Metals form metallic bonds which explain many of their chemical and physical properties.

This is most familiar in the solid state but metallic properties are still quite recognizable in the liquid state e.g. mercury and molten iron.

However, gaseous metals are much less familiar and it is not obvious whether they would still have a distinct metallic behaviour. For example, is there a distinction between gas and plasma for a metal?

This article mentions that the metallic bond persists in the liquid state but says nothing about the gas state: Metallic bonding at Chemguide.

So, do metals retain any distinct metallic behaviour in the gas state?

Note that I am not asking only about the electrical properties. My main question is whether the gas consists of neutral atoms, small molecules, or positive ions in a sea of electrons?


2 Answers 2


No, gaseous metals do not retain metallic bonds, nor metallic conductivity, nor luster, nor any other metallic properties. They are no different from other gases. True, they typically require pretty high temperatures to form, but then again, they are hardly special in this regard, as many non-metallic substances require the same.

See, all metallic properties are in fact collective effects. They are caused by metallic bonding, and not just of two atoms, but of an entire piece of metal (like a thousand, only much more). You don't have bonding in gases. The particles (lone atoms in this case) are basically free to go. Maybe they bump into each other, two or three at a time, but definitely not a thousand.

Plasma is an entirely different thing, unrelated to your initial question. Yes, plasma is pretty different from a gas; to obtain it, you'll have to heat the gas until it ionizes.

So it goes.

  • 3
    $\begingroup$ There could be molecules/small clusters, but that would not be enough for properties we'd recognize as metallic. $\endgroup$ Nov 16, 2018 at 12:12
  • 1
    $\begingroup$ Thanks. I suspected something like that but wanted confirmation. So, we can have gaseous iron which is not a plasma? Just neutral iron atoms (or small clusters) rather ions in a sea of electrons. $\endgroup$
    – badjohn
    Nov 16, 2018 at 12:19
  • 1
    $\begingroup$ Yes we can.$\,$ $\endgroup$ Nov 16, 2018 at 12:25
  • $\begingroup$ About molecules of metal atoms in the gas phase, I know that Li2 (real dilithium) exists, since it's discussed in my chemistry textbook and plenty of other sources. A look at the citations of this page on Li2: cccbdb.nist.gov/exp2x.asp?casno=14452596 leads me this free access paper: aip.scitation.org/doi/10.1063/1.2436891 where, in addition to Li2, Na2, NaLi, Mg2, and Si2 are apparently studied experimentally. $\endgroup$
    – H. H.
    Apr 17, 2021 at 9:52
  • $\begingroup$ Actually, I think that paper just references experimental studies of those molecules, and is not an experimental study itself, but rather a compilation and theoretical analysis based on that data. Still, it clearly is said to be "experimental" data. $\endgroup$
    – H. H.
    Apr 17, 2021 at 10:19

I'm not an expert, but I think some metals boil directly into the plasma state, and that other metals (and most other gasses in general) start out as almost entirely unionized gasses when they boil and then slowly become more and more ionized as the temperature goes up. Thus, the difference between gas and plasma is more of a matter of degree than a first-order (sudden) phase change. https://en.wikipedia.org/wiki/Degree_of_ionization

There are two pieces of "evidence" I want to present for what I'm saying.

  1. Certain metals have extremely high boiling points. For example Tungsten and Rhenium both have boiling points that are slightly higher than the surface temperature of the sun. Both elements have lower ionization energies than Iron (https://www.lenntech.com/periodic-chart-elements/ionization-energy.htm), which is mostly ionized at these temperatures even though Hydrogen and Helium aren't (https://www.astro.princeton.edu/~burrows/classes/204/stellar.atmospheres.HR.pdf , pg. 18), so Tungsten and Rhenium should be ionized at these temperatures.


  1. I've recently been using this USGS paper as a reference. https://pubs.usgs.gov/bul/2131/report.pdf (It does mysteriously repeat several pages with different page numbers at one point, but it's probably mostly trustworthy.) Starting on page 67, this paper contains constant-pressure heat capacities (Cp) at different temperatures for various elements, most of which are metals. This notably includes gaseous data for monatomic calcium, cadmium, cesium, mercury, potassium, lithium, magnesium, sodium, strontium, and zinc.

    Cp is defined as the minimum amount of energy that needs to be put into a given amount of a substance in order to raise it by certain temperature. In this paper, it is how many joules are required to raise one mole of a substance by one kelvin. Temperature itself is related in a definite way to the average translational kinetic energy of a particle, which is $\frac{3}{2}$kBT. In a monatomic ideal gas, translational motion is the only significant way that the energy of heat added to a system is stored, so the amount of energy you put in is directly related to the change in average translational kinetic energy, and therefore to the change in temperature. Thus, Cp has the same value for all monatomic ideal gasses (provided it is per mole rather than per unit mass). That value is 20.78615$\frac{J}{mol K}$. However, other types of matter can store some of the energy in other ways, and therefore have higher heat capacities. For example, polyatomic particles (molecules) can store more energy in vibrations and rotations; solids, liquids, real gasses, and supercritical fluids store potential energy in overcoming attractive (and repulsive) forces between their particles that are significant even when not colliding in such cases; and plasmas can store energy by knocking electrons out of their atoms. https://en.wikipedia.org/wiki/Equipartition_theorem https://en.wikipedia.org/wiki/Molar_heat_capacity#Physical_basis_of_molar_heat_capacity

    Returning to the USGS paper, it can be seen that Cadmium, Mercury, Potassium, Lithium, Magnesium, Sodium, and Zinc all have a Cp of 20.79$\frac{J}{mol K}$ just above their boiling point; Cesium mysteriously has a Cp of 20.78$\frac{J}{mol K}$ just above it's boiling point (possibly an error, though I doubt that); and Calcium and Strontium both have a Cp of 20.84$\frac{J}{mol K}$ just above their boiling point. (Their identicalness is probably mostly a coincidence.) After boiling, Cesium, Potassium, Lithium, and Strontium all show an increase in Cp as temperature increases up to the point where the data stops, which is 1800K in all cases here mentioned. Since all of these substances are described as "ideal monatomic gas, P = 1 bar" by the paper, they shouldn't be able to store energy through rotations, vibrations, or interatomic forces (at least not significantly), so the most likely place for this energy to be going is into ionization, or at least into exciting electrons if not fully removing them from atoms. This would explain why only very electropositive, low-ionization-energy metals show this increase below 1800K. (Other metals probably show the same increase at higher temperatures.) The main hole in this theory is that sodium still has Cp=20.79$\frac{J}{mol K}$ at 1700K and 1800K, where Lithium and Strontium show an increase, even though these latter 2 metals have lower ionization energies. In light of this, let's take a closer look at the relevant data from the paper compared to the ionization energies agreed on by that Lenntech page I linked to, nist.gov, and my CRC (although my CRC seems to have a rounding error on Cadmium, giving 8.993 eV, rather than 8.994):

    enter image description here

    When it is taken into consideration that the paper generally only gives data at 298.15K, 300K, 100K increments above that, and phase-/polymorph-change temperatures, it can be seen that sodium is the only thing about this table that is not consistent with the idea that elements with lower ionization energies show the increase in heat capacity above 20.79$\frac{J}{mol K}$ at lower temperatures, and it is entirely possible that these values were not actually measured for sodium but were just calculated based on the assumption that it was an ideal gas. However, given how little evidence I had for my theory to begin with, this inconsistency actually throws serious doubt onto it. Thus, I tried to find other sources for the heat capacity at constant pressure of gaseous sodium as a function of temperature, but I could only find calculations, no experiments, and the values they gave were so completely different from that USGS paper that I feel I must be missing something. (Moreover, all sources I found seem to be related to cooling systems for nuclear power plants and focus mainly on liquid Na or Na-K alloy.) For example, this paper ( https://www.ne.anl.gov/eda/ANL-RE-95-2.pdf ) gives calculated heat capacities in $\frac{J}{kg K}$ for "sodium vapor" from 400K to 2500K on page 15. Converting 20.78615$\frac{J}{mol K}$ to $\frac{J}{kg K}$, based on the atomic weight of sodium being 22.98976928$\frac{mol}{g}$, gives 0.9041478$\frac{J}{kg K}$, which this paper passes between 400K and 500K for Cp, with Cp continually rising with temperature according to them. (Noting that Cv is heat capacity at constant volume, which should be lower at 12.47169$\frac{J}{mol K}$ = 0.1075126$\frac{J}{kg K}$ for sodium)

So, it looks like I'm still not really sure what heat capacities of gaseous metals say about their ionization, but I still think it's pretty likely that high-boiling-point low-ionization-energy elements like Hafnium, Uranium, Zirconium, and Tungsten are quite ionized as soon as they boil and don't really have any stable unionized gas phase at atmospheric pressure except as vapor in equilibrium with solid or liquid metal. That being said, the type of ionization I've been talking about for most of this post is probably quite low levels of ionization in mostly neutral gasses, so it probably wouldn't be fair to say that strontium, for example, is a plasma as soon as it boils, since it's probably just a mostly neutral gas that has a very small, but non-trivial, amount of ionization. Also remember that it's probably possible that most or much this early increase in heat capacity comes just from exciting electrons into higher energy orbitals within atoms rather than actual ionization.

  • $\begingroup$ There is copernicium which supposedly is a gas at room temperature,before it decays radioactively. $\endgroup$ May 31, 2021 at 22:34
  • $\begingroup$ I've just realized that the "something" I was missing about sodium vapor might be related to pressure. Also, @Bruh Moments , that does raise the (perhaps somewhat philosophical) question of whether such a thing as Copernicium gas at standard temperature and pressure can exist. I think it probably pretty clearly could (briefly), given that Cn-285 has a half-life of (a whole) 28 seconds, but I imagine it would heat itself up at a pretty significant rate. Hard to say how quickly, though, since most of the radiation would probably escape your gas sample. $\endgroup$
    – H. H.
    Sep 1, 2022 at 9:13
  • $\begingroup$ @BruhMoments The bigger question is how you could get enough Copernicium to notice (or care) generated AND cool it down and pressurize it to STP before too much of it decayed, and, even if you had quite a bit left, that would affect the purity of your sample, not that that matters too much with ideal-like gases. $\endgroup$
    – H. H.
    Sep 1, 2022 at 9:16
  • $\begingroup$ I said the "something" I was missing about "sodium vapor" having very different heat capacities from sodium gas (presumably at standard pressure) might be related to the difference in pressure between standard pressure and vapor pressure. However, that really shouldn't affect heat capacity (assuming their not averaging in the heat capacity of the air the sodium vapor is mixed with into it's heat capacity (which would probably make it's heat capacity DROP with increasing temperature)), EXCEPT that it would REDUCE ionization below the SP boiling point. These are not the differences I found. $\endgroup$
    – H. H.
    Sep 1, 2022 at 9:23

Your Answer

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

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