Four kilometers down on the ocean floor, in total darkness, rocks are quietly making oxygen. Not through any biological process. Not through anything that requires sunlight. Just through chemistry—the kind that happens when certain metallic minerals sit in seawater.
This shouldn't be possible. For as long as we've studied Earth's atmosphere, oxygen production has been the exclusive domain of photosynthesis: plants, algae, and cyanobacteria that harness sunlight to split water molecules and release oxygen as a byproduct. That's the story we learned in school, and it's been the foundation of how we understand both the origin of life on Earth and the conditions necessary for life elsewhere. Oxygen is a biological product. Until now, apparently, it wasn't.
According to a 2024 study published in Nature Geoscience, researchers discovered that polymetallic nodules—potato-sized lumps of manganese, iron, cobalt, and nickel scattered across abyssal plains—generate oxygen through a process called electrochemical seawater splitting. In laboratory conditions mimicking the deep ocean, oxygen levels in the surrounding water tripled in just 48 hours, despite the complete absence of light and any living organisms. The mechanism is straightforward: the different metallic elements in the nodules act like a natural battery, with different oxidation potentials creating an electrical gradient that splits water molecules (H₂O) into hydrogen and oxygen.
The implications cracked something open immediately. If oxygen can be produced abiotically—without life—then the conventional narrative of Earth's early atmosphere starts to look incomplete. For decades, the story went like this: Earth's early oceans were anoxic, and oxygen only accumulated when cyanobacteria evolved around 2.4 billion years ago, triggering the Great Oxidation Event. But if inorganic rocks have been passively generating oxygen on the seafloor all along, that timeline and those mechanisms need reexamination. According to Chemistry World's reporting on the discovery, the research team found that these nodules have likely been producing oxygen for millions of years in the deep ocean, creating localized oxygen-rich pockets in an otherwise oxygen-poor environment. The question now is how much of Earth's atmospheric oxygen history involves biology, and how much was already there, courtesy of geology.
The mechanism underlying this discovery is essentially geochemical serendipity. Polymetallic nodules form in deep ocean sediments through slow accretion of metallic hydroxides and oxides over millions of years. Different metals have different tendencies to lose electrons (oxidation potentials), so when they're physically bound together in the same nodule and bathed in conductive seawater, they spontaneously form micro-scale electrochemical cells. The seawater itself becomes the electrolyte. This isn't a new kind of chemistry—it's the same principle that powers a battery in your flashlight—but finding it occurring naturally at such scale and in such abundance was a genuine surprise. Deep-sea conservation researchers note that these nodules exist by the millions across abyssal plains, particularly in the Clarion-Clipperton Zone between Hawaii and Mexico, potentially producing oxygen across vast stretches of the planet's least-explored realm.
The kicker is what this means for astrobiology. If oxygen can be generated on Earth without any biological involvement, then the search for life on distant worlds might need to reconsider what counts as a biosignature. A planet with a rocky ocean floor and mineral-rich seawater could theoretically develop an oxygen-containing atmosphere even without a biosphere. That doesn't mean the oxygen came from life—it might just be geology doing its thing. It's a humbling reminder that we've been operating with incomplete information about our own planet. The deep ocean, where we've barely set foot, apparently still has fundamental chemistry lessons to teach us. And those lessons might completely reshape how we think about where life can emerge, and where we should look for it.