Why are sparks blue?

I am not sure if First Excited State posted this as a blog entry, but it was mentioned on twitter. Question: why are sparks blue? My first gut response was that this is the blackbody color. Wrong for several reasons. The short answer is that sparks are blue because of the colors given off from nitrogen and oxygen when they are excited.

In order to make this post longer than necessary, let me say something about blackbodies. A blackbody is an object that emits radiation due only to it's temperature. Since it does not reflect anything, it looks black at room temperatures. You can make a black body, it's not hard. Simply take a closed box with a small hole in it. Look at the hole, it will appear black regardless of the actual color inside. Here is an example of one I made. Ok, I can't find a picture of that box. I will post it later because it is pretty cool. Instead, here is a diagram:


Essentially, light goes in, but doesn't come out (like thunder dome). When light goes in, it reflects off the surface but some of it gets absorbed. Each time it reflects, some gets absorbed. By the time it finally makes it out of that tiny hole, there is essentially nothing left. What does come out of the hole is light that is produced by the thermal activity of the material (and not by reflected light). It looks black to you because all of this blackbody radiation for this temperature is in the infra red spectrum.

Some other examples of blackbodies that you are probably familiar with:

  • Incandescent light bulb filament while on.
  • The sun (while on).
  • A hot stove element.

All of these objects give off radiation that is related to the temperature of the object. The higher the temperature, the more light given off at shorter wavelengths. These objects actually give off radiation (note that I am using light and radiation interchangeably) at essentially every wavelength. This is usually called a continuous spectrum. If you looked at it through a spectral slide or a prism, you would see all the colors of the rainbow. The best way to see this is with this awesome applet from PhET.


Blackbodies and other types of radiation are very complicated (quantum mechanically speaking). What is the difference between blackbody radiation and other stuff that gives off light? If you looked at a fluorescent light through a spectral slide, you would not see the rainbow. Instead you would just see some colors. If you have not done this before, you should get one of these spectral slides or glasses. They really are cheap. Just don't use it to look directly at the Sun (regardless of what Phil Plait says because it would suck if he was wrong). This is usually called an emission line spectra (as opposed to continuous)

What is the difference here? An emission line spectra is created when there is an excited gas. By excited, I mean that the electrons in the gas jump up to higher energy levels, and then fall back down. When they drop down they give off light. The frequency of the light produced is related to the change in energy levels. That is as much detail as I want to go into here, but if you are interested, see this post. So, different gases have different energy levels and thus produce light of different frequency.

Why don't blackbodies do the same thing? How come the light only depends on temperature and not the material that it is made of? (for example a gas of excited iron vs. a block of iron) The reason is that the energy levels in a block or iron are completely different than the energy levels in atomic gas of iron.

Ok. Back to sparks. The light could not be blackbody radiation because it's a gas. The light is actually given off when free electrons recombine with air ions (air ions means oxygen or nitrogen molecules missing an electron). To examine the spectra from a spark, I am going to put one of these spectral slides from Educational Innovations and put it in front of my video camera. Then I can use Tracker Video to analyze the spectrum. Here is a picture of the same thing with Hydrogen gas.


And using tracker, I can get the intensity of the light along that purple line I drew in there.


Now for comparison, here is the same thing done with a spark.


And here is a graph of the intensity.


No analysis, but that does not look like a continuous spectrum.

Finally, some other interesting things about sparks (for more details on this, see the excellent analysis of sparks in Matter and Interactions Vol II by Chabay and Sherwood).

  • A spark occurs in air in the electric field exceeds 3x106 Newtons/Coulomb.
  • It is NOT because charge is jumping from one object to the other.
  • Free electrons in the air are accelerated in the opposite direction to the electric field. These electrons collide with molecules and free other electrons creating an electron avalanche.
  • The light comes from electrons recombining with air ions (as stated above).
  • The electric field is not strong enough to pull electrons from the air molecules. These electrons had to already be there. (and they are from radioactive sources and cosmic rays).
  • In a vacuum, you wouldn't see a spark (no air). Also, no one can hear you scream. (I know I keep using that joke, I am sorry).

As a final plug for Matter and Interactions they have an order of estimate calculation for how large an electric field would have to be to accelerate electrons to the speed that they knock out other electrons. They compare this to the experimental value of 3x106N/C. Cool.

More like this

The latest snowstorm is wreaking some havoc on my plans for the day, which means I'm going to lift another question and answer from the Physics Stack Exchange, with some modification. This one is a question about thermal radiation: What are the quantum mechanisms behind the emission and absorption…
Last week, I did a post for Forbes on the surprisingly complicated physics of a light bulb. Incandescent light bulbs produce a spectrum that's basically blackbody radiation, but if you think about it, that's kind of amazing given that the atoms making up the filament have quantized states, and can…
Alas for Michael Jackson. Talented musician, deeply broken human being. Most of my knowledge of him was through cultural osmosis rather than his actual music, and I'm young enough so that I can't really remember the time before he was a punchline about changing skin color, disastrous plastic…
Are you ready for this? This week's Friday Galaxy is Mk509: DSS2 image from Skyview Admittedly, the image of this galaxy does not rank very high on the "wow, what a cool and pretty looking galaxy" scale. However, this is an interesting galaxy because of what's going on at the nucleus. Like all…

Ah, but were everything that simple...

Older style fluorescent lights did have discrete colour bands. But, the more modern CFLs (and, even some of the long fluorescent tubes now) have a "broad emission spectra", and tend not to have such pronounced spectral lines. I'm not real sure how they do this (I've been out of the electro-optical world for a number of decades now.), whether it's a blend of phosphors, or whether it's nanocrystals.


It might be interesting to run some spectral analysises on a variety of fluorescent tubes. And, just for grins and giggles, it wouldn't hurt to run some on Neon lamps, which should let you use the Rydberg Formula
to analyze the lines that you see (e.g., Lyman series, Balmer series, Paschen series, Brackett series, Pfund series, Humphreys series, etc.):


As for gas emission spectral lines, you didn't mention line broadening due to thermal motion, nor line splitting due to magnetic fields.

Oh, and this can also be applied to solar spectra, and gets used in astro-physics (e.g., red shift):


As for sparks, note that the emitted spectral components of a spark may vary in time, especially if the electrodes begin to be vapourized (thermal effects? sputtering?) and contribute their own spectral components to the emitted light.

For that matter, the emission of light from phosphors isn't necessarily instantaneous (e.g., They may exhibit phosphorescence rather than just
fluorescence.), and the decay period is affected by many things, such as phosphor particle size, doping levels, temperature, incident radiation, etc. For a really interesting demonstration of this, use a bottle of typing correction fluid (e.g., "White-out") to paint a white stripe on the blades of a black muffin fan. Then, allow the fan to spin up (or down) while being illuminated by a fluorescent lamp. You should begin to see colour bars as the rotational speed of the fan approaches synchronization with the light given off by the lamp (AC line excited, obviously). One thing to be wary of, though, is that you may also see Fechner colours (Pattern Induced Flicker Colours, PIFC), as exhibited by Benham's top:


For an interesting application of fluorescent lamp illumination, and variable persistence of phosphors, look up US Patent 4715687. :-)

As for the electron excitation of gas molecules, that's a HUGELY complex field itself. Consider, for example, that the gas mixture in a discharge tube is usually not a pure gas, but, is instead a Penning mixture:


Now, consider the atmosphere where a spark will occur, and how closely that may be to a Penning mixture.

And, since most atmospheric gasses aren't monomolecular, you get to
factor in contributions to the line spectra based on the motion of the component atoms (e.g., vibrational energy levels).

Additionally, there's the issue of the work function of the metal electrodes which can spew electrons from the surface of the metal.


To really make your head spin, the work function not only depends upon the metal, but also upon it's crystallization type and orientation! Plus, alloys of metals may have different work functions from their constituents. Yes, you're thoroughly and deeply into solid state and quantum physics at this point.

So, what seems to initially have been a very simple question has involved many subtopics of physics and chemistry. :-)


Yeah! Bruce Sherwood is coming to my local AAPT meeting this weekend to tell us all about Matter and Interactions! Of course, since I already took his course, I already had to do that problem with the spark and the electric field. I learned SO MUCH in that class. I recommend to all you HS teachers reading this, you should take the Matter and Interactions for Secondary Teachers distance learning class from NCSU. Sign up this summer for the fall course, which will be on E&M. You will learn a LOT, and feel a LOT more confident when teaching about electricity and magnetism!

Some interesting answers to the blue sparks, but - and I may be being a bit simplistic here- when sparks occur, some ozone is usually formed from the arcing. Ozone is a blue gas- is there enough ozone to give the blue??

what about mechanically generated sparks?
like hitting (certain types of) rock with a hammer, or using a power grinder on (certain types of)metal?

do the little fragments that break off have so much energy that they blackbody glow? (i guess from the broken bonds recoiling), or is this little metal fragments burning in oxygen


Those types are sparks are really hot little pieces of stuff - which you guessed (correctly).

Gracias, muy interesante...

I found a rock in my garden when I was weeding, it looked strange like a melted stone only about a one & a half inches around a little rusty & blackened I got a magnet & it sticks to the magnet, I put it on the bench grinder & blue sparks came off of it I am wondering if it is a small meteorite I don't know.