The off-gas leaving a membrane upgrader at a renewable natural gas plant is better than 90 percent CO₂. That is the entire purpose of the machine. It is also why the stream will not burn. Methane's lower flammability limit is 5 percent in air, the slip carried in that off-gas is closer to one, and CO₂ is a fire suppressant in its own right. So the thermal oxidizer handling that stream gets supplemental natural gas, metered and paid for, to hold temperature.
An RNG plant is buying fuel to destroy the cleanest CO₂ stream on the property.
The Oxidizer Is Not a Mistake
Nothing about that arrangement is an oversight. A membrane or PSA upgrader recovers most of the methane out of raw biogas, which runs 35 to 45 percent CO₂ before treatment, but it does not recover all of it. Whatever slips through leaves with the CO₂, along with whatever trace organics the feedstock carried in.
On a landfill gas project the requirement is written down. NSPS subpart XXX and the subpart Cf emission guidelines require a control device to cut non-methane organic compounds by 98 percent by weight, or bring them to 20 ppmv as hexane. A regenerative thermal oxidizer sitting around 1,500°F clears that without difficulty.
On a dairy or wastewater digester there is usually no federal control device requirement at all, and the pressure arrives from the other direction. Methane slip lands in the carbon intensity score the RIN and LCFS pathways are built on, and a released molecule scores worse than a destroyed one. Different driver, same equipment, same 8,760 hours a year.
What the Fuel Bill Is Buying
A regenerative oxidizer recovers most of its own heat. Ceramic beds trap the exhaust energy and preheat the incoming gas, and a well-run unit gets 90 to 95 percent of it back. The remainder has to come from somewhere. On a stream with real heating value, the stream covers it. On an upgrader's off-gas it cannot. CO₂ has no heat of combustion left in it, and a percent or two of methane in a bath of CO₂ is not a fuel. The gap gets filled with purchased gas, every hour, for the life of the plant.
That is a line item an operator can find in about ten minutes. Very few have ever priced it against the alternative, because until recently there was no alternative worth pricing.
Dilution Does Not Reverse
The fuel is the visible cost. The larger one is what combustion does to the stream. Air goes into the oxidizer to support the flame and hold the chamber in its operating band, and nitrogen rides in with it. A stream that arrived above 90 percent CO₂ leaves the stack several times more dilute, mixed into nitrogen, oxygen and water vapor.
Concentration is the one property of a CO₂ stream you cannot buy back cheaply. Pulling CO₂ out of a dilute flue gas is the reason amine systems exist, with an absorber, a stripper, reboiler steam and solvent losses to manage. Taking it off a stream that is already above 90 percent is purification work. Those two sit at opposite ends of the cost curve, and the oxidizer is the piece of equipment that moves a site from one end to the other.
So the decision lives at the tie-in point. A capture train has to come off upstream of the oxidizer or it never sees the concentrated stream at all. Every design question after that one is downstream of a choice already made in the piping.
Purity Is Still the Work
Ninety percent CO₂ is not a product. A brewer, a bottler or a food processor buys against a beverage grade specification: 99.9 percent CO₂ minimum, benzene held under 20 parts per billion, sulfur compounds below the level a person can taste, and a certificate of analysis tied to the batch riding on that truck. Nothing short of that gets through a distributor's receiving check.
Biogas off-gas arrives short of the spec in ways that depend on where it came from. Hydrogen sulfide from the digester. Siloxanes if the feedstock includes landfill or wastewater gas. Oxygen and nitrogen picked up during upgrading, both of which the spec limits and neither of which is easy to strip. The contaminants that decide whether a load qualifies are present in parts per billion, and bulk separation equipment does not touch them.
Closing that gap is what our cryogenic purification, now patent pending, was built to do. CleanCycleCarbon entered continuous commercial beverage grade production in December 2024 on a biogenic stream, which is the part of this that took the longest to prove. Capturing the CO₂ was never the hard step. Making it good enough to drink was.
The oxidizer does not leave when a capture skid shows up. It still handles startup, upset conditions and the hours when the capture train is down, and on a permitted landfill site it has to stay regardless. What changes is that it stops being the only path the CO₂ has.
What to Check on Your Own Site
Start at the fuel meter on the oxidizer. Annual supplemental gas consumption is what the plant currently spends to make the stream disappear, and on most sites that number is sitting in the historian, trended and never read.
Then pull composition and flow on the off-gas itself, ahead of the oxidizer rather than at the stack. Flow converts to tons per day, which decides whether there is a project at all. Composition tells a purification designer what has to come out and how hard it will be. A landfill-derived stream and a dairy stream can read the same percentage CO₂ and still be two entirely different problems once you get down to parts per billion.
The permit matters too. If a destruction efficiency is written into it, the oxidizer stays and the capture project gets designed around it. If the driver is a carbon intensity score instead, there is more room to move, and the slipstream taken for capture is methane that never reached the stack in the first place.
None of that needs a feasibility study to begin. A fuel meter reading and a gas analysis are enough to tell an operator whether the plant is paying to throw away something a bottler two hours down the road is trucking in from four states away.
