Going deep on next-gen geothermal

Geothermal needs heat, water and permeability in one place, and only the third is genuinely scarce. It is hot everywhere if you drill deep enough and nearly the whole planet is saturated with water at depth, but the rock also has to let that water move. That is why the industry clusters on volcanically active margins, where recent volcanism both brings high temperatures near the surface and leaves the rock fractured. Horne also declines the premise that geothermal is expensive; in California it is one of the cheapest sources of electricity. What costs money is uncertainty. A bank knows exactly how large a solar farm will be and does not know how big a geothermal resource is or how long it will produce, so it lends at a higher price, and the developer spends real money gathering information to narrow that range. Cost, on his framing, is downstream of geology rather than a separate problem.

Enhanced geothermal attacks the scarce ingredient directly by manufacturing permeability: fracture hot rock to create a flow path between an injection well and a production well, then run the result like a conventional plant. When Kann suggests this is fracking borrowed from oil and gas, Horne corrects him. The idea was postulated at Los Alamos in the 1960s and projects were built from the 1970s, before the shale boom. What is new is the transfer of oil and gas practice into it, specifically horizontal wells, casing the well from top to bottom and multi-stage plug-and-perf stimulation, none of which was normal practice in geothermal. The rock differs, since oil and gas fracture relatively pliable sedimentary rock while geothermal fractures brittle volcanic rock, which Horne says is harder in some ways and easier in others. On his account the physics is no longer the question, and what remains is doing it at size.

The other two approaches sit further back, for opposite reasons. Closed loop replaces fractures with drilled holes connected into a loop or a network, and its problem is heat transfer rather than feasibility. A fracture sweeps water past a large surface area, while a borehole has very little, so once the rock immediately around the hole has cooled you are waiting on conduction through rock that is a poor conductor. The consequence is that you have to drill a great deal of hole, and cost follows; Horne cites work at the National Renewable Energy Laboratory and elsewhere finding it hard to get closed loop down to a competitive number. He confirms directly that the obstacle is economic rather than technical, and notes it does not widen the map much, since you still have to find hot rock. Super-deep is the approach most likely to be misunderstood. Kann twice frames it as drill deep enough and geothermal works anywhere; Horne assigns that promise to the other two and says super-deep is a thermodynamic play. Conventional and enhanced systems run at roughly 200 to 300 degrees Celsius, far below the 800 to 900 degrees of a coal or nuclear plant, and reaching 500 or 600 degrees converts much more of the heat into electricity per well. Not geothermal anywhere, in his words, but better geothermal where it is already good. The blocker is materials: supercritical water carries enormous quantities of dissolved solids and can be strongly acidic, and one Icelandic supercritical well produced black steam because it was dissolving its own steel casing and bringing it up with the fluid. Neutralizing such water is possible and sometimes done, but a thousand tons a day of sodium hydroxide is not a business.

Cutting across all three is drilling, which is where Horne puts the biggest recent gains. Batch or factory drilling from shale has arrived in geothermal, with Fervo drilling eight wells in Utah in stages rather than one at a time and cutting drilling times by a factor of two or three, and polycrystalline diamond bits are giving longer runs between trips out of the hole. He suggests drilling costs are heading toward half what they were five to ten years ago, which matters because the $80 per megawatt-hour target is a fixed post: cheaper drilling simply moves more of the map inside it. The same transfer compresses schedules. Conventional geothermal carries roughly a ten-year lead time dominated by exploration, and enhanced systems need less of that exploration because they are not hunting for a rare combination of conditions, which Horne thinks could bring the cycle down to five years and possibly three or four. Exploration does not disappear, since you only learn how hot the rock is by drilling the first well, but much of the planet has already been characterized by oil and gas or by earlier geothermal surveys, and ground rejected as not quite good enough for conventional geothermal may be fine for the enhanced version. Which is why his closing answer is about developers rather than technology.