The frontier this week is a gel that eats carbon. In Nature Communications (s41467-025-58761-y), Mark Tibbitt’s group at ETH Zurich reports a printable, living hydrogel seeded with cyanobacteria that pulls roughly 26 mg of CO₂ per gram of material out of the air over about 400 days. Photosynthetic microbes have been coaxed into building products before. What earns this ETH result a second read is where the carbon actually ends up.

Two ledgers, one gel

Living carbon-capture materials usually bank their CO₂ as biomass — the cells grow, the carbon rides inside them, and the day the culture dies the ledger reopens for renegotiation. The Nature Communications work runs a second pathway alongside the first: the cyanobacteria shift the local chemistry enough to precipitate solid carbonate minerals inside the matrix. That mineral fraction is what matters to anyone thinking in centuries, because a carbonate does not exhale when the organism stops. Under the arithmetic, the durable store is the mineral; the biomass is the fast, reversible float sitting on top of it.

Why now, and not five years ago? Because the geometry finally cooperates. The Tibbitt team prints the gel as an open lattice so light and dissolved CO₂ reach the interior cells rather than starving them in a dense block. That printed lattice follows the same logic PAZ’s Voronoi concept panel lays out for additive-manufacturing infill — seed a scalar field, thicken the cells where the work happens. Here the field is light and gas exchange instead of stress: pack surface where the microbes need photons, open channels where CO₂ has to travel.

Where the lab bench stops

So here is the honest boundary. A 26 mg/g uptake measured on gram-scale samples in a controlled chamber is a lab result, roughly three years upstream of anything you would specify into a real wall. Scale it and you inherit questions the paper does not pretend to close: keeping a living culture alive and lit inside a façade for decades, the water budget, what happens the first winter the cells go dormant. My generation scaled more than one wonder-material on a demo and met its real tolerance in production, expensively. Read the error bar before the headline.

←TODAY: 2026 — an ETH-printed hydrogel banks ~26 mg CO₂ per gram over 400 days, part of it locked as solid carbonate. →3012: façades that keep photosynthesising for a century, their carbon ledger written in mineral rather than marketing. Fulcrum: the store only counts if the mineral pathway outlives the cells — arithmetic, not awe.

Atelier: Offices that have lived through green-material procurement report the same trap — a beautiful uptake figure with no duration and no phase attached. Set one policy this week: any “carbon-negative façade” claim entering your materials review must arrive with mg CO₂ per gram, the measurement window in days (the ETH figure is 400), and the fraction stored as stable mineral — no line, no entry into the BEP.

Hack: Size the carbon ledger of a living panel before you trust any brochure. The 26 mg/g figure is small; multiply it by real panel mass and the number stops being magic. Run this and you learn whether a façade skin banks kilograms or grams.

uptake = 26e-3              # g CO2 / g, ~400 d (Tibbitt, ETH, Nat. Commun.)
rho    = 1100               # kg/m3, hydrogel ~ water
panel  = 0.03 * 2.7 * 1.2   # m3: 30 mm skin, 2.7 x 1.2 m
mass   = rho * panel        # kg
print(round(mass * uptake, 2), "kg CO2 / panel / 400 d")  # -> 2.78

Under three kilograms per panel over more than a year. That is real, and it is modest — useful context the next time a slide promises a building that “absorbs its own emissions.”

The move

Read the paper’s methods, not the press summary, and carry one number in your pocket: the mineral fraction. When a bio-material rep next pitches a living skin, ask what stays banked after the culture dies — and specify nothing until the independent replication lands. Awe is fine; awe without the arithmetic is how you get a beautiful, brittle future.

Source: nature.com