Reply #135: Coal ash consists mainly of compounds like silicates, alumina, and iron rust:

[struggle4progress]

a rather glassy or ceramic material, which has been formed at high temperature in an oxidizing environment, so it won't be very reactive. The first challenge is to extract a trace element from it

Coal ash is (say) 10 ppm natural uranium. A good quality uranium deposit is about 20% U308 -- say, 20 000 times richer in uranium than coal ash. The chemical problem of extracting an element, from a sample which is 20% of that element, is quite different from chemical problem of extracting an element, from a sample which is 0.001% of that element. Coal ash contains almost everything at low concentrations, so in the initial stages of a separation attempt, you're going to get a "soup" that contains all manner of stuff at very low concentrations. To overcome the entropic barrier presented by the extreme dilution, you will need some very favorable reactions

Weapons-grade uranium is about 85% U-235, with a critical mass of some tens of kilograms. Natural uranium is about 2% U-235 49% U-238, and 49% U-234. Thus, you need to start with at least 40x more natural uranium than the amount of weapons-grade uranium you hope to obtain

What's it going to take to produce ten kilograms of weapons-grade uranium from coal ash? At 10 ppm natural uranium, you can't get more than 10 g natural uranium from a metric tonne of coal ash, so 10 kg of natural uranium requires at least 1000 metric tonnes of coal ash; multiplying by 40, you'd need at least 40 000 metric tonnes of coal ash to produce ten kilograms of weapons-grade uranium. The actual numbers will be much worse, since you cannot expect quantitative extraction of a trace element, and you can't expect easy isotopic separation. You're actually contemplating a very substantial industrial enterprise

For perspective, consider this: the average abundance of uranium in crustal rock is about 2.5 ppm. If you can figure out a feasible way to extract uranium from coal ash, you can probably figure out a feasible way to extract uranium from most rocks: there's only a factor of about four in the trace concentrations