Related to this question: Gaseous metals?

The top answer clarifies that even with metals, the gaseous state is distinct from a plasma.

However, my question is, is this actually true in all cases? Or is there an element and a pressure for which heating it to a gas will necessarily also be hot enough for it to immediately become a plasma? Or is there a situation in which heating can bring about the transition from solid or liquid straight to plasma while skipping gas?

  • $\begingroup$ Well, usually plasma is considered different state of matter then gas, just a nitpick ;) That being said elements with high boiling points under high pressure could go directly from liquid to plasma, I think. There are also more "arcane" ways to make plasma, even very cold one. $\endgroup$
    – Mithoron
    Commented Nov 17, 2018 at 1:51
  • $\begingroup$ The line between neutral gas and plasma is a blurry one. Gases have different amounts of ionization, from almost entirely neutral to almost entirely ionized, and thermal ionization gradually goes up as temperature increases. Also, you may want to look back at "Gaseous metals?" if you haven't since I wrote my answer, since I discussed this a bit and argued (perhaps not very clearly) that heat capacity data probably showed evidence for some ionization or at least electronic excitation in gaseous calcium and strontium at their 1 bar (normal) boiling point. $\endgroup$
    – Mr. Nichan
    Commented Apr 17, 2021 at 10:33

2 Answers 2


In principle, all that is required is for a substance to evaporate from a liquid into a plasma directly is for its ionisation energy to be significantly smaller than the average kinetic energy of its particles at its "boiling point". Or in other words, as a very rough estimate this is true when $\mathrm{k_BT_{boiling}≳EI}$.

Going further down in the periodic table, elements tend to have higher boiling points and lower ionisation energies. After reaching the superheavy element region (Z>100) things get much more complicated due to strong relativistic effects, but the elements from Rf to Hs (Z=104 to 108) are expected to have the highest boiling points at ambient pressure of any known chemical substance, possibly well past $\mathrm{6000\ K}$. However, their calculated ionisation energies of $\mathrm{6-8\ eV}$ suggest they would only ionise considerably at temperatures 10 times higher still ($\mathrm{1\ eV/k_B= 11605\ K}$), meaning there is still a long way to go.

As others have mentioned, increasing the pressure is a reliable way to push boiling/sublimation points higher, up to where they can match and surpass the energy regime required for ionisation. However, this would require absurdly high pressures; for example, the boiling point of tungsten at $\mathrm{1000\ bar}$ is still a mere $\mathrm{9000-10000\ K}$ (ref). Thus, a very crude extrapolation would suggest the direct conversion of liquid (or more likely solid) tungsten to a plasma would happen around $\mathrm{90000\ K}$ under a pressure of $\mathrm{20\ Mbar}$ (20 million atmospheres!).

  • 2
    $\begingroup$ The 90000 K 20 Mbar boiling of Tungsten you imagine is probably impossible, since that is probably far beyond the critical point of Tungsten, based on the less crude extrapolations of Leitner and Pottlacher: researchgate.net/publication/… and previous literature they mention. In the supercritical regime, it would just gradually shift from liquid-like to gas-like density. $\endgroup$
    – Mr. Nichan
    Commented Apr 17, 2021 at 11:01

Yes! If you look at a phase diagram of an element, you will see that with extremely high pressures, extreme heat will keep it a liquid. Eventually, theoretically with some elements, a phase change could occur directly from liquid to plasma, but the pressure would have to be incredibly high. This is caused from the electrons moving too far away from the nucleus.


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.