Flames To Ashes: Campfire Science
    By Enrico Uva | November 26th 2012 12:00 AM | 27 comments | Print | E-mail | Track Comments
    About Enrico

    I majored in chemistry, worked briefly in the food industry and at Fisheries and Oceans. I then obtained a degree in education. Since then I have...

    View Enrico's Profile
    Birch bark, which is found in many parts of the U.S., Canada, Europe and China, is a great way to start a campfire. Rich in terpenoids, the paper-thin material ignites easily. The heat it releases provides enough activation energy to set small twigs ablaze, which of course should be placed in a tee-pee arrangement, so as to let in more oxygen. All of this should take place in a pit surrounded by stones, not to let wind take heat away from the young fire and not to burn the forest down.
    The hues of a flame are rarely constant for a second, a hint that something complex is occurring within them. There is a set of chain mechanisms involving intermediate molecules that are needed for subsequent steps. Many of the in-between products are radicals, reactive molecules with unpaired electrons. Radicals are often created in the high temperature regions of the flame, but they diffuse back into the colder regions where they are needed to generate the final products. To reveal more details, laser-based investigative techniques have been used so as not to disturb the flame. Even when burning as simple a molecule as diatomic hydrogen, radicals like O, H, OH and HO2 form. When the combustion of hydrocarbons like cellulose and lignin in wood takes place, we get a greater variety of radicals, some of them carbon-centered. C2 and the radical CH arise in excited form and release blue and green light. Lignin-derived radicals involving benzene structures are not the healthiest things to inhale, but mere occasional exposure is probably nothing to worry about, unless there's a concentration of wood-burning stoves in a particular area.  All of this underlines the fact that whenever we write an overall equation for a fuel consumed in a fire, it's like we are seeing only the ingredients of a recipe and the final product without witnessing the cooking.

    Hot, gaseous products of combustion expand and rise, stretching flames vertically. The ascension leads to pressure gradients, and fresh air is pushed into the fire. The circulation supplies it with more oxygen, the electron-thief that campfires depend on to release heat as more tightly bonded products like water and carbon dioxide are created. There's energy needed to drive molecular fragments of cellulose apart, in the same way that you need to exert force against gravity if you want to push a ball up a hill. But once at the top, the ball can roll further down on the other side. With chemicals, it's not the combination of mass, gravity and varying heights that accounts for differences in potential energy but Coulombic forces acting over a variety of distances between positive atomic nuclei and valence electrons.

    Why is a wood flame predominantly yellow-orange? It has been proposed that it's not the result of electron transitions; what's supposedly responsible is incandescence of particles at about 1100 to 1200 o.  Since the combustion of wood is incomplete, the flame's soot particles, some of which are elemental carbon(others are polymers), emit part of their vibrational energy as photons. How fast the molecules vibrate depends on their temperature, and the hotter the surface, the higher the frequency of the photons emitted. The same mechanism would account  for the red glow of logs at the base of a fire. But the temperature is lower, in the neighborhood of 700 to 1000oC, hence a color of a longer wavelength and a lower frequency.  Different parts of the charcoal emit light of slightly different frequencies, intermittently and in different directions. A point on the surface of the charcoal particle that has just emitted photons will have lost energy and cooled slightly. Although exothermic reactions quickly compensate, from that same spot, the temperature will not necessarily be identical, especially in light of air movement and the exact frequency of photons is not necessarily replicated. 

    It's well known that the ease of ignition and burning rate of wood vary greatly with moisture content. Specifically, a drop in moisture content of 10% results in an increase of 20–30% in burning rate. When wood is too dry the combustion rate increases, but an inadequate oxygen supply leads to more undesirable emissions. The combustion rate also depends on boundary conditions and the species being burnt. Why does it vary with tree type? Wood composition is not constant.  Wood is essentially a matrix of cellulose and other carbohydrate fibers (hemi cellulose) reinforced by the adhesive binding action of lignin. But hardwoods can have anywhere from 18 to 25% lignin along with varying amounts of hemicellulose, usually a partly acetylated, acidic xylan. Softer woods have other hemicellulose fibers and more "binder", 25 to 35% lignin. There is also an assortment of oils and other secondary products present.

    The different wood recipes not only affect kinetics but thermochemistry. Softwoods, compared to hardwoods, release on average an extra 5% of heat, a maximum of 21 instead of 20 MJ/kg, to be precise. From the point of view of carbon dioxide emissions, it's not a good idea to rely on wood as a primary fuel. For every MJ of heat obtained from wood, on average, 80 g of CO2 are emitted. In contrast, natural gas combustion only puts out 50 g per MJ.

    Finally, when the fire dies and we're left with ashes, what exactly are we staring at?  In general the ash is of an alkaline nature, with a pH of about 12, mostly due to the presence of carbonates of calcium and potassium, specifically CaCO3 and K2Ca(CO3)2. At higher temperatures about 1300 oC, calcium and magnesium oxides are ashes' main compounds. Those alkaline compounds were not originally present in plant tissue. Neutral metals weren't either, so the carbonates must have formed indirectly, perhaps ions   precipitating with carbonic acid, derived from water and carbon dioxide. 

    In the same way that the nature of flames and soot depends on the type of wood, the more detailed composition of ash also varies. According to a fairly recent study, it matches the needs of trees growing in a particular area. Among ashes of all conifers and broadleaves tested, that of birch trees, whose bark started this discussion, has the most calcium, the 2nd most phosphorus and is the only one without aluminum. But birch ash only has a fraction of the potassium and magnesium ions found within that of poplars and maples, respectively. Like cremated people and exploding suns, trees leave a signature in their ashes.

    • Kurt Nassau. The Physics and Chemistry of Color: the 15 Causes of Color. Wiley-Interscience. 1983 


    It's not the result of electron transitions; what's responsible is incandescence of particles at about 1100 to 1200 oC   ... emit part of their vibrational energy as photons. How fast the molecules vibrate depends on their temperature, and the hotter the surface, the higher the frequency of the photons emitted. The same mechanism accounts for the red glow of logs at the base of a fire. But the temperature is lower, in the neighborhood of 700 to 1000oC, hence a color of a longer wavelength and a lower frequency.

    Are you sure about this? The color temperature of yellow starts at around 3000 degrees while molecular vibrations usually radiate in the infrared. The typical flame-yellow, as far as I know, comes from the D-lines of sodium inside all organic material.
     The typical flame-yellow, as far as I know, comes from the D-lines of sodium inside all organic material.
    I've seen the sodium idea thrown around in a few sources but the incandescence explanation shows up in more reputable references and it also applies to a candle. I've also seen the associated temperatures quoted in several places, including "15 Causes of Color" and other sources listed at the the end of the article.
    There's very little sodium in plant material; since there's a lot more potassium,  if gas emissions were the main cause of the wood fire's color, wouldn't it be predominantly violet?Also if you don't adjust a Bunsen burner properly, you easily get an orange flame instead of the usual blue one caused by C2, CH emissions , and there are no metals present in methane. Instead, the soot produced by the incomplete combustion is responsible for the orange color.
     ...while molecular vibrations usually radiate in the infrared
    Molecular vibrations account for water and ice's blue colors, so vibrations are not restricted to infrared.

    The D lines are very bright; so you do not see the K if a little Na is also present. At least that is what they taught me, but then it was chemistry guys teaching, so it must be taken with caution, ha ha. Water is blue because of an overtone that is absorbed in the red, so the actual vibration happens still in the infrared. The problem I have with your explanation is that the temperature is not enough to get any such transition started without some extra trick (like the Na being ionized). If it is vibration (~ heat), what kind of effect makes that vibration 2000 degrees hotter than the thermal bath? As far as I can see with a very quick web search, it is still electronic transitions of the radicals.
    While the Sodium lines might be bright, in many cases you can greatly reduce the amount of yellow by increasing oxygen into to combustion process. While there may be more advanced methods now, for a long time this was how furnaces, grills, torches were all adjusted.
    Never is a long time.
    The problem I have with your explanation is that the temperature is not enough to get any such transition started without some extra trick
    There all sorts of reactions in the flame that can get the transitions started without implicating the scarce (or in the case of poorly burning methane) non-existent sodium.....
    Written later:

    I just started a little birch bark orange-flame fire in an ash tray and aimed a spectroscope at it. I don't see any Na line in it. 
    Yes, we are beyond the Na now - forget the Na. The main point already in the first comment is that you write explicitly that the vital transitions are not electronic.
    Perhaps the answer is that my color temperature chart is misleading.  I looked at color temperature, so yellow seems to need 2000 to 3000 C:

    However, looking at subjective color from incandescence confirms your statements as written in your post:

    Subjective colour to the eye of a black body thermal radiator

    °C (°F) Subjective colour[5]
    480 °C (896 °F) faint red glow
    580 °C (1,076 °F) dark red
    730 °C (1,350 °F) bright red, slightly orange
    930 °C (1,710 °F) bright orange
    1,100 °C (2,010 °F) pale yellowish orange
    1,300 °C (2,370 °F) yellowish white

    I am confused about how these two are consistent. Somebody know what is going on?
    Thor Russell
    If the RGB detectors in your eye are not equally sensitive, then you could perceive the peak at the wrong place. e.g. if there are 2* as many red photons as green, but your green receptor is more sensitive, it could look more green than red?
    Thor Russell
    I like this. A wikipedia writer has that problem and starts to give different temperatures according to the hue experience he remembers from when his eyes were still normal while heating up a piece of iron or so and not looking at any standardized color maps, thus telling the world differences of more than 1000 degrees. Any other ideas?
    Thor Russell
    ? not following. In your thermal radiation link you can see metal appearing yellowish, but it is less than 3000 C because it would have presumably melted otherwise.I meant something like this:

    "So a piece of hot metal will appear white because although it's peak emmission will be in the infrared it is still emitting a lot of power at shorter wavlengths and your eye is more sensitive to blue than red."

    Thor Russell
    you can see metal appearing yellowish, but it is less than 3000 C
    Yes, and now tell us how the other article gets a "color temperature" of 3000 K for yellow, having in mind that these are based on fundamental concepts like black bodies and standardized concepts like hue.

    Or are my eyes so messed up that the yellow is not where the 3000 is, combined with the coincidence that the monitor pixels and my eye problems conspire to still let me see the same yellow hue on the photos of flames?

    So, seriously now - how does carbon sooth (pretty darn black - even more so than iron!) lead to a color that belongs to a 3000 K black body via vibrational degrees of freedom excited at 1200 C? Something here is tricky, and it is not my eyes.
    The soot could still be burning, glowing like the end of a lit cigarette, but that doesn't really explain a piece of steel glowing at the same "Temp".
    Never is a long time.
    Hi Sacha

    „Color“ is a strictly subjective („psychological“) feeling of the perception of electromagnetic radiation. I. Newton wrote in his “Opticks”: „For the Rays, to speak properly, have no Colour”. Color is not a physical concept.
    The identical feeling of “yellow” can be produced by the perception of electromagnetic radiation with a wavelength of about 525nm or of a mixture of electromagnetic radiation with wavelength of 420 nm and of 570 nm. Look at the yellow part of a rainbow and you see the “525nm-yellow”, look at your computer screen with a picture of a rainbow and you see the “420nm + 570nm – yellow”. If you see those two “yellows” differently, you probably have not the standard visual system of humans (eg some kind of “color blindness”). There are many other combinations of different wavelengths which will produce the same feeling of “yellow”.

    „Color temperature“ is a propriety of a mixture of electromagnetic radiation of usually many different frequencies. If a “black body” is heated to 3’000 K, it emits a mixture of electromagnetic radiation (according to Plancks formula) which is perceived as “yellow”. Any other mixture of electromagnetic radiation which is perceived as the same “yellow” has a color temperature of 3’000 K. Usually one does not speak of color temperature, if the radiation consists only of one or a few wavelengths.

    That the two concepts of “color” and “color temperature” are different, you can see in the diagram you provided. The possible colors are represented at the outer boundery of the “colored” part of it. Along the upper part the “colors of the rainbow”, perceptions that can (but not “must”) be produced by electromagnetic radiation of any single wavelength from 380 nm to 700 nm. Along the lower, straight line (named “purple line”) the colors that can only be perceived when at least two different wavelengths of radiation are present. There is no “purple” part of a rainbow. The possible perceptions of a hot “black body” by a human beeing are along the curved line in the center of the diagram and are going from red to yellow to white (5’800 K, the temperature of the “surface” of the sun) to blue. Here you have no “green”, you can heat a black body to any temperature you like, it will never be percepted as glowing “green”.

    Discussions about “color” are usually corrupted by mixing up physical concepts (wavelength), physiological concepts (light receptors in the eye, signal processing) and psychological concepts (color).

    You have explained exactly zero news to me and not even touched on the actual question!  What is wrong with you people?  The question is about how the candle can get there via incandescence (which is nothing other than saying that it is from black body radiation, especially with carbon particles whose surface cannot even be said to be very far from black, because, carbon is - here it comes - BLACK).
    especially with carbon particles whose surface cannot even be said to be very far from black,

    Soot is not all carbon particles. While in the flame, some of the soot consists of polymerized fuel.
    Yes, which is the only answer that is left possible I think if the color on the 3000K-graph should be correct.  Because something is clear since the very start of this discussion: It simply cannot be thermal (vibrations) and at the same time a black body (BB), because there is no other dependence but temperature in that case.  So if it is thermal (not electronic etc), it cannot be a black body, i.e. the emissivity of the surfaces cannot be =1.  The problem is only that your link actually does not agree with this:
    "This radiation appears reddish-orangish-yellowish. Chemical reactions in the flame plasma also emit radiation, so the emission spectrum of a complete candle flame can be quite complex. However, the characteristic continuum spectrum of the blackbody radiation from the soot is the dominant feature."
    So - if the dominant feature is BB radiation, then why is it looking like 3000 K if it is only 1200 C? Still do not get how this can be consistent.  Something is missing every time.
    Sascha, maybe if you place a thermometer in the yellow part of the flame, it reads 1200 C, but the soot is actually 3000K because it's still burning.
    Never is a long time.
    Now this is finally something plausible, in case the temperature of 1200 C is the average temperature of the gas/soot mix, which is obviously not in a static equilibrium but a rapid flow equilibrium.  The misleading part would then be (from Enrico's article):
    "what's supposedly responsible is incandescence of particles at about 1100 to 1200 oC ."
    Let's find out how they measured the temperature of 1200 C.  If they used a small thermocouple, you may be correct.  If they used the infrared emission, we are back to square one.
    Dang, I burnt out my fine wire K-type thermocouple.
    Before I fried it, I stuck it into the edge of the bright yellow top edge of a bic butane lighter flame. It maxed out the display at 999C, but it did struggle a bit to get there. With just the tip in the edge it was in the 700-800C range.

    So, it would seem to me that the temperature values of a flame that are quoted are as measured by thermometer.

    The thermocouple wires were glowing various shades of red, orange to a little yellow.
    Never is a long time.
    Assuming that the heat drawn away by the wires is negligible and that you reached the max (i.e. it would have stayed at 999C if you had measured longer), it seems you may have solved the problem.  Especially I like that you also tell us the color of the thermocouple.

    It amazes me again and again: how many articles have we collectively looked at now, all repeating the same mistake plus adding a few more?
    Thor Russell
    This article claims that the soot is not like a simple black body so it appears hotter:

    (although it says candle flames are 700F)

    Have little idea why the color temp of a candle flame is put down as orange-red however, it looks pretty yellow here and in real life even outside in the sun.
    Even so, still not sure how the 3000K figure can be correct because doesn't molten metal (a good BB?) appear yellow at a lower temp than that?
    Thor Russell
    From your astronomy link:
    "The color does not come from black body emission because 'red' would indicate black body temperatures of 2000 K, orange of 3000 K and yellow of 6000 K."
    Totally wrong temperatures!  Astronomers worse than chemists? ;-)
    Thor Russell
    Whatever the answer is to this sorry mess, I think your consistency is out of the picture now unless it involves consistently wrong answers being given by people who shouldn't.
    Thor Russell
    Thanks for the above charts. I don't know what's going on with the subjective interpretation.
    Sascha, here's a specific reference to the hypothesis that the red to yellow colors of a log fire are due to incandescence.
    (see page 131 of the article)
    If it's not valid, it made it past the editors of Scientific American at a time when they were at their best.

    Thank you Enrico.  The article seems fine.  I have already, as you can see in the comments above, come to the conclusion that the confusion is about that I either should not trust the colors in a wikipedia article, or there is something that I do not understand about the definition of "color temperature".  As you can see, the yellow 3000 K, well, something is "non trivial" with that. ;-)
    Thor Russell
    There are no end of images here claiming a candle flame is at least as much red as orange:
    Seriously, WTF?

    Thor Russell
    Where do these people buy their candles; what kind of candles?