There are lots of videos on YouTube showing sodium, potassium, etc. exploding when dropped into water (this, for example).

I understand that when an alkali metal is exposed to water, a violent exothermic reaction occurs where a hydroxide and hydrogen gas are produced, but why and how does the sample of metal end up detonating and fragmenting? Physically speaking, how can a block of metal seemingly blast apart from the inside when the reaction occurs on the surface of the sample?

The Wikipedia article on alkali metals explains this, but I still don't seem to understand how this would result in an explosion. Diagrams would be very helpful.

The key of the question is how does the release of hydrogen gas and energy result in an explosion? Is there thermal runaway, and is the metal physically destabilized in some way during the reaction?


4 Answers 4


Until recently the answer was unknown, but a short time ago it was discovered that the reaction is in fact a Coulombic explosion. The rapid exchange of electrons between the sodium and the water causes the surface of the sodium droplet to become positively charged, and the ions repel each other. This behaves very like a negative surface tension, and the surface of the droplet increases in area rapidly forming a spiny, porcupine like shape as fingers of the molten metal are shot into the liquid at astonishing speed. The larger the surface area gets the faster the reaction occurs, leading to a runaway effect. The study was published in Nature Chem. (Ref.1).


  1. Philip E. Mason, Frank Uhlig, Václav Vaněk, Tillmann Buttersack, Sigurd Bauerecker, Pavel Jungwirth, "Coulomb explosion during the early stages of the reaction of alkali metals with water," Nature Chemistry 2015, 7, 250–254 (https://doi.org/10.1038/nchem.2161).
  • 4
    $\begingroup$ Science also published an article on this after the study came out in Nature and provided a link to a video here: news.sciencemag.org/chemistry/2015/01/… Absolutely incredible finding after all these years. $\endgroup$
    – docscience
    Mar 17, 2015 at 4:42
  • 1
    $\begingroup$ For those stuck behind paywalls or that like videos: High speed camera reveals why sodium explodes! $\endgroup$
    – A.K.
    Dec 10, 2018 at 18:02
  • $\begingroup$ However, a note of caution. The experiments dealt with a liquid sodium/potassium alloy. The sciencemag.org link reports that the behavior of solid sodium is inconsistent, and often produces no explosion at all. $\endgroup$ Jan 21, 2019 at 21:09

EDIT/Retraction: This answer was written prior to the publication of the research described in the accepted answer, and is now known to be incorrect. Science!

I assume this is due to mesoscopic defects -- if you think about a piece of metal on a molecular scale, it's not going to be presenting a totally flat sheet of atoms at its surface. Hydrogen gas forming within cracks in the metal can force those cracks to open wider, and this then exposes more surface to react, and produces more cracks.

This process will result in the force of the generated hydrogen ripping the bulk metal apart if it happens quickly enough. If the heat generated from the reactions then also starts boiling water in those same cracks and defects, that'll only add to the effect.

  • $\begingroup$ So it seems that as the reaction proceeds, the hydrogen pushes into the cracks and the increased surface area for the reaction produces even more hydrogen and heat, causing the cracks to expand rapidly in a feedback loop and leading to an explosion. This means there is thermal runaway, since more heat leads to higher rate of reaction, and this combines with the increased hydrogen production and surface area for reaction to tear apart the metal, possibly at a rate in excess of the speed of sound. Am I right to say this? $\endgroup$
    – bwDraco
    Nov 25, 2012 at 8:22
  • $\begingroup$ @DragonLord: As I understand it, yes. $\endgroup$
    – Aesin
    Nov 25, 2012 at 12:07
  • 2
    $\begingroup$ Additionally, the amount of heat produced is enough to melt sodium and potassium, and the interior gets super heated. $\endgroup$
    – Ben Norris
    Nov 25, 2012 at 17:03
  • $\begingroup$ @BenNorris: Now that's thermal runaway. $\endgroup$
    – bwDraco
    Nov 26, 2012 at 3:15

I don't think any clever effects need to be invoked to explain this phenomenon. The simple fact that the heat of the reaction is always enough to melt the blobs of metal involved is enough. Once you have a mobile molten drop of metal (which is often on fire), very little energy is required to cause the drop to split up so the violence of the continuing reaction is often quite sufficient to cause the drop to split and disperse.

You can easily verify that drops of liquids don't need much energy to cause them to disperse and, if you were still allowed to play with mercury, you could also prove this applies to liquid metals. Drops of mercury will easily split into smaller drops when small amounts of energy are added (e.g. letting a drop fall a short distance onto a surface). These smaller drops are a lot harder to contain than their parent (which is one reason why mercury contamination is so prevalent and hard to remove in old labs).

So in short, exothermal reaction + liquid metal = readily dispersed drops of burning metal.

PS While what I say in this answer is broadly correct, the recent research (see accepted answer) add additional explanatory power to the detail of the reaction. The reaction is more violent that my simple mechanism would explain and the high-speed camera footage demonstrates that more factors were at wok than most people expected.


The reaction between alkali metals and water produces hydrogen, which will rise up and mix with the oxygen in the air above. If the reaction generates enough heat, it can ignite this mixture - hydrogen and oxygen will detonate when mixed in the correct ratio. This explosion will either directly spray out the metal (which by then will probably be molten) or force it down into the water where it will produce even more hydrogen gas and steam, propelling it back up and outwards.


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.