Chemical Soup: The Mystery of Detonation

By Dr. Robin Tuluie Ph.D.

Detonation is the result of an amplification of pressure waves, such as sound waves, occurring during the combustion process when the piston is near top dead center (TDC). The actual "knocking" or "ringing" sound of detonation is due to these pressure waves pounding against the insides of the combustion chamber and the piston top, and is not due to 'colliding flame fronts' or 'flame fronts hitting the piston or combustion chamber walls.' Let's look in some detail at how detonation can occur during the combustion process: First, a pressure wave, which is generated during the initial ignition at the plug tip, races through the unburned air-fuel mix ahead of the flame front. Fuel 1 Typical flame front speeds for a gasoline/air mixture are on the order of 40 to 50 cm/s (centimeters per second), which is very slow compared to the speed of sound, which is on the order of 300 m/s. In actuality, the true speed of the outwards propagating flame front is considerably higher due to the turbulence of the mixture. Basically, the "flame" is carried outwards by all the little eddies, swirls and flow patterns of the turbulence resident in the air-fuel mix. This model of combustion is called the "eddy burning model" (Blizzard & Keck, 1974). Additionally, the genus of the flame front surface - that is the degree of 'wrinkling' - which usually has a fractal nature (you know, those weird, seemingly random yet oddly patterned computer drawings), is increased greatly by turbulence, which leads to an increased surface area of the flame front. This increase in surface area is then able to burn more mixture since more mixture is exposed to the larger flame front surface. This model of combustion is called the "fractal burning model" (Goudin, F.C. et al. 1987, Abraham et al. 1985). The effects of this are observed in so-called "Schlieren pictures," which are high-speed photographs taken though a quartz window of a specially modified combustion chamber (Fig. 1, above). Schlieren pictures show the various stages of the combustion process, in particular the highly wrinkled and turbulent nature of the flame front propagation (initially called the flame 'kernel'). A higher degree of turbulence, and hence a higher "effective" flame front propagation velocity can be achieved with a so-called squish band combustion chamber design. Fuel 2 Sometimes a swirl-type of induction process, in which the incoming mixture is rotating quickly, will achieve the same goal of increasing the burn rate of the mixture. As a general rule-of-thumb the pressure rise in the combustion chamber during the combustion phase is typically 20-30 PSI per degree of crankshaft rotation. Once the pressure rises faster than about 35 PSI/degree, the engine will run very roughly due to the mechanical vibration of the engine components caused by too great of a pressure rise. Sometimes, the pressure wave can be strong enough to cause a self ignition of the fuel, where free radicals (e.g. hydroxyl or other molecules with similar open O-H chains) in the fuel promote this self ignition by the pressure wave. However, this can still occur even without the presence of free radicals; it just won't be quite as likely to happen. This is why high octane fuels, with fewer of these active radicals, can resist detonation better. However, even high octane fuel can detonate - not because of too many free radicals - but because the drastic increase in cylinder pressure has increased the local temperature (and molecular speed) so high that it has reached the ignition temperature of the fuel. This ignition temperature is actually somewhat lower than that of the main hydrocarbon chain of the fuel itself because of the creation of additional radicals resulting from the break-up of the fuel's hydrocarbon chains in intermolecular collisions. Detonation usually happens first at the pressure wave's points of amplification, such as at the edges of the piston crown where reflecting pressure waves from the piston or combustion chamber walls can constructively recombine - this is called constructive interference to yield a very high local pressure. If the speed at which this pressure build-up to detonation occurs is greater than the speed at which the mixture burns, the pressure waves from both the initial ignition at the plug and the pressure waves coming from the problem spots (e.g. the edges of the piston crown, etc.) will set off immediate explosions, rather than combustion, of the mixture across the combustion chamber, leading to further pressure waves and even more havoc. Whenever these colliding pressure fronts meet, their destructive power is unleashed on the engine parts, often leading to a mechanical destruction of the motor. The pinging sound of detonation is just these pressure waves pounding against the insides of the combustion chamber and piston top. Piston tops, ring lands and rod bearings are especially exposed to damage from detonation. In addition, these pressure fronts (or shock waves) can sweep away the unburned boundary layer (see figure 2 above) of air-fuel mix near the metal surfaces in the combustion chamber. The boundary layer is a thin layer of fuel-air mix just above the metal surfaces of the combustion chamber (see figure 2, above). Physical principles (aptly called boundary conditions) require that under normal circumstances (i.e. equilibrium combustion, which means "nice, slow and thermally well transmitted") this boundary layer stays close to the metal surfaces. It usually is quite thin, maybe a fraction of a millimeter to a millimeter thick. This boundary layer will not burn even when reached by the flame front because it is in thermal contact with the cool metal, whose temperature is always well below the ignition temperature of the fuel-air mix. Only under the extreme conditions of detonation can this boundary layer be "swept away" by the high-pressure shock front that occurs during detonation. In that case, during these "far from equilibrium" process of the pressure-induced shock wave entering the boundary layer, the physical principles allured to above (the boundary conditions) will be effectively violated. The degree of violation will depend on (a) the pressure fluctuation caused by the shock front and (b) the adhesive and cohesive strength of the boundary layer. These boundary layers of air-fuel mix remain unburned during the normal combustion process due to their close proximity to the cool metal surfaces and act as an insulating layer and prevent a direct exposure of metal to the flame. Since pressure waves created during detonation can sweep away these unburned boundary layers of air-fuel mix, they leave parts of the piston top and combustion chamber exposed to the flame front. This, in turn, causes an immediate rise in the temperature of these parts, often leading to direct failure or at least to engine overheating. Scientists and engineers have recently begun to understand combustion in much greater detail thanks to very ambitious computer simulations that model every detail of the combustion process (Chin et al. 1990). Basically, a complete computer model includes a solution to the thermodynamical problem, that is a solution to the conservation equations and equation of state, as well as a mass burning rate and heat transfer model. In addition, a separate code (called a chemical kinetics code) models the chemical processes which occur during combustion and sometimes juggles several thousand different chemical species, some in vanishingly small concentrations! Needless to say these codes require huge amounts of memory and CPU time that only the largest supercomputers in the world can provide. They are far beyond the reach of the private individual and usually only employed by large research institutions or major car manufactures.
Here's a brief recital of the question we received:
Someone asked:
Rob, I read your "Chemical Soup: The Meaning of Gasoline." Quick question if you have the time... You mentioned that "flame propagation is greatly enhanced by turbulence." Should this be a consideration when an engine is ported? Can turbulence be enhanced by porting without losing the intake flow?
Unless the ports are specifically designed for a strong swirl-type induction process, the turbulence created during the intake process is not very affected by porting. This is true as long as one sticks with the same general port layout. However, drastic porting changes may increase or decrease the turbulence in the combustion chamber, but it is quite difficult to say anything definite. I think that any improvement gained by porting the engine is likely to be far greater than any possibly detrimental effect the porting may have had on turbulence.
As far as I know there is only one motorcycle engine that uses a highly turbulent intake process of the swirl type. It's a "homebuild" single cylinder racing engine from Switzerland that uses cross-scavenging and has two pairs of diagonally opposed intake and exhaust valves. Most conventional ports do induce a very small amount of swirl, but this is not important as far as generating much turbulence. Rather, the biggest benefit is obtained by reducing the squish band to it's safe minimum (about 0.020-0.040 in, depending on the particular engine used). This will have a far greater effect on increasing the turbulence in the combustion chamber than any other modification.
Mike Meagher ( wondered about the effects of the squish band.
It is important to realize the two important functions of reducing the squish band clearance: (a) to enhance turbulence due to rapid ingestion of gas into the combustion chamber, hence increasing the burning rate of the mixture and (b) to reduce the volume of the unburned gas in the boundary layer of cool gas near the piston top and cylinder head surfaces. Typically, gas trapped in the squish area doesn't burn, even if the squish band clearance is relatively large. The cooling effects of the large surface-area-to-volume ratio of this region will prevent any ignition of the fuel-air mix therein, even if the squish band clearance is rather large. Hence any gas caught in the squish band will not be burned near TDC when it does the most good, but later during the combustion process when one cannot extract as much work from the late-burning gases. The amount of gas trapped in the squish band can actually be a substantially greater amount than just the relative volume of the squish band because the pressure wave from the ignition process literally crams a lot of the unburned gas into crevice areas like the squish band. Reducing the squish band clearance will decrease the amount of unburned gas substantially, leading to more complete and faster combustion, lower emissions and improved power. It is one of the few "all gain with no pain" modifications one can carry out on racing or even street motorcycles.

Someone wondered: Is the extra cooling of the squish band less than the added heat?
Basically the mixture in the squish region is in thermal contact with the cylinder wall and piston top and at roughly the same temperature, which is quite lower than the burn temperature. Reducing squish will decrease the amount of the cool gas in the squish region and increase the amount of hot gas in the burn region. A reduced squish clearance will increase temperatures a little even if the compression ratio is held constant. There is no "extra cooling" mechanism if you reduce the squish band clearance. The cooling rate of the gas in the squish zone depends on the thermal conductivity of the gas-metal interface, on the total surface area of this interface and the temperature difference between gas and metal. Note that these factors are all essentially constant at TDC and don't depend on the squish clearance. Hence the cooling rate is the same for large squish clearances and for small squish clearances. Thus there is no "extra cooling" mechanism if you reduce squish band clearance.

David Goodenough ( asked:
Suppose I mix one gallon of 87 octane pump gas, and one gallon of 92 octane pump gas. Are you telling me that instead of two gallons of 89.5 octane gas I have something closer to 92 (like between 90 and 91)?
The mixed gas' octane rating will in general not be a linear function of the original constituents' octane ratings. Neither will it be a simple function in most cases. Rather, the octane rating becomes a quite complicated, non-linear function of some very small amounts of free radicals, such as hydroxyl and hydroxen peroxide, in the fuel. Essentially, there is no simple analytic way to predict the final octane rating of a fuel; rather, extensive tests with a calibrated engine are necessary (see MON and RON explanations in the last article).
David also asked:
While I'm at it, how does the energy per ounce of mixture react?
As mentioned before, the "energy per ounce" (more exactly the Specific Energy for an stoichometric [an ideal] mixture) does not vary much at all between different kinds of pump gasoline or even racing gasoline.
Ramon Hontiveros ( wrote:
, I got the article and read it, now some questions: Isn't the fuel already in gaseous form due to carburation?
Ramon, the air fuel mix as it flows into the combustion chamber is not perfectly atomized, that is the fuel vapor droplets consist of larger droplets of fuel molecules surrounded by air. It takes additional energy to further atomize this vapor, that is to break the hydrostatic forces (the surface tension of the fuel droplet). This additional energy can be taken from a hot surface (such as the piston crown, etc.), which then leads to a cooling of the piston. The additional energy can also be imparted via large turbulences and pressure waves, as in a squish band-type motor, which will help to further atomize the fuel. Note that the term "atomize" is actually misleading since the molecules are still left intact, that is the hydrocarbon chains (and oxygen bonds for alcohols) are not broken.
Ramon also wondered:
does carburation just "spray" the gas into the air flow as tiny droplets which are thus still in liquified form?
Ramon also asks:
Also, if the fuel does evaporate quickly and creates additional pressure - thus reducing the amount of fresh charge - then the engine will produce less horsepower, right?
Correct. The horsepower will depend on the volumetric efficiency of the engine which is a function of the pressure difference between ambient air and cylinder pressures. If additional fuel is vaporized inside the combustion chamber the pressure in the cylinder will rise, and, while the valves/ports are still open, reduce the volumetric efficiency, and thus the power output.
So 2-strokes would benefit from using fuel that has a _lower_ heat of vaporisation rating?
Correct. A fuel with a lower heat of vaporisation will "atomize" easier and thus improve engine cooling, but decrease power somewhat.
So which type of fuel has a lower heat of vaporisation? Leaded or unleaded?
A fuel's heat of vaporisation does not depend on the it's lead content. Rather, it depends on the fuel's main hydrocarbon chains; iso-octane verses n-heptane, for example. Since pump gas can consist of up to 20 different components with a wide range of individual boiling points (We were serious when we called it "chemical soup!") one should look at the specifications sheet for each fuel separately. For racing fuels these are available from the manufacturer.
Lastly, I gather from the article that it's okay if you end up mixing some leaded fuel with the remaining unleaded fuel in the tank?
Most fuels, pump or racing, will give about the same energy release, so when switching from pump to racing fuel (in general) do not expect a drastic increase in power. Mixing leaded fuel with the remaining unleaded fuel in the tank has no advantage and will give inconsistent plug readings; hence I wouldn't do it on a race bike.


Abraham, J. et al., 1985, "A Discussion of Turbulent Flame Structure in Premixed Charges", SAE paper 850345 Blizzard, N.C. and Keck, J.C., 1974, "Exp. and Theo. Investigation of Turbulent Burning Model for Internal Combustion Engines", SAE paper 740191 Chin et al., 1990, "Diagnostics and Modeling of Combustion in Internal Combustion Engines," JSME, Tokyo, p. 81-86 Goudin, et al., 1987, "An Application of Fractals to Modeling of Premixed Turbulent Flames", Combustion and Flame 68, p.249-266

Disclaimer: The information contained on this page and on this site is condensed from the combined wisdom of the members and contributors of the Orange Crush Forum. The contributions are reprinted here exactly as posted by the contributors. The spelling, syntax, grammar, etc have purposely not been corrected in order to retain its original flavor. The contributors are from throughout the World, and English may very well not be their native language. Don't be an ass and complain about the lexicon. It is mostly subjective, with a little objectivity thrown in for seasoning, based on the experiences of the contributors. Use this info at your own risk. The site owner is not responsible for its accuracy or validity. None of the procedures described should be taken as recommendations by anyone. Take anything you read or hear anywhere, but especially on the World Wide Web with a very large dose of salt. The cognoscente is a skeptic.