When the Sun is blazing and the wind is blowing, Germany’s solar and wind power plants swing into high gear. For nine days in July 2023, renewables produced more than 70% of the electricity generated in the country; there are times when wind turbines even need to be turned off to avoid overloading the grid.

But on other days, clouds mute solar energy down to a flicker and wind turbines languish. For nearly a week in January 2023, renewable energy generation fell to less than 30% of the nation’s total, and gas-, oil- and coal-powered plants revved up to pick up the slack.

Germans call these periods Dunkelflauten, meaning “dark doldrums,” and they can last for a week or longer. They’re a major concern for doldrum-afflicted places like Germany and parts of the United States as nations increasingly push renewable-energy development. Solar and wind combined contribute 40% of overall energy generation in Germany and 15% in the US and, as of December 2024, both countries have goals of becoming 100% clean-energy-powered by 2035.

The challenge: how to avoid blackouts without turning to dependable but planet-warming fossil fuels.

Solving the variability problem of solar and wind energy requires reimagining how to power our world, moving from a grid where fossil fuel plants are turned on and off in step with energy needs to one that converts fluctuating energy sources into a continuous power supply. The solution lies, of course, in storing energy when it’s abundant so it’s available for use during lean times.

But the increasingly popular electricity-storage devices today – lithium-ion batteries – are only cost-effective in bridging daily fluctuations in sun and wind, not multiday doldrums. And a decades-old method that stores electricity by pumping water uphill and recouping the energy when it flows back down through a turbine generator typically works only in mountainous terrain. The more solar and wind plants the world installs to wean grids off fossil fuels, the more urgently it needs mature, cost-effective technologies that can cover many locations and store energy for at least eight hours and up to weeks at a time.

Engineers around the world are busy developing those technologies – from newer kinds of batteries to systems that harness air pressure, spinning wheels, heat or chemicals like hydrogen. It’s unclear what will end up sticking.

“The creative part… is happening now,” says Eric Hittinger, an expert on energy policy and markets at Rochester Institute of Technology who coauthored a 2020 deep dive in the Annual Review of Environment and Resources on the benefits and costs of energy storage systems. “A lot of it is going to get winnowed down as front-runners start to show themselves.”

Finding viable storage solutions will help to shape the overall course of the energy transition in the many countries striving to cut carbon emissions in the coming decades, as well as determine the costs of going renewable – a much-debated issue among experts. Some predictions imply that weaning the grid off fossil fuels will invariably save money, thanks to declining costs of solar panels and wind turbines, but those projections don’t include energy storage costs.

Other experts stress the need to do more than build out new storage, like tweaking humanity’s electricity demand. In general, “we have to be very thoughtful about how we design the grid of the future,” says materials scientist and engineer Shirley Meng of the University of Chicago.

Reinventing the battery

The fastest-growing electricity storage devices today – for grids as well as electric vehicles, phones and laptops – are lithium-ion batteries. Recent years have seen massive installations of these around the globe to help balance electricity supply and demand and, more recently, to offset daily fluctuations in solar and wind. One of the world’s largest battery grid storage facilities, in California’s Monterey County, reached its full capacity in 2023 at a site with a natural-gas-powered plant. It can now store 3,000 megawatt-hours (MWh) and is capable of providing 750 MW – enough to power more than 600,000 homes every hour for up to four hours.

Lithium-ion batteries convert electrical energy into chemical energy by using electricity to fuel chemical reactions at two lithium-containing electrode surfaces, storing and releasing energy. Lithium became the material of choice because it stores a lot of energy relative to its weight. But the batteries have shortcomings, including their fire risk, their need for air-conditioning in hot climates and a finite global supply of lithium.

Importantly, lithium-ion batteries aren’t suitable for long-duration storage, explains Meng. Despite monumental price declines in recent years, they remain costly due to their design and the price of mining and extracting lithium and other metals. The battery cost is above $100 per KWh – meaning that a battery container supplying one MW (enough for about 800 homes) every hour for five hours would cost at least $500,000. Providing electricity for longer would quickly become economically unfeasible, Meng says. “I think four to eight hours is really a sweet spot for balancing cost and performance,” she says.

For longer durations, “we want energy storage that costs one tenth of what it does today – or maybe, if we could, one hundredth,” Hittinger says. “If you can’t make it extremely cheap, then you don’t have a product.”

One way of cutting costs is to switch to cheaper ingredients. Several companies in the US, Europe and Asia are working to commercialise sodium-ion batteries that replace lithium with sodium, which is more abundant and cheaper to extract and purify. Different battery architectures are also being developed – such as “redox flow” batteries, in which chemical reactions take place not at electrode surfaces but in two fluid-filled tanks that act as electrodes. With this kind of design, capacity can be enlarged by increasing tank size and electrolyte amount, which is much cheaper than increasing the expensive electrode material of lithium-ion batteries. Redox-flow batteries could supply electricity over days or weeks, Meng says.

US-based company Form Energy, meanwhile, just opened a factory in West Virginia to make “iron-air” batteries. These harness the energy released when iron reacts with air and water to form iron hydroxide – rust, in other words. “Recharging the battery is taking rust and unrusting it,” says William Woodford, Form’s chief technical officer.

Because iron and air are cheap, the batteries are inexpensive. The downside with both iron-air and redox-flow batteries is that they give back up to 60% less energy than is put into them, partly because they gradually discharge with no current applied. Meng thinks both battery types have yet to resolve these issues and prove their reliability and cost-effectiveness. But the efficiency loss of iron-air batteries could be dealt with by making them larger. And since long-duration batteries supply energy at times when solar and wind power is scarce and more costly, “there’s more tolerance for a little bit of loss,” Woodford says.

Spinning wheels and squished air

Other engineers are exploring mechanical storage methods. One device is the flywheel, which employs the same principle that causes a bike wheel to keep spinning once set into motion. Flywheel technology uses electricity to spin large steel discs, and magnetic bearing systems to reduce the friction that causes slowdowns, explains electrical engineering expert Seth Sanders of the University of California, Berkeley. “The energy can be stored for actually a very substantial amount of time,” he says.

Sanders’ company, Amber Kinetics, produces flywheels that can spin for weeks but are most cost-effective when used at least daily. When power is needed, a motor generator turns the movement energy back into electricity. As the wheels can switch quickly from charging to discharging, they’re ideal for covering rapid swings in energy availability, like at sunset or during cloudy periods.

Each flywheel can store 32 KWh of energy, close to the daily electricity demand of an average American household. That’s small for grid applications, but the flywheels are already deployed in many communities, often to balance fluctuations in renewable energy. A municipal utility in Massachusetts, for instance, has installed 16 flywheels next to a solar plant; they supply energy for more than four hours, absorbing electricity during low-demand times and discharging during peak demand, Sanders says.

A different kind of mechanical facility stores electricity by using it to compress air, then stashes the air in caverns. “When the grid needs it, you release that air into an air turbine and it generates electricity again,” explains Jon Norman, president of the Canada-based company Hydrostor, which specialises in compressed-air storage. “It’s just a giant air battery underground.”

Such systems usually require natural caverns, but Hydrostor carves out cavities in hard rock. Compared to batteries or flywheels, these are large infrastructure projects with lengthy permitting and construction processes. But once those hurdles are passed, their capacity can be slowly scaled up by carving the caverns more deeply, at pretty low additional cost, Norman says.

In 2019, Hydrostor launched the first commercial compressed-air storage facility, in Goderich, Ontario, storing around 10 MWh — enough to power some 2,100 homes for more than five hours. The company plans several much larger facilities in California and is building a 200-MW facility in the Australian town Broken Hill that can supply energy for up to eight hours to bridge shortfalls in solar and wind energy.

Storing energy as heat and gas

Around the world, there are efforts afoot to make use of excess renewable electricity by using it to heat up water or other heat-storing materials. This can then provide climate-friendly warmth for buildings or industrial processes, says Katja Esche of the German Energy Storage Association.

Heat can also be used to store energy, though that technology is still being developed. Energy storage and systems expert Zhiwei Ma of Durham University in the United Kingdom recently tested a pumped thermal energy storage system. Here, the main energy-storing process occurs when electricity is used to compress a gas, like argon, to a high pressure, heating it up; electricity is generated when the gas is allowed to expand through a turbine generator. Some experts are skeptical of such thermal storage systems, as they supply up to 60% less electricity than they store – but Ma is optimistic that with more research, such systems could help with daily storage needs.

For even longer-duration storage – over weeks – many experts put their bets on hydrogen gas. Hydrogen exists naturally in the atmosphere but can also be produced using electricity to split water into oxygen and hydrogen. The hydrogen is stored in pressurised tanks and when it reacts with oxygen in a fuel cell or turbine, this generates electricity.

Hydrogen and its derivatives are already being explored as fuel for ships, planes and industrial processes. For long-duration storage, “it looks plausible that that would be the technology of choice,” says energy expert Wolf-Peter Schill of the German Institute for Economic Research who coauthored a 2021 review on the economics of energy storage in the Annual Review of Resource Economics.

The German energy company Enertrag is building a facility that uses hydrogen in both ways. Surplus energy from the company’s 700-MW solar and wind plant near Berlin is used to make hydrogen gas, which is sold to various industries. In the future, about 10% of that hydrogen will be stashed away “as an emergency backup measure” for use during weeks without sun or wind, says mechanical engineer Tobias Bischof-Niemz, who is on Enertrag’s board.

The idea of using hydrogen for electricity storage has many critics. Similar to heat, up to two-thirds of the energy is lost during reconversion into electricity. And storing massive quantities of hydrogen over weeks isn’t cheap, although Enertrag is planning on reducing costs by storing it in natural caverns instead of the customary pressurised steel cylinders.

But Bischof-Niemz argues that these expenses don’t matter much if hydrogen is produced from cheap energy that would otherwise be wasted. And, he adds, hydrogen storage would be used only for Dunkelflauten periods. “Because you only have two or three weeks in the year that are that expensive, it works economically,” he says.

A question of cost

There are many other efforts to develop longer-duration storage methods. Cost is key for all, regardless of how much is paid for by governments or utility companies (the latter typically push such costs onto consumers). All new systems will need to prove that they’re significantly cheaper than lithium-ion batteries, says energy expert Dirk Uwe Sauer of Germany’s RWTH Aachen University. He says he has seen many technologies stall at the demonstration stage because there’s no business case for them.

Developers, for their part, argue that some systems are approaching that of lithium-ion batteries when used to store energy for eight hours or more, and that costs will come down substantially for others when they are manufactured in large volumes. Maybe many technologies could, ultimately, compete with lithium-ion batteries, but getting there, Sauer says, “is extremely difficult.”

The challenge for developers is that the market for long-duration technologies is only beginning to take shape. Many nations, such as the US, are early in their energy transition journey and still lean heavily on fossil fuels. Most regions still have fossil-fuel-powered plants to cover multiday doldrums.

Indeed, Hittinger estimates that the real economic need for long-duration storage will only emerge once solar and wind account for 80% of total power generation. Right now, it can often be cheaper for utilities to build gas plants – fossil fuels, still – to ensure grid reliability.

One important way to make storage technologies more economical is a carbon tax on fossil fuels, says energy systems researcher Anne Liu of Aurora Energy Research. In European countries like Switzerland, utilities are charged up to about $130 per metric ton of carbon emitted. California grid operators, meanwhile, have spurred storage development by requiring utility companies to ensure adequate energy coverage, and helping to cover the cost.

Market incentives can also help. In the Texas energy market, where electricity prices fluctuate a lot, electricity customers are saving hundreds of millions of dollars from the build-out of lithium-ion batteries, despite their costs, as they can store energy when it’s cheap and sell it for a profit when it’s scarce. “Once those power markets have incentive, then the longer-duration batteries will be more viable,” Liu says.

But even when incentives are there, the question remains of who will foot the bill for energy storage, which isn’t considered in many cost projections for transitioning the grid off fossil fuels. “I don’t think there’s been enough time spent studying how much these decarbonisation pathways are going to cost,” says Gabe Murtaugh, director of markets and technology at the nonprofit Long Duration Energy Storage Council.

Without interventions, Murtaugh estimates, California customers, for instance, could eventually see a threefold increase in utility bills. “Thinking about how states and federal governments might help pay for some of this,” Murtaugh says, “is going to be really important.”

Saving costs and resources

Cost considerations are prompting experts to also think of ways to reduce the need for storage. One way to strengthen the grid is building more consistently available forms of renewable energy, such as geothermal technologies that draw energy from the Earth’s heat. Another is to connect the grid over larger regions — such as across the US or Europe — to balance local fluctuations in solar and wind. Ensuring that storage technologies are as long-lived as possible can help to save costs and resources.

So can being smarter about when we draw electricity from the grid, says Seth Mullendore, president of the Vermont-based nonprofit Clean Energy Group. What if, rather than charging electric cars when getting home from work, we charged them at midday when the Sun is blazing? What if we adjusted building heating and cooling so the bulk would happen during windy periods?

Mullendore’s nonprofit recently helped to design a program in Massachusetts where electricity customers could sign up to get paid if they responded to signals from their utilities to use less energy – for instance, by turning their air-conditioning down or delaying electric car charging. In a smart grid of the future, such tweaks could be more widespread and fully automatic, while allowing consumers to override them if needed. Governments could encourage programs by rewarding utility companies for designing grids more efficiently, Mullendore says. “It’s much less expensive to have people not use energy than it is to build more infrastructure to deliver more energy.”

It will take careful thought and a worldwide push by engineers, companies and policymakers to adapt the global grid to a solar- and wind-powered future. Tomorrow’s grids may be studded with lithium-ion or sodium-ion batteries for short-term energy needs and newer varieties for longer-term storage. There may be many more flywheels, while underground caverns may be stuffed with compressed air or hydrogen to survive the dreaded Dunkelflauten. Grids may have smart, built-in ways of adjusting demand and making the very most of excess energy, rather than wasting it.

“The grid,” Meng says, “is probably the most complicated machine ever being built.”