The Powerhouse: Inside the Invention of a Battery to Save the World (20 page)

The exercise of disassembling a selection of all the major vehicles on the planet taught Hillebrand that almost every new automotive technology followed a long adoption curve, including disc brakes, fuel injection, and the automatic transmission. He could think of no big advance that achieved immediate acceptance. For electric cars, the largest unknown was outside the control of the inventors or manufacturers. It was oil prices. If they were comparatively moderate for a sustained period, electrics would be even more hobbled than they already were. But if prices climbed and stayed high, creating buying anxiety at the gasoline pump despite improved internal combustion efficiency, they could motivate more concentrated research on better batteries and success in a decade. “We could do it if we have to,” he said. “The problem has been that we haven’t had to.”

I
f the odds of beating gasoline were so low, how could the battery guys hope to win? To find out, Sujeet Kumar, Jeff Chamberlain, and dozens of private energy executives filed into a darkened conference room at UC Berkeley, invited by the Department of Energy for a two-day gathering just before Orlando. The morning keynote speaker was Vinod Khosla, the most aggressive clean-energy investor in Silicon Valley. The New Delhi–born Khosla had graduated from the elite Indian Institutes of Technology and gone on to cofound Sun Microsystems. In recent years, he had invested more than $1 billion of his own and investors’ money in solar power, biofuels, and batteries. He was blunt-spoken and usually dressed in black.

Khosla began by challenging the premise of gasoline’s invincibility: recent history showed that it was as vulnerable as any other incumbent technology, he said. That the odds of beating gasoline were low was precisely why he was invested in doing so—it made the potential financial gain astronomical. “Experts as a group speak knowingly of the 2008 financial crisis, but in June 2008 none predicted it,” he said. The same was true for energy. Experts spoke with great clarity as to the future of shale gas. In 2008, none forecast its arrival.

“You say that I have a hundredth of a percent shot at success?” he said. “I’ll take the odds.”

A slide appeared on the wall. “For those who can’t read it, the probability of success is on the vertical axis and the chance of destructive impact on the horizontal,” he said. “What I am saying is that when there is more than a 90 percent chance of a technology failing, that is when you tend to have the most disruptive potential.” For venture capitalists, the destruction of an established pattern of business—and hence the creation of a new way—tended to bring by far the largest profit.

That was not how most venture and angel investors worked, he said. Instead, venture capitalists sought to reduce risk to a point at which the difference between the consequences of failure and of success was incremental.

“I am suggesting the exact opposite,” Khosla said. One should welcome acute risk and the potential upside the risk offered.

Khosla knew that his message, while heard by most or all of those sitting before him, would be heeded by very few because “only unreasonable and naïve people can attempt things that are near impossible.” Insults might at least discomfit them. “Experts who can always tell you why something won’t work—they can always scare reasonable, rational people from attempting these crazy ideas,” he said.

ExxonMobil did not entirely rule out Khosla’s scenario. That was clear on page 48 of its 2040 outlook.

“Technology also can be unpredictable,” ExxonMobil’s futurists said. “A breakthrough in low-cost, large-scale storage of electricity would greatly improve the prospect for wind and solar for electricity generation. Faster-than-expected drops in battery costs would likely make electric cars more of a factor through 2040 than we expect them to be.”

In other words, the Argonne group—and any of the teams in the battery race—could confound ExxonMobil’s prognosis. The outlook was where the company and Khosla converged. His core investment principle was the oil giant’s nightmare scenario.

N
ot everyone in Orlando was dejected. Kumar, for one, was not. Neither were the Japanese present at the conference. They believed the battery industry was not so much doomed as situated in the much-dreaded “valley of death.” This was the gaping and long chasm between the completion of a product and its arrival in the marketplace. How gaping and long was unpredictable. The space of time could be months, decades, or unbridgeable. Start-up companies routinely failed during this stage, often for lack of cash or conviction. They lost to the market. It was why Amine said that many—perhaps most—of the electric pioneers were destined to this fate.

But the valley had another meaning as well, according to a number of voices at the conference, which was that those willing and able to hang on three, five, or nine more years could find themselves in a very different market.

One pathway to a seriously powerful battery was to improve the anode rather than the cathode. The anode was the staging point for the lithium. From there, the lithium shuttled to the cathode, providing the current that propelled an electric. Anodes were judged by how much lithium they could store and the rate at which it could be extracted, which was what delivered distance and acceleration. The standard was a graphite anode developed by Bell Labs back in the 1970s. Kumar was engaged in an industry-wide competition to replace it with an anode made of silicon, a metal that could absorb a much larger ratio of lithium atoms. A graphite anode absorbed one lithium atom for every six carbon atoms; but each silicon atom could accommodate four lithium atoms. Next to pure lithium metal, that made silicon the most energetic possible anode. Such an anode had the potential to deliver an order of magnitude better performance than graphite, whose discharge capacity was about 400 milliampere-hours per gram. Silicon had a theoretical capacity of
4,000
milliampere-hours per gram. You could not hope to attain that maximal peak in practice, but 1,400 or 1,600 were conceivable and, if achieved, would more than triple the graphite anode’s performance.

But silicon had a problem. For use in an automobile, you needed an anode to withstand at least 1,000 charge-discharge cycles. As you intercalated lithium into a silicon anode, it expanded tremendously. Graphite more or less maintained its shape while absorbing the lithium, but silicon blew up three or four times in size. As you charged and discharged again and again, the anode kept expanding and contracting, until it finally pulverized and killed the battery.

This was not news. The virtues of silicon had long been discussed, but no one had yet managed to resolve the expansion issue. Researchers would reach two or three dozen charge-and-discharge cycles, and the anode would break up. Everyone knew that. But Kumar wanted to try his hand. The motivation was a competition run by ARPA-E, the Department of Energy’s new funding unit for radical technologies. The grants, for ideas promising profound leaps in energy technology, ranged from $500,000 to $10 million. Considering that Envia’s entire first round of funding was just over $3 million, the sum in play was beguiling. The prestige of an ARPA-E grant could also attract yet more money and industry attention to Envia’s.

Kumar’s best chance seemed to be twinning a silicon anode with NMC 2.0. Khalil Amine had an interesting concept for silicon, so Kumar applied for the ARPA-E competition with the Argonne scientist as a partner. The submission was straightforward. It named the Amine silicon concept plus a few others that, if coupled with the NMC 2.0, could result in a 400-watt-hour-per-kilogram battery, sufficient to enable a three-hundred-mile car. That seemed to Kumar to meet ARPA-E’s requirement for a transformational breakthrough.

It was a bold proposal—the generally accepted physical limit of a lithium-ion battery using a graphite anode was 280 watt-hours per kilogram. No one had ever created a 400-watt-hour-per-kilogram battery. In all, ARPA-E received some 3,700 submissions for $150 million in awards. Thirty-seven were selected. Envia was among them—Kumar won a $4 million grant.

For the subsequent year, Kumar’s team worked through the handful of silicon anode concepts he had proposed until it settled on one. Kumar said Amine’s anode, a composite of silicon and graphene, pure carbon material the thickness of an atom, had failed to meet the necessary metrics. Instead, the best anode was made of silicon monoxide particles embedded into carbon. Kumar’s team built pores into this silicon-carbon combination measuring between 50 nanometers and 5 microns in diameter, and filled them with electrolyte. Carbon in the shape of fibers or nano-size tubes were also mixed into the anode, thus creating an electrically conductive network. The silicon’s expansion was thus redirected and absorbed. Even if the silicon broke apart immediately, the carbon fibers and tubes provided a path across which the lithium ions could pass on their way to and from the cathode.

Kumar said the results were excellent but that there was a disadvantage to working at nanoscale. This path to the better battery was expensive. You started with a vacuum reactor and a costly substrate, sometimes using platinum, a precious metal. Then you grew nanowires and nanotubes. What resulted was like pixie dust—you derived just milligrams of material each time while what was required was bulk powder. The process might decline in cost over time, but for now it could not be justified.

Perhaps there was a cheaper way. What if his team skipped the vacuum reactor and the platinum and instead employed a conventional furnace to transform a precursor of cheap silica into good-quality, ten-gram lots of powder, just enough to make coin cell samples? That would significantly reduce the cost. Kumar’s team would try.

The jerry-rigging worked. Together, the cheaper components—including Kumar’s version of NMC 2.0 on the cathode side—were delivering the milestone energy density of 400 watt-hours per kilogram.

When he heard of the success, Arun Majumdar, the director of ARPA-E, already a fan of Envia’s, was elated. He said that Kumar should now seek independent verification. This was a very big deal and Kumar would want to validate his claim through an unimpeachable expert or body.

Kumar sent the material to Crane, a respected Indiana-based evaluation division of the Naval Surface Warfare Center. Crane came back with a stamp of approval—it had put the cell through twenty-two charge-discharge cycles and confirmed the 400-watt-hour-per-kilogram energy density. Kumar had reached the milestone.

It was a considerable achievement, perhaps big enough, Majumdar said, to justify its announcement at the annual ARPA-E Summit in Washington, which was just two months away. Former president Bill Clinton and Microsoft cofounder Bill Gates would both be keynote speakers. Majumdar said he would let Kumar know his decision.

The prospect of such attention animated Kumar and his partners. The spotlight would be of incalculable promotional value considering their aspirations for an IPO. Awaiting Majumdar’s decision, Kumar traveled for a dress rehearsal at the Orlando conference.

 • • • 

Kumar walked on stage at ChampionsGate wearing a dark blue suit. Flipping through a slide deck, he said that what he was describing did not involve a typical laboratory experiment—it was not grams of material encased in a nickel-size coin cell, but a standard 45-amp-hour electric-car battery, vacuum-sealed inside a manila-envelope-size pouch. It was a breakthrough that could finally enable the long-range, affordable electric car.

Kumar recounted some facts about NMC 2.0—if you pushed the charging voltage to 4.5 volts, the blockbuster cathode, twinned with a graphite anode, delivered 280 watt-hours per kilogram, double the performance of the lithium-cobalt-oxide material contained in the audience’s smart phones and laptops. But you achieved a wallop when you swapped in the silicon-carbon anode—400 watt-hours per kilogram, which was “a world record,” Kumar said. In addition, the advance reduced the cost to just $250 per kilowatt-hour, half that of lithium-cobalt-oxide. Just two years in the future, he said, the cost would be down to $180 a kilowatt-hour.

The battery was only a prototype—he had charged and discharged it just three hundred times. Experts in the audience knew that Kumar would have to more than triple the number of cycles before the battery could be used in a car. Kumar himself would tell you that climb would be “very tough. Very complicated.” Another $4 million or $5 million in R&D spending might be required to reach all the way there. Challenges remained. But Envia was “breaking the barrier,” he said.

One by one after the presentation, the representatives of GM, Dow Chemical, and Hyundai approached Kumar. They wanted a private word.

Anderman, the conference organizer, said Kumar was exaggerating. His talk had served a purpose, he said—it had demonstrated to the shell-shocked industry that companies were still trying. He had invited Kumar to speak for precisely that reason. “But there is no breakthrough at Envia,” he said. Like everyone in the industry, Kumar was working to stop silicon from swelling, but he had not done so as yet. Neither had he achieved 400 watt-hours per kilogram. He was hyping his achievements.

Jeff Dahn, the Canadian researcher at Dalhousie University in Halifax, felt differently. One of the most original minds in batteries, Dahn was notorious for ripping into the ideas of his colleagues—publicly and usually with precision. He pointed out flaws that most battery guys, knowing how hard it was to make an advance of any type, typically kept to themselves. Dahn was with Anderman in the belief that battery scientists often cherry-picked their results in order to postulate nonexistent advances. That did not make him a pessimist—he was a true believer, confident that scientists would eventually get it right. To reach that point, they needed first to stop doctoring the results and be honest with the world and themselves. Dahn had recently delivered blunt PowerPoint presentations that, spliced with videos of exploding batteries, accused fellow scientists, including Khalil Amine, of camouflaging the risk that their inventions could catch fire. But Kumar, he said, was not a member of this group of embellishers.

Dahn was unusually complimentary to Kumar. What the Envia man unveiled was not necessarily elegant—it was really an engineering feat, packed together efficiently. But it also worked. “It looks like it can get to four hundred,” he said. “I am very familiar with the materials that he is talking about and I think it is doable.”

Dahn was acting coy. Back in 1999, the 3M Company had filed a patent application for his version of NMC just a few months after Argonne. In the subsequent years, Thackeray and Dahn bickered over the precise atomic structure of NMC—was it a saucy amalgam of nickel, cobalt, manganese, and lithium (Dahn’s position), known as “solid solution,” or a more structured composite with a discernible chemical architecture (Thackeray’s)? For the motorist, the difference seemed to be immaterial. But it could prove crucial should NMC 2.0 become part of a massively best-selling battery. Thackeray had managed to win the crucial original American patent while the 3M Company had grabbed patent rights in China, Japan, and South Korea. 3M had gone to war with pilferers of NMC—Sony, Matsushita, and Panasonic—and won. The details of the settlements were sealed, but the outcome upheld the Dahn patents. 3M had filed no suit against Argonne, but Dahn made it sound like one was possible. He said, “I think Argonne and 3M are not on the best terms.” It was only a matter of time before 3M went after GM, too. In Dahn’s view, it was his IP and not Argonne’s that was contained in the Volt. “I think that Argonne is just using this composite argument to stay outside the 3M patent,” he said.

When Dahn was younger, he could get worked up over who was tromping on his perceived turf, but he said he welcomed Kumar’s work on NMC. There was an enormous gulf between achieving three hundred and one thousand charge-discharge cycles. “That is going to take some time,” Dahn said. But for starters, three hundred cycles “look pretty darn good” if one was aiming at smart phones and laptops. Such performance would give longer life at the same cost as current batteries. “I am sure Apple would love to make the iPhone lighter and thinner,” he said.

Was his position on Envia a more positive, new Jeff Dahn?

“No. I am being realistic,” he said.

Do the math, Dahn said. The basic NMC-spinel battery in the GM Volt delivered about 100 watt-hours per kilogram. Since GM overengineered the battery to maintain a margin for error, about 37 percent of it went unused—the excess was there just in case added capacity was needed. So it was effectively running at about 66 watt-hours per kilogram. If you now doubled the capacity using the Envia formulation and slimmed down the unused capacity, you would triple your range—rather than 40 miles, the Volt would travel more than 120 miles on a single charge.

Alternatively, GM could stay with the 40-mile range and cut about $10,000 off the price of the car. “You have your choice,” Dahn said. “This is why people are fighting for higher energy and longer life. It is what it is all about.”

Dahn had questions. For example, why didn’t Kumar report more data? What happened after three hundred cycles? But he was not worried—they were familiar questions. The first lithium-ion battery—Sony’s 1991 technology—was “a piece of junk.” But since then, its performance had improved by a factor of two, making lithium-ion tower over anything that existed previously. Envia’s three hundred cycles would increase. “How long and how fast? Nobody knows,” Dahn said. “But you can bet your bottom dollar it is going to get better.”