Is Fire a Plasma?

If not, what is it then?

If yes why, don't we teach kids this basic example?

enter image description here

UPDATE: I probably meant a regular commonplace fire of the usual temperature. That should simplify the answer.

  • 22
    $\begingroup$ Home experiment: put a strong (i.e. rare earth) magnet on the end of a pole and see if it affects the behavior of a flame more than a pole without the magnet does. Why? Because the magnet will affect ions and free electrons in motion much more strongly than neutral components. $\endgroup$ Commented Apr 9, 2012 at 20:56
  • 10
    $\begingroup$ Don't have a fireplace here nor a strong magnet. The answer is? $\endgroup$ Commented Apr 9, 2012 at 21:11
  • 6
    $\begingroup$ As far as I know ordinary flames are not significantly ionized: they're just hot gas with chemistry going on in them. But I'm not certain enough to post an answer, so don't quote me. $\endgroup$ Commented Apr 9, 2012 at 21:18
  • 4
    $\begingroup$ A similar question is the basis for a recent project in science communication. Check it out: flamechallenge.org $\endgroup$
    – Greg P
    Commented Apr 10, 2012 at 2:20
  • 13
    $\begingroup$ @dmckee What about a simpler experiment of conductivity: can a circuit be closed with a flame? If there is plasma there will be free electrons and ions so two conducting leads to a flame ccould be closing a circuit and current would appear. $\endgroup$
    – anna v
    Commented Apr 10, 2012 at 4:49

5 Answers 5


Firstly, 'Fire', according to numerous comments and answers [here][1] is a 'process', in which case, the answer to the question will be 'no', since plasma is a state of matter. It would be unfair to leave it there by blaming the semantics, and given the abundant references to 'flame' region, I am going to assume that that is what the question meant to ask. I am also assuming that proving a candle flame constitutes plasma is enough to sufficiently answer the question.

From some papers (a quick google search gave me [2,3]) that flames have ionised content and that they are electrically conductive. My suspicion was that not all flames are conductive, but [3] includes the statement:

It has been known for a long time that flames possess a high electrical conductivity and can be distorted by an electric field.

Sources [4] and [5], and numerous other sources, including a video on YouTube [6] claim that a candle flame is ionised and that's what causes the flame to be affected by electric field.

Now is it plasma?

The 'Plasma Coalition', which is a coalition of many reputed institutes around the world [7], says that ionisation alone is not enough, but enough atoms have to be ionized to significantly affect the electrical characteristics of the gas, in order for it to be called plasma. In one of its documents [8], it expands on this description in great detail.

It actually has a paper dedicated to this question, [8], which says that some flames contain plasma, whilst others don't. It expand further in sufficient detail, claiming that the answer depends on the region, what's being burned, the temperature, etc.

It also acknowledges that the current knowledge about flames is quite limited to conclusively ascertain the charged particle densities at a particle location in the flame, as of 2008.

A wide variety of sources that claim that a flame (like a candle flame) is plasma is referring to the fact that it is ionised.

Francis F Chen's book [10] includes an exercise on page 12 that connotes a typical flame being plasma. This claim is repeated in [4] and [5] (refers to candle flame).

My Conclusion

I understand that the Plasma coalition paper [8] says that the temperature of a candle is too low for much ionisation to occur, but technically, the experiments cited above [2,4,6], demonstrating the significant effect of flames in an electric field, coupled with the theoretical predictions [3,10] seem to imply that the flame is indeed a plasma. Even by the condition stated by the Plasma Coalition [11] itself!

I found it interesting that an old paper [3] proposes to explain the excessive amounts of ions formed in hydrocarbon flames by suggesting that it is in part due cumulative excitation or chemi-ionisation. I do not know if it is still relevant today.

$\ \ $ [1] Is fire matter or energy?, Physics Stack Exchange.

$\ \ $ [2] Electrical Properties of Flames: Burner Flames in Longitudinal Electric Fields. Hartwell F. Calcot and Robert N. Pease. Ind. Eng. Chem. 43 no. 12, pp 2726–2731 (1951).

$\ \ $ [3] Mechanisms for the formation of ions in flames. H.F. Calcote. Combust. Flame. 1 no. 4, pp. 385–403 (1957).

$\ \ $ [4] Waves in Dusty Space Plasmas. Frank Verheest (Kluwer Academic, 2000, The Netherlands).

$\ \ $ [5] Sun, Earth and Sky. Kenneth R. Lang (Springer, 2006, Berlin).

$\ \ $ [6] What's in a candle flame, Veritasium YouTube Channel.

$\ \ $ [7] About the Coalition for Plasma Science.

$\ \ $ [8] About Plasmas. Coalition of Plasma Science, 2008.

$\ \ $ [9] Plasma State of Matter. Lecture notes for PX384 Electrodynamics at Warwick University, chapter IV. Erwin Verwichte, 2013.

$\ $ [10] Introduction to Plasma Physics and Controlled Fusion. Francis Chen. Available here for the moment.

$\ $ [11] What is Plasma?. Coalition for Plasma Science, 2000.

  • 5
    $\begingroup$ What a great answer! I think the main point to be extracted from ref [8] is that a region of gas is considered a plasma if its density of ions is high enough so that most of the region is shielded from electric fields. By this criterion some hot flames are plasmas. $\endgroup$
    – user4552
    Commented Aug 24, 2013 at 3:21
  • 1
    $\begingroup$ Nice insights and organization of thought here. Do you think it is safe to say that unlike a vaporization point, the transition between gas and plasma is much more gradual? It seems there is a threshold where it behaves fully 'plasma-like' and the confusion is about states in between where electrical properties may be expressing but not as strongly. $\endgroup$ Commented Aug 3, 2015 at 3:13
  • $\begingroup$ Percent of plasma aside, ions and conductivity don't define plasma. A vapor of sodium and chlorine ions is not a plasma, yet it's ionized and conductive! Also, emission lines don't imply plasma: the light may be chemical fluorescence; only the orbital transitions. Actual "Plasma" goes beyond these, and requires stripping electrons from atoms, producing a mixture of free electrons and positive ions, with emission lines of recombination. If those (UV) wavelengths aren't present, then it's not a plasma. $\endgroup$
    – wbeaty
    Commented Sep 27, 2017 at 7:06
  • $\begingroup$ Another confusing issue: exposing high voltage and/or hot metal to a flame will produce plasma, at quite low voltage compared to sparks in air. Plasma in flames may be pure experimental artifact. I find many misleading online videos with big carbonized candle-wicks exposed to kilovolt plates. Certainly plasma is produced, but it's just a spark from an HV power supply and carbon electrode. If we convert a flame to plasma, we're not allowed to conclude that all flames are plasma, or that they include any plasma. Avoid HV supplies, metal gas jets, and black candle wicks. Detect UV lines instead $\endgroup$
    – wbeaty
    Commented Sep 27, 2017 at 7:26
  • $\begingroup$ @wbeaty High temperature pushes electrons into higher orbitals thus we get positive ions and electrons. The pneumonic is F O Cl N Br I S C H . F is the most Electronegative so F+ is even more electronegative than that. They typically use Br as flame retardant so Br+ is less electronegative than O+ but more than C+ and H+ (but the reverse at high temperature) $\endgroup$
    – ChemEng
    Commented Aug 11, 2022 at 12:50

Broadly speaking, fire is a fast exothermic oxidation reaction. The flame is composed of hot, glowing gases, much like a metal that is heated sufficiently that it begins to glow. The atoms in the flame are a vapor, which is why it has the characteristic wispy quality we associate with fire, as opposed to the more rigid structure we associate with hot metal.

Now, to be fair, it is possible for a fire to burn sufficiently hot that it can ionize atoms. However, when we talk about common examples of fire, such as a candle flame, a campfire, or something of that kind, we are not dealing with anything sufficiently energetic to ionize atoms. So, when it comes to using something as an example of a plasma for kids, I'm afraid fire wouldn't be an accurate choice.

  • 18
    $\begingroup$ I don't think so--- the fire itself has a significant proportion of ionized atoms, it is not just hot gas, because the glow is due to the recombination in particular lines which are dependent on the chemical emission lines (you can see this in burning salt). $\endgroup$
    – Ron Maimon
    Commented Apr 10, 2012 at 4:01
  • 2
    $\begingroup$ @Ron, are you sure it's recombination or maybe it's just transitions? I don't have a strong opinion on that, the reason I ask is that I think this might make the difference between fire and plasma (if any). $\endgroup$ Commented Apr 11, 2012 at 14:54
  • 1
    $\begingroup$ @LevLevitsky: I am not sure--- it might be just outer shell transitions--- but once you ionize once, it is so much easier to eject the electron out teh atom. I am confused now on the issue--- I have found sources going both ways. I though the best thing would be to test conductivity, but somebody did, and said he didn't see conductivity (but this might be a low conductivity), the sharp boundary of the flame, the lack of gradual cooling, suggests that it is some sort of steady-state phase of combustion. It is hard to say if steady state non-equilibrium stuff is this or that. I don't know. $\endgroup$
    – Ron Maimon
    Commented Apr 11, 2012 at 15:35
  • 1
    $\begingroup$ @Anixx: Of course, but why the sharp change in color? Does the gas abruptly cool down? Why there? I figured it was ionization transition of some kind, but I am not sure anymore. $\endgroup$
    – Ron Maimon
    Commented Sep 4, 2012 at 20:49
  • 2
    $\begingroup$ The assumption of "glowing gas" is physically wrong. If you gas would state to glow, it must emit black-body-radiation. This would be in infrared or maximum red - like glowing metal. But you see line-emission - easy verifiable by looking at it trough a prism. The fact that you see coloured flames is due to the fact of recombination and the following transitions. See this experiment for example. $\endgroup$
    – Fips
    Commented Nov 5, 2013 at 22:35

Back of the envelope calculation:

The Saha equation for a Hydrogen plasma says

$$\frac{N_i^2}{N_H} = V \left(\frac{2 \pi m_e k_b T}{h^2}\right)^{3/2} \exp\left(\frac{-R}{k_b T}\right)$$

where $N_i$ is the number of ions, $N_H$ is the number of Hydrogen atoms, $V$ is the volume of the plasma, and $R$ is the Hydrogen ionization energy (13.6eV).

Defining the degree of ionization $\xi = N_i / N_0$, where $N_0 = N_i + N_H$ is the total number of atoms in the system, this can be written

$$\frac{\xi^2}{1-\xi} = \frac{V}{N_0} \left(\frac{2 \pi m_e k_b T}{h^2}\right)^{3/2} \exp\left(\frac{-R}{k_b T}\right)$$

A candle burns at 1000 Celsius, and the flame has a volume of around 1cm^3, with probably 10^20 atoms in the flame. For simplicity, let's assume it's mainly Hydrogen in the flame (the ionization energy of other elements is of the same order of magnitude anyway, so we won't be far off). Then I make the right hand side of the equation (we'll call it $f$) to be around 10^-54. Then we can solve $\frac{\xi^2}{1-\xi} = f$ using the quadratic formula:

$$\xi = \frac{\sqrt{f^2 + 4f} - f}{2}$$

This gives us $\xi = 10^{-27}$: none of the particles in a candle flame are ionized (remember, we guessed there were only 10^20 particles). This makes perfect sense, because 1000C is only around 0.1eV, a good two orders of magnitude less than the ionization potential. The particle density is too low to make up for that.

If you think any of my approximations don't apply (personally, I'm not too sure about the particle density) then please correct me in a comment!

  • 1
    $\begingroup$ Your error lies in the fact you assume the ionization to start from the fundamental n=1 state. In a flamme, electrons first climb by collision the energy ladder up to some excited states. When their bounding energy is close to the thermal energy they get ionized. $\endgroup$
    – Shaktyai
    Commented Sep 4, 2012 at 13:23
  • $\begingroup$ I see what you mean, but I don't think you're quite right. The vast majority of atoms in a plasma are either ionized or in the ground state. This is because the continuum has a huge statistical weight compared to the bound states (the density of states is large). You can actually use (a slightly more general form of) the Saha equation to find out how many atoms are in the excited states in a Hydrogen plasma; it turns out it's safe to ignore all except the ground state. $\endgroup$ Commented Sep 4, 2012 at 14:31
  • $\begingroup$ I did run a kinetic (Fokker Planck) atomic code for fusion plasmas at the edge of Tokamaks. At 0.1 eV all the gas is ionized. The energy difference between the n=6 and n=7 is already of the order of 0.1 eV. And collisions pump up level by level electrons from ground state to the continuum. The problem is extremely complex because the system is not in thermodynamical equilibrium (because of radiations)and one needs to introduce different temperatures for the upper levels, the lower ones and the free electrons. $\endgroup$
    – Shaktyai
    Commented Sep 4, 2012 at 16:07
  • 5
    $\begingroup$ Flames can have non-negligible ionization due to chemical reactions involving free radicals (starting, in hydrocarbon flames, with ${\rm CH+O\rightarrow CHO^{+}}+e^{-}$). Check this old survey about the topic for more information. $\endgroup$
    – mmc
    Commented Sep 5, 2012 at 2:40

Nope. Fire is a thermal phenomenon, plasma is more of electrical.

What's plasma?

Plasma is the state when you strip off electrons/add electrons to a gas--so plasma consists of charged gas ions. It usually glows due to electron transitions and whatnot.

What's fire?

In a flame, you basically have hot soot/&c molecules flying up. Any hot material emits photons, which are usually in the infrared range for normal temperatures. At higher temperatures, they can go into the visible range.

One way to explain this is by blackbody radiation-- the soot must emit photons since it has a nonzero temperature.

What's actually going on is that the electrons are "thermally excited"--they have extra energy and are prone to making transitions. Transitions lead to absorption/emission of light, and this is what causes the color.

You can see that there aren't any ions involved in fire, so it's not plasma. But ionization will occur if you heat it to even higher temperatures, and it can become plasma.

  • $\begingroup$ Would an oxy-acetylene (welding) torch be a plasma $\endgroup$ Commented Apr 10, 2012 at 1:59
  • 12
    $\begingroup$ "Fire is a thermal phenomenon, plasma is more of electrical." I can't say I like this formulation much...heat alone is enough to ionize atoms if things get hot enough. $\endgroup$ Commented Apr 10, 2012 at 2:23
  • $\begingroup$ @dmckee: But ionization will occur if you heat it to even higher temperatures, and it can become a plasma. I know, but it's not hot enough in fire.. $\endgroup$ Commented Apr 10, 2012 at 3:19
  • $\begingroup$ @MartinBeckett in some instruments yes : directindustry.com/prod/farley-laserlab/… $\endgroup$
    – anna v
    Commented Apr 10, 2012 at 4:43
  • $\begingroup$ @annav, yes a plasma cutter is a plasma! Oxy-acetylene is about 3500C which is the bottom end of a K type star surface $\endgroup$ Commented Apr 10, 2012 at 14:46

Fire is a plasma. There are two kinds of plasmas: hot plasmas relevant to astrophysics or fusion are indeed a mixture of totally ionized gas. In cold plasmas ( northen lights, Neon tubes,flamme) the ionization degree is less than one but the mixture typically exhibit collective behaviour and a zoo of waves one do not encounter in gases. The most famous is the plasma oscillation and the Alfven wave but they are many others. poorsod's calculus assume the ionization takes place between n=1 and n=infinity. In reality, the atoms are first excited by collisions, their electrons jump on higher n until their bounding energy is lower than thermal energy of free electrons. For 0.1 eV more than 99% of the atoms are ionized (I did work on a computer model to analyze this problem). Though the equilibrium Saha approach is known to be false (the electron distribution function is not Maxwellian), you can get a preety good idea of the problem if you split your neutral atoms population into atoms in the fundamental, n=2, n=3, etc.. and use Saha equation for each population.

  • $\begingroup$ interesting, can other people please confirm this too? $\endgroup$ Commented Sep 7, 2012 at 1:35
  • 5
    $\begingroup$ I am sorry, as mentioned by @Mitchell Fire is generally not plasma. it rare for fire to be plasma. A single electron or ion or few of then dose not qualify for being plasma, it has to gather in sufficient amount so that it can show collective behavior. $\endgroup$ Commented Jan 4, 2013 at 8:33
  • $\begingroup$ Fire is not plasma: presence of ions doesn't define plasma. Instead, the presence of free electrons indicates plasma: a mixture of electrons and positive ions. Note well that conductivity doesn't require plasma. If only ion-pairs are created by chemical reactions, but no electrons freed, then a flame will have 'electrolytic' conductivity, similar to molten salt and salt/acid/alkaline solutions, where conductivity is based on mobile ion populations only. $\endgroup$
    – wbeaty
    Commented Sep 27, 2017 at 6:46

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