Smelting using focused solar radiation against a spinning carbon crucible would be cheap, mechanically simple, and effective. I’m not sure why you think you can’t refine in space.
It does not work for asteroid mining.
There you have just pieces of iron with a low content of alloying elements.
Among the alloying elements, nickel is the most abundant (17.3 times less than iron), then cobalt (20.9 times less than nickel), then germanium (20.4 times less than cobalt).
The precious metals that would justify the mining operation are present as a few grams each for a ton of iron.
Melting the alloy will never separate the metals by itself.
However perhaps some kind of floating zone melting could enrich the proportion of precious metals in a part of the iron, but it is very unlikely that a high enough enrichment could be achieved by a reasonable number of zone melting passes.
On earth the cheaper metals could be dissolved by an acid solution, but on metallic asteroids you have neither water nor acids.
The SciFi solution would be to vaporize and ionize the metal alloy and separate the metal ions by their specific charge, like in mass spectrometry.
This would need a lot of energy, but at least it does not need reactants and it is the only method that achieves almost perfect separation.
However for now, the throughput of such a ionic separator is extremely small, too small for industrial production. Perhaps it will become possible to scale such ionic separators to acceptable productivities.
There are no materials from which you could make a centrifuge for gaseous osmium and iridium.
Vaporizing the input metal with an electron beam and ionizing the vapors allows after that contactless interaction with the ions, using electric fields and magnetic fields, guiding them into separate condensation chambers (which need strong cooling).
The ionic current of such separators must be increased several orders of magnitude over those currently existing, for this to become a viable separation technology.
Of all metals that could be extracted, osmium and iridium are those for which there is the greatest difference in abundance between the surface of the Earth and the metallic asteroids.
If the temperature is not high enough to vaporize the platinum-group metals, they will remain as solid grains, which are likely to damage any centrifuge.
Making a centrifuge from a ceramic material like hafnium carbide is unlikely to work, due to its fragility and low tensile strength, especially at very high temperatures. A ceramic coating of the parts in contact with the hot gas might work, but even if there is a lot of experience in making such things nobody has made parts working at temperatures so high as needed for this application.
The problems for making such a gaseous centrifuge are similar to those for making a high temperature gas turbine.
During the last century, tremendous resources have been dedicated for increasing the maximum temperature of gas turbines. The working temperatures have been slowly increased, but more and more slowly in recent years and there is very little hope that it is possible to increase the working temperatures much beyond what has been already achieved.
A metal separation centrifuge would require working temperatures not higher by 10% or by 50%, but temperatures at least 3 to 6 times higher than for the existing gas turbines.
Based on the existing experience in improving gas turbines, I believe that this separation method can be safely dismissed.
While a centrifuge is not feasible, there are chances to use a part of your proposal.
Heating the metal alloy at a temperature high enough so that iron will sublimate quickly (together with nickel, cobalt, gold etc.) while platinum-group metals will sublimate very slowly could produce an alloy highly enriched in platinum-group metals with a mass many times lower than the starting mass.
This method cannot separate any individual metals and it would lose the gold and other possibly useful components, like germanium, gallium or cobalt, but it could reduce the mass enough so it may make sense to take the concentrated alloy with platinum-group metals away from the asteroid, to a place where it could be processed with more selective methods.
While such a method has some small chances of being profitable, it is very wasteful. Real asteroid mining must separate the metal alloy in all its components, because all can be very valuable, less for being brought back on Earth, but for building any kind of structures in space or on other planets/asteroids.
Another idea, taking advantage of the vastness of space and relative masses in another way. Rather than containing the gaseous material could you not vaporize a large amount of material and linearly accelerate it at the same time by a fixed amount. Maybe even a high powered laser could accomplish both at once without physical contact, or a laser combined with a magnetic field. The mass of the heavier metals will mean their ultimate velocity will be significantly less than the lighter ones. Over some distance they would condense back to liquid then solid but would have striated and continue to separate in distance as the relative velocities continues to pull them apart in distance. You could even do this by producing a pulsed beam of material moving towards earth from the mined asteroid. Closer to earth you would collect material in order of arrival and separating them into bins by expected relative arrival time by elemental mass. Would this not lead to a pretty refined mixture and require no physical contact?
To your point about the sublimation points being different slowly heating the material while applying force to the vapor would also increase the separation in space, and leaving some highly concentrated high vapor point platinum group residual alloy to be refined on earth - maybe this would be considerably less wasteful as you would capture everything at the collection point relatively separated with no exotic materials or centrifuges?