Fusion, and other fusion-level bets: Part III

The post inverts the series’ order and takes the challenge first, because what has to be beaten is formidable. Hydrocarbons score high on three things at once, energy density, stability and cost, and they are two things at once as well, both the primary energy and the package it travels in, so the packaging is effectively free. Out of that he draws three jobs, each of which he counts as a civilization-scale prize in its own right: long-range heavy transport, where density matters more than price; ultra-long-duration storage, of the kind the United States gets today from having weeks of coal, gas and oil sitting in stockpiles; and moving energy between regions, so that places with extraordinarily cheap sunshine could export it the way the Middle East exports oil. The third is the hardest, because it demands very high density and very low cost in the same substance, which is the rarest pairing there is.

He then takes the strongest objection to his own premise, that we may already have the universal battery in hand. Lithium-ion is the jack of all trades: no other chemistry performs well across so many parameters, it has a fifteen-year head start and a supply chain nobody can match, lithium itself is theoretically the best combination of weight and reactivity available, and the family has already branched into a dense lineage for vehicles and a cheap one for grids. That sets the bar for a successor, which would have to be denser than the dense branch and cheaper than the cheap one, and on current trajectories he does not see it. The density bets that absorbed most of the investment, silicon and lithium metal anodes and solid-state electrolytes, look set on even the most ambitious projections to deliver about half again on today’s cells, which he says would begin to open up heavy trucking if the extra range did not cost much more, and still a drop in the bucket against the density of fossil fuel. The cost side is worse, because cells are approaching a materials floor while the installed system costs many times the cell, and he thinks even halving system cost leaves anything beyond ten or twelve hours untenable. So he stops asking for one battery and asks for two, one uber-dense and one uber-cheap.

On the dense limb he lists three long shots at the edge of lithium chemistry, lithium-sulfur, interhalogen and lithium-air, two of them with results logged by outside laboratories and the third so far only theoretical, each still fighting the same problem of how many times it can be recharged. Then he concedes the point that matters: even a thousand watt-hours per kilogram, roughly triple the best cells today, would not move aviation or shipping. On the cheap limb he is more confident, and this is where he revises himself. Sodium-ion, which he wrote about in January, now looks better to him than it did then, because the industry has converged on one low-cost cathode and because a system built around sodium’s tolerance of heat and cold can drop the cooling equipment lithium-ion needs. He adds a four-day iron-air battery and thermal storage in hot bricks, both far cheaper per unit of energy than lithium-ion, both backed by his firm. None of them, he says, stockpiles months of energy or has the density to cross an ocean, and he concludes that probably no battery of any kind ever will.

That gap is what the last third of the post is written into. Metals oxidize and give off heat exactly as hydrocarbons do, and several abundant ones carry more energy per unit of volume than coal while being far denser and more stable than any battery. The catch is that nature hands us metals already oxidized, since ore effectively means oxide, so the metal has to be stripped of oxygen first. That makes metal fuel a way of storing energy rather than a fuel anyone digs up, and it needs two leaps existing metallurgy does not offer: reduction as practised today is a century-old industrial process that wastes much of the energy it stores, and burning the fuel back into electricity through a conventional steam turbine throws away most of the rest. Solve both and you get the thing the whole essay has been circling, renewable energy loaded onto ordinary railcars and ships, stockpiled for as long as you like, with the spent oxide returned to be recharged in a closed loop. He is candid that he cannot say how it would be done, because the company his colleagues incubated to do it has not yet disclosed even which metal it uses.