Bacteria in the Copper Heap

Published October 5, 2026

The Deep Harvest, Chapter 6: Biology Eats Mining

At the Sossego copper mine in Brazil, about 90 million tonnes of tailings sit where the mill left them. Tailings are the crushed rock a mine throws away once it has taken the copper it could get. The Sossego tailings still hold about 0.07 percent copper, or 0.7 kilograms in every tonne. The Deep Harvest's main research report on biomining puts the copper in that pile at about $1.4 billion. That is a gross figure: the report deducts nothing for the cost of getting the copper out, so the pile's real worth to anyone is lower and unknown.

The same report puts the average copper ore grade at 3 percent in 1920 and 0.7 percent now, so the Sossego tailings hold about a tenth of what a typical working mine digs today. To get copper out of rock that lean, the industry hands part of the job to bacteria.

What the bacteria do

The process is called bioleaching: microbes help dissolve metal out of ore into a liquid, and the metal is collected from the liquid. The main workers are acid-loving microbes that live at pH 1.5 to 2.5, in liquid strong enough to dissolve metal.

One of them, Acidithiobacillus ferrooxidans, feeds on iron and sulfur. It turns ferrous iron into ferric iron and turns reduced sulfur into sulfate. Ferric iron and sulfate dissolve sulfide minerals, the metal-and-sulfur compounds that carry copper, gold, nickel, and cobalt. In the common copper mineral chalcopyrite, the reaction runs like this: ferric iron takes apart the copper-iron sulfide, releasing dissolved copper, ferrous iron, and sulfur. The ferric iron is used up doing that, and the bacteria turn the ferrous iron back into ferric iron so the attack keeps going. The report also describes a second route, where the microbes oxidize the mineral directly where they touch it.

A second microbe, Leptospirillum ferriphilum, tolerates ten times more ferric iron than Acidithiobacillus. In a heap of ore, Leptospirillum moves in first and builds the biofilm, the thin living film that coats the rock surfaces, and other microbes follow it in a predictable order.

The bacteria do not take the copper themselves; they keep restoring the ferric iron that dissolves it. Nuton, the copper process covered in the book's second report, grows its own proprietary microbes in bioreactors to make ferric iron and sulfuric acid, then uses that liquid to dissolve copper from primary sulfide ore.

Three machines, three speeds

Operators run the same chemistry in three kinds of plant, and each one trades speed against cost.

The cheapest is the heap. Heap leaching means piling broken ore on the ground, running solution through it, and collecting the copper-bearing liquid that drains out. The main report says heap bioleaching handles millions of tonnes a year at minimal capital cost and stays workable below 0.3 percent copper. Ore sits in the heap for 30 to 60 days to reach 50 to 70 percent recovery. Recovery is the share of the metal in the rock that the plant actually collects; the rest stays in the rock.

The fast one is the stirred tank. Ore or concentrate goes into tanks where the microbes and liquid are mixed. The report gives 94 percent gold recovery after four days in the tank. The tanks cost more to build than a pile on the ground, so operators keep them for high-value metals and for refractory concentrates, ore whose gold is locked inside sulfide minerals. At Chuquicamata in Chile, thermophilic stirred-tank bioleaching at 45 to 60 degrees C produces 20,000 tonnes of copper a year. Thermophiles are heat-loving microbes. The same report, in a general passage on heat-loving species such as Sulfobacillus thermosulfidooxidans and Sulfolobus archaea, gives their working range as 60 to 80 degrees C, using heat released by the sulfide oxidation itself to speed the reaction. The report gives both ranges and does not reconcile them.

The third way leaves the ore where it is. In-situ bioleaching pumps the liquid into the orebody underground, so nothing is dug, hauled, or dumped as tailings. In the BIOMOre project at the Rudna Mine in Poland, about one kilometre down, engineers injected liquid from a bioreactor into fractured ore and kept copper coming out for a sustained period.

Where it already pays

The main report and the book's research plan both put bioleaching at about 20 percent of global copper production. BIOX biooxidation, where microbes break down the sulfide minerals that lock gold in so the gold can be recovered, reaches 94 percent gold recovery at commercial plants. The US Department of Energy awarded Freeport-McMoRan $80 million for bioleaching piles warmed by geothermal heat.

The clearest numbers come from Nuton at Lion Copper and Gold's Yerington project in Nevada. In a program of more than 40 test columns, Nuton's microbes recovered 73.2 percent of the copper from sulfide ore, against 60 percent from oxide ore under conventional heap leaching. Across the whole planned life of the mine, 506.5 million tonnes at 0.21 percent copper, the blended recovery comes to 67.4 percent. The plan calls for 120 million pounds of copper cathode a year, the finished plates of refined copper, for 12 years.

The main report gives a different recovery figure, 85 percent from primary chalcopyrite, without tying it to the Yerington columns, and neither report explains the gap. The main report's heap range of 50 to 70 percent in 30 to 60 days describes heaps in general, not either of those tests.

The two reports also disagree about Johnson Camp in Arizona. The main report says Rio Tinto's Nuton technology reached its first industrial-scale deployment at the Johnson Camp Mine in December 2025. The Nuton report describes Gunnison Copper's Johnson Camp project, with 25 million pounds a year of capacity and production beginning in September 2025. Who is credited and which month production began differ between the two files, and the book's files do not settle it.

None of the sources behind these figures has been independently checked yet. The book's source lists mark every one of them unverified, and the main report itself warns that some of its sources are market research or company promotion.

The cost of waiting

The Deep Harvest's outline calls the slow clock the "time-value penalty of biomining," and the chapter brief calls it a "kinetic time tax." Kinetics just means reaction speed. The penalty is working capital: the money a company has already spent mining, crushing, and stacking ore that will not pay back until the copper comes out of the heap weeks later. No file behind the book puts a number on that cost.

The main report does not see it as a penalty. It calls biomining the "dominant extraction paradigm" and argues that for deposits at 0.3 to 0.7 percent copper, slower kinetics are "acceptable precisely because conventional processing becomes economically unviable at these grades."

The grade numbers do not line up neatly either. The main report models the break-even grade for copper rising from 0.5 percent now to 0.7 percent by 2035, and says that figure is a model. The same report says heaps stay workable below 0.3 percent, and Yerington is planned at 0.21 percent. The files leave those claims side by side without explaining how they fit together.

The Nuton report measures grade a different way. Copper projects still in development average 0.39 percent, against 0.53 percent at operating mines, while mining costs rose from $14 a tonne in 2007 to $21 a tonne in 2017. That comparison and the 1920-to-now decline come from different measures and do not make one trend. On cost, the main report credits bioleaching with savings of $0.20 to $0.40 per pound of copper.

Beyond copper

The same chemistry reaches other metals, but most of the results stop short of a working plant. Biooxidation frees refractory gold from the iron sulfides pyrite and arsenopyrite, as BIOX already does. For laterite, the weathered rock that holds over 70 percent of accessible nickel reserves, the main report gives 53 percent nickel and 46 to 60 percent cobalt in 7 to 11 days, using Acidithiobacillus thiooxidans and A. ferrooxidans. It does not say whether that was a lab test or a plant.

For rare earth elements, the report gives bioleaching efficiencies of 99.5 percent for lanthanum, 95.8 percent for neodymium, and 93.5 percent for yttrium, with electronic waste above 62 million tonnes a year as a possible feed. It does not state the scale of those tests either. One trade-press article describes engineered bacteria that pull gold selectively out of printed circuit boards "at economically viable rates," and adds a condition: "If biological approaches scale."

What changes at the mine

The main report gives 0.5 to 1 tonne of carbon dioxide equivalent per tonne of copper for biomining, against 2 to 4 tonnes for smelting, and about 80 percent less water than flotation, the usual way of concentrating copper ore before the smelter. The Nuton report gives 2 to 4 megawatt-hours of electricity per tonne of copper for biological leaching plus solvent extraction and electrowinning, the steps that pull copper out of the liquid and plate it as cathode, against 15 to 20 megawatt-hours for flotation, smelting, and refining. A separate report lists bioleaching as one of the leach routes that use less acid.

The bigger change is in what the mine leaves behind. Acid mine drainage is the acidic, metal-laden water that runs off mine sites. The main report says biomining prevents that drainage rather than treating it afterward, with remediation taking 5 to 15 years instead of treatment that never ends.

Engineered microbes

The main report says engineered combinations of microbes could raise yields by 35 to 40 percent. Its own limitations section says that number is a computer prediction and has not been tested at industrial scale.

The space version

The High Ground, the companion book about building in space, includes what its outline calls a bio-mining fork in Chapter 4, Built from Dust. The space evidence behind it is one experiment. ESA's BioRock, run on the International Space Station from 2019, is the only published biomining experiment with more than one kind of microbe done in microgravity. It put three microbes, including Sphingomonas desiccabilis, on basalt for 21 days. Under reduced gravity the microbes pulled out up to 283 percent more vanadium than the same setup on Earth, which the researchers did not expect.

The samples were grams of basalt in millilitres of liquid. No published study has tested biomining on actual lunar regolith, or on a high-fidelity copy of it, in microgravity.

Weeks of waiting

At Yerington, Nuton's microbes are planned to turn ore at 0.21 percent copper into 120 million pounds of cathode a year. In a bioleach heap, ore spends 30 to 60 days before 50 to 70 percent of its copper comes out, and the money spent mining and stacking it waits in the heap the whole time. The book's outline calls that wait a penalty, the main report calls it acceptable at these grades, and no file behind the book gives the working-capital figure that would settle it. That answer decides whether piles like the 90 million tonnes at Sossego stay waste or become feedstock, and The Deep Harvest carries that mine waste forward into the materials that build things.

Read the full argument in The Deep Harvest, Chapter 6: Biology Eats Mining.