I want to make a plasma speaker. I'm worried about the amount of ozone being produced by the ionization of the air and it possibly not being a safe amount. Even if it doesn't produce a dangerous amount, I don't like the smell of ozone. I'd like this all to be continuous. I am very unfamiliar with the chemistry here but I have read that UV light, around 250nm, can cause $\ce{O_3 \to O_2 + O}$ and also $\ce{O_2 \to 2O}$

But also without any help it seems that $\ce{2O \to O_2}$ and also $\ce{O + O_2 \to O_3}$. So it seems that if I shined UV light of the right frequency on the spark I'd make a lot of $\ce{O}$ and perhaps that would cause more $\ce{O_2}$ than $\ce{O_3}$ but I'm not really sure.

I was also thinking that it might oxidize some sort of metal filter really well and I could just keep replacing those but I wasn't sure. I could also heat the filter slightly to speed up the reaction if necessary.

Even if the UV light idea doesn't work out very well I still want a way to do this. Is the amount of ozone produced by a continuous electrical current though air harmful? Is there a way of getting rid of the ozone to make some odorless non-harmful chemical?

  • $\begingroup$ We've banned CFC's precisely because they're very effective at the reaction you want. But they're banned so it's not really an answer. $\endgroup$
    – MSalters
    Apr 10, 2015 at 8:11

4 Answers 4


Even if the UV light idea doesn't work out very well I still want a way to do this

With some photochemical background, I suggest to forget about running such a UV lamp

  1. Remember that your plasma will be formed in air. For every molecule of ozone formed, there are much more harm- and odourless $\ce{O2}$ molecules around. If you cleave $\ce{O3}$

\[\ce{O3 + O2->[\mathrm{h}\nu][\lambda\ \lt\ \mathrm{300\ nm}] [O2 + O* + O2] -> O2 + O3}\]

the resulting oxygen atom will be captured by the next $\ce{O2}$ to form ozone again.

The overall process will not reduce the number of ozone molecules, unless $\ce{O*}$ reacts with $\ce{N2}$, which might lead to some other unpleasant species ;)

Note that these processes are fortunately effective in the ozone to protect us from short wave UV radiation, which brings us to the next aspect of the setup.

  1. In order to photolyze ozone, you'll neeed a light source that emits UV at $\lambda\lt300\,\mathrm{nm}$. It's not that these light sources would not be available, but this UV range is not that of a tanning booth lamp! Remember that the ozone layer protects our DNA from running havoc! The least you can expect from short-wave irradtion is the sun burn of your life! Consequently, you will have to shield the whole setup properly!

In summary, I absolutely like the idea of building a plasma speaker yourself and there even seem to be some DIY kits around, but I suggest to forget about the UV part, which can be much more dangerous than the ozone formed.


Burak Ulgut brought up a very interesting point. Back in the days, plasma speakers using other gases (or gas mixtures) were tried, and even went commerical.

This site in German describes a helium driven plasma speaker system :)


I would say it would be much easier to try to keep the area oxygen free. Atomic oxygen (i.e. $\ce{O}$) is a highly reactive radical. This would react either with another $\ce{O}$, or an $\ce{O_2}$. It is hard to avoid this unless you can get rid of the $\ce{O_2}$.

Not sure what kind of a geometry you are thinking of, but purging the volume with $\ce{N_2}$ gas might be an easy way to achieve this.

  • $\begingroup$ The mention of an alternative gas is an interesting point! $\endgroup$ Apr 10, 2015 at 7:44

You were wondering about a metal filter. That's a viable idea, but not all metals are created equal. Iron indeed would require frequent replacements, it rusts rapidly when in contact with ozone.

Copper is a fairly noble metal (oxidizes slowly), still reasonably cheap, and it works as a catalyst to break down ozone. The oxygen radicals on the copper surface often recombine to form O2, instead of forming Cu2O/CuO. (Copper(I) oxide and copper (II) oxide).

There are more copper-oxygen reaction products that can form. The initial reaction forms CuO3 (basically the ozon molecule sticking to one copper atom) and CuO4 (when ozon sticks to an existing CuO molecule). Another observed reaction product is Cu2O3. With all those combinations possible, a lot of pathways for chemical reactions open up.

In comparison, a free Ozone molecule can't simply "lose" one of its three oxygen atoms. Creating a free oxygen atom costs energy (that's why ozone is stable), so you need two ozone molecules to react with each other. That in turn means the decomposition rate is quadratic in the ozone concentration, not linear. So with low ozone concentrations, the reaction rate is very low.

  • 3
    $\begingroup$ Could you expand your answer a little more, in its current state it is more a comment than an answer. For example, in which state has copper to be present to have the described feature. How would this reaction take place? What would the turn over be for such a reaction? How could you force the ozone to even come in contact with the copper? $\endgroup$ Apr 10, 2015 at 10:35

Your right about the UV light. A UV light at 254 nm will efficiently destroy ozone. Therefore if this were on, and in the vicinity of the ozone production, you should keep the overall ozone production relatively low. Just make sure your UV light is focused at 254 nm as lower wavelengths of UV light do a great job producing ozone: http://www.oxidationtech.com/ozone/ozone-production.html#answer2

For metals, aluminum oxide would do a good job destroying ozone. It is commonly used for ozone destruction.

The other easy bet is a carbon filter. Carbon breaks down ozone quickly to CO and CO2. The only downside is that the carbon will become consumed over time and the carbon filter would need to be replaced.


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.