Spy Schools

by Daniel Golden

  Exasperated, Smith finally asked Nathan Kundtz, another member of the team, to figure out the method. Two weeks later, “I had the tool, and presented it to the group,” Kundtz recalled. “I was not thinking, I’ll show Ruopeng. I was blissfully ignorant that anyone would care.”

  One listener did care. As Kundtz spoke, Liu “sank in his chair, looked incredibly angry, and was silent,” Smith said. Liu was not only chagrined that a colleague could replicate his technique and frustrated that he couldn’t keep it for himself, but also brooding about a shift in the group’s pecking order. Kundtz, a graduate student in physics who had joined the lab in January 2008, was eclipsing him as the star student.

  When Kundtz came to the lab, he told me, Liu was “the most charismatic” member, and published the most. “He had his own de facto subgroup, and he would run the meetings like a professor.” Initially the two were “very friendly,” Kundtz said. But tensions developed. Kundtz was just as ambitious, stubborn, and aggressive as Liu, and a more painstaking researcher.

  Liu began sulking, and his research contributions diminished. “Certainly Ruopeng was unhappy that David was taking a liking to Nathan and spending more time with him,” Gollub said.

  Smith, in fact, was increasingly disenchanted with Liu’s grandiose approach to science. Liu had developed what he called a theory of metamaterials and was determined to publish it, but Smith wouldn’t let him, because the analysis was flawed. “As time progressed, his theory fell apart,” and Liu couldn’t correct it, Smith said. Finally, Smith fixed it himself in a published article.

  “Ruopeng would often find interesting things,” Smith said. “The theory wasn’t right. He’d present it, and look so confident, that a lot of people thought he was brilliant. It was hard to tell in half an hour that it was all nonsense.”

  Soon after Kundtz unveiled Liu and Ji’s invisibility cloak technique, several postdoctoral students arranged a confidential lunch with Smith. They complained that Kundtz was stealing ideas and urged Smith to expel him from the group.

  One of them, Aloyse Degiron, told me that their demand had nothing to do with Kundtz reconstructing the cloak technique. He said Kundtz’s abrasiveness had antagonized his colleagues: “It was very difficult to work with Nathan. He has this very strong personality. Clearly he had a very high opinion of himself. If someone said something stupid, he wouldn’t hesitate to say it was stupid.”

  However, Gollub said that the postdocs felt “a certain sympathy” for Liu. “People felt some of these projects that were originally pushed along by Ruopeng were getting unfairly shifted to Nathan.”

  Smith believes that Liu incited the mutiny: “He was so mad about Nathan spilling the beans on his theory.” When Smith went around the table, asking each postdoc what ideas Kundtz had stolen, “one after the other of the postdocs said they didn’t know firsthand but heard it from someone else,” he recalled. He told them that they should blame him instead, because he had assigned Kundtz to get a handle on the projects.

  “A bunch of smart people were allowing themselves to be manipulated, and I was pretty irritated,” Smith said.

  * * *

  THE JANUARY 16, 2009, publication of “Broadband Ground-Plane Cloak” in Science, with Liu and Ji as lead authors and Smith and Cui among the coauthors, should have been a triumph for Smith’s group. It drew wide attention, including a joke in Jay Leno’s Tonight Show monologue: “Scientists are developing a cloak that bends light to make you invisible—like being an actress over forty in Hollywood.” Bada boom.

  Smith heralded the advance in a Duke publication. “The difference between the original device and the latest model is like night and day,” he announced. “The new device can cloak a much wider spectrum of waves—nearly limitless—and will scale far more easily to infrared and visible light.”

  For Liu, the article generated both prestige and profit. As another incentive for technological achievement, China’s government rewards Chinese scientists for bylines in premier journals. It paid Liu more than ten thousand dollars, according to Smith.

  But the celebration was short-lived—and all because of a footnote. It cited funders for the research: not only Smith’s patrons, such as the Air Force Office of Scientific Research and defense contractor Raytheon, but also Cui’s, which included the National Basic Science Foundation of China, the National Basic Research Program of China, and Project 111. In late January, Smith’s project manager at the Air Force sent him “a very harsh email” telling him that the Pentagon wanted to know why he was taking money from China. Presumably the Pentagon had invested in Smith’s research because it hoped to mask a fighter jet or other weapons from enemy radar. It didn’t want China, a potential adversary, to have the same capacity.

  Alarmed, Smith pulled out of the collaboration with Cui’s group and Project 111. “It all just fell apart after our sponsors became so concerned,” he said. Smith told Liu not to work with Cui’s group any longer, and to distinguish clearly in his dissertation between his own research and that of collaborators at Duke and in China.

  * * *

  AS LIU PLUGGED away at his doctoral dissertation on “Designing and Building Microwave Metamaterials” in 2009, he began preparing for his career after Duke. He was torn between two paths: academia and business.

  He had a knack for entrepreneurship. One night in the lab, he talked with a colleague about technology opportunities in Africa, such as manufacturing cheap cell phones. But he could also see himself as a professor, guiding his own group’s research. “Ruopeng wanted to do something innovative and real, and also to be a business success,” Da Huang, his friend from Zhejiang University and Smith’s group, told me.

  Armed with recommendations from Pendry and Smith, Liu applied to top U.S. universities. “I believe that Ruopeng did make a positive scientific contribution and one or two of his ideas I would describe as creative,” Pendry told me. “I wrote positive references for him.” He added that their “limited interactions were entirely scientific … although I certainly heard the rumours connecting espionage.”

  The Massachusetts Institute of Technology, which had an opening for an assistant professor, invited Liu to give a presentation. Smith agreed to help him by glancing over it beforehand. On April Fool’s Day 2009, over lunch at a restaurant on the Duke campus, Liu ran through the slides for his job talk on his laptop. One of them caught Smith’s eye. It depicted an elaborate website, complete with contact phone numbers and email addresses, the details of Liu’s experiments, and an article from a Duke science publication.

  Smith pointed out that Liu needed Duke’s permission to reprint the article. Liu answered that the site wasn’t real; he had only created it for the MIT interview.

  Liu had been shirking his Duke responsibilities, both in Smith’s lab and in a course on electromagnetism, where he was supposed to be a teaching assistant but rarely showed up. “You’re doing no work for the group, no one can find you, and yet you put all this effort into a website?” Smith asked.

  Sobbing, Liu insisted that he had cobbled the site together quickly. “I came here at twenty-two years old,” he said. “I’m naïve.”

  “I was on to him,” Smith told me. “I said, ‘We both know that’s not true. You know what you’re doing.’” It turned out that the website was hosted on a server in China. It was the venture, Smith concluded, for which Liu had hoped to save the cloaking technique.

  Similarly, Liu’s commercial ambitions in China seemed to explain why he sometimes sought to delay publication of Duke research. “Ruopeng would show the material to Cui in advance of publication, and then slow down the publication,” Kundtz told me. “It was a timing game.”

  Despite those slowdowns, Liu was startlingly prolific. Scouring the Internet a few days after their lunch, Smith took a close look at Liu’s publications, which the postdoctoral fellow had warned him about in 2008. He found at least forty-three scientific articles that Liu had coauthored at Duke. A typical graduate student has between zero and
five publications; a remarkable student, perhaps a dozen. He had amassed his astonishing total largely by inspiring other students with his visionary babble, and then feeding off their ideas and data. He hadn’t told Smith about at least a dozen papers that he wrote with Cui’s group, although several listed Smith as a coauthor, exploiting his name and reputation. One of Liu’s published collaborations with the Chinese addressed a topic on which Smith had assigned him to work with a Duke student. That was a “clear violation” of academic ethics, Smith told me.

  Smith summoned Liu to his office. The professor sat behind his desk, with the offending publications stacked on his lap, hidden from Liu’s view. Then he trotted out one paper, as if it alone concerned him, and reminded Liu that they had agreed to sever ties with Cui’s group. Liu at first protested that he had told Smith about the article, but then admitted he had forgotten, just that once.

  Smith pulled out another publication, and another. Liu grew very quiet. He eventually blamed his Chinese collaborators and “made a big show of telling everyone that they shouldn’t do that again,” Smith said.

  Liu similarly misled Smith about a different sort of collaboration. After the 2008 Nanjing conference, Liu suggested publishing its proceedings, which is customary in academia. He asked Smith to compose an introduction. Smith said he was too busy to write or edit fresh material, which wasn’t needed anyway. Liu agreed and said he would write the abstracts for each presentation by Smith’s group. “There was no mention of this being a book, and I was so busy that I just ignored all of the subsequent correspondence,” Smith said. “Later, Ruopeng began to call it a book, and I’d stop him and I’d say, ‘This is just a collection of conference papers, right?’ and then he’d nod and agree.” One day, a box of books showed up at Smith’s door, and he saw himself listed as the volume’s co-editor, and coauthor of six chapters, which Liu had written. “It was filled with broken English and not very high quality,” Smith said. Liu had again exploited Smith’s fame to fulfill his ambitions and boost his reputation.

  Smith’s tolerance was exhausted. On April 21, 2009, he took away Liu’s key to the lab and told him to finish his dissertation at home. Later that year, Smith also nixed Liu’s prospects for a postdoctoral fellowship at Princeton by telling an old friend there about Liu’s rogue collaborations with Cui’s group.

  Understandably, Liu feared an even more severe blow: that Duke would deny him a doctoral degree. Duke’s graduate school standards of conduct prohibit academic dishonesty, such as “representing someone else’s work as your own.” According to a person familiar with the procedure, “There usually has to be a panel assembled to evaluate the evidence if formal charges are brought. That takes weeks and weeks to get done. It would have delayed his graduation certainly.”

  Instead, Liu defended his dissertation on November 30, 2009, without incident. The discussion was purely technical: Smith didn’t even tell the other professors on the committee about Liu’s pipeline to China. “It would have been very difficult at the time to show that he had done anything unambiguously wrong,” Smith told me. The website in China, for example, could be “excused by a Chinese student not really understanding the rules well enough,” or even praised as entrepreneurial zeal. Smith discussed his suspicions with some Duke administrators and professors, but received no guidance, he said.

  Duke may have wanted to avoid a brouhaha for another reason: it was forging ahead with plans for a branch campus in China. A week after Liu’s dissertation defense, Duke trustees approved continuing negotiations with Chinese officials to build a campus in the city of Kunshan, which would supply the land and facilities for free. “Duke, like the other universities that aspire to be world class, recognize that they need to have a presence in China” if they are “going to get the best students, the best resources, the best faculty,” said Dan Blue, chairman of Duke’s board. It would have been an awkward time to punish a Chinese researcher already renowned in his homeland for his prestigious Science article.

  Liu received his doctorate on December 30, 2009. Like Princeton, MIT had turned him down, but China was beckoning. His mentor, Cui, had ties to a brain-gain program, and Liu could bring Duke’s metamaterials research and some of its most promising China-born scientists back with him. He likely could have named his price. He told a friend over dinner that, through one of its ventures to lure overseas talent, the government offered him as much as $100 million to start a metamaterials center. “They would pay for the equipment, they would hire as many engineers as he needed,” the friend said. “There were companies in China that would be in the pipeline to productize the technologies that he would invent. Before he left America, he knew of this institute that he was going to be leading.”

  Liu returned to China in January 2010 with his wife, Weizi Huang. She had been working on her own dissertation at Duke—on statistical models for identifying genetic risks for ovarian cancer—and planned to earn a doctorate in computational biology in another year. One day she told her adviser, Edwin Iversen, that she had to go back to China with her husband. She settled for a master’s degree.

  “She wasn’t happy to leave,” Iversen told me. “She wanted to complete her PhD.” Her consolation prize was to be a founding member of Kuang-Chi Science.

  * * *

  AT MY REQUEST, and armed with a series of questions that I had prepared, a Shenzhen-based freelance journalist interviewed Liu in June 2016. They met in the billionaire’s office on the second floor of Kuang-Chi’s Shenzhen headquarters. The office, which overlooks a nearby pond, is spare and functional: white walls, whiteboard, file cabinet, desk with a globe and a photo of himself, sofa, conference table with oranges in a plastic bowl, coffee table with some of his awards and a model helicopter.

  Cornering Liu wasn’t easy. He’d canceled a prior appointment after a two-hour wait, and then left for Singapore and Beijing. Back in Shenzhen, he was an hour late for the rescheduled interview, but finally materialized, smiling and affable, wearing a light gray blazer and a white shirt.

  During the ninety-minute conversation, he acknowledged that Smith had accused him of “stealing stuff” and taken away his key to the lab, that he had been “a little worried” that the university would withhold his doctorate, and that “the FBI went to Duke to investigate all the stuff after I left.” Still, he said he did nothing wrong, and “nothing was stolen,” because all of the research was basic and open, and sharing of ideas is integral to academic collaboration.

  “I’m not working in any kind of classified lab,” he said. “I worked in fundamental research and published papers and they can be seen by everyone in the world. Everyone right now can download the paper and see what I have done; it is all transparent.”

  He maintained that he started Kuang-Chi with fifty thousand dollars borrowed from his mother; China’s government did not send him to snatch American technology and did not pay him to return. The Duke research “is not all that valuable.… We cannot solve any problems in reality and it is all in academic environment and everything is transparent. Anyone can look at anything we have done and it is published, so if China and Shenzhen funded as a reward this kind of fundamental research, I think it would spend all of its money.”

  Asked about criticism from his former Duke colleagues of his scientific prowess, Liu clapped his hands together twice, and then laughed. “We’ll make things work,” he said. “Comments cannot change the world. Build the machine. Make it fly.”

  He suggested that Smith had turned against him for fear of losing research support after the Pentagon backlash. “Probably what people can do is prevent future funding,” he said. “That’s why I can understand that Dave had concerns with that, and he should have concerns with that; otherwise it would influence the funds for the whole team.”

  He recalled the argument with Smith in the Chicago airport: “Some people wanted to have a tour, and then no one would be able to chair the seminar, so I said no.… We just needed someone to chair the global academic session.
” He blamed Chunlin Ji, his coauthor from the statistics department, for starting the China-based website that featured Duke’s research, and T. J. Cui’s team for publishing articles, without Smith’s knowledge, that listed Liu and sometimes Smith as coauthors. “They wanted to put the other party, the team at Duke, because they do think that they have worked together a lot with Dave.”

  Liu was less forthcoming on other episodes that alarmed members of Smith’s lab. He had a habit of responding, “That’s interesting,” and then equivocating. For example:

  Did visitors from Cui’s lab take pictures of the equipment so they could reproduce it in China? “That’s interesting. So how could people go to the lab? Duke has to invite people, then people can get the visa, and they come to the lab, it is an academic exchange.”

  Did he mislead Smith about Project 111? “If he says I set him up for Project 111 without telling him, then how could he spend time in China? I have no right to kidnap him. Right?”

  Did he dodge Smith’s request to explain how the invisibility cloak worked because he was saving the method for his business in China? “What’s the proof? So if we do not put all the algorithms and the stuff out, how can we publish a paper?”

  Was Liu upset that Kundtz figured out the technique, and did he stir up the protest against him? “That’s interesting. Actually, when Nathan came, I was about to graduate. He wanted to work together with me on some projects but at the time I was about to graduate.… I was not concerned about who actually replaced me as a star student because I was leaving as a star student.” In fact, Kundtz joined Smith’s group twenty-three months before Liu graduated.

  Did he write the chapters for the metamaterials book, and list Smith as coauthor, after assuring the professor that it would only be a collection of conference proceedings? “That’s interesting. At that time I was still his student and wrote up the conference proceedings and put my adviser’s name.… It was basically from the conference, but also with some new type of stuff.”