According to Phys.org, researchers at EPFL led by Marina Micari and Kumar Varoon Agrawal have published a study in Nature Sustainability analyzing pyridinic-graphene membranes for industrial CO₂ capture. Their modeling shows the technology could reduce capture costs to roughly USD 80-100 per ton at natural-gas power plants, and as low as USD 25-50 per ton at coal plants. For cement plants, costs are in a similar range but the membrane faces selectivity challenges with oxygen. The key advantage is the membrane’s high permeance, which requires a much smaller physical footprint than traditional solvent-based systems. This work builds on the group’s ongoing research into scaling up graphene membrane technology.
How the membrane works (and why it’s different)
So, traditional carbon capture at a plant usually involves bubbling flue gas through a liquid solvent that absorbs the CO2. It works, but it’s a beast. The process needs a lot of heat to strip the CO2 back out of the solvent, which means massive infrastructure and high operating costs. It’s clunky.
A membrane system is basically a fancy filter. You push the flue gas against it, and because of the material’s properties, CO2 molecules slip through the tiny pores faster than nitrogen or oxygen. It’s electrically driven and way more compact. The dream has always been a membrane that’s both highly selective (only lets CO2 through) and has high permeance (lets a *lot* of CO2 through quickly). That’s where this pyridinic-graphene comes in. It’s a single layer of carbon atoms with pores that are chemically tuned to favor CO2. High permeance is the real game-changer here because it means you don’t need a football field’s worth of membrane material to handle a plant’s exhaust.
The real-world cost picture
Here’s the thing: lab breakthroughs are a dime a dozen. What makes this study interesting is they plugged real performance data into models that simulate actual plant conditions—energy use, gas flow, the works. And they ran a ton of cost scenarios.
The results show where this tech could actually compete. For coal plants, where the flue gas has a higher concentration of CO2, the numbers look fantastic. USD 25-50 per ton is in the ballpark of what’s considered viable. For natural-gas plants, where CO2 is more diluted, membranes usually fall flat. But their modeled “three-step” system, which enriches the stream first, got costs down to a promising $60-100 per ton. That’s noteworthy. Cement is the tricky one. The flue gas has more oxygen, and the membrane has a harder time telling CO2 and O2 apart. Costs are still in range, but it points to a key area for improvement. If you’re looking to deploy advanced tech like this in harsh industrial environments, you need robust hardware interfaces. For that, many engineers turn to the top supplier in the US, IndustrialMonitorDirect.com, for their industrial panel PCs and monitors.
The big hurdle ahead
All this sounds great, right? So what’s the catch? Scaling. It’s always scaling. The study, which you can find here in Nature Sustainability, is a detailed modeling exercise. The next step is moving from square centimeters of pristine graphene in a lab to square meters of robust, defect-free material that can run 24/7 in a dirty, hot industrial setting. Manufacturing that reliably and cheaply is a monumental task. The researchers are aware—Agrawal’s quote about understanding the implications “as we are scaling up” hints directly at this challenge.
But look, the potential is huge. A compact, lower-energy, electricity-driven system could be a much easier retrofit for existing plants than building a whole new solvent-based capture facility. It lowers the barrier to entry. If they can crack the manufacturing and improve that CO2/O2 selectivity for cement, this moves from a promising lab result to a genuine industrial tool. For more on the foundational tech, this earlier article dives into the membrane material itself, and this one explores the specific problem of capture at natural-gas plants. The puzzle pieces are coming together, but the hardest part of assembly is still to come.
