Hybrid electric cars need much better batteries--and A123, a plucky Massachusetts start-up, says it's got them
We knew from the start that we wanted to do auto batteries,” says Ric Fulop, a 30â''something entrepreneur with an electrical engineering degree and a curly mop of brown hair. ”But we also knew that automakers only buy from companies with volume production and real customers.”
It’s a version of the old chicken-and-egg problem that has confronted would-be tech entrepreneurs for decades. But Fulop and company came up with a novel solution: ”We had to do power tools first.”
In 2001, Fulop, then 26, set up A123 Systems in Watertown, Mass., with three partners, taking the position of vice president of business development. Late last year the company’s new design for lithium-ion batteries hit the market in a line of power tools aimed at professional builders from the DeWalt Industrial Tool Co. The batteries operate at 36 volts, twice the voltage of their predecessors, and hold 130 watt-hours per kilogram--twice as much as standard nickel-metal-hydride cells.
Lithium-ion cells are poised to take an increasing share of the auto battery market, just as electric drive seems set to begin a long, slow climb to become, at last, a serious power-train option. But what’s rarely understood is how much that second revolution depends on the first.
The auto industry transformation began modestly enough a decade ago with the Toyota Prius, the now wildly successful gasoline-electric hybrid. And if A123 and dozens of like-minded companies and research groups can deliver on the promise of lithium-ion batteries for vehicle propulsion, in four to 10 years plug-in hybrids could be capable of going substantial distances on electricity alone. Enthusiasm for the plug-ins being tested now, along with the 15- to 65-kilometer pure-electric range projected for their successors using lithium-ion battery packs, has raised hopes. Some analysts dare to contemplate the re-emergence of a mass-market electric car, perhaps within a decade.
Chalk it up to changing attitudes as much as breakthrough inventions. High gasoline prices have given regulators and drivers alike a reason to smile on hybrids. And investors in the currently fashionable green tech sector love new energy-storage technologies, so critical to electric-drive vehicles.
There are plenty of technical challenges--in the cells themselves, in the battery packs where they reside, and in the cars that will have to be engineered around them. The first to meet the challenges will be in the driver’s seat of tomorrow’s cars. A123, with its modest staff of 300 scientists and engineers, says its unique proprietary technology gives it a shot [see photos, "Battery Factory."
”The first vehicles to use lithium-ion batteries will come in 2009,” Fulop declares. ”In 2010, there’ll be several. By 2015, most of the world’s hybrids will use them.”
A123 already has contracts to supply batteries to several European and American automakers, Fulop adds coyly, declining to identify the companies. He points out that early this year A123 received one of General Motors’ first commissions for R&D work on lithium-ion batteries.
In fact, in June, GM raised the stakes, announcing two more R&D contracts: one to Compact Power of Troy, Mich., which plans to use cells from Korean battery maker LG Chem, and the other to a division of the German auto parts maker Continental, which plans to build battery packs incorporating A123’s cells.
Experts agree that lithium-ion cells will power coming generations of cars--hybrid, plug-in hybrid, and pure electric. At firstï»' the car companies will put the new batteries in just a few standard hybrids, to test the waters, or they’ll use them to fill market niches, like the one for such dazzlingly fast sports cars as the Tesla Roadster [see sidebar "Tesla: Not for Geeks Alone,"]. Later they’ll put them in plug-ins--at first, in standard parallel designs, which drive the wheels with either the motor or the engine or some combination of the two. Then, perhaps, they’ll move on to the more radical series design, in which the electric motor drives the wheels, leaving the engine no other role than to recharge the batteries.
Cars won’t come until the batteries are affordable, and batteries won’t be affordable until the automakers purchase a lot of them. This year, though, the world’s top two automakers made firm commitments to lithium-ion technology.
Toyota, the world’s biggest and most profitable car company, said that late next year it will put lithium-ion batteries in an unspecified hybrid vehicle. It will also test a fleet of plug-in hybrids, using nickel-metal-hydride cells, that are able to run a few kilometers on batteries alone. Today’s Prius can do that for only a couple of minutes, and then only at speeds of less than 50 km/h.
General Motors is playing catch-up--but with a vengeance. Late this year, it expects to finally launch its first hybrids able to run in all-electric mode, if only for a minute or two. GM recently said it will ”soon” offer a true plug-in, with an all-electric range of 16 km (10 miles), although it hasn’t committed to a launch date, saying that the batteries aren’t yet ready. GM is also planning to build a true series hybrid--the Chevrolet Volt, first shown as a concept vehicle in January.
Why aren’t the batteries ready for prime time? There are lots of reasons, including cell life and cost, but perhaps the biggest of all is safety. Remember last year’s vivid videos of flaming laptops? Nobody was hurt, but the resulting recall of millions of lithium-ion batteries was a black eye for Sony and other major vendors. If a lithium-ion powered minivan carrying a family were to burst into flames, the resulting fiasco could set the industry back a decade. And it’s no use arguing that something like 250 000 gasoline-powered cars catch fire every year in the United States alone. New products are held to a higher standard.
Safety is key, and it all comes down to preventing fires and explosions. These catastrophes happen when a cell shorts out, gets hot, and starts an exothermic oxidizing reaction that kicks the temperature to hundreds of degrees Celsius in a fraction of a second. The heat then shorts out adjacent cells to produce a runaway thermal reaction that can be spectacular (just ask Sony). And, unlike a gasoline fire, the conflagration can’t be smothered, because it gets oxygen from the cell’s intrinsic chemistry.
Field failures occur once in every 5 million to 10 million of the most common lithium-ion cells, those known as the 18650 design, according to Brian Barnett, a technology analyst at Tiax, a consulting firm. Of course, the more cells there are in a battery pack, the greater the chance of a problem. Although it’s clear that impurities introduced during manufacturing are largely to blame, the mechanism remains unclear.
There are several ways to make the new technology safe enough for cars. One, perhaps transitional, approach is to link large numbers of small cells in networks--as the Tesla does--with safeguards to ensure that a problem in one cell cannot propagate to others. A123 and some other start-ups instead chose to focus on the fundamental reactions in the cell.
First, a little chemistry. (Don’t worry--it’s so straightforward that chemists like to call lithium-ion ”the physicist’s battery.”) Like any battery cell, this one has two electrodes sitting in an ion-rich solution, the electrolyte [see diagram, "Anatomy of a Cell”]. The electrodes are typically very close, so a polymer film, called a separator, prevents contact and a possible short circuit. A switched external circuit connects the electrodes to draw power, and the electrochemical reaction begins.
Ionized elements in the anode--in this case, including lithium--are tugged by the electric potential that is inherent in their chemical relation to elements in the opposing electrode, the cathode. The ions move through the electrolyte and the separator. Those arriving at the cathode give up electrons; those coming to the anode accept them. Electrons travel through the external circuit, producing a flow of charge complementary to the flow of ions. During recharge, current is forced into the cell, reversing the process.
Cell shapes vary widely, from thin discs hardly larger than a pinhead to high-power specimens the size of a small fire extinguisher. The consensus view in the auto industry is that battery packs will consist of up to 100 large-format cells of 20 to 50 ampere-hours apiece--each cell perhaps 50 millimeters wide and 200 mm long--grouped into modules that include sensors and electronics. The modules feed data to an electronic battery management system, which performs the crucial function of enabling cells of varying power and voltage to work together as a unit.
The Tesla alternative, packaging thousands of inexpensive commodity cells, requires far more sensors and control software than would be practical for mass-market vehicles.
There is no one lithium-ion battery. Several chemical designs compete, each with advantages and drawbacks. ”No chemistry will be the perfect one,” says Klaus Brandt, the chief executive of Gaia, a German subsidiary of Lithium Technology Corp., of Plymouth Meeting, Pa. ”Different chemistries will find different niches that vary with cost, performance, and safety.”
The anode is typically made of graphite, but the cathode composition varies widely from design to design, and as much as any other factor it determines a battery’s capacity. The critical feature is the rate at which the cathode can absorb and emit free lithium ions; this parameter in turn largely determines the power density. Each of several competing cathode materials has a different mix of cost, durability, susceptibility to temperature, and so forth. Cobalt is more reactive than nickel or manganese, meaning it offers high electrical potential when paired with graphite anodes, permitting higher voltage. However, cobalt, like nickel, is expensive. Manganese is cheaper, but it is slightly soluble in electrolytes--which means a shorter useful life.
The most important cathode contenders:
Cobalt dioxide is the most popular choice today for small cells. It has been on the market for 15 years, so it is proven and its costs are known. It has high electrical potential and the highest energy density--up to 600 Wh/kg. On the other hand, when fully charged, it is the most prone of all the cathode alternatives to oxidation and subsequent thermal runaway. Its internal impedance--the extent to which it ”pushes back” against an alternating current--also increases with both calendar time and cycling. Cobalt dioxide cells are manufactured by dozens of Chinese, Japanese, and South Korean companies.
Nickel-cobalt-manganese is somewhat easier to make. Substituting nickel and manganese for some of the cobalt raises the electrical potential only slightly, but it’s enough to let manufacturers tune the cell either for higher power or for greater energy density, though not both at the same time. (Remember that total energy determines the vehicle’s range, whereas available power determines its acceleration.) It is susceptible to thermal runaway, though less so than cobalt dioxide. Its long-term durability is still unclear, and nickel and manganese are both pricey at the moment. Manufacturers include Hitachi, Panasonic, and Sanyo.
Nickel-cobalt-aluminum is similar, with lower-cost aluminum replacing the manganese. Companies manufacturing NCA cells include Toyota and Johnson Controls–Saft, a joint venture between a Milwaukee company and a French firm.
Manganese oxide spinel offers higher power at a lower cost than cobalt, because its three-dimensional crystalline structure provides more surface area, permitting more ion flow between the electrodes. The drawback is an energy density only slightly better than 450 Wh/kg. GS Yuasa, LG Chem, NEC-Lamilion Energy, and Samsung offer cells with such cathodes.
Iron phosphate may well be the most promising new cathode, thanks to its stability and safety. This is what A123 is using in its batteries. Other manufacturers include Gaia and Valence Technology, in Austin, Texas. The compound is inexpensive, and because the bonds between the iron, phosphate, and oxygen atoms are far stronger than those between cobalt and oxygen atoms, the oxygen is much harder to detach when overcharged. Therefore, when it fails, it does so without overheating.
Unfortunately, however, iron phosphate doesn’t conduct well; to compensate, engineers have to add dopants. Even then, the cells work at a lower voltage than cobalt, so more of them must be chained together to drive a motor. That means iron phosphate battery packs need more interconnections and sensors to control the system.
One way around that problem is A123’s use of nanostructures in the cathode. This proprietary method produces better power and longer life than earlier generations of iron phosphate cells, says Andy Chu, a researcher at A123.
As phosphate molecules in the cathode acquire and give off lithium atoms--undergoing lithiation and delithiation--the phase boundary between the two states shifts, just as the boundary between cold water and ice does during freezing. In A123’s nanostructures, Chu says, the molecular lattices of the two states are structurally more similar to each other than in other phosphate cells, so atoms need less time to rearrange themselves. That means lithiation can proceed faster, delivering more power.
Moreover, because the lattice spacing of the two phases is closer, the physical stress on the cell is reduced, especially in deep discharge and charge. The cells should thus last longer.
”The batteries have performed as advertised by A123,” said a third-party tester who requested anonymity because he wasn’t authorized to speak to the news media. He noted that even when the cells were subjected to severe abuse, including extreme overcharging, they failed in a ”relatively benign fashion.”
One shadow hanging over iron phosphate chemistries is the extent of the coverage of patents for work done by the pioneering researcher in the field, John Goodenough, now of the University of Texas at Austin. A123 insists that its work does not violate the patents. Gaia, on the other hand, purchases only materials manufactured under license to the patent owner.
One characteristic flaw of lithium-ion batteries, anode plating, comes when a recharging cell dumps lithium ions faster than the anode can absorb them. This problem can be caused either by low temperatures, which slow the rate of diffusion, or by overcharging, which slows the rate of absorption. One of the jobs of the battery management system is to keep overcharging from ever happening.
Plating is bad for a number of reasons, particularly because it further reduces absorption, increasing the concentration of carbon ions until they begin to react with the oxygen in the electrolyte. The oxidation--equivalent to that in a burning lump of coal--creates a lot of heat, which in turn increases the rate of deposition.
A123 says its carbon anode combines the high rate of charging provided by graphitic carbon with the long life of nongraphitic types. It won’t give details of its proprietary formulation, saying only that it fine-tunes the size and structure of the particles.
Altair Nanotechnologies of Reno, Nev., wards off plating by coupling standard cobalt oxide cathodes with anodes made of lithium titanate spinel rather than graphite. The spinel won’t react with oxygen, and it also charges fast and lasts long. However, the energy density--at the current, early stage of development--is only half that of standard cobalt cells, and it is little better than that of nickel-metal-hydride cells.
The second-toughest problem after thermal runaway is limited life span, as measured by both the calendar and the number of charge-discharge cycles. A123’s Fulop says the cycle-life goals are easy to meet, but the calendar-life ones will be harder.
Cobalt-based cells for portable electronics lose as much as 20 percent of their capacity each year, starting from the day of manufacture. That may be tolerable for cellphones and other portables that are replaced every three or four years, but not for a car, which is expected to last 15 years.
The California Air Resources Board requires a vehicle’s power train to last for 10 years or 150 000 miles (240 000 km) with the original components. GM has said, meanwhile, that it expects battery packs for its Volt concept car to last for at least 4000 full-discharge cycles. That’s good but might not be good enough. At one charge-discharge cycle per day, the pack would last for 11 years--though it’s the rare car that runs 365 days a year for a decade.
Worse yet, auto and battery makers don’t have the luxury of spending 10 years testing lithium-ion packs. ”Ideally,” says Mark Verbrugge, director of GM’s materials and processes laboratory, ”we’d have half the life span to test it. But we don’t, so there’s no clean answer.” Meanwhile, automakers are ”oversizing” their battery packs to ensure they’ll power the car even after projected degradation. Of course, that strategy adds cost and weight.
THEY PUT THE JOLT IN THE VOLT
A123’s dynamic trio pose beside Chevrolet’s Volt, a serial hybrid concept car that uses their company’s batteries.
Then there’s the final hurdle: cost. At the moment, 12-V lead-acid batteries cost US $40 to $50 per kWh. Nickel-cadmium and nickel-metal-hydride cells for portable electronics cost $350/kWh; lithium-ion cells for the same market go for $450/kWh. Move to hybrid vehicles, though, and the price for longer-lived, more rugged nickel-metal-hydride batteries shoots up to about $700/kWh. That’s more than double the $300 target set by the U.S. Advanced Battery Consortium for automotive lithium-ion packs.
Manufacturers expect to reach that target by 2015, but in the earlier stages of production the price will likely be several times higher. How low must the price fall before a manufacturer will commit to even a low-volume purchase? No one will say, though every manufacturer surely has a threshold in mind. As GM’s Verbrugge summarized with a straight face, ”Cost lower--always better.”
World politics plays a role in some of those costs, especially prices of the raw materials. Lithium is not a ”strategic metal,” unlike nickel--whose price is surging as demand for stainless steel grows--so the cost of the metal per kilowatt-hour is lower for lithium than for nickel-metal-hydride. Right now, Chile and Argentina supply much of the world’s lithium carbonate, but Bolivia and China also have large reserves.
Geography does matter in another way, one that may give A123 an advantage: its headquarters and research labs are in the United States. No automaker wants to depend on a supplier in a distant land, especially one whose loyalties lie with a competitor. Take Ford: it purchased nickel-metal-hydride battery packs for its Escape Hybrid SUV from Japan’s Sanyo Electric Co., which had developed them for Toyota. But if battery supplies get tight, Sanyo’s ties to Toyota surely will outweigh Ford’s needs.
With North American and European companies intent on nurturing battery companies in their own backyards, A123 is focusing its sales and marketing efforts in those regions. And clearly that focus has started to pay off.
A123 is confident it can compete with the big boys. It is fully global, concentrating its research and development in North America and manufacturing in Asia. Already Europe’s Continental will build A123 cells into battery packs, and Fulop itches to provide details of A123’s other contracts.
How, then, is development going? The goal remains the same: raise cell power beyond today’s 3.5-V maximum, as high as the company’s nanostructure phosphate chemistry will permit, while working toward cell life beyond 10 years and 5000 full-discharge cycles. If A123’s cells can deliver the power and energy of the best cobalt varieties with far less danger of a spontaneous meltdown, the company could carve out a big, profitable share in the auto components industry. A123 won’t reveal the details of its R&D, of course, but it has said it expects further significant improvements in chemistries and molecular structures.
For a company whose first products reached consumers less than a year ago, A123 is on a fast track. DeWalt cites builders who say they can finally replace corded tools, because the 36â''V line now provides all the torque they need. In fact, A123 has ignored most inquiries from potential buyers, instead focusing on existing customers and future markets. That’s a luxury few companies enjoy.
Clearly the next five years are critical. Fulop notes that at this year’s conference on advanced auto batteries, several companies declined to discuss current and future developments. ”The submarines have gone underwater and turned on their sensors,” he says cheerfully. ”Everyone’s preparing their attack.” Asked what keeps him up at night, Fulop is startled. ”The big challenges are behind us,” he says. ”We’re well capitalized. We’ve got numbers of customers. Our products are good for this application.”
Then his voice rises. ”What excites me is to finally see our batteries in actual vehicles, after so many years of work,” he says with a grin. ”That’s cloud nine.”
About the Author
JOHN VOELCKER, IEEE Spectrum’s automotive editor, has covered automotive technology for 25 years.