In space, it's much more difficult to get rid of heat and keep heat flowing out from the cold side. RTG's aren't power dense, but can work for decades at a time making them simpler to support. Any sort of fission reactor is going to need much more labor to keep in good nick. They also have many more bits so there's more that can go wrong. Just having to shut down a reactor due to a minor fault can be exceptionally bad since it could mean people needing to return to Earth in short order if it can't be put back into service very quickly.
