The Powerhouse: Inside the Invention of a Battery to Save the World

by Steve LeVine

  How did some of the best minds in batteries overlook a defect this basic? Voltage fade was deeply pernicious, Kang said. It was what Chamberlain said—if you were employing the standard measuring tools, determining a battery’s stability by checking its capacity, you would notice nothing wrong with the NMC 2.0. From cycle to cycle, you observed a stable composition. That is what Thackeray and Johnson saw and reported in their invention. Voltage fade became conspicuous only when you incorporated gauges of stability that, while familiar in industry, were highly uncommon in research labs. Only then did you understand that NMC 2.0 was profoundly flawed.

  • • •

  Peter Faguy, a bearded and garrulous Department of Energy official, dropped in on the battery guys. He was one of the officials who had met with Sun-Ho Kang. Now he sat with Chamberlain and Thackeray in a small conference room at Argonne.

  “You’ve gotta put a big team on this,” Faguy said. He meant voltage fade. “We’ve got to sort this out in three or six months.” There was no time to waste.

  The Argonne guys stared at Faguy. They agreed more or less with his assessment, but the deadline he suggested was “laughable.” “It just didn’t seem to be realistic,” Thackeray said later. They did not challenge him, but once he was gone, they quickly forgot about the directive.

  Considering the egregiousness of the flaw, the battery guys—at least on the surface—seemed oddly relaxed about it. It was the old divide—once an invention was out of the laboratory, especially if it was already licensed, the scientists were on to the next project. Thackeray appeared unconcerned about the reputation of his signature material. He showed no sign of turmoil as he worked with a postdoctoral assistant to try to fix the fade.

  Months passed. Chamberlain and Tony Burrell, his new deputy, were summoned to Washington.

  “Solve this friggin’ problem,” said David Howell, Faguy’s boss and head of battery development at the Department of Energy.

  Chamberlain said he had not been entirely ignoring Faguy’s prior instructions. He had a plan to attack voltage fade, he said. He would put together a team at Argonne and its members would concentrate solely on the problem with NMC 2.0. It would cost a few million dollars, split between the Department of Energy and one or more of the licensing companies. He believed it stood a good chance of working.

  “You have plenty of money in your budget,” Howell said. “Use it.” He seemed exasperated. Chamberlain should repurpose his budget into a program that made good on NMC 2.0 by solving voltage fade. He was not allocating another dime.

  Later, speaking privately with Howell, Chamberlain asked whether he had been pressured by the companies to come down so hard. Howell said that, no, he hadn’t. Chamberlain saw that Howell was worried about the prospect of corporate disenchantment. He wanted to avert an angry call from GM or anyone else. Given the administration’s agenda and the hoopla over the NMC licenses, Howell might not have been surprised if the president himself rang.

  If Argonne solved the problem, there would never be such calls. If there ever were, the next best thing to a solution would be to say that he had dispatched all his troops on a monumental effort to find an answer.

  A month later, Burrell called Howell with the Argonne plan. The lab would reshape the entire Battery Department around the effort. Almost every researcher in the department would be reassigned to the problem, and other specialists brought in from other parts of the lab as well. They would plumb why voltage fade was happening and seek a solution. The effort would cost $4 million a year, all of it taken from existing Argonne programs funded by the Department of Energy. Argonne anticipated a slog—the answer would not come in weeks or even months. It could take three or four years of work. The effort might not succeed, and if failure seemed likely, they would pull the plug early.

  Faguy was not satisfied. You were building a house of bricks and you did not lay down a new brick until the previous one was settled in place. Voltage fade was a loose brick in the edifice of the electric car. He thought that NMC 2.0 could underpin big new industries in an otherwise laggard American economy. Car companies around the world felt it had the best energy density, capacity, and voltage of any material currently within reach. “This is the best chance that we have of supporting U.S. industry for the next five years,” Faguy said. But only if it actually worked as advertised. The Department of Energy—Argonne—“has an obligation” to fix it. He tried to sound undemanding, but Burrell’s potential four-year time line was too long. He said, “We don’t have an expectation that it will be solved in the next few months,” but “we also certainly don’t expect to put money into it in the next three years.” The middle ground was eighteen months—by then he wanted either a solution or, at the very least, a description of “the foundation of a solution.”

  Voltage fade was now one of the field’s defining challenges.

  27

  An Engineering Solution

  The wall of Envia’s conference room was decorated with a photograph of piles of lithium carbonate in a field, white pyramids of powder that resembled cocaine. A plaque marked the 2009 collaboration between Envia and Argonne on the NMC. In a corner were Envia shoulder bags and on the table a selection of the company’s cylindrical and flat batteries. On the wall, a plaque denoted “Envia’s Values,” which included “problem solving, not just program management” and “develop products, not just patents.”

  Kumar stood with three scientists. “This is the core R&D group,” he said. Chad Bowling, a blond and serious Louisville, Kentucky, native, had graduated from Stanford with his bachelor’s in chemical engineering two years earlier, in 2009. Stanford engineering graduates were typically snapped up by the high-tech industry, but with the economy in recession, Bowling couldn’t find a job. He was subsisting on dinners of bean burritos and had started applying to restaurants to wait tables. Then a former professor recommended him to Kumar. Envia was seeking to hire a Ph.D. scientist. After speaking with Bowling, Kumar agreed to try him out as an intern. Six months later, he promoted Bowling to full-time technician.

  “We call him ‘Doctor Chad,’” Kumar said. Normally one required a doctorate for the work that Bowling was carrying out, but “Chad is getting his Ph.D. at the Envia School.”

  “Right, exactly,” Bowling said. His voice was firm, his manner still.

  Bowling worked under Mani Venkatachalam, a physicist and materials scientist from the southern Indian region of Tamil Nadu. Venkatachalam was Kumar’s chief electrochemist. Kumar directed him to “go out and be ahead of everybody else—get the highest capacity, the highest power, maximize the voltage. Don’t worry about the customers. Do what you want to do.”

  One of their aims was to navigate voltage fade. Like everyone else, Kumar had a blind spot until Riley called. The A123 man was right—it needed to be brought under control. Kumar was secretive about Envia’s precise approach to fade—how precisely Bowling and Venkatachalam, the company’s voltage fade team, were progressing. You could see that Bowling would make electrodes, laminates, and coin-size test cells and take the results to Venkatachalam and the rest of the research group, whose members would remark on his data. They would speak of “nano-coatings” and an element or two of which they were made. But because of the value should they succeed, none would discuss details such as the specific quantities of the ingredients that went into the concoction. No one working on the problem anywhere was speaking that openly of their work.

  Some two dozen men filed into the white, spare conference room. All toted laptops, which they put down, open, on the long rectangular table, and sat. Kumar handed over the meeting to Bowling.

  “I just got these numbers like an hour ago,” Bowling said, displaying his slide deck on the screen. “With the normal process, I mixed for five hours, then put on an oxide coating where we would mix two hours, over seven hours in all. Then do it again for the next composition.”

  “So this is not fired?
” one of the scientists asked.

  “No, it just aims to establish one layer, and then another layer,” Bowling replied.

  A compound doped with aluminum curtailed voltage fade best, he said. The question was in what proportions and at what thickness. He was using a scanning electron microscope to scrutinize the aluminum impurities and determine their optimal size.

  “The ten-K magnification makes it pretty clear,” he said.

  “Nice fibers,” Kumar replied.

  Bowling flipped through more slides and revealed that when he applied thicker layers of the coating to the electrode, voltage fade seemed almost to flatline. Kumar’s voice revealed increased excitement. “This is kind of a breakthrough,” he said.

  Bowling replied evenly, “This is definitely the most drastic of all the average voltage improvements. And it’s showing no sign of a top in terms of improvement. I mean, as high as we’ve gone, it’s continuously improving.”

  “We can go more,” Kumar said. He sat back in his seat. Everyone was silent.

  “The whole world is here,” Kumar said, pointing to one section of the latest slide. “They have more than ten percent fade. Now we have solved the issue by at least eight percent. And we’re going to solve the last two percent, right? The last two or three percent?”

  “The last one percent now,” someone said.

  “That’s good. It’s really nice,” Kumar said. “So we are very, very close to solving this problem.”

  After the researchers filed out, Kumar said, “Chad has one of the best experimental hands I have seen. There are lots of top scientists trying to solve this problem. We’re ahead of everybody.” He said that once one of the samples showed an absence of voltage fade, “we will know we have solved the problem, and we can go back and say, ‘Okay, what really solved the problem?’ and then science is done.”

  He meant that you could tinker in order to reach a favorable outcome, but tinkering didn’t explain why the atoms were behaving a different way. When you went further and discovered the reason, you added to the field. The answer probably did not matter to any of their customers, but it was important to be able to explain the result, if only to themselves.

  This was Kumar seeking an engineered solution to voltage fade. For a start-up, and really any private company, there often was not the time to figure out the atomics of a problem and root it out—what Argonne was attempting. You instead used your instincts to circumvent a difficulty and get a system to work as well as possible. The result would fall short of ideal. But it would be sufficient to start to sell Envia’s version of NMC 2.0.

  • • •

  Tempers were fraying over NMC 2.0. Envia had won contracts to collaborate with Argonne and Berkeley along with a couple of universities on the battery challenges, and the Department of Energy was closely watching its adaptations of NMC 2.0. But Kumar also had to carefully compartmentalize his work so as not to compromise the deal with GM or the other carmakers with whom he was conducting research. A few months earlier, a Department of Energy official had remarked to Kumar, “You guys should bring a solution in sooner rather than later. You owe us a response.” Kumar had responded, “My guys are already going fast. I want them to work without any more pressure.” Then he had told his men, “Don’t worry about any deadline, any time line. Do whatever you feel like. Let’s try this crazy idea, that crazy idea.” Envia was a start-up, not a national lab. The priorities differed.

  Kumar thought back a year earlier, when he first began attacking the problem. His team hashed over the different theories as to why the material was unstable when you juiced the voltage. Was the change happening deep within the material, or on the surface? If the change was in the interior, you could dope the material to try to trigger a different reaction. If it was on the surface, you could paint on a coating to try to alter how it reacted in contact with the electrolyte. Initial data pointed to the surface, so the team started a list of every possible variety of nano-coating that might create stability. Some of the coatings would inadvertently prevent the lithium from shuttling properly. But then Bowling and Venkatachalam combined two different coatings and found that, regardless how thick they painted the cathode, the lithium moved just fine.

  Kumar understood that the answer was a composite of coatings. But he still didn’t know what the composite was arresting or why it succeeded in doing so. He didn’t have the instruments that could figure it out. Lawrence Berkeley National Laboratory, twenty minutes north of Envia, had a powerful transmission electron microscope that could peer deeply into the material at high resolution. The lab charged $400 an hour to use it. Kumar would be studying numerous samples, at $5,000 to $10,000 each. He imagined the bill might run over $100,000, too much merely to “do science,” he said.

  Alternatively, he could write a proposal to persuade LBL to do the work out of its federal funding. But in that case, Kumar would have to publish his findings. That was the law—keeping the results secret would be unfair because the public was paying for the work.

  Since Kumar sought to keep Envia’s intellectual know-how confidential, the best solution was to pay. But that violated Kumar’s and Kapadia’s tenets of hoarding cash.

  Kumar approached a friend who worked at LBL. “I want to bring you some of my material, to analyze it, but take all the material back and keep my IP,” Kumar said, adding that he did not want to pay. “How can I do that?”

  He couldn’t.

  • • •

  Though they said otherwise, Thackeray and his team at Argonne seemed bothered that Envia—a start-up licensee—might find the solution to the voltage fade conundrum first. The industry talk was that Envia was on the verge of an answer. Outwardly, all of them said they wished Envia well. Its success would be the same as Argonne itself triumphing. But an Envia victory could be viewed as a mark against Argonne, a sign that it couldn’t clean up its own mess.

  Faguy thought Envia was not alone in preparing to settle for an engineering solution. Short of actually solving the fade, the other NMC commercial licensees also seemed to be looking for a work-around that made the material usable. Even if it didn’t perform as promised, it would still improve on current chemistries. Having paid for the NMC, they would find a use for it in smart phones and other electronics. At that stage, “the case will be closed for them,” he said.

  But as far as he and Howell were concerned, that wasn’t good enough. The Department of Energy wanted “a bedrock understanding of what the hell is going on” within the cathode, along with “the whole gamut of possible solutions.”

  Further in the future, Faguy saw the problem as a dress rehearsal for nightmares to come. The battery race would involve a series of unforeseen, terrible problems that you simply could not recognize in the tiny volumes and coin cells produced in the national labs. You needed a ton of the material and hundreds of cells, and you had to charge and recharge them again and again before the problems surfaced. Only then could you think about the solutions necessary to get the technology into a car. Voltage fade was a test run for how the Department of Energy—how Argonne—would resolve the crises to come.

  28

  Going Deep on the Fade

  Thackeray went to California to see what precisely Kumar was up to, taking along his postdoctoral assistant, Jason Croy. Just how close was Kumar to solving voltage fade?

  Croy, a slight, thirty-seven-year-old physicist with cropped blond hair and a frequent smile, had finished graduate school just a year earlier. He grew up in Frankton, an Indiana town of 1,800 people, and came late to physics. For nine years after high school, he, his older brother Johnny, and three friends toured the Midwest in a heavy-metal band called Connecticut Yankee. Croy sang lead and Johnny played bass on two records of original music, including a song called “F=G(M1M2/D2),” Newton’s equation for gravity. Science was always there. He built a telescope and took an astronomy course at Ball State. One day, a lecturer descri
bed the possible use of plasma physics to understand nuclear fusion, the process that powers the sun in which two nuclei combine, releasing energy, and Croy “thought it was the greatest thing I had ever seen.” He enrolled full-time and went on to earn a Ph.D. from the University of Central Florida. Croy’s aim was “something that would really have an impact.” In 2011, Thackeray advertised a postdoctoral position on Argonne’s Advanced Photon Source, the 3,600-foot electromagnetic loop that produces intense X-rays that researchers used to examine the materials they were creating. They called it the “beam line.” When they were working on the beam line, they called it “beam time.” It was a highly prized tool, and a coveted job working with it, because of the deep images it produced of the atomic structures of their work. Though a decade older than most other applicants, Croy won the position.

  Croy’s years on the road left him uninhibited and quietly commanding. Thackeray was a powerful public speaker, but after watching Croy’s delivery a couple of times, he seemed almost to prefer that the younger man present their findings. In the Envia conference room, Kumar introduced the Argonne men to his assembled researchers and said that without Thackeray, Envia would not exist. The Envia guys were visibly elated to be near the inventor of the NMC. Now they all turned to Croy, who was standing. Croy said the slides assumed two ways to understand voltage fade: it was either repairable or forever unmanageable, the latter because of the immutable laws of thermodynamics, the most basic physics of energy. The answer, he said, was actually both—voltage fade challenged the limits of fundamental physics, but there could be a fix. To get there, he and Thackeray had used the beam line to explore the bowels of the NMC.