Can wind turbines self-replicate? (And how will we make beer?)

The slogan the blog is named for describes a substitution: instead of burning fuel continuously, we make a large up-front investment in physical structures, trading operating expense for capital expense and carbon emissions for a much heavier draw on raw materials. Lubershane’s point is that the substitution currently runs one way only. Wind and solar need enormous quantities of steel and cement to deploy, and the electricity they produce cannot yet be used to make either. He separates two problems that sound alike. This is not an acute supply-chain squeeze, because the steel and cement markets are large enough to absorb transition demand without flinching, unlike the materials markets feeding electric vehicles. It is a long-term decarbonization problem, because these are the world’s two largest construction materials by weight and each carries a high-single-digit share of global emissions. The loop has to close.

Steel is his case study for why closing it is hard, and the difficulty has two independent parts. The first is chemistry. Iron comes out of the ground bound up with oxygen, so making steel requires stripping that oxygen away, and in a blast furnace the fossil fuel does double duty: it supplies the heat, and its carbon and hydrogen atoms carry the oxygen off. Electricity substitutes easily for the heat, which is why about a quarter of today’s steel demand is already met by melting scrap in electric arc furnaces, but electricity alone cannot perform the chemical half. The second part is economic and applies even where the chemistry is not in the way. Coal and gas delivered to an industrial site are extremely cheap when all you want from them is heat, and typical wind and solar power costs substantially more per unit of energy, before adding transmission or the storage needed to cope with intermittency. So an electric route cannot win on the sticker price of typical clean energy, and he sees three ways new technology might overcome that disadvantage.

The first is electrochemistry, which uses an electrical potential to drive the reaction itself rather than burning something to make heat. That can skip much of the high-temperature heat and satisfy the chemistry at the same time, and the efficiency advantage over a blast furnace is what lets it run economically on low-cost clean power; being modular rather than enormous also lowers the risk of each new plant. Cement is the parallel case, because roughly half of its emissions come from the limestone rather than the fuel, so electrifying the kiln would solve only half the problem. The catch is that electrochemical processes suit new build rather than retrofits, and the world has already sunk trillions into plants, many of them built in China and India in the past two decades with at least a decade or two of life left. A great many processes also genuinely just need heat. There the lever is a heat pump, which moves heat instead of generating it and so can deliver more heat energy than the electricity it consumes. Its constraint is lift, the temperature gap it works across: a refrigerator’s lift is small and its efficiency correspondingly high, while industrial processes need heat at or well above boiling, implying lifts several times larger. A machine that delivers that range at a decent efficiency would address about two thirds of all the heat used in American manufacturing, which is where beer, potato chips and dyed clothing sit. Two routes lead there: recover waste heat, which shortens the lift and is already sold commercially but is limited in quantity, costly to plumb in and sometimes available at the wrong time; or build an air-to-steam machine from scratch.

The third option attacks the price rather than the process. Where the wind or sun is strong, power can be procured cheaply enough to be roughly competitive with coal and gas, but industry needs energy continuously and renewables do not arrive that way. Heat turns out to be among the cheapest ways to store energy and the most efficient, since the losses from any storage cycle are heat anyway. That turns the operator’s question into a siting question: not only whether cheap renewable power exists, but how close to the fence line it can be had, because a large enough plant with enough nearby land might route power straight into thermal storage and skip a costly grid connection altogether. His conclusion is a portfolio rather than a winner. Which route fits depends on the regional resource, the state of the existing industrial base and site-specific variables, the best answer will often combine them inside one facility, and on that basis he says the loop closes and wind turbines will self-replicate.