Renewable Natural Gas From Anaerobic Digestion: Why Projects Stall at the Injection Meter

The first time I watched an anaerobic digestion RNG skid hit nameplate methane recovery and still get turned away at the pipeline, it was an oxygen analyzer that did it. In 2023 a mid-size digester-gas project I was reviewing lost four months of pipeline injection to an oxygen trip. The membrane train was producing 97% methane, dry, with hydrogen sulfide under 4 ppm [commissioning data], clean biomethane production by any datasheet. Yet the utility's point-of-receipt meter kept failing the gas on oxygen at around 0.6%, against a tariff cap of 0.2% [utility tariff]. The air was leaking in through a cracked digester cover seal. That's the part of renewable natural gas that never makes the feasibility deck.

I've commissioned enough waste projects to hold a strong prior here, and it's this: anaerobic digestion biogas projects don't fail at the digester. The biology is the most solved part of the whole chain. They fail downstream of the upgrader - at the gas-quality meter, in the interconnection queue, and at the offtake. If you're scoping an RNG project and your risk register spends three pages on digestion and half a page on injection, you've got it backwards.

Methane purity isn't the spec that bites you

Everyone quotes the methane number. Pipeline tariffs want roughly 96-98% methane for injection (per the EPA's 2024 overview of renewable natural gas from biogas), and any competent biogas upgrading skid clears that easily. So methane purity is rarely what stops you. The specs that bite digester gas are the ones nobody frames first: oxygen, Wobbe index, water dew point, and total inerts.

Oxygen is the quiet killer. Most US pipeline tariffs cap O2 somewhere between 0.2% and 10 ppm depending on the operator [typical tariff range], and anaerobic digestion is structurally prone to oxygen ingress. Covers, foam, pumps, and the small air dosing some plants use for biological H2S control all push trace O2 and nitrogen into the raw gas. So why does a perfectly tuned upgrader push you out of spec? Because CO2 removal concentrates whatever O2 and N2 you started with. Strip 40% of the CO2 out of the stream and the residual oxygen fraction climbs, it doesn't fall. Your upgrader can be working exactly to datasheet and shove you out of spec at the same time.

Then there's Wobbe index. US distribution systems generally want a Wobbe number in the 1,330-1,400 band [typical tariff range]; biomethane sits comfortably there, but if you over-reject CO2 or under-reject nitrogen you can drift to the edge. And water: pipelines typically hold you to a 7 lb/MMscf dew point or tighter. None of these limits is exotic. The trouble is that digester gas composition moves with feedstock and temperature, so the spec you pass in commissioning week isn't the spec you pass in February. The first commissioning week is when the assumptions meet reality, and gas quality is where they meet it hardest.

Upgrading is a recovery-versus-oxygen tradeoff

The four mainstream upgrading routes (membrane, pressure-swing adsorption, water scrubbing, amine) all get sold on methane recovery. That's the headline number, and it does matter, both for economics and for your carbon-intensity score. But recovery is not the axis that decides an AD project. Oxygen and nitrogen rejection is.

Read that table and the pattern jumps out: none of the volume upgrading technologies remove oxygen. A multi-stage membrane system will give you methane recovery above 99.5% (per a 2024 review in MDPI's Energies), and membrane-PSA hybrids reach over 96% purity at over 99% recovery while trimming roughly 15% off upgrading cost (per the same review). All true. And none of it touches your O2 problem. If your raw gas carries oxygen you need catalytic deoxygenation or a nitrogen-rejection unit bolted on, and that's a line item I've seen left out of more budgets than I can count. This isn't thermal waste-to-energy technology where the gas cleanup is the obvious headline cost. Here the cleanup you forgot is the one that holds up your injection permit.

Interconnection is a queue, not a pipe

Say your gas is in spec. You still have to get it into a pipe that will take it, and that's a slower, more political process than the engineering suggests. An RNG interconnection has two pieces, as the EPA lays out: a point of receipt (metering, gas-quality analyzers, odorization, and a shut-in valve that rejects off-spec gas) and a pipeline extension to physically reach the main. The utility builds and owns most of it, on the utility's timeline.

Two things stall projects here. First, takeaway. Tap a distribution line instead of a transmission main and that line has to absorb your flow - and distribution demand collapses in summer. A line that swallows your gas in January can back-pressure you in July, which means you either compress up to a transmission tap (more capital) or you flare. I've watched a developer discover this after financial close, which is exactly the wrong time to learn your offtake has a seasonal floor. Second, the study queue. Utility interconnection studies routinely run 12-24 months before you turn a valve [developer experience], and that clock usually isn't on the critical path anyone shows the investment committee. Actually, that 12-24 month range is the optimistic case; in congested utility territory I've seen it stretch past three years.

This is the same lesson the rest of the sector keeps relearning across global waste-to-energy projects: the conversion step is rarely the bottleneck. The connection to whatever takes your output (grid, pipe, or buyer) usually is. Waste conversion technology has gotten reliable; the interconnect hasn't gotten faster.

The molecule's value lives in the credit stack

Here's where my own scar tissue shows. The commodity gas itself is nearly worthless to an RNG project; pipeline gas might fetch $3-4/MMBtu. The reason anyone builds these is the environmental credits stacked on top. RNG from landfills, wastewater digesters, and agricultural digesters generates a D3 (cellulosic) RIN under the federal Renewable Fuel Standard, while RNG from waste streams like food-waste and MSW-organics digesters generates a D5 [per EPA AgSTAR]. The D3 RIN alone was worth on the order of $42/MMBtu of RNG in early 2023, per industry reporting on dairy projects, and it moves around a lot. Add California LCFS credits on top for a vehicle-fuel pathway and the stack dwarfs the gas by an order of magnitude.

So the whole project thesis rests on a policy instrument with a volatile price and an approval process of its own. The EPA pathway petition that makes your gas RIN-eligible can take a year or more, and it's feedstock- and process-specific. Whether your molecule even counts as clean depends on feedstock and accounting, the same question that decides when renewable energy from waste actually counts as clean power. That's the offtake risk that kills AD projects, and it's why I get nervous when a developer treats the RIN like a commodity with a tidy forward curve. On a Bogotá feasibility study in 2021 we killed the project ourselves at the offtake stage (no creditworthy buyer at the kWh price we needed) and I'd rather make that call twenty more times than carry one project to financial close on a credit price I can't underwrite. If the offtake math doesn't close, kill it at feasibility, not at financial close. The build-out is happening: the EPA's AgSTAR program counted 191 manure-based AD systems producing RNG by mid-2024, with another 69 under construction. But the projects that close are the ones that locked offtake before they poured concrete.

None of this holds uniformly. Below roughly 300-400 scfm of raw biogas, the upgrading and interconnection capital rarely pencils at all, no matter how clean your offtake is; small farm digesters usually do better burning gas on-site for heat and power than chasing pipeline injection, and a vehicle-fuel or CNG offtake can beat injection where a fueling anchor already exists. The oxygen problem I've hammered on is worst for covered-lagoon and manure systems and milder for well-sealed wastewater digesters. And jurisdictions without an LCFS-style credit (most of the world) change the offtake math entirely; the federal RIN may be all you get, which narrows the set of zero-waste-to-landfill solutions that clear an investment committee. Read the tariff and the credit program for your specific interconnect before you trust any number in this memo.

The digester is the easy 80% of an anaerobic digestion RNG project. The molecule's path from the upgrader outlet to a buyer's meter (oxygen spec, interconnection queue, RIN pathway) is the 20% that decides whether you've built an asset or an expensive science experiment. That cracked cover seal cost four months. The seal was cheap. The four months weren't.

Sources & Notes

1. EPA, "An Overview of Renewable Natural Gas from Biogas" (EPA 456-R-24-001, January 2024) - basis for the 96-98% methane pipeline-injection range and the interconnection structure (point of receipt plus pipeline extension).
2. EPA AgSTAR, "Renewable Natural Gas from Agricultural-Based AD/Biogas Systems" - 191 manure-based AD systems producing RNG and 69 under construction as of mid-2024; D3 versus D5 RIN classification by feedstock.
3. "Biogas Upgrading and Bottling Technologies: A Critical Review," Energies (MDPI), 2024 - multi-stage membrane methane recovery >99.5%; membrane-PSA hybrid >96% purity at >99% recovery with ~15% cost reduction.
4. "Current status of biogas upgrading for direct biomethane use: A review," OSTI/NREL - comparative upgrading-technology performance and methane-slip basis.
5. Biomass Magazine, "Aemetis sells D3 RINs generated by dairy RNG project" - industry reporting basis for the ~$42/MMBtu D3 RIN value cited for early 2023. RIN prices are volatile; treat as illustrative, not a forward curve.