14
$\begingroup$

Liquid chromatography is used to analyze and purify different chemical samples in a solvent. It is also used for industrial scale chemical purification. I was wondering if something similar exists for liquid (generally molten) metals.

I have not been able to find any examples online, but I may be missing some key vocabulary. It seems there would be some big advantages to purify metal in this way. Lead for example has impurities of "arsenic, antimony, bismuth, zinc, copper, silver, and gold" per the lead wiki article. It also goes on to state that the impurities are typically removed via pyrometallurgical processes; which appear to be very energy intensive. I imagine liquid chromatography type purification process would work well for other metals with a lower melting point such as gallium, mercury, zinc, tin, and cadmium.

I imagine the solid material that would fulfill the role of the resin could be any material with a high amount of wettable surface area and can handle the elevated temperatures. Potentially could be zeolites, activated carbon, graphene, metal powders, fiberglass? I imagine selecting a material that is properly wetted by the liquid metal would be the trick.

Does this exist? If not, any thoughts on why it might or might not work?

Improve this question
$\endgroup$
14
  • 5
    $\begingroup$ I have to say this is a curious idea, but it appears to be fraught with very difficult problems, both from an implementation and a concept standpoint. For one, chromatography isn't all that useful for purifying substances deeply, but it is good at purifying them broadly. In a crude sense, in synthesis there are a combinatorially-exploded number of possible molecular substances which require differentiation (among small subsets of the whole, but very very many subsets) , but there aren't so many metals, especially if you're talking about just the elements or simple combinations of them. $\endgroup$
    – Nicolau Saker Neto
    Commented Jan 23 at 8:15
  • 2
    $\begingroup$ Chromatography is a means of partitioning a mixture into various components. It can be applied to attempt to isolate one component, and once isolated it might be deemed sufficiently pure. Washing steps such as in organic synthesis (eg separatory funnel) might be closer to what you are thinking of. $\endgroup$
    – Buck Thorn ♦
    Commented Jan 23 at 9:10
  • 5
    $\begingroup$ Metals that are easy to liquefy (mercury, gallium, alkalis metals...) are easy to purify by distillation, so why bother with chromatography? Metals that are not are going to need a lot of heat, so the process will use just as much energy as pyrometallurgical processes so there is no clear energy benefit. $\endgroup$
    – matt_black
    Commented Jan 23 at 11:16
  • 2
    $\begingroup$ Chromatography is used to purify substances, IF there is no other way to do so, and the target material is extremely valueable. Not 30€ per gram, more like per milligram. Or scientifically valueable. $\endgroup$
    – Karl
    Commented Jan 24 at 20:42
  • 2
    $\begingroup$ @matt_black, Looking at the example of Lead: "Specific heat of Lead is 0.13 J/g K. Latent Heat of Fusion of Lead is 4.799 kJ/mol. Latent Heat of Vaporization of Lead is 177.7 kJ/mol." There are clearly other difficulties with this idea as pointed out by everyone here, but there is enormous energy savings potential. $\endgroup$
    – ericnutsch
    Commented Jan 26 at 5:50

4 Answers 4

Reset to default
12
$\begingroup$

An obstacle of implementation is that the blend of metals has to be kept liquid all the time the chromatography runs, top to bottom of the column, while the equipment around has to sustain the mechanical strain and stress at elevated temperatures without reacting with the sample (which could cause new contamination, e.g., by formation of alloys). With cryoscopy in mind, it is likely metals then purified have a higher melting point, than the submitted sample; potentially calling for additional heating to keep the process flowing.

Activated carbon and graphene you mention as candidates for the stationary phase could oxidize if the process is not anaerobic ($\ce{C + O2 -> CO2}$, and $\ce{CO2 + C <=> 2 CO}$), and already traces of carbon can substantially affect mechanical properties of metals (cf. $\ce{Fe}$ vs. $\ce{Fe3C}$ and its subsequent products). Money aside, not every gas can be a suitable inert gas either (nitrides, e.g., $\ce{3 Mg + N2 -> Mg3N2}$). Metal powders: potential creation of solutions, alloys, and new compounds. There may be obstacles with zeolites and glass (what about the silicon in these silicates?), while fiberglass is a composite of glass fiber and some plastic polymer, the later possibly burning away, too.

A technique however used in e.g., semiconductor industry to further purify metals in the liquid state is zone melting (or zone refining). This may work if thermodynamics are favorable (prompt: phase diagrams) and the starting material already is of somewhat high purity. E.g., silicon which underwent a purification via distillation of intermediate $\ce{SiCl4}$. Contrasting to chromatography, zone melting usually does not even attempt to split a mixture into multiple separate fractions A, B, C, D. Yet, the directed crystallization of zone melting does not allow you to pass beyond an eutectic minimum (point C below), for instance to transform cast iron into steel, let alone pure iron.

enter image description here

(source)

Improve this answer
$\endgroup$
2
  • $\begingroup$ Presuming you could address the engineering difficulties associated with heat, structure, oxidation, stationary phase selection, and alloying; would it be reasonable to expect to see a difference in migration rates of the different metals? $\endgroup$
    – ericnutsch
    Commented Jan 27 at 1:02
  • 1
    $\begingroup$ @ericnutsch I speculate this taps into recording diffusion coefficients (e.g., diffusion coefficient of liquid potassium into liquid sodium) e.g., by use of radioactive potassium as a marker, in contrast to e.g., diffusion of iron in liquid sodium. Perhaps there is an easier way to monitor the advance of a concentration profile in melts, larger atoms might advance slower than smaller ones as they experience more dynamic resistance and have a greater inertia. $\endgroup$
    – Buttonwood
    Commented Jan 27 at 18:00
9
$\begingroup$

Molten liquid chromatography does not exist currently. However the problems you mentioned for purifying metals is very conveniently done using ion-exchange chromatography. A lot of difficult to separate metals (rare earths) are separated or historically separated that way.

Given the high viscosity/ density of liquid metals diffusion of "analytes" would be slow, which means mass transfer kinetics between mobile phase(=liquid metal) and your solid refractory resin would be very slow. If we had a liquid metal as a mobile phase, the LC separation would be horribly bad and poor for these reasons.

Improve this answer
$\endgroup$
1
  • 1
    $\begingroup$ Could you possibly put some ballpark numbers to it? or direct me to where to calculate the diffusion? As long as the transfer kinetic reduction was on the same order of magnitude as the diffusion reduction, I think there could still be some potential. $\endgroup$
    – ericnutsch
    Commented Jan 27 at 1:08
4
$\begingroup$

Not really, but you can identify metallurgical processes that rely on principles of phase separation and immiscibility of different components in molten salts, for instance during the production of sodium metal and chlorine gas from NaCl. In that process molten NaCl is subjected to electrolysis and the less dense molten sodium produced rises to the top and flows out of the chamber one exit while chlorine gas is evacuated separately. This is illustrated here:

enter image description here

Improve this answer
$\endgroup$
1
2
$\begingroup$

Liquid chromatography is used to analyze and purify different chemical samples in a solvent... I was wondering if something similar exists for liquid (generally molten) metals.

The detail to remember is that, in molten metal, the metal atoms are separate individual molecules rather than being parts of larger complex compounds. As such, a Lead atom's molecular size and shape is near-identical to that of a Bismuth atom, and so its travel-time through a Liquid Chromatography column would be imperceptibly time-differentiated from that of the impurity.

This idea would essentially require an isotope-differentiating level of discrimination from a organic-molecule-scale liquid-phase column technology - basically trying to do Mass Spectrometry separations in a benchtop Gel Electrophoresis type device.

If the metal atoms weren't actually "molten" but rather just complexed with organic compounds in such a way as to remain mobile in the liquid phase then you could remove the high-temperature engineering issues and would be working with larger molecules better suited to the technology.

Of course, if the same single metal-complexing molecule is used to hold both the pure metal and impurity atoms, then you're back to a situation of needing isotope-differentiating resolution on your column (to separate the MoleculeA-Lead complexes from the MoleculeA-Bismuth complexes). On the other hand, if you use multiple complexing agents that will preferentially attach to pure metal and impurity separately (e.g. MoleculeA-Lead and Molecule-Bismuth) then you already have a chemical process that can discriminate the target molecule from impurities and could likely engineer a large-scale phase-separation process more efficiently than trying to do separation by a HPLC column time-of-flight differentiation process.

Improve this answer
$\endgroup$
3
  • 1
    $\begingroup$ Chromatography doesn't just separate by size and shape, it can also separate by physical properties. (e.g. Organic HPLC typically separates mainly based on hydrophobicity rather than size or shape.) If the interaction of the lead atoms with the stationary phase were different than bismuth atoms, then they could potentially be separated. (As the one which interacts more tightly would be delayed more as it flowed through the column than the one which interacts less.) $\endgroup$
    – R.M.
    Commented Jan 23 at 19:36
  • $\begingroup$ @R.M. You are absolutely correct, but I believe that "size and shape" was perfectly reasonable for an ELI5-like first-order approximate explanation. $\endgroup$
    – DotCounter
    Commented Jan 23 at 19:45
  • $\begingroup$ @AmateurDotCounter Beside chromatography mostly based on adsorption/desorption, and partition between mobile/stationary phase, there equally is the size exclusion chromatography. Here, larger polymer molecules typically elute ahead of the smaller ones. But the sizes of the polymer molecules in question (and their differences) are larger, than individual metal atoms of a melt. $\endgroup$
    – Buttonwood
    Commented Jan 23 at 21:13

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.