When the sun is shining, solar power in California is so abundant that the price for electricity from photovoltaics sometimes falls to zero. Indeed, California at times has had to pay its neighbors to take some of its surplus solar energy. But what if that energy could be stored for nighttime use, or for days or weeks when cloudy periods curtail production?
The problem of storing massive amounts of electricity becomes increasingly urgent as more and more intermittent renewable electricity is delivered to the power grid. About 13% of total US electricity supply, excluding hydroelectric power, comes from renewable sources, according to the Energy Information Administration. Even at that modest level, “the grid is 40 seconds from blackout all the time,” says Robert Abboud, CEO of Beacon Power, which produces flywheels for short-term energy storage. “The only way we stay sane is to balance the grid all the time between generation and demand. If we don’t do it correctly, it just falls down.”
Barring a renaissance in nuclear power, a breakthrough in fusion, or dramatic reductions in the cost of carbon capture and storage, meeting the Biden administration’s aggressive goal of a carbon-free electricity supply by 2035 will require a mainly renewable energy system. Storage remains too expensive for massive amounts of renewables to compete against legacy electricity sources such as coal and natural gas. Recognizing that reality, the Department of Energy in January 2020 announced a goal to lower by 90% the levelized cost of storage on an electricity grid by 2030, to around 5 cents per kilowatt-hour (¢/kWh). Levelized cost is defined as the nationwide average of capital and operating costs summed over the lifetime of the asset.
An artist’s rendering of Energy Vault’s grid-scale energy-storage system adjacent to an array of solar cells. Inside the building, a bank of freight elevators move composite blocks up and down to store and generate power.
An artist’s rendering of Energy Vault’s grid-scale energy-storage system adjacent to an array of solar cells. Inside the building, a bank of freight elevators move composite blocks up and down to store and generate power.
Similarly aggressive cost-reduction targets have been achieved for solar and wind energy, notes Daniel Kammen, who heads the Renewable and Appropriate Energy Laboratory at the University of California, Berkeley (see Physics Today, June 2021, page 27). Kammen codeveloped a two-factor model that correctly predicted the plunge in solar costs that has occurred during the last four years. Using that model as a basis, he says the US is already ahead of DOE’s storage price target and likely will get there by 2025. “We think that with the 1¢/kWh to 3¢/kWh that solar is already at, we’ll have baseload renewable power systems that are much cheaper to install and operate than it is to operate existing fossil-fuel plants,” he says.
DOE has included $1.16 billion in its fiscal year 2022 budget request for R&D on all forms of energy storage, more than doubling the $400 million it’s spending this year. That proposed outlay indicates the priority that the agency attaches to storage, says Eric Hsieh, director of grid systems and components in DOE’s Office of Electricity.
Water and chemistry
Pumped hydroelectric storage has been employed by electric utilities for more than a century, and today it still accounts for about 95% of all grid-scale capacity. Electricity generated during low-demand periods is used to pump water from one reservoir to another at a higher elevation. When demand is high, water is released through turbine generators to the lower pool. Although efficient and suitable for long-term storage, pumped hydro is severely restricted in its further deployment by availability of suitable locations, multibillion-dollar capital costs, and environmental concerns.
More recently deployed technologies provide short- and medium-duration energy-storage needs. Capacitors and flywheels address the short term. Beacon Power’s flywheels, for example, can boost power nearly instantaneously to provide load-leveling services—such as voltage and frequency regulation—to PJM, the nation’s largest regional transmission organization. Although flywheels can’t provide long-term storage yet, Abboud says that will come soon with the development of higher-strength carbon-fiber materials and magnetic bearings made with high-temperature superconductors, both of which will dramatically increase the rate of spin and mitigate energy lost by friction.
For storage of two to six hours, lithium-ion batteries dominate. Tesla, the leading supplier of grid-scale batteries, surpassed 3 GWh of deployments in 2020 alone. In California, where investor-owned utilities are required by law to provide a total of 1325 MW of power capacity by 2024, PG&E, the state’s largest utility, has already contracted for more than 1400 MW of lithium-ion deployments by 2023, nearly triple its 535 MW share of the state mandate. One of those projects, located at an electric substation in Monterey County, will be among the largest utility-owned lithium-ion projects in the world when it becomes operational later this year with a capacity of 182 MW. Consisting of 256 Tesla Megapack battery units, the facility will be capable of dispatching a total of up to 730 MW of energy over four hours. It will enhance grid reliability by addressing capacity deficiencies resulting from increased local energy demand, according to the company. It will also provide energy and load-leveling services to the electricity markets served by the California Independent System Operator.
But lithium-ion’s economics and physical properties limit its storage duration to eight hours of discharge. “If you take a lithium-ion system, charge it, and leave it for three months, it will self-discharge,” says Vincent Sprenkle, technical group manager for the Electrochemical Materials and Systems Group at Pacific Northwest National Laboratory (PNNL).
The energy storage company BoxPower has built several solar-powered microgrids of up to 250 KW with lithium-ion batteries to service PG&E customers in remote, wildfire-endangered locations. Those systems require backup generation from propane. “Battery storage is currently the limiting factor,” says CEO Angelo Campus. “In order to get to a truly 100% renewable solution, you have to start thinking about multiday or even seasonal range.”
Lithium is less than abundant, and the cobalt used in some batteries is even more scarce. (See Physics Today, December 2019, page 20.) The battery’s solvent electrolyte is flammable, as was illustrated by the fire at a Tesla Megapack grid-scale storage site near Melbourne, Australia, in late July. The solvents could also create a hazard in the event of a leak.
Most industry experts agree that lithium-ion batteries’ cost floor is too high, says Hsieh. The most promising battery technologies for long-term applications can decouple power from energy capacity, he says. In such battery types, energy-storage capacity can be increased by adding to the volume of electrolytes stored in the tanks. The battery’s electrochemical cells can be electrically connected in series or parallel to determine the power. (See the article by Héctor D. Abruña, Yasuyuki Kiya, and Jay C. Henderson, Physics Today, December 2008, page 43.) Within that category, redox flow batteries appear to be the leading candidates. Flow batteries consist of tanks of aqueous chemical redox pairs—electroactive compounds that can reversibly undergo reduction and oxidation—that flow through a fuel-cell-like device to generate electricity.
Tesla Megapack lithium-ion batteries located at a PG&E substation in Northern California. When completed later this year, the facility will be capable of dispatching energy to the grid at the rate of 182 MW for up to four hours.
Tesla Megapack lithium-ion batteries located at a PG&E substation in Northern California. When completed later this year, the facility will be capable of dispatching energy to the grid at the rate of 182 MW for up to four hours.
Redox flow-battery chemistries can vary, but the most common pair is vanadium–vanadium, in which that element’s four valence states are utilized. “[Flow batteries] have a nice body of scientific inertia behind them,” Hsieh says. “Most of the people working on them think there is a viable pathway to achieve very low costs.” Organic electrolytes that could further reduce flow batteries’ cost are being developed, says Wei Wang, director of PNNL’s Energy Storage Materials Initiative.
Unlike lithium-ion technology, aqueous-electrolyte batteries use abundant, nontoxic materials, says Esther Takeuchi, a materials scientist at Stony Brook University and Brookhaven National Laboratory. She notes that R&D investments in large-scale storage batteries have paled in comparison to the attention that’s been devoted to lithium-ion for electric vehicles and consumer electronics. Although energy densities and voltages of aqueous cells are lower than those of lithium-ion cells, they might be 20% of the price. Takeuchi and her colleagues have been investigating an aqueous-electrolyte stationary battery with a zinc anode and manganese oxide cathode.
Because of the pumps, tanks, and power electronics they require, however, flow batteries have a “not insignificant cost floor,” says Hsieh. Costs will come down as the manufacturing process is optimized, says Sprenkle. Flow batteries are less efficient (70–85% in the case of vanadium) than lithium-ion (90–95%), so marginal improvements in efficiency should have a big impact on lowering their levelized cost.
DOE is taking another look at the more-than-century-old lead–acid battery for grid-storage applications, says Hsieh. Today’s lead–acid cells discharge only 20–30% of their theoretical potential, and research on basic material properties is addressing how much more of that capacity could be used. “You could take advantage of a well-developed fully closed recycling-life-cycle supply chain in the US to get more performance out of an existing technology,” he says. Lead–acid’s lifetime of around two years needs improvement to compete against lithium-ion’s seven years, notes Sprenkle.
DOE has research programs underway on other battery chemistries, including sodium ion and sodium metal. “Batteries making it into the market is a long slog,” says Takeuchi. “But things are changing quickly. One change is our ability to probe batteries effectively to really understand the mechanisms. That’s one of the things we can do so well at Brookhaven. It takes years off the front end of R&D.” Improved computational power also is speeding up the development process. “Historically batteries were developed in a very Edisonian” trial-and-error way, she notes.
Heat, gravity, hydrogen
Nonchemical storage options include hydrogen, gravity- and thermal-based systems, and compressed air. “The real wild cards are the completely new technologies,” says Hsieh. “Every week I hear about a different type of thermal storage: storing heat in fluids or in concrete blocks, or in carbon-fiber blocks, with various ways of converting heat back into electricity.” Those systems are attractive in part because their base materials are so cheap. Only a couple of them should be needed to meet DOE’s cost-reduction goal, he notes.
Some concentrated solar arrays today focus solar energy onto a molten-salt medium during the day, storing the heat to generate electricity in the evening or during cloudy weather (see Physics Today, June 2021, page 27).
As electric vehicles become more prevalent, their collective batteries can become a grid-scale storage medium, says Kammen. “Every grid-enabled [equipped for bidirectional charging] F-150 truck and Tesla could be part of a roving fleet of storage.”
Some schemes are conceptually simple—others not as much. Swiss-based Energy Vault’s technology of raising and lowering 35-ton blocks made of dirt and polymer can store energy for just 60% of lithium ion’s levelized cost, according to a 2020 report from the forecasting firm BloombergNEF. The company predicts that by 2025 that number will drop to 51%, which would make Energy Vault’s storage system the least expensive among a dozen new alternatives ranked by Bloomberg. Other ventures included WattJoule’s redox flow battery; Ambri’s liquid-metal battery, composed of a liquid calcium-alloy anode, a molten-salt electrolyte, and a cathode of solid particles of antimony; a cryogenic air-storage system from Highview Power; and Siemens Gamesa’s thermal storage system, in which rocks are heated resistively and the stored energy is then discharged by steam turbine.
Energy Vault CEO Robert Piconi says the company has redesigned its product since its first grid-connected commercial deployment in Switzerland last year. In that system, computer-controlled cranes mounted atop each other stacked and unstacked the blocks. The redesign came in response to customer demand for higher power and shorter-duration storage. The new model employs freight elevators instead of cranes to raise and lower the blocks inside structures that resemble 30-story office buildings. Piconi says the company has eight orders across four continents totaling 1.2 GWh. One contract, with the renewable energy giant Enel Green Power, also calls for Energy Vault to recycle worn-out fiberglass wind-turbine blades into its storage blocks.
The Canadian company Hydrostor offers a fossil-fuel-free system where compressed air is stored in purpose-built underground caverns. A closed-loop water reservoir maintains the system at a near-constant pressure throughout the charge‒discharge cycle. The heat generated from compression is extracted and stored. As the cooled air enters the cavern, it displaces the water, pumping it to the surface. During discharge, hydrostatic pressure forces the air to the surface, where the stored heat expands it to drive a turbine.
Hydrostor’s first commercial plant, located in Goderich, Ontario, has been in operation since 2019, providing up to 10 MWh of peaking capacity to the grid operator. A number of other projects, in Australia, Chile, the US, and Canada, are in various stages of development.
Analyses performed at PNNL have shown that compressed-air pressurization of an underground cavern is one of the most cost-effective storage solutions, says Sprenkle. But it’s not necessarily a green technology. And Hsieh notes that compressed-air storage is geographically limited.
Hydrogen produced through water electrolysis using surplus renewable energy can be a long-term energy-storage medium, though costs remain a challenge, says Hsieh.
