Near Baotou, a freckle of brown concrete blocks on the edge of the Mongolian steppe, is what used to be a lake. Animals and people once lived around its banks, but they died or moved away. All that remains of civilisation are the skeletons of halffinished towers and rows of pipes, each one churning black sludge into a hole where the water once sat. As the mud settles, it leaks sulphur and toxic dust back towards the city.

Places like this exist all over China, wherever prospectors find terbium or neodymium or any of the other 15 rare earth elements (REEs) that make society work. As at Baotou, the results can be hellish. Excavations leave behind piles of filth, radioactive grime and clouds of ammonia. In Shandong, an eastern province facing South Korea, rare earth mining has infected cereal crops. Local babies have lower IQ rates than their parents did at the same age, and suffer from neurological problems. Around Baotou itself, rates of diabetes and osteoporosis have skyrocketed.

The mining continues anyway, and the reasons why are obvious. REEs remain vital to dozens of modern technologies, with the industry as a whole expected to reach $20 billion by 2024. Every time you swipe your iPhone, or pick up your laptop, you hold rare earths in the palm of your hand.

Their reach is now so wide that even environmentally progressive industries lean on them. The most efficient wind turbines use several REEs in their powerful permanent magnets – despite the fact that their extraction can cause chaos.

The irony of using rare earths in a green industry is not lost on enemies of wind power. As the Institute for Energy Research, a think tank that has received funding from several major fossil fuel companies put it, wind energy is not nearly “as clean [as] wind lobbyists want you to believe”. But recent developments have pushed against these criticisms, especially compared with when they were made in 2013.

The extraction and recycling of REEs have become far more efficient over recent years, while new wind turbines are less reliant on rare earths than older models.

Forged in a crisis

The contemporary REE market was forged in 2010–11, when a squeeze on Chinese exports combined with high demand to overwhelm the industry. With the People’s Republic controlling 95% of global supply, the price of rare earths soon exploded.

If the gearbox is going to be failure-prone, up in a very high tower and offshore, then maintenance costs start to mount up. You need a more reliable turbine so manufacturers are moving towards building more direct drive wind machines.

Costs rose between 100 and 900%, bottlenecking orders and delaying shipments. Though the situation has since settled, REE-reliant industries were shocked into action, explains Alex King, director of the Iowa-based Critical Materials Institute.

“A lot of effort went into finding ways to work around using rare earths, rather than being held hostage by shortages of supply,” he says. “When the price goes up, people are incentivised to reduce their demand.” He is quite right. For example, Toyota recently announced that its latest generation of hybrid electric motors would contain half the amount of rare earth materials, while other Japanese compaies are following closely behind.

But the situation for wind turbines is rather more fraught, a problem centred on the lack of decent alternatives to REEs. Doubly-fed induction generator (DFIG) turbines swap REE permanent magnets for regular gearboxes, which convert the turning speed of the blades into the rotations that the generator needs to generate electricity. But their size (they weigh up to 2t) and noise make them unattractive, especially in populated areas where locals might cause a fuss.

The infamous unreliability of DFIG turbines hardly helps matters, King adds. “If you look at a wind farm, and you see a turbine that isn’t turning,” he says, “the chances are that the reason for that is a gearbox failure.”

With these difficulties in mind, you might expect DFIG turbines to be nowhere. In fact, over 99% of US turbines rely on them, with only about 230 using REE-based direct drive magnets, even though the lack of a gearbox makes them far more reliable. King explains this disparity by highlighting historic “fears” about the cost and availability of REEs, pointing out that a single direct drive turbine needs perhaps 170kg of rare metals to work.

Wave goodbye to DFIG

These pressures have traditionally pushed wind manufacturers towards an unappealing choice: either funnel cash into direct drive machines or stick to flawed DFIG models. But if landlocked turbines offer frustrating possibilities for manufacturers, caught as they are between DFIG and direct drive, the future of seaborne wind farms is far more straightforward.

This will be a world without DFIG, King believes. Because sea turbines are three times larger than their landlocked cousins, the correspondingly large gearboxes needed in DFIG machines are simply too unreliable so far out to sea. “If the gearbox is going to be failure-prone, up in a very high tower and offshore, then maintenance costs start to mount up,” he says. “You need a more reliable turbine, so manufacturers are moving towards building more direct drive wind machines.”

Indeed, recent enthusiasm for offshore wind farms seems to have tilted the industry decisively towards REE-based direct drive machines. For example, GE and Siemens recently announced their next two generations of offshore farms would be direct drive. Bolstered by general healthy growth – Europe built 560 new offshore turbines in 2017 – it’s little wonder King expects a “significant resurgence in demand for rare earth metals”.

The revival in seaborne direct drive turbines has dovetailed with technical changes to make REEs cheaper and cleaner to extract for the machines that need them. Green recycling is one hope. Though King admits that “some uses require you to have some acid in the system,” the latest techniques use “quite benign” chemical processes to extract useful materials from car engines or hard-disk drives.

Other methods take advantage of ionic liquids, which can separate worthless metal oxides from neodymium magnets.

At a more basic level, some companies take REE magnets wholesale and fit them into direct drive engines.

If environmentally friendly recycling is an “aspirational solution” to the supply and cost problems that have blighted REEs, King is similarly confident about recent advances in mining. Using the same methods as recyclers, prospectors from Australia to Malaysia are making REE extraction far cleaner than workers at Baotou could ever imagine.

At the same time, breaking the Chinese monopoly on mining forestalls another 2010-style crunch, keeping prices down. In fact, King notes, costs have stabilised so much that “many of the 400 mining projects being considered around the world during the great price spike have been shelved because of the low prices. These days, it is not clear that they could make a profit.”

More broadly, manufacturers are “developing materials and design approaches that reduce the need for [REEs]”, King says. In practice, this involves making magnets that work at cooler temperatures, avoiding the need for heat-resistant materials like dysprosium, the shiny element used to make high-powered direct drive magnets. For their part, mid-speed turbines are being made smaller, cutting the amount of rare earth material needed. Though King expects turbine manufacturers to keep using REEs, he thinks their quantities can be reduced. “I don’t think we’ll be able to reduce that number faster than the number of wind turbines is rising, but we’ll make inroads,” he says. Energy companies are proving this already.

Last year, Siemens announced its new range of offshore direct drive turbines were free from either dysprosium or terbium (but still need neodymium), while the industry generally is developing permanent magnets that push dysprosium levels down from 3–6% to near 1%.

Many of the 400 mining projects being considered around the world during the great price spike have been shelved because of the low prices. These days, it is not clear that they could make a profit.

Cleaner and cheaper

The way forward is not completely clear, however. Mining rare earths is now cleaner and cheaper than before, but the process itself is still fiddly. “Dysprosium is really the first pinch point,” says King. “The name ‘dysprosium’ derives from the Greek for ‘hard to get’ and was challenging to extract in the late-19th century, when it was first discovered. It still is now, making it a very scarce material.”

But the wind industry has come a long way since the Institute for Energy Research launched its broadside back in 2013, and is clearly keen to keep up momentum. They are still being researched, but future generators might use high-temperature superconductors, which would drastically lower the weight of direct drive turbines.

King and his team at the Critical Materials Institute are focusing on sharpening what he calls the “front end” of the recycling process, reducing the cost of getting recyclable materials to reprocessing centres. Otherwise you risk “creating a system in which the cost of bringing all the material into one place exceeds the value of what you can recycle”.

Given all the industry has achieved over recent years, there is every reason to think King and his colleagues can keep reducing their reliance on REEs, and just as well. By 2025, the wind energy sector will be worth $25 billion, and by 2030, turbines will generate over 20% of global electricity.

With these kinds of numbers, the status quo is simply unsustainable. Not that old damage can be undone. Toxic dust will keep floating across the Baotou lake for years to come, landing on the blocky apartments and the people of the town, even as the wind industry moves onto cleaner alternatives.