In our solar system light and heat are, of course, provided by the sun. I was wondering if a comparable effect might be had by positing a burning moon instead? So rather than a true star, you'd have a much smaller but relatively close satellite with a surface that's just engulfed in flames.

Please assume that we can play with the orbit, size, and heat until reaching the desired effect, and that the following problems are not within the scope of the question, and can be dealt with artificially.

  • The time scale. (It doesn't have to support evolution, just be habitable.)
  • The presence of enough fuel. (In terms of the science, I'm more asking about the likely effects of the moon once it's already burning rather than whether it could plausibly ignite. I'm equally happy with answers that assume a fuel source not found in our universe, and answers that are hard science all the way through.)

Assuming all that, if this moon did occur, could it temporarily sustain a habitable world? And even tweaked to the ideal conditions, would the effects be in some ways necessarily and noticeably different from the perspective of the surface?

This question is inspired by ancient cosmologies in which the sun orbited the earth. I'm basically trying to suss out whether a science-friendly version of that is possible.

Thanks for your time.

(The question has been edited to clarify my aim in response to some of the answers and comments below; thanks for your responses and help.)

  • 36
    $\begingroup$ Setting the moon on fire is generally considered a bad idea. $\endgroup$
    – CaM
    Commented Apr 28, 2017 at 17:23
  • 3
    $\begingroup$ Well, I mean, for it to burn requires fuel, and for it to burn long enough to provide the energy needed for civilization to happen requires a lot of fuel and oxidizer, and I don't have any numbers to back it up, but I suspect the mass of whatever fuel is required to make that happen is enough to collapse the moon into either an inert ball, or something that ends up using itself for fuel in fusion/fission, i.e., a star. So my gut says that anything you try outside of magic will end up being a star, or being unsuitable. But I don't know for sure. It doesn't seem scientifically feasible though. $\endgroup$
    – Jason C
    Commented Apr 28, 2017 at 17:29
  • 4
    $\begingroup$ Related: Can a habitable planet have mini-suns (i.e. solar satellites or glowing moons)?, and How can I set a moon on fire?. $\endgroup$
    – HDE 226868
    Commented Apr 28, 2017 at 17:41
  • 2
    $\begingroup$ Something burning (in the conventional sense) requires oxygen. If the object is in space (such as a moon), it must contain oxygen within its own material in order for that material to burn. But, combustion of most flammable materials is (I think) moderated by the rate at which oxygen convects to the flame; instil oxygen into the stuff itself, and it can all just go up at once. Think gunpowder. Perhaps patterned strata of separate combustables and oxidizers could control the reaction, similar to the patterning in solid rockets...? $\endgroup$
    – Maxander
    Commented Apr 28, 2017 at 22:40
  • 2
    $\begingroup$ Could it be a large starship burning up, caught in orbit? That way you can include man made energy sources. Perhaps it could be organic; e.g. a massive laviathon in caught in orbit, burning blubber. Not sure of the physics/chemistry, but maybe those alternate moons/satellites work around some of the issues outlined below. $\endgroup$
    – JohnLBevan
    Commented Apr 29, 2017 at 8:56

11 Answers 11


A burning Moon does not have enough fuel to sustain life on Earth for more than a few years, and it would have to be so hot it would instantly blow itself apart. Here's the rough calculations.

Energy Density

This is a problem of energy density, and an important historical one.

Before nuclear fusion was discovered, the new science of geology was in conflict with the known physics and chemistry. Geology said the Earth had to be very old, hundreds of millions or billions of years old! But there was no known mechanism to have something like the Sun burning for that long. Even radioactive decay and nuclear fission were insufficient. Only nuclear fusion could provide energy for the 5 billion required years. If you want to read more about that, I'd suggest Bill Bryson's pop-sci book A Short History Of Nearly Everything.

The most energy dense, naturally occurring, chemical reaction is the oxidation of hydrocarbons: ie. burning methane, oil, fat, kerosene, etc... that's why we use them in cars, it's very energy dense and very safe. Burning hydrocarbons have an energy density of roughly 5e7 J/kg.

In contrast, uranium and thorium in a nuclear reaction have an energy density of 8e13 J/kg. The energy density of fusion is even higher, 6e14 J/kg.

So you can see, a burning Moon is roughly 10 million times less energy dense than a star. This will effect how long it can burn, and how much energy it can give off.

Next problem is one of surface area. While we could, theoretically, ensure our burning moon is fully oxidized and can burn all the way to the core, only the energy on the surface will radiate into space. The moon's surface area, and thus its radius, limits how much energy it can radiate.

And finally, its distance from the Earth is important. Since the burning moon radiates in all directions, only a tiny fraction of its energy will reach the Earth. The closer the Moon is to the Earth, the larger percentage of its energy will reach the Earth.

How Long Can A Burning Moon Heat The Earth?

Let's try this out for our Moon. Let's assume the Moon magically became a burning ball of hydrocarbons and oxygen. How much energy would reach the Earth, and how long would it last?

First, some important attributes of our Moon. I'll use approximate numbers to make the calculations simpler.

  • Radius: 1700 km
  • Surface area: 3.8e7 km^2
  • Mass: 7e22 kg
  • Distance: 3.8e5 km

We can preserve the Moon's mass, or the Moon's radius. I'll do mass, it's easier and since the Moon is about 5 times more dense than gasoline it provides more fuel giving this a better chance of working. 7e22 kg at 5e7 J/kg is 3.5e29 J of energy available. This assumes the entire mass of the Moon burns.

How much energy is that? Using the handy Orders of Magnitude (energy) list, we find that the Earth receives about 5e24 J of energy from the Sun per year. This is nearly 100,000 times that, so it's potentially enough energy to heat the Earth for 100,000 years. But not so fast, that's the total energy output by the Moon, but how much reaches the Earth?

To know that we need to know what percentage of the Moon's sky the Earth covers. We can figure that out by imagining a sphere around the Moon with a radius that's the distance to the Earth, that's all the Moon's energy radiating outward into space. The surface of that sphere is 4πr^2. r is the distance from the Earth to the Moon, 3.8e5km, giving us a surface area of 1.8e12 km^2.

The Earth can be thought of as a disk on the surface of that sphere. Its surface area is πr^2. r is 6.4e3 km giving us a surface of 1.3e8 km^2.

1.3e8 km^2 / 1.8e12 km^2 is 7e-5. The Earth receives only 1/14,000th of the burning Moon's radiated energy. This means of the 3.5e29 J available, only 2.5e25 J will reach the Earth ever.

The Earth needs 5e24 J per year to sustain its current environment. Only 2.5e25 J will reach the Earth. The burning Moon can only heat the Earth for 5 years and that's an upper bound.

Now that I've done the calculations once, you can change the parameters and do them again. Moving the Moon closer or making it larger will help.

How Hot Does The Surface Of The Moon Have To Be?

The next problem is just how hot the surface of the Moon would have to be. How much power is each square meter of the Moon radiating? Is it feasible? Since the Moon is so small, it might need to be absurdly hot.

Let's, again, assume everything is the same as now, and the Earth is receiving its 5e24 J per year. Power is normally measured in Watts which is J/s. 1 year has 3.15e7 seconds, so the Earth receives about 1.6e17 Watts from the Sun.

We calculated above that's just 1/14,000th of what the Moon is putting out, so the total power output of the burning Moon is about 2e21 Watts. That's about 1/1000th of what the Sun produces, or roughly the same as a very, very small red dwarf star.

The Moon has a surface area of 3.8e7 km^2 giving us a power per unit area of 5e13 W/km^2. Is that a lot? The Sun has a surface area of 6e12 km^2 and puts out 3.8e26 Watts giving a power per unit area of 6e13 W/km^2. Nearly the same!

Somehow the burning Moon has to put out the same energy per per unit area as the Sun using fuel that's 10 million times less energy dense. That's a problem, no fire is going to burn that intensely.

Worse, that level of energy output will produce a great force wanting to blow the Moon apart. Stars cope with this by being very, very massive; gravity balances this force wanting to blow it apart. Our burning Moon has nothing like the gravity of the Sun and will be instantly blown apart, the Earth will be showered with extremely hot fragments of burning hydrocarbons.

This part is particularly important because it means no amount of high-energy unobtanium will work. A burning Moon has to radiate too much energy for its gravity to hold itself together; it will blow itself apart.

Again, now that I've done the calculations, you can play with the parameters to try and make it work.

  • 10
    $\begingroup$ "it means no amount of high-energy unobtanium will work" — You are obviously not familiar with Phlebotinised Hyperunobtainium. This not only has an incredible energy density, it also has a large handwave charge which provides an additional large attractive force, but which only applies to other handwave-charged particles (so the effect on normal matter, like the Earth is composed of, is only the gravitational one), A ball of Phlebotinized Hyperunobtainium is almost impossible to torn apart due to the large handwave force. $\endgroup$
    – celtschk
    Commented Apr 29, 2017 at 19:48
  • 1
    $\begingroup$ @celtschk You fell victim to one of the classic blunders! The most famous is never get involved in a flame war on the Internet! But only slightly less well known is this: the third law of Worldbuilding says the science-based tag will always balance out the handwave force! $\endgroup$
    – Schwern
    Commented Apr 29, 2017 at 20:00
  • 1
    $\begingroup$ Wow, you invested all the effort to work out the basic numbers and put them into prose. This is a nice answer. $\endgroup$
    – Durandal
    Commented Apr 30, 2017 at 18:35
  • $\begingroup$ @Durandal I've learned a lot of physics answering these questions. :) $\endgroup$
    – Schwern
    Commented Apr 30, 2017 at 18:36

Depending on what timescale you're aiming for it may be remotely possible.

One historic theory tested for the energy source of the sun was coal burning. Turned out this could power the sun, but only for a few thousand years. Thus the theory was falsified by observation: the sun has been shining for much longer than coal could provide energy.

With a moon being closer, and having much less surface area, this could work. Any chemical process generating enough head would do, but only for very limited amounts of time.

Still, no there are probably no naturally occuring large bodies in space that contain a large amount of "burnable" chemicals. The reason is that when a large body is formed it will almost surely be heated above the ignition point of said chemicals.

  • $\begingroup$ I read a book where they used a black hole to cause phobos to incandesce, as it is a more efficient method(glancing a bunch of matter off the event horizon) than burning. It isn't burning though. You might be able to do this in a more reasonable way by making the moon out of a non-fissionable isotope with a relatively small half-life and covering it in a bunch of phosphorescent material. It'd really glow in the dark. $\endgroup$ Commented Apr 29, 2017 at 2:57

Impossible with known physics

There are three ways to get energy: chemical bonds (electromagnetic force), nuclear bonds (strong force) or matter-antimatter reactions.

To have a burning moon with chemical bonds will require enormous amounts of fuel and oxidizer. Not impossible to get on a moon sized scale but difficult.

A burning moon based on fusion will require star like masses which is not a moon. Fission based burning requires man made products here on earth so I think improbable that conditions on this moon will naturally create these products.

Anti-matter powered burning requires exceptionally rare anti-matter. While this can be man made, we don't usually find it floating around.

Overall, without some kind of super-organizing force to provide fuel/oxidizer for burning or provide artificial gravity to compress fusion fuel, I find this moon impossible.


I'm going to quote this from Scientific American:

"Jupiter is called a failed star because it is made of the same elements (hydrogen and helium) as is the Sun, but it is not massive enough to have the internal pressure and temperature necessary to cause hydrogen to fuse to helium, the energy source that powers the sun and most other stars."

If a body the size of Jupiter lacks sufficient mass to self-ignite, then a moon would be even harder pressed to reach sufficient mass.

I suppose a body of sufficient and correct elements might theoretically be ignited in some way, but it would burn only temporarily (on cosmic scales), before the combustible materials were consumed and the flames died out.


Two wormholes end in the nucleus of the moon. One, for some reason, flows oxygen into it. The other, more commonsensically, flows hydrogen. Boom! Light, heath, beautiful moonsets.

... the problem then is what to do with all that water. I guess a third wormhole takes it away into another universe.


In a story universe I’m tinkering with, the planet is lit by lamps in orbit, not a star. The key is that the lamps don’t contain all the fuel they need for a lifetime of billions of years. Rather, a large mass stores all their energy. But the mass isn’t squandering energy like a normal star — it’s a stockpile that will last them orders of magnitude longer; e.g. only lighting the planet with a spotlight, not shining into space in all directions.

Energy is tranferred from central storage, either by regular shipments of mass, or wirelessly like with tuned resonance inductance.

  • $\begingroup$ That would work; I remember this plot device being used in Arthur C. Clarke's The Sands of Mars, with the moniker of Project Dawn. $\endgroup$
    – LSerni
    Commented Apr 30, 2017 at 11:16

First, stay away from our moon. Seriously. While it has no oxygen atmosphere so I doubt you could set anything on fire there, that doesn't mean you need to try.

As to your question, you are going to have to get the moon's surface to the same temperature as the sun (5778 K). It's not just about getting light from the sun/moon/whatever you have in the sky. You are also going to need it to deliver the right wavelength of light (at least for terrestrial life). Too red and your body won't get enough Vitamin D. Too blue and you need a LOT of sunscreen. Some of that is a product of evolution, but since I don't think your lunar fire is going to last long enough for a new species to evolve, that's going to be a constraint.

Now will that produce the right amount of energy for the planet (around a kilowatt per square meter)? It's much smaller than the sun, though on the other hand it's much closer. Well those cancel out by some weird astronomical coincidence. It's the same reason the sun and moon appear about the same size in the sky.

That being said, do NOT interpret that coincidence as some hint from the gods that this little plan of yours is a good idea. I know it's tempting, I mean what are the odds that we would have a moon the exact right size and the exact right distance that it could act as a surrogate sun? Doesn't matter. We need the moon. It regulates tides and seasons and other things we need that for some reason are more important than having a cool looking second sun.


How about a glowing hot moon?

enter image description here from http://www.gh-ia.com/images/heating-ball.jpg

Burning, as in oxidizing, will of course eventually consume reactants and go out. But I could imagine a planet with an extremely strong magnetic field and a moon made of a high resistance conductor. As the moon moves through the magnetic field, induced currents and consequent resistance heats up this moon to glowing.

Ultimately the energy to heat the moon comes from whatever process is generating the planet's magnetic field.


Super advanced aliens could always built a giant artificial satellite that artificially provides light and heat to the planet it orbits. It could have giant fusion generators to generate the energy to power the artificial lights that light up the planet.


Let's start with the energy requirement of Earth.

Solar irradiance amounts to around 1360 W/m2 averaging on the whole surface, which is 510.100.000 km2, or 0.5E+15 m2. The total output must then be 0.7E+18 W, way less than the output of the Sun which is 0.38E+26 W.

There are several geometric limits that we need to overcome. For example we cannot simulate an isotropic radiation source because, being much nearer than the Sun, what correctly heats the Equator will let the Poles freeze, and the correct irradiance at the Poles from a point source equivalent will make the Equator inhabitable.

So our Moon can't "burn". It must be a "fly's eye" of emitters, each pointed in different directions.

A power of 0.7E+18 W means 0.7E+18 W J every second. Since one megaton is 0.4184E+16 J, we need 167 megatons every second.

Matter-antimatter annihilation yields around 43 megatons per kilogram, or about the yield of the Tsar Bomba. We need around four of those exploded every single second; 3.89 kg of antimatter (and as much of ordinary matter), for a total of around 670 tons of fuel every day supposing a conversion efficiency of 100%.

Consider that antimatter requires careful containment. Our satellite will require to be quite sizeable.

At that level of output, other things get difficult. For example, the radiation will likely be generated in several different annihilation plants, and then conveyed on the appropriate area of Earth through mirrors. Mirror reflectivity needs to be as close as 100% as possible, for even a thousandth of a percent would result in the mirror overheating and possibly in a catastrophic meltdown.

The "satellite" will likely be a lattice of burner-emitter units, set enough apart to provide some emergency insulation against mishaps, capable of redundancy if we have to scram some of them.

Incidentally, this makes the satellite double up as a planetary control device; should all mirrors focus on the same planetary area, they would pack a considerable punch. Depending on how small an area they can focus on, this goes from rendering Australia uninhabitable to subjecting the whole Texas to one single continuous kiloton-level nuclear bombardment, hold the ionizing radiation.

(This means that deploying such a satellite would be a political and security nightmare).

Where do we put the satellite: obviously not at the distance of the Moon. We could place it at the L1 point, so that it's "still" relative to the Earth. Since the Earth rotates, it will "see" the satellite rising and setting just like the sun. There is the problem of supplying the satellite hovering at 1,500,000 km from Earth.

Otherwise we put it at about 60.600 km of altitude, so that it orbits in about 48 hours around the Earth's barycenter; since the Earth rotates once every 24 hours, it will see the satellite follow a widdershins orbit once every 24 hours. This is completely unfeasible if there is also a Sun available, as it would give you two superimposed "days" flipping around one another, with a long day, a short day, a long night and a short night all in the same 24 hours.

We don't want to put it too near, otherwise it will be unnaturally low on the horizon at the Poles.

This right there presents a snag: the radiation being sent out has a pressure. The satellite won't be able to keep a passive orbit. The thrust is given by 0.7E+18 W divided the speed of light in m/s, which is 0.3E+9, and comes out at 0.23E+10 N, of the same order of magnitude of the thrust of the first stage of a Saturn V rocket.

We might want to, say, place the satellite in a slower, unstable orbit (which means putting it on a naturally faster orbit, that is, nearer the Earth, and having it go slow) - and stabilize it with its own energy emission. But the gravitational acceleration provided by the Earth decreases with the distance squared, and the thrust of the emitters needs to be divided by the total mass of the whole satellite to get an acceleration figure. If the mass of the satellite is one million tons (1E+9 kg), those 2.3E+9 N of thrust boil down to an acceleration of 2.3 m/s^2.

This is the acceleration at a distance (from the center of the Earth) of about two Earth radii, so 13.000 km, or 6500 km of altitude. If the satellite is stationary in respect to the Earth, it can "float" on the thrust of its emitters. At that distance, though, the parallax effect is noticeable - the "Sun" would be way lower on the horizon than the real Sun, and invisible from the "poles".

A heavier satellite could keep station farther out. We would want to have it librate by about double the planet's original axial tilt to replicate the seasons, however, otherwise either one pole would always be in darkness, or (if we kept station on the rotation plane) both poles would see a "midnight sun" just below the horizon all year long.



The closest to what you are asking, from my point of view, is having a planet really close to a white dwarf star. Too big for being accounted as a moon (the typical size would be roughly the size of Earth itself), in an orbit too far away for being called a moon, and will probably tidal lock the planet (a side with burn while the other would freeze as in deep space).

Serious doubths on how this tidal lock would affect planet core rotation (and the magnetic field protecting the planet from solar wind... another problem added to the tidal lock). However, should it be made possible, it has an interesting side effect: it would last forever.

Ok, not forever. Everything will end. Even the universe. But in a future so far away that it defies description, a white dwarf will still emit light and heat, therefore it would offer a chance for life. A corpse of a star would literally burn to the end of times and quite possible be the last light emiting thing to shut down in the universe.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .