Just after sunrise on a cloudless morning last June, two propellers started spinning on a slender aircraft sitting on a runway in the New Mexico desert. Chocks securing bicycle-sized wheels were removed, and the plane teetered down the tarmac. The wheels delicately detached from the plane as it lifted into the air and gradually ascended into the stratosphere.
The 150 kg plane, made by BAE Systems, is designed to stay aloft for weeks or months, which makes it useful for military operations, mapping, and communications. During the day, the plane draws electricity from the solar panels covering its wings and fuselage. But at night, it switches to battery power.
The long flights require an extremely light craft, so BAE has chosen a lithium-ion battery developed by Amprius Technologies that stores lots of energy without adding much weight. The key to this high-capacity battery is an anode made of silicon.
Silicon can store far more energy than graphite—the material used in the anode, or negatively charged end, of nearly all lithium-ion batteries. Silicon-dominant anodes are used in niche applications, such as BAE’s drone, but so far their high cost has kept them out of electric cars, a much larger market. Silicon anodes also swell significantly during charging, which reduces the battery’s longevity.
A number of companies now have technologies to control silicon’s swelling and are moving closer to large-scale manufacturing of anode materials. Sila Nanotechnologies and Group14 Technologies are building big silicon anode materials plants in Washington state. Amprius and OneD Battery Sciences also have plans for major facilities in North America.
In the electric vehicle (EV) industry, silicon-rich anode materials are initially slated to go into batteries for luxury SUVs. Some analysts say the materials will never be cheap enough for mass-market vehicles. But silicon proponents argue that expensive cars and electric planes are just a starting point that will demonstrate the technology’s value. They expect that scaling up will make the materials affordable for all types of cars.
“Right now, a lot of EVs are still pretty mediocre,” says Sila CEO Gene Berdichevsky. “If you can provide a huge range increase for a small premium, that’s worthwhile. Long term, that’s the way to lower costs.”
In a lithium-ion battery, lithium ions shuttle from the positively charged cathode into the negatively charged anode during charging. Anode materials that can accept more of these lithium ions can store more energy.
Graphite stores lithium ions between sheets of carbon, at best caching one lithium ion for every six carbon atoms. Silicon forms an alloy with lithium ions—a process that can store more than four lithium atoms for every silicon atom. The additional lithium atoms, and a lack of space to store them, cause silicon to expand by about 300% during charging. Graphite swells by only about 10% during a charge.
Silicon’s swelling can cause the anode to lose contact with a foil that collects the battery’s electric current. It also cracks a lithium-containing film that forms between the electrolyte and the anode. As the anode expands and contracts, the film repeatedly reforms, consuming lithium from the cathode and the electrolyte and degrading the battery’s performance.
Researchers have found several ways to tame this swelling. The start-ups Ionblox and NanoGraf are betting on silicon oxides: they store less energy than pure silicon but also swell less, expanding about 100% when infused with lithium. In addition, the film that forms when silicon oxides react with the electrolyte tends to be more robust, says Shirley Meng, a University of Chicago battery researcher and adviser to Ionblox.
Small amounts of silicon oxides have been mixed with graphite in some EV batteries, but it’s hard for battery makers to use anodes that contain over 7% of the oxides because they still swell too much, according to NanoGraf CEO Francis Wang. “You’re now seeing a second generation of silicon oxides,” he says.
NanoGraf’s material contains a metal-doped silicon oxide mixed with several additives that reduce swelling and form a more stable interface between the silicon and electrolyte. This allows the company to load the anodes of its batteries with up to 35% silicon oxide; graphite fills in the remainder.
Ionblox is aiming for even higher levels of silicon oxide. The firm uses an elastic, polymer binder to ensure that its oxide doesn’t lose contact with the battery’s current collector during expansion. It has also added pores in the oxide to accommodate swelling, as well as carbon nanotubes to increase conductivity. Using these techniques, Ionblox can make anode materials in which silicon oxides represent 60% or more of the active ingredients. “We believe that you have to use all these approaches . . . if you want to make silicon-dominant anodes,” CEO Sujeet Kumar says.
Meanwhile, Group14 and Sila are using another technology. Group14 puts nanometer-sized silicon particles inside a carbon-based matrix that leaves empty space into which the silicon can expand. Berdichevsky says Sila, which uses a similar concept, coats its anode material with a shell that prevents the repeated formation of the anode-electrolyte film.
The anode material of both Group14 and Sila is about half silicon, according to a report from the Volta Foundation, a nonprofit supporting the battery industry. Most of the companies’ customers plan to mix these products with graphite. That reduces the amount of energy the anode can store but makes it easier to incorporate into a battery maker’s existing process. “It really helps with adoption,” Berdichevsky says. “Customers don’t necessarily like to jump all the way there.”
Another approach is to form the silicon into nanowires. OneD Battery Sciences, for example, grows silicon nanowires in the pores within conventional graphite. The pores have enough room for the spaghetti-shaped nanowires to swell. CEO Vincent Pluvinage says the company has created materials that are 50% silicon by weight, though most customers are requesting 3–30% silicon content.
Amprius also uses silicon nanowires, growing them directly onto the battery’s current collector to ensure they stay connected. At a microscopic level, CEO Kang Sun says the firm’s nanowire material looks like a carpet with lots of fibers. The space between the nanowires gives the material room to expand.
Unlike other firms, Amprius doesn’t mix its nanowire material with graphite powder and thus increases the amount of energy it can store. But the novel manufacturing process can’t easily be retrofitted on battery plants that use graphite anodes. Instead, the company is producing its own battery cells.
While new technology has resolved some of the issues related to swelling, the high cost of silicon weighs against its use in EVs, according to several battery industry analysts.
To break into car batteries, companies will have to show that $1 of silicon can store more energy than $1 of graphite, says Charlie Parker, founder of the battery advisory firm Ratel Consulting. He doesn’t think they are there yet. “The cost penalty doesn’t outweigh the performance gain,” he says. “This stuff is great. It’s also expensive.”
Several firms are hoping to ease into the auto industry by first placing their materials into devices for which customers are willing to pay a premium.
Group14 and Sila are deploying their anode material in consumer electronics, where batteries represent a small fraction of the product’s cost. NanoGraf’s beachhead market is the military, which is prepared to pay top dollar to reduce the weight of batteries that soldiers carry. And the aerospace company Airbus says its drones could not achieve extended stratospheric flight without the high-performance batteries it buys from Amprius.
“It’s the extreme-use cases where you need this incredible performance,” Parker says. “In those applications where it is entirely necessary, you just pay it and go about your day.”
The hope is that these smaller markets will allow companies to boost manufacturing volumes and reduce costs through economies of scale. James Willoughby, an anode materials analyst with the consulting firm Wood Mackenzie, says larger facilities should eventually make silicon cheaper than graphite as a way to store energy. “The major constraint on the cost at the moment is really just the scale,” he says.
And silicon materials firms are trying to scale up. Amprius is planning a major facility in Colorado. Sun claims that the larger capacity will bring the cost of Amprius’s material close to that of graphite.
NanoGraf opened a 35-metric-ton-per-year plant in Chicago in December and has already signed a lease for a larger one down the road. The former is large enough to fulfill the company’s contracts with the military, but Wang predicts that production volumes will have to reach 1,000 metric tons per year to compete with graphite.
“Scaling is always rough,” he says. “They call it production hell, and that’s where we’re heading. It’s totally doable.”
Both Sila and Group14 say the large facilities they are building in Washington will be able to produce materials at the price carmakers are demanding—at least for luxury vehicles. Group14 expects its anode materials to be used by Porsche’s battery subsidiary by year-end. And Sila anticipates that Mercedes will use its material in the EQG SUV, which Car and Driver predicts will cost $150,000 when it debuts sometime in the next few years.
Sam Adham, a battery materials analyst with the consulting firm CRU Group, says customers buying such premium vehicles can probably absorb silicon’s higher cost. “It’s also a test bed for proving the technology in a real-world application before it proliferates,” he says.
Berdichevsky says the initial output from Sila’s Washington plant will go into the Mercedes EQG. But as production increases, he expects to power more mass-market vehicles. In December, Sila agreed to sell silicon material to Panasonic, which supplies batteries to Tesla. “We’re starting to think about . . . bigger projects. The EQG is not a high-volume car,” he says. “The goal is to ramp into very large-scale production.”
In contrast, OneD hopes to go straight into affordable cars. Pluvinage says the luxury EV market isn’t growing, so it doesn’t make sense to target those cars. “Our customers are telling us, ‘We are only interested in silicon technology if they can decrease cell cost,’ ” he says.
In addition to offering high energy density, silicon is gaining momentum because battery makers are increasingly concerned about China’s domination of the graphite supply chain. But silicon is no panacea.
Amprius, Group14, OneD, and Sila all use silane gas as a starting material. Willoughby at Wood Mackenzie says much of the world’s silane gas is produced in China, where it’s primarily used to make polysilicon for solar panels and semiconductors.
Group14 and Sila’s facilities will both be in Moses Lake, Washington, home to the silane gas producer REC Silicon. The problem, Pluvinage says, is that REC sells most of its silane to the solar industry and is charging a premium for the remaining supply.
OneD hopes to open a pilot facility in Moses Lake later this year but wants cheaper silane for commercial-scale production. The company recently announced that it will work with the engineering firm Koch Modular to construct silane plants adjacent to its future production sites. The partners are planning one such complex in Quebec.
Group14 expects that REC will have excess silane available for silicon anode producers in the near term, though the firm notes that the battery industry will need new silane producers in the long run. In July, Group14 acquired the German silane producer Schmid Silicon. Group14 chief technology officer Rick Costantino Costantino says Schmid’s technology makes silane inexpensively and allows his firm to integrate silane production directly into anode materials plants.
Other companies are looking for ways to avoid silane altogether. The start-up Advano is investigating making silicon anode materials from recycled solar panels (J. Power Sources 2023, DOI: 10.1016/j.jpowsour.2023.233245). And Ionblox CEO Kumar argues that silicon oxides, which are essentially made by processing sand, will be easier to procure than silane.
Even if firms source cheap raw materials and bring down costs, Aaron Wade, a battery price analyst with CRU Group, says it’s unlikely that silicon will totally replace graphite. Instead, he expects to see a variety of anode technologies—just as cathode chemistries are optimized for energy density, safety, or cost, depending on the application.
“The battery industry . . . is continually diversifying,” he says. “We’re going to have applications for graphite, applications for silicon, and applications for lithium metal.”
But Costantino is convinced that silicon will be the first alternative to graphite to be commercialized. He says the technology is ready now and will only get better as other battery components, such as electrolytes and additives, are redesigned to work with silicon. “Folks have been using graphite for 40 years. They’ve had time to optimize and improve everything else around it,” he says. “We’re just getting started with silicon.”