This question on NaCl crystalization actually got me wondering: are there any ionic amorphous solids? Like ionic crystals are crystalline materials of electrostatically-attracted ions, can ions form an amorphous phase? I can see no reason why not, but I cannot think of any example either.

  • $\begingroup$ Do you mean are there any that will naturally form, or are there ways to take substances that would normally form crystalline solids and make them form amorphous solids instead (second is definitely possible) $\endgroup$
    – soandos
    Commented Apr 26, 2012 at 1:12
  • $\begingroup$ @soandos examples of both would be nice… but I was thinking of the first. In fact, I think I do already have an example of the second, as you sure can freeze ionic liquids into a glassy amorphous state… $\endgroup$
    – F'x
    Commented Apr 26, 2012 at 1:15
  • $\begingroup$ that is along the lines of what I was thinking. I would tend to doubt the existence of the first type. $\endgroup$
    – soandos
    Commented Apr 26, 2012 at 1:17

6 Answers 6


I assume you meant amorphous solid phases, since you can always melt any ionic compound to form an essentially amorphous liquid phase.

Googling "amorphous salts" brings up a fair number of results in the primary literature, such as this and this and this. The defining characteristic of such amorphous solid phases is that they exhibit no measurable melting transition, i.e. no latent heat of fusion can be measured in a calorimetry experiment. There also examples of amorphous semiconductor oxides and amorphous metal oxides and amorphous metals - feel free to search for these and see for yourself. So yes, they do exist, and in fact are of research interest.

  • $\begingroup$ The link to springerlink.com is broken. I am not able to find any copy saved on the Wayback Machine, either. $\endgroup$
    – user124612
    Commented Jun 3, 2022 at 12:30

The difficulty with creating an ionic bond glass is that creating true glasses depends on a combination of directional bonding between atoms (or molecules) and a rate of cooling that doesn't give those bonds time to move into optimal positions. The result is a bit like piling together sticky spheres that become locked before they can find an optimal stacking arrangement.

Pure ionic bonds in contrast are more like slippery spheres that are attracted towards each others' centers, but don't have any particular preferences about which parts of their surfaces make contact.

Given that, the most likely result by far from very rapid cooling of a molten ionic solid will be to produce micro or nanocrystalline ionic solids that may superficially resemble glasses, but which under X-ray or neutron diffraction crystallography would prove actually to have crystal structure at a very fine scale.

With that said, you nonetheless can come up with a recipe for trying to make a truly glassy ionic solid, especially since there's no such thing as a pure ionic bond. You would want to cool a very thin layer of molten salt at at an incredibly high rate, at least millions of degrees per second if I recall glassy metal cooling rates rightly. For ionic, you would also want the result to be pretty close to absolute zero, since whatever degree of "directional stickiness" you can get out of ionic solids is likely to be less even than metallic bonds.

Finally, if you really want to try to make such a thing as a serious project, you of course must check the literature in detail. My rule of thumb is that if you dig deep enough, there are almost no decently plausible ideas that you can think of in the physical sciences that someone has not at least written a paper on in some journal, and maybe even done some experiments.

Still, what you are asking is an intriguing idea for a real experiment, and it wasn't that many years ago that both metallic glasses and Penrose-tiled quasicrystals were new ideas.

  • 1
    $\begingroup$ On the contrary, there are numerous examples of amorphous salts, amorphous metal and amorphous metal oxides in the research literature that do not require cooling rates of "millions of degrees per second" to be formed. $\endgroup$ Commented May 12, 2012 at 7:18
  • $\begingroup$ Examples welcome! Organic salts will be much easier, but if you can find a reference to a simple elemental salt that shows no micro or nano crystalline structure under X-ray diffraction ("amorphous" can be a bit ambiguous, and often includes such not-real-glass states), that would be especially interesting. $\endgroup$ Commented May 12, 2012 at 8:41
  • 1
    $\begingroup$ Also, a "just in case" clarification: "millions of degrees per second" was just the unit used in some very early paper I read. The equivalent unit of "degrees per microsecond" captures the modest ranges a lot better. Such work used very thin films chilled very quickly indeed to keep the bonds cattywampus (technical term, sorry) down to the atomic level. Nanocrystals can be tough to prevent, so the earliest glassy metals were limited to very thin films. That got fixed as folks developed better alloys, ones that leave the atoms scratching their heads when deciding quickly who to bond to next. $\endgroup$ Commented May 14, 2012 at 22:37
  • 1
    $\begingroup$ Just add lots of different ionic units with complicated ratios to form an intricate salt lattice and the cooling rate can be reduced to 1000C/s form a glass. After all that's the basis of metallic glasses. $\endgroup$ Commented May 30, 2014 at 12:52

The following is not a proof, but an idea that would imply that it is not possible to have such a material.

The reasoning goes as follows:

In any ionic "sea" of atoms, there are two types of situations:

  1. The atoms have enough energy to move freely, disregarding the attractive and repulsive forces of the other ions in the sea. This would cause a fluid, as the particles are not maintaining a position anywhere.
  2. They do not have that much energy, and therefore they will be impacted by the attractive and repulsive charges in the "sea." It then follows that there is an optimal configuration that will stabilize all of these forces. While I suppose that it might be possible for the optimal configuration not to be periodic (and thus not crystalline) I would tend to believe that it is not possible (it would seem from tessellations that if one assumes some basic characteristics, they must be regular). In a similar vein, I believe that it is not possible to have degrees of freedom in such an optimized state (though there may of course be more than one locally optimal state possible) as each atom is at a local minimum of energy, and it lacks the ability to "move up" the gradient (if it had more energy, it would be a liquid).

I therefore believe that it is improbable that such things exist.

  • $\begingroup$ By your argument, how do you explain the existence of a glass? As I'm sure you're aware (this is more for other people), the use of "proof" should never leave a chalkboard dedicated to mathematics. "Proof" does not belong in a physical science. $\endgroup$
    – Chris
    Commented Apr 26, 2012 at 6:14
  • $\begingroup$ Glasses cool too quickly is my understanding. Yyou can create an amorphous ionic compound of any type by cooling it rapidly (see comment to question) $\endgroup$
    – soandos
    Commented Apr 26, 2012 at 8:10

I would advise those who doubt in the existance of ionic solids to google ionic liquids which are widely commercialised. They probably solidify at sub-zero temperatures.


Many pharmaceutical salts can be obtained in the solid amorphous state by spray-drying, freeze-drying or milling.


The textbook example are crystals isotropized over geologic timescales through the effect of ionizing radiation. Heat them up, to overcome the activation barrier, and they release the accumulated energy and turn crystalline again.

The relevant keyword is "Wigner effect".


One type of ionic compound that is likely to appear in amorphous form is silicates. Fu and Krawczynski 1 studied Raman spectra of natural and artificial silicate glasses. According to the authors:

Amorphous silicates are common in extraterrestrial chondrites of petrologic types 1 and 2. In addition, high percentage[s] of amorphous components and poorly crystalline phyllosilicates were found in mudstones at Gale Crater by the CheMin instruments on board of Mar[s] Curiosity rover...

Such compounds might be expected to be more common on bodies where less geological activity (which often involves high temperature processes) is available to promote crystallization.

Naturally occurring amorphous silicates tend to have polymerized or macromolecular anions, which hinder the crystallization process. Some naturally occurring examples given by Ref. [1], from the Earth and Moon, are below. Note that the anion has a low enough oxygen content to form polymeric/macromolecular structures in all cases. [Dear readers: I would like to make a table, but the syntax is not working for me.]


Principal oxides(*): $\ce{SiO2,Al2O3, Na2O,CaO}$

NBO/T(#): 0.01

Hawaii basalt glass

Principal oxides: $\ce{SiO2,Al2O3, CaO,MgO,FeO,TiO2,Na2O}$

NBO/T: 0.73

Apollo 15 brown glass

Principal oxides: $\ce{SiO2,MgO,FeO,Al2O3,CaO,Cr2O3}$

NBO/T: 1.66

Apollo 16 green glass

Principal oxides: $\ce{SiO2,Al2O3,CaO,MgO,FeO}$

NBO/T: 0.43

*Oxides listed account for at least 99% of the oxygen atoms in the material.

#Ratio of Non-Bridging Oxygen to Tetrahedrally coordinated atoms, the latter being $\ce{Si,Al,P}$; a value less than 2.0 allows macromolecular anions to form and lower values indicate more extensive polymerization.


  1. Fu, X., A. Wang and M. J. Krawczynski (2017). "Characterizing amorphous silicates in extraterrestrial materials: Polymerization effects on Raman and mid-IR spectral features of alkali and alkaline earth silicate glasses". J. Geophys. Research Planets 122, 839-855, https://doi.org/10.1002/2016JE005241.

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.