Bacteria in the Bedrock, 5.3 Kilometers Under Sweden

What We Know

Thomas Gold was an astrophysicist at Cornell, not a petroleum geologist, when he talked Swedish investors, the state utility Vattenfall, and the U.S. Gas Research Institute into bankrolling a well that would reach nearly seven kilometers down through granite in central Sweden. The year was 1986. He was chasing a theory that remains a minority position in petroleum geology: that oil isn’t fossilized plants and animals at all, but a substance the planet keeps generating on its own, deep in the crust, from carbon left over since Earth formed. He picked his site for a specific reason. The Siljan Ring is a 377-million-year-old meteorite crater, and Gold argued its impact fractures ran deep enough for that primordial gas to migrate upward.

The well, Gravberg-1, never produced commercial oil. In 1987, roughly 900 barrels of drilling lubricant vanished into the hole, and the crew hauled up a hundred liters of black sludge that Gold announced contained both oil and the remains of archaebacteria. Petroleum geologists picked the claim apart fast: geochemist Paul Philp found the sludge’s chemistry matched surface shale oil, meaning it had most likely migrated down the borehole with the lost lubricant, not up from the depths. Gold called the episode a “complete fiasco” and switched to water-based drilling fluid for what came next.

What came next outlasted the theory that paid for it. In 1994, a separate team of microbiologists (Uwe Szewzyk, R. Szewzyk, and T.A. Stenström) published a peer-reviewed isolation of living, thermophilic bacteria from water drawn out of Gravberg-1, this time with the contamination question built into the method from the start. The well was cased down to 5,278 meters and open to the rock only below that line; the water they cultured from carried a salinity signature that could only have entered at or below it. At an estimated 70°C, they grew pure cultures of anaerobic, fermenting bacteria related to Thermoanaerobacter — the deepest confirmed life ever recovered from continental rock. Gold’s oil theory remains a minority position. His well’s bacteria did not.

The Roadblock

Proving that something alive really lives at the bottom of a multi-kilometer borehole, rather than having ridden down with the drill bit, is close to the whole problem. Drilling fluid, casing, tools, even the crew, carry surface microbes downward by the billions, and a fractured rock formation two or three miles down offers no wall a sloppy sampling method can’t breach. Gold’s own sludge is the field’s textbook failure case: a real sample, a real claim, undone by a chemistry that gave away exactly where it had actually come from.

Ruling contamination out takes more than a clean core barrel. Szewzyk’s team needed a salinity gradient stable enough to prove the water hadn’t mixed on its way up, and a casing depth precise enough to say which part of the hole any given sample could and couldn’t have touched. Every deeper record since has needed some version of the same argument, rebuilt for a different site, a different fluid chemistry, a different set of tracers. There’s no universal test for “this cell was always down here.” Each borehole has to earn its own proof, from scratch, and most of this field’s disputed claims have died on exactly that hook.

Best Guesses

The 122°C heat ceiling

No organism has grown above roughly 122°C, the point where Methanopyrus kandleri strain 116, pulled from a Kairei hydrothermal vent, tops out under pressure, close to where proteins and DNA themselves fail. A 2016 expedition drilled unusually hot seafloor sediment off Japan’s Cape Muroto to test that number and found hyperthermophiles active almost exactly up to it, with cell counts already collapsing a hundredfold above 45°C. Lab and field land near the same wall. The catch: that wall sits at very different depths depending on the rock, 3.5 km on an average gradient, past 5 in cooler Siljan granite, under 1.2 beneath Muroto’s hot seafloor.

Starvation before heat

Long before rock gets hot enough to matter, argue Tori Hoehler and Bo Barker Jørgensen, most deep cells are already starving. Their 2013 review found organisms recovered from deep sediment running on power budgets orders of magnitude below what growth or division costs, enough, perhaps, to repair molecular damage, and little else. The biosphere’s edge, by this account, is less a wall than a slow fade into a state barely distinguishable from death: too little energy to reproduce, evolve, or even register on the instruments built to detect it. A cell too starved to grow looks, to most methods, exactly like a cell that never was.

Pores too small for cells

Junghee Park and J. Carlos Santamarina compared cell counts from 116 drilling sites worldwide and found a pattern with nothing to do with heat or hunger: pore size. As sediment compacts, the spaces between grains narrow, and in fine clay and shale they can shrink past roughly one micrometer, the size of a typical bacterial cell, in rock that might otherwise sit at a survivable temperature. Deep rock, on this reading, can be sterile for the same reason a keyhole won’t pass a doorknob, whatever the thermostat says. It has far less to say about fractured granite like Siljan’s, where the record-holding bacteria were still moving through open pore space at 5.3 kilometers.

A drilling limit, not a life limit

Every well-funded, carefully controlled expedition that has pushed into new depth or heat has come back with life closer to the theoretical edge than the last record implied: Szewzyk’s Sweden well in 1994, the Muroto core in 2016. To some researchers, including much of the Muroto team, that pattern reads less like a true floor and more like a limit set by drilling technology and money, receding as both improve. It’s the hardest position to pin down, since it can absorb any future record as confirmation without predicting where the real edge sits. Nobody has drilled a hole clean, deep, and hot enough to prove it wrong, either.

The Verdict

Gravberg-1 got close on the pore-space side: fractured Swedish granite, still open at 5.3 kilometers. Expedition 370 got close on the temperature side: hyperthermophiles active almost up to 122°C, off Muroto. No hole has done both at once — reached rock that hot with its pore space confirmed open, sampled with contamination controls as rigorous as Szewzyk’s salinity trick, and checked for both cell activity and the specific power budget those cells are running on. Nobody has drilled that specification yet — not for lack of a plan, but because rock that hot, that permeable, and that reachable without a runaway budget hasn’t turned up in the same borehole.

Each new record still gets claimed by whichever theory happens to fit it: the thermal ceiling, the energy budget, the pore throat, or simply better drilling. The confirmed floor only sinks as far as the next well can afford to dig.