How Combustion Actually Works

Why this matters

The chemistry of burning a hydrocarbon is fixed, published and about a century and a half old. Nothing about it is in doubt, and nothing in the field fails because the chemistry misbehaved. What fails is the delivery: whether the right quantity of air reached the fuel, whether the two mixed before they reached the flame front, and whether the reaction had enough time at temperature to finish before something cold interrupted it. Once you hold combustion as a mixing-and-timing problem rather than a chemical one, the whole diagnostic set reorganises, and a burner that makes carbon monoxide with a perfectly correct gas pressure stops being a mystery.

Before you stand in front of an operating burner

Flue gas contains carbon monoxide, which is colourless and odourless at every concentration that will hurt you. Wear or place a personal carbon monoxide monitor and have it running before you fire the appliance, not after. OSHA's general industry permissible exposure limit is 50 ppm as an 8-hour time-weighted average at 29 CFR 1910.1000 Table Z-1, and NIOSH publishes a 35 ppm recommended time-weighted average with a 200 ppm ceiling. If ambient carbon monoxide in the occupied space climbs toward those figures, or anyone reports headache, nausea or confusion, everyone leaves the building immediately, nobody operates a switch on the way out, the space is ventilated from outside, and nobody re-enters until an instrument reads it clear.

If you smell gas at any point, that is a different instruction and it is not a diagnosis: everyone out immediately, no switches, no lights, no phone used inside, shut the gas off at the exterior shutoff only if it is on your path out, and call from outside. Never test for a gas leak with a flame; use a listed combustible gas detector or a leak-detection solution.

Before you open a burner compartment, close the appliance gas shutoff and let the assembly cool, then handle it with gloves rated for contact heat, because a heat exchanger and burner rack hold enough stored heat to cause a deep contact burn long after the flame is out. And if the appliance carries older flue cement, pipe lagging or gasket material of unknown composition, do not cut, drill or abrade any of it: those materials can contain asbestos, the route is inhalation, and the duty sits at 29 CFR 1910.1001 in general industry and 29 CFR 1926.1101 in construction. Stop and have it identified.

Fuel, oxygen, ignition, and the leg that gets dropped

The fire triangle gives you fuel, oxygen and an ignition source, and every tech can recite it. The tetrahedron adds the self-sustaining chain reaction. Neither of them tells you why a burner that has all four makes carbon monoxide.

The version that does the work in the field is the one combustion engineers use, and it is three conditions rather than three ingredients:

• Temperature. The reaction has to reach and hold its ignition temperature throughout the burning volume, not just at the flame front.
• Turbulence. Fuel and air have to actually mix at the molecular scale. Adjacent is not mixed.
• Time. The reaction chain runs through intermediate products, and it needs enough milliseconds at temperature to run to completion.

Every field combustion fault is a shortage of one of those three. Fuel supply and air supply are inputs you can measure and set. Temperature, turbulence and time are what the burner and the combustion chamber do with those inputs, and they are where the appliance either finishes the job or does not.

The chemistry, stated once, so you can stop guessing at it

For methane, which is the dominant component of natural gas as delivered:

CH4 plus 2 O2 gives CO2 plus 2 H2O.

Air is about 20.9% oxygen by volume, the rest overwhelmingly nitrogen. So the two volumes of oxygen have to arrive inside 2 divided by 0.209, which is 9.57 volumes of air. That drags 7.57 volumes of nitrogen into the appliance that take part in nothing, absorb heat, and leave up the flue carrying it.

Take the water out, which is what a flue gas analyser does when it samples dry, and the products are 1 volume of carbon dioxide in 1 plus 7.57, or 8.57 volumes. That is 11.7% carbon dioxide, and it is the theoretical ceiling for methane. You will never measure higher than that on a methane-dominant fuel, and a reading near it means almost no excess air.

For propane the same arithmetic runs C3H8 plus 5 O2 gives 3 CO2 plus 4 H2O. Air required is 5 divided by 0.209, or 23.9 volumes per volume of propane, nitrogen along for the ride is 18.9, dry products are 3 plus 18.9 or 21.9, and the theoretical carbon dioxide ceiling is 13.7%. Different fuel, different ceiling, same method. If your analyser is set to the wrong fuel, every derived number it prints is wrong by that difference, and it prints them without complaint.

The balance sheet for one real appliance

This is the artifact worth keeping in your head, filled in for a 100,000 Btu/h input natural gas appliance. Heating value for natural gas as delivered commonly runs near 1,000 Btu per cubic foot, but it does vary by supplier and season, so confirm it with the gas utility before you treat any derived flow as exact.

Three things fall straight out of that sheet, and they are the reason to build it.

Air dwarfs fuel. For every cubic foot of gas the appliance must be handed about 10 cubic feet of air at exact stoichiometry and about 14 as it is actually run. Nobody sizes a gas line by eye and nobody should size an air path by eye either, yet the air path is the one that routinely gets left to whatever the room happens to have. What the building has to do to deliver that air is its own subject and has its own article.

The flue moves more than the burner consumes. About 36 actual cubic feet per minute of hot products leave a 100,000 Btu/h appliance, against 1.67 CFM of gas going in. Roughly 22 volumes leave for every volume of fuel. A vent that is undersized, blocked or fighting the building is not a minor restriction on a small flow.

Nitrogen is the tax. Of the 1,388 cubic feet per hour of air, only about 200 cubic feet per hour is oxygen that reacts. The rest is inert gas you heat to flue temperature and discard. That is the physical reason excess air costs efficiency, and what it costs is worked out in its own article rather than here.

Why the theoretical ratio never happens

Set an appliance to exactly 9.57 volumes of air per volume of gas and it will make carbon monoxide, sometimes a lot of it. The chemistry says the oxygen atoms are all present. The burner cannot get them to the right place fast enough.

Mixing in a real burner is not molecular. Gas issues from a port at high velocity, entrains some air on its way to the flame, and the rest arrives from around the flame envelope. There are always local pockets running rich and local pockets running lean, and a pocket running rich has no oxygen available regardless of the room average. This is why every practical appliance is run with surplus air: the surplus is not there to be consumed, it is there to make sure the lean pockets outnumber the rich ones.

Where the reaction gets interrupted

The chain from methane to carbon dioxide runs through intermediates, and carbon monoxide is one of the last ones before completion. It is the reaction stopping one step short. Three field conditions stop it, and all three are quench mechanisms rather than chemistry:

• Flame impingement. The flame touches a surface cooler than the ignition temperature, usually a heat exchanger wall, and the reaction stops on contact. This is the single most common field cause and it is a geometry problem: a burner sitting wrong, a cracked or distorted exchanger, a lifted flame reaching further than it should, a sooted surface that has changed the flame's shape.
• Insufficient residence time. The products leave the combustion zone before the chain finishes, which happens when a burner is overfired, when a vent pulls too hard, or when a chamber has been changed by a repair.
• Local oxygen starvation. The mixture is rich in the place that matters even when the room has ample air. Blocked ports, a shifted orifice, a partly obstructed air path serving one burner in a rack.

Notice that none of these are cured by adding more air to the room. The distinction between a room-air problem and a burner-geometry problem is what a flue gas measurement is for, and reading that measurement is its own article.

What changes the numbers

Fuel. Switching between natural gas and propane changes air demand per volume of fuel by roughly two and a half times, changes the carbon dioxide ceiling from 11.7% to 13.7%, and requires different orifices. Set the analyser's fuel selection before you take a single reading.

Altitude. Air density falls with elevation, so the same volumetric air flow delivers less oxygen mass. Appliance derating and orifice changes above roughly 2,000 feet are manufacturer- and code-specific; read the installation instructions for the elevation rather than assuming a percentage.

Combustion air temperature. Air pulled from a cold outdoor duct is denser than air from a warm mechanical room, so the same duct delivers more oxygen mass in January than in July. That is one reason a burner set up on a mild day can run rich on the coldest night of the year, and why setup readings carry the outdoor condition with them.

How to verify you understand the appliance in front of you

Build the balance sheet for the actual appliance before you touch an adjustment. Take the nameplate input, confirm the fuel and its heating value with the utility, and write down the gas volume, the stoichiometric air, and the flue volume at temperature. Then clock the meter to confirm the appliance is firing at its nameplate rate rather than at whatever it drifted to, because every line below the first is wrong if the input is wrong. If clocked input and nameplate input disagree by more than a few percent, that discrepancy is the first thing to resolve, and no combustion adjustment made before resolving it will hold.

References

• 29 CFR 1910.1000 Table Z-1, general industry permissible exposure limit for carbon monoxide
• 29 CFR 1910.1001 and 29 CFR 1926.1101, asbestos duties in general industry and in construction respectively
• NFPA 54 (ANSI Z223.1), National Fuel Gas Code, for appliance installation and gas supply provisions
• Manufacturer installation instructions for fuel type, orifice sizing and high-altitude derating
• See related: What Excess Air Does to Efficiency; How to Read a Flue Gas Result