[Kryptid (OP)]

I don't expect there to be any naturally-occurring stars that are powered by nuclear fission (fissile materials are simply too rare and dispersed in the universe), but could one be constructed in theory? Imagine some kid from a million-year old, ultra-advanced alien civilization wants to make one for a science project or something. What would be a good way to proceed? Create a giant sphere of fissile uranium? Would it auto-ignite? Would it simply explode from the exponential fission of more and more atoms? Could this be controlled by mixing the uranium with a non-fissile material?

I tried to do some research on the subject of nuclear fission myself, but I came up with a problem: I can't quite figure out what circumstances make a simple criticality event (such as during the accidents with the "demon core":http://en.wikipedia.org/wiki/Demon_core) different from a genuine nuclear explosion. Surely it's the rate of fission that makes the difference, but how do the initial circumstances lead to the different rates? A star is basically an explosion that is contained by gravity, so the fission star would have to be exploding in the same sense as a nuclear bomb. If it is merely "critical", then it will only be a giant, solid sphere of uranium-235 with a surface temperature of a few hundred degrees Celsius.

I also realize that the feedback loops which make a fusion star stable would not operate the same with a fission star. In a fusion star, gravity and fusion balance each other. If the star becomes larger, then the temperature and pressure drops so that no fusion occurs and gravity begins to contract it sufficiently to reignite fusion. If gravity compresses it too much, then the heat of fusion causes it to expand back out.

It's different with fission, as heat and pressure are not requirements for its initiation. In fact, large amounts of heat can reduce the rate of fission. Interestingly, this itself could be used as a feedback loop. If the star becomes too hot, fission drops and it cools down. What temperature would that be at? Much lower than a typical star?

[alancalverd]

To some extent, you have answered the underlying question already! Some water-moderated fission reactors (I think CANDU was one such system) are inherently stable as any temperature rise reduces the efficiency of the moderator, so reducing the thermal neutron flux. You can set the operating temperature by design. Not a star in the conventional sense, but a source of heat and ionising radiation nevertheless. What more could one ask for?

[CliffordK]

One of the regulators in a large scale nuclear fission system would be nuclear (neutron) poisons.  It is possible, however, if a star was sufficiently large, that to some extent the core would layer by density which might help purify the fissile material to the point where it would reach criticality...

Perhaps fission would tend to disturb the layering of the core, and thus cause a feedback loop.

[evan_au]

You need a mechanism to create large quantities of fissile material. Normal stars will only fuse hydrogen to produce elements up to iron which is very stable; it actually consumes energy to produce elements heavier than iron.

But there is such a known mechanism - its a supernova. During the explosion of a supernova, elements blasted outwards by the star are bombarded with neutrons and protons, leading to a rapid increase in atomic number, almost certainly beyond the uranium and radium which occur naturally on Earth, perhaps up to element 130.

Astronomers study the light curve of supernovae, and during part of the decay, the brightness of the supernova decays at the rate you would expect from decay of cobalt 60. Other, earlier parts of the light curve are no doubt dominated by rapid breakdown of even more unstable heavier elements. This really is a star (or stellar remnant) powered by decay of heavy elements, rather than fusion of light elements.

The nuclear soup contained in the expanding shell of gas around the star, and on the remnant of the stellar core would participate in a complex series of nuclear reactions, but probably quite dissimilar to our controlled nuclear reactors.

[bem13]

Hi, this was an idea that interested me also.
My nuclear physics is very weak but I wanted to add what I think are some common sense points and then blather a bit.

Firstly, definition. By a fission star, I would mean something that could shine brightly enough to illuminate a planet but is powered by fission. Also it has no machinery, it is just a big ball of molten ingredients. And it can run without supervision or repair for at least thousands of years.

Size isn't too relevant. Fusion powered stars are only massive because fusion requires it. It could be moon sized and at a moon-like distance from the planet. This would be a great practical joke to play on any budding young astronomers below.

It should radiate at a visible wavelength. It is actually quite common for planets to release more energy than they receive and one of the reasons is internal radioactivity. If we included things that radiated any wavelength at all due to internal fission then you could call pretty much anything a fission star.

A lot of the discussion so far considers how to moderate a fission reaction but I think mere nuclear decay could probably suffice. Because volume grows faster than surface area, I think a radioactive decay that merely warms a cubic meter of ore could create very high temperatures at the surface of a planet-sized object. It is probably not hard to do the math to calculate the sort of heating for a planet of a given radius to create visible wavelength temperatures. You would not need to worry about runaway reactions in this case, I think.

Lastly, this would be a lot more interesting if we could find a natural phenomenon to produce such an object.

I guess what you want to do is really pound an object with neutrons until it glows in the dark while not blowing it totally away. This is totally uninformed speculation, but perhaps some sort of freak celestial mechanics with neutron stars and Roche limits could create this environment?

Another random speculation: Just as planets tend to lose helium and hydrogen most easily, perhaps some freak balance of heat and gravity could reduce a planet to only the heaviest elements? My assumption is that biasing towards the heaviest elements increases the proportion of radioactive elements. This may be sufficient given the discussion on radioactive decay and volume above.

[PmbPhy]

I don't know about a star per se but naturally occurring nuclear fission was found  in 1972 at Oklo in Gabon, Africa, by French physicist Francis Perrin.

See http://en.wikipedia.org/wiki/Natural_nuclear_fission_reactor

[Notso]

Simplified case of a uranium fission reactor:

Nuclear fission reactors split uranium atoms with thermal neutrons. On average, the by-products are two lighter atoms and 2.5 thermal neutrons (number recalled from A-level physics textbook, don't quote me on it). In order to maintain a stable reaction, the number of thermal neutrons which go on to cause other uranium atoms to split must be kept at an average of exactly 1. Any more and the reaction rate goes exponentially up and we get a nuclear bomb, any less and the reaction slows exponentially to a standstill.

The problem is that the atomic products of the fission reaction will inhibit the passage of thermal neutrons which would need to go on to split other uranium atoms. If we start with a "perfectly" distributed star of fissionable material that would allow an average of exactly one neutron from each fission to create another fission event, then the reaction will eventually falter.

[evan_au]

this would be a lot more interesting if we could find a natural phenomenon to produce such an object.

This year*, it is thought that elements much heavier than iron (eg gold, uranium and the other radioactive elements) could be sprayed into space during the colission of two neutron stars. This would produce a huge array of nuclear debris spanning a huge range of isotopes (most of which would quickly decay into the familar elements). Of interest to nuclear enthusiasts, the fissile U235 would initially be as common as the more stable U238.

Neutron stars are thought to start with temperatures in the millions of  degrees. After an energetic colission like this, they would be back to a temperature of millions of degrees, which would certainly be enough to give you a rapid suntan (and the X-Rays would also tan your internal organs...)

It is conceivable that some of the debris from the colission could coalesce in orbit around the now-merged neutron star. This could form a mass the size of a moon, a planet or perhaps even a star. This matter would be very hot, and would be continually generating considerable heat due to nuclear decay (fission of heavy elements).

However, this would not form a nuclear chain-reaction. There are a few isotopes that will promptly release more neutrons than they absorb. But most isotopes will absorb more neutrons than they produce, and most daughter nuclei will absorb neutrons. Forming a steady-state nuclear reactor requires isotopically enriched fuel.

From this I deduce that you could form a white-hot glowing moon, powered by nuclear fission. But it would release most of its energy by spontaneous fission of heavy nuclei after their half-life, rather than the decay being prematurely triggered by a self-sustaining nuclear chain reaction. But such a moon would go dark long before a small star ran out of hydrogen.

* The previous theory, that heavy elements were formed in a supernova explosion fell out of favor because there apparently aren't enough neutrons released in a supernova to form stable isotopes of these heavy elements.

[evan_au]

By the way, if there were a lot of fissile material floating around in space which started to get together, every time it approached a critical mass, it would promptly disassemble itself in a sub-critical explosion.

Depending on the conditions, critical mass could occur for bodies of a few kilograms up to a few tons - far too small for gravitational confinement.

[yor_on]

http://www.thenakedscientists.com/forum/index.php?topic=52835.msg444626#msg444626

maybe this guy?