Ellingham diagram by Wikipedia user DerSilberspiegel, CC-BY-SA 4.0
I was reading about cold plasmas after learning about rosining chips. They're already used in many applications (1 ppm ionization is becoming commonly used for sterilizing in medicine, for example) but I was thinking about a few more, including a possible low-temperature mostly electrolytic route to rutile reduction that doesn't have the low cathode efficiency problems that plague traditional molten-salt electrolysis.
I'm pretty sure a cold plasma of air can eat epoxy, and I suspect that a cold plasma of just oxygen could do so. Ozone is mostly known to attack olefin double bonds, which by itself would not be sufficient to eat epoxy, but I suspect ozone and the other reactive oxygen species in the cold oxygen plasma would be sufficient.
Ionizing air generates oxides of nitrogen, which tend to not only oxidize things (more aggressively even than ozone) but also to nitrate them. I suspect this would be fine for decapping chips; aluminum, silicon, and copper are all relatively inert to such materials. However, at temperatures below 100°, zirconia may be vulnerable, and perhaps thus also hafnia. Typical arcs in electric arc furnaces can reach a few hundred ppm of oxides of nitrogen, and rarely go below a few tens of ppm. This suggests that to produce the ten grams or so needed to decap a chip you'd need to run tens of kg of air through your plasma pencil, most of which of course won't react, so you'd probably need tonnes. This seems slow but perhaps a feasible approach.
(Of course if you have a drop of water on the object, that will facilitate the attack of oxides of nitrogen on many materials, including copper, zirconia, and hafnia; though, perhaps, at first, before the water has absorbed much of the gas, it might have a protective effect instead.)
I've written before about using an air cold plasma with simply a glow discharge for selective functionalization, for example for selective electroless metal plating or selective wetting of otherwise nonreactive surfaces.
A cold plasma of steam would be easier to produce than that of oxygen, and it would have a similar oxidizing effect on some materials. However, I think the hydrogen ions would prevent other materials, such as many metals, from oxidizing. This could be used in some cases for more selective oxidation.
A potentially more interesting application of cold plasmas is reduction with just hydrogen. It's been routine since at least WWII to anneal the iron powder in hydrogen before using it in powder metallurgy, to reduce the oxide film from its surface and enable it to cold-weld and sinter. Similarly experiments have successfully extracted oxygen from lunar regolith simulant by reducing it with hydrogen. Surely a similar effect can be achieved with a cold hydrogen plasma.
For example, you could pack a powder bed with a powder of an iron
oxide, infuse it with hydrogen at above 100° to drive out all the air,
insert an iron electrode into it, and apply a high-frequency high
voltage to the electrode. This would produce a hydrogen plasma in
contact with some of the powder, and some of the iron oxide would be
reduced, producing steam, which would diffuse away from the electrode
and be replaced by fresh hydrogen to continue the process. The iron
electrode would thus gradually grow dendritically through the oxide
powder, though not converting all of it. Some flow of hydrogen would
be useful to flush out the steam and prevent it from re-oxidizing the
iron; alternatively some kind of desiccant such as calcined alabaster
could sequester the water, or a more easily oxidized metal could
reduce it.
Ellingham diagram by Wikipedia user DerSilberspiegel, CC-BY-SA 4.0
Looking at an Ellingham diagram with hydrogen to steam ratios, it seems that for hydrogen reduction of magnetite to hematite, the equilibrium favors reduction as long as there is less than about ten thousand parts of steam to one part of hydrogen, almost independent of temperature over the usual ranges; and for hydrogen reduction of iron oxide to iron, about 10× as much hydrogen as steam is needed, again almost independent of temperature. Unless I'm reading this diagram wrong.
Copper, cobalt, and nickel seem similarly simple. Other metals are a bit less so; zinc seems to require a bit more hydrogen than water, and that only at 1100°. Getting down to below zinc's melting point requires a million-to-one hydrogen-to-steam ratio, and getting down below 200° requires a trillion to one. At silicon's melting point of 1460° (XXX Pyrolysis 3-D Printing says 1414°, maybe this is the wrong temp) we should be able to reduce it with hydrogen at ten thousand times the concentration of steam, and at a trillion to one this reduces to a balmy 600°. Rutile at its melting point above 2300° can be reduced with hydrogen at about 3000 times the concentration of steam, but at a more comfortable 1000° it's a billion to one, and at a trillion to one we're down to about 600°, a bit higher than silex. (Damned Tuftean plots with no grid lines.) Even at these trillion-to-one levels, sapphire doesn't yield to hydrogen's persuasion until past 800°, magnesia alba until 1000°, and lime until 1200°.
(Maybe in some cases you'd get the hydride of the metal rather than the metal itself.)
Actually, lime is particularly interesting in this connection because, as in the Pidgeon process, it can combine with silex into the stabler larnite. This reaction can work as a sort of desiccant, in effect allowing metallic silicon (or ferrosilicon) to reduce most metals, even very active ones like magnesium, again as in the Pidgeon process. Facilitating this reaction with ionized hydrogen rather than hellish temperatures seems potentially useful.
Ferrosilicon is normally obtained by carbothermal reduction. I wonder if you could go further, passing a cold plasma of hydrogen contaminated with steam over carbon at some more everyday temperature, to increase its hydrogen-steam ratio before passing it over the metal powder again? You could use the carbon itself as an electrode to ionize the water.
A more quotidian approach to removing water would be to cool the hydrogen–steam mixture, pass it over a garden-variety low-temperature desiccant such as alabaster, muriate of lime, or quicklime, and then heat it up again to the reaction temperature before reionizing it. Carrying out the cooling and heating steps with countercurrent heat exchangers or regenerators would eliminate their unnecessary energy consumption.
In such a case, where does the energy come from? We apparently have hydrogen circulating in a closed loop at constant pressure to reduce an oxide, say silex or rutile, to its base metal, which entails adding energy to it — if we burn the metal we will get the energy back. The desiccants are losing energy by being hydrated, but not nearly enough to reduce the metals. The mystery is solved, though, when we observe that we must continually add hydrogen to the system if we are to prevent its pressure from dropping. The energy to reduce the metals was the chemical potential energy in the hydrogen.
In the case of the desiccants, one is left to wonder what to make the heat exchangers out of if the reaction gas is corrosive at only 800° to even sapphire and tends to oxidize metals. Zirconia, perhaps, or some kind of highly refractory carbide or nitride: BN, TiN, WC, HfC, the usual suspects. Or maybe just aluminum, if it's countercurrent: pipes with enough water vapor will be lined with amorphous sapphire the way aluminum normally is, while pipes without will eventually just be bare aluminum, but neither will corrode. Other metals that form similar passivation coatings serve just as well at higher temperatures that aluminum can't handle.
So, as an example, maybe you can reduce a packed bed of powdered rutile at, say, 800°, in a hydrogen atmosphere at atmospheric pressure, by applying a high-frequency electrical charge through an electrode to ionize the hydrogen, while maintaining the hydrogen very dry (better than 100 billion to one ratio to the water) by passing it through a desiccant such as alabaster at a lower temperature such as room temperature, say 20°, with a countercurrent heat exchanger in between the rutile chamber and the alabaster chamber to maintain them efficiently at different temperatures, plus some additional heating to maintain the rutile at its high temperature and some additional heatsinking to maintain the desiccant at its low temperature, while continually adding new hydrogen to the system to replace the hydrogen absorbed as water in the desiccant. The energy to reduce the rutile comes principally from the hydrogen, whether that is produced by electrolysis or, for example, from natural gas.
Hmm, shit, that desiccant probably can't get the hydrogen that dry. So you probably need a somewhat higher temperature. Or maybe if you cool the desiccant more, or use a more aggressive desiccant, you can get down to those levels. Obvious candidates include sodium (as in the Hunter process), calcium (as in the Kroll process), ferrosilicon with lime (as in the Pidgeon process), and of course lithium, magnesium, or aluminum.
Solid magnesium and aluminum have the annoying problem of forming an adherent solid oxide film when oxidized, preventing them from reducing any further water; as outlined in Petrovic and Thomas's 2008 "Reaction of Aluminum with Water to Produce Hydrogen: A Study of Issues Related to the Use of Aluminum for On-Board Vehicular Hydrogen Storage", approaches to solving this problem for aluminum include "hydroxide promoters such as NaOH, oxide promoters such as Al₂O₃, and salt promoters such as NaCl", but I think all of those are necessarily in aqueous solution, and would thus produce too much contaminating steam of their own. Other approaches might include maintaining the desiccant metal molten, with a layer of flux salts on top to keep the oxides molten, and bubbling the gas through it to deoxidize it. Magnesium melts at 650°, aluminum at 660°, their eutectic of about 65% aluminum at a pleasant 437°, but as explained above, even at these temperatures, not much oxygen will escape to reoxidize the hydrogen. Alternatively, you could disrupt the oxide layer with plasma, perhaps using the very same hydrogen gas or perhaps using a more conventional sputtering gas like argon, which would of course then be mixed in with the hydrogen.
In this form, you're essentially performing an aluminothermic (or magnesiothermic) reduction of rutile, but with hydrogen acting as a catalyst (instead of, as previously, a fuel.)
Higher pressures would tend to increase the reaction rate, but I don't think they'll affect the equilibrium of the metal reduction much, because both hydrogen and steam are gaseous, with the same number of moles as the resulting water. If you instead used a reducing gas with a different number of hydrogens, such as methane, ammonia, maybe hydrazine if it can stand the heat (silane can't), nitric oxide, or even plain nitrogen, you might be able to use pressure to shift the equilibrium. (Probably by reducing the pressure.) This might allow you to tolerate a larger percentage of water vapor in the system before the metal stopped reducing.
However, higher pressures do beneficially affect the relationship between the oxide equilibrium and a desiccant equilibrium! That's because at a certain temperature, a certain percentage of desiccant being hydrated corresponds to a certain partial pressure of steam, and this is unaffected by the partial pressure of hydrogen. So, by increasing the pressure, you proportionally increase the proportion of hydrogen in the desiccated gas.
Zircon, zirconia, coltan, and molybdenite could likely be reduced in the same way.
If you have an unlimited supply of sufficiently dry hydrogen, the unused hydrogen contaminated with steam could be disposed of by simply passing it through a flame. Ozone and oxides of nitrogen are perhaps not as easy, though natron water should serve to some degree, and automotive catalytic converters are a common solution to precisely this problem. Olefins such as ethylene, propylene, or hexene would probably also eliminate ozone and oxides of nitrogen, and could then be burned with impunity. 2-methyl-2-butene is used in such a way as a free radical scavenger.
Diesel engines commonly spray urea water ("diesel exhaust fluid" or "AdBlue" or "Azul 32") into the exhaust to eliminate nitrogen oxides instead of using a catalytic converter. This also eliminates ozone. Urea is nontoxic and quite cheap; the solution is sold by the ten-liter bottle. The final resulting gas mix still contains nitrogen dioxide, though.