The Second Kind of Impossible

by Paul Steinhardt

  But by this time, the manuscript had been sent to the printers. The journal declined to perform a partial edit and refused to remove the acknowledgments. The only alternative the editors offered was to scrub the article altogether.

  So I had arrived at the tipping point. Pulling the article at the last moment would be a consequential decision. Withdrawing from publication is usually interpreted as a sign of trouble, and bad news has a way of spreading quickly within the scientific community. I knew stopping publication would raise suspicions and threaten the future credibility of the project.

  As I wrestled with the final decision, I revisited the events of the last five months. More than any scientific venture I had ever been involved with, the natural quasicrystal investigation that Luca and I had pursued had been an unpredictable affair with an inordinate amount of gut-wrenching twists and turns. I could certainly appreciate that Lincoln and Glenn were accustomed to more disciplined investigations.

  I could also understand why the red team was suspicious of fakery. In their view, metallic aluminum was simply impossible in the natural world. But in my opinion, the endless hours of laboratory work in Florence, D.C., and Princeton had substantially eliminated every idea about how the sample could have been artificially produced. A mineral of natural origin, even if we could not explain all the details surrounding it, stood as the most plausible explanation.

  Every time I went over the arguments, I came to the same conclusion: We should proceed with the paper. It was responsibly written, and the cost of withdrawing it from publication had become much too high. There was no doubt in my mind that I was making the right decision. I firmly overruled the objections of the red team.

  Our paper, titled “Natural Quasicrystals,” was published in Science magazine on June 5, 2009. It was coauthored by Luca Bindi, Nan Yao, Peter Lu, and me with acknowledgments to Lincoln Hollister and Glenn MacPherson.

  The aftermath was bittersweet. On the one hand, the Science magazine article received a substantial amount of international attention. There was no sign of any skepticism, which was surprising given the controversial nature of what we were reporting. Luca and I were ebullient. At the same time, I could not help noticing that the red team had suddenly gone silent. Lincoln and Glenn gave us no indication that they were aware of the publication and chose not to comment in any of the many follow-up articles that were published in multiple outlets all over the world.

  * * *

  FLORENCE, JULY 3, 2009: About a month after the publication of our discovery of the first-ever natural quasicrystal, and two years after we started working together, Luca and I finally had the chance to meet in person. I was able to visit him in Florence during a trip to deliver a series of lectures in Europe on a different topic.

  We greeted each other with a warm hug. After hundreds of daily Skype chats about everything from scientific problems to family issues, meeting Luca felt like a reunion with an old friend. It was clear that the intensity of our working relationship had forged a strong bond between us. Luca was taller and even more athletic than I had imagined. The bubbly enthusiasm and sincere warmth he managed to convey over the Internet were all the more evident in person. The two of us are from different cultures, different generations, and different scientific backgrounds. But we had discovered ourselves to be truly kindred spirits.

  Luca took me on a brief tour of the university museum and proudly showed me the beautiful new mineral displays he had designed. We settled down to work in his office, and spoke nonstop for several hours about the state of the investigation and what should be done next.

  Although the Science article was now successfully published, we were both keenly aware that we had fallen short of the high expectations of the red team. It made little difference that the scientific journals, the general press, and the readership had accepted our result. We would not be satisfied until one of two things happened: We either had to convince Lincoln and Glenn that we were right, or they had to convince us we were wrong. So that meant our investigation would continue. Luca and I agreed we would sift through more grains, look for more clues, conduct more experiments on the grains we had already studied, read more about aluminum alloys, and explore more theories about how the Florence sample may have formed.

  I admitted to Luca that I had anticipated that the Science article would help prime the pump. I had hoped, for example, that it might motivate geologists from around the world to check their own mineral collections for natural quasicrystals or, better yet, send us their samples to study. But to my disappointment, no one reached out to us. We would need to do the work ourselves.

  Our intense discussion continued through lunch, until we reluctantly had to go our separate ways. I affectionately hugged Luca goodbye and left Florence with an even greater appreciation and admiration for my Italian colleague.

  I had begun to sense a growing coldness from Lincoln and Glenn in our emails, which had now become sporadic. And as it turned out, my instincts were correct. A few weeks after I returned home from Europe, Lincoln expressed his frustration to me in a stern note:

  I believe that the sample you have been working with is not natural. I feel I am up against a wall of diminishing returns to determine its origin.

  Lincoln explained that he did not want to continue working with us unless we could somehow find a completely fresh sample from some other source. Glenn’s withdrawal from the project was implicit.

  I felt sad and discouraged as I read Lincoln’s email. It was certainly true that the red and blue teams had experienced their fair share of disagreements. But like all good scientific disagreements, our debates had always remained civil and had never devolved into personal attacks. I always believed that Lincoln and Glenn were crucial to our investigation so I was determined to keep them involved by one means or another. The fact that we were currently weathering a difference of opinion did not alter my respect for them one iota.

  But where could the next breakthrough come from? And in the absence of any imminent news, what could I possibly do to bring Lincoln and Glenn back on board?

  THIRTEEN

  * * *

  THE SECRET SECRET DIARY

  FLORENCE, SEPTEMBER 2009: It had been nearly three months since our Science article had announced our discovery. A full summer’s worth of investigation had passed and no one on our team had anything remotely interesting to report. Luca, Nan Yao, and I had spent long hours toiling away in the lab but were not making any progress.

  Then, just when it seemed that our project might grind to a permanent halt, the most unexpected thing happened. It did not surface in the lab. Or at a conference. Or through an exchange with other scientists. The catalyst was wine and pasta.

  Luca was enjoying dinner in Florence with his sister Monica and her friend Roberto and was regaling them with the most dramatic highlights of our story. By now, it was a long tale: The invaluable sample of khatyrkite Luca found tucked away in his museum’s storage room, the unexpected discovery of a natural quasicrystal in the Princeton lab with Nan Yao, the embarrassingly fake samples we discovered in private collections, the untouchable holotype locked away in a St. Petersburg museum, the untrustworthy Russian scientist we tracked down in Israel, the inexplicable mix-up with the famous Allende meteorite, along with endless rounds of inconclusive testing and debate.

  Luca explained that we had traced the original khatyrkite sample in the Florence museum back to a Dutch mineral collector. But unfortunately for us, the trail ran cold in Amsterdam. Luca’s dinner companion, Roberto, lived in Amsterdam and was intrigued by that detail. He nodded when he heard the collector’s last name was Koekkoek. It was not so surprising that he was hard to locate, according to Roberto, because that was a relatively common last name. In fact, Roberto had a neighbor named Koekkoek. The elderly woman lived just down the street, and he often helped carry her packages home from the grocery store. He promised Luca he would ask for her advice.

  By now, Luca and I had already spent months scouring Amst
erdam. There was little chance, Luca figured, that a random acquaintance of Roberto’s, someone with no connection to our story, could be helpful. But he was wrong.

  Within twenty-four hours, Roberto had returned to Amsterdam and dashed off an email to Luca. Not only did his neighbor know of Nico Koekkoek, the two of them had been very well acquainted. She was, in fact, his widow!

  The unexpected news hit us like a thunderbolt. Luca immediately bought a ticket for the two-hour flight to Amsterdam and shot me a quick email to say he was going to try to interview the elderly woman as soon as possible.

  “I feel myself as a CIA agent,” Luca wrote.

  * * *

  AMSTERDAM, SEPTEMBER 2009: Luca swooped into Amsterdam the next day with high expectations. He excitedly headed for Roberto’s neighborhood and the Koekkoek apartment, where, to his surprise, he and Roberto were abruptly stopped in their tracks by a stubborn brick wall named Debora Koekkoek. The eighty-year-old was apparently taken aback by the unsolicited visit and, to Luca’s dismay, categorically refused to cooperate. She was unwilling to share any of her family’s private information with an unknown Italian, no matter how charming or how persuasive he may be.

  To his credit, Roberto tried his best to salvage the situation. The only choice, he determined, was for Luca to disappear from sight so that he could try to speak privately with his neighbor. Luca reluctantly agreed, and waited sullenly at a nearby café.

  How could Roberto be expected to discover anything useful, Luca wondered. Two days earlier, he had never even heard about our quest.

  As Luca steamed, the discussion between Roberto and Debora turned into a battle of wills. Whenever Roberto asked Debora about her husband’s collection, she insisted that she knew virtually nothing about it. She was willing to admit that her late husband had traded in minerals and seashells. She also knew that he had liquidated everything in his mineral inventory in 1990 in order to focus exclusively on collecting seashells, which he found more appealing. That was it. That was all she knew. End of story. No matter how many different ways Roberto tried to ask about the mineral collection, Debora pleaded ignorance.

  Finally, perhaps because Roberto said something that jogged her memory or perhaps to put an end his persistent questions, Debora shyly volunteered a crucial piece of new information. Even though her husband had sold off his mineral collection, she told Roberto, he had never disposed of the secret diary in which he kept a record of his purchases. And she still had it.

  After some gentle persuasion, Debora agreed to let Roberto peek at the secret diary. Sure enough, he quickly found an entry for khatyrkite, which Nico Koekkoek simply described as “ore from Russia.” He had also dutifully recorded that he had obtained the sample on a trip to Romania.

  The entry further explained that Koekkoek purchased the sample in Romania in 1987 from someone named Tim. There was no last name or contact information.

  A mineral dealer named Tim? In Romania? Tim the Romanian?

  Roberto jotted down some notes, said his goodbyes to Debora, and conveyed the news to Luca, who passed it on to me. Luca and I surmised that Tim was most likely a mineral smuggler. He and Koekkoek were apparently doing business together in the late 1980s, a time when Romania was still in the grip of a communist dictator and considered a Soviet satellite. Sneaking natural minerals out from behind the Iron Curtain at that time would probably have been considered a serious offense.

  * * *

  PRINCETON AND ROMANIA, OCTOBER 2009: The next step was sure to be simple, I thought. Compared to finding Leonid Razin in Israel or the widow of a Dutch mineral dealer in Amsterdam, tracking down Tim the Romanian was going to be a snap.

  After all, I thought, how many smugglers could there be in Romania named Tim?

  My optimism was ill-founded. Although we sent out an all-points bulletin to contacts in Romania and collectors worldwide, it seemed that no one had ever heard of Tim the Romanian.

  As we continued to widen our search for Tim, there was a glimmer of hope on a different front.

  One of the problems we had been grappling with from the onset of the investigation was that we had only two tiny specks of grains to study and an unusually limited amount of information about the rock they came from. There was one highly magnified image of a single slice of the original sample, which showed a complex configuration of aluminum-copper alloys and silicate minerals. But after taking that image, Luca had pulverized the slices in order to try to extract the specks that he sent to Princeton for me to examine. And of course, those were the specks that turned out to include the first-ever natural quasicrystal.

  Lincoln Hollister was constantly complaining that there was only one image to study. Whenever we would meet, he would repeatedly emphasize that Luca had made a big mistake by pulverizing the Florence sample, especially without first having taken a more extensive series of images at different magnifications with his electron microscope. The images could have shown that the quasicrystal and other aluminum-copper alloys were enmeshed with, or had multiple contacts with, silicate minerals known to be natural. By identifying the contacts and perhaps finding examples where the metal and silicates reacted chemically with one another, we would have strong evidence that the quasicrystals were natural, too. Unfortunately, after everything was pulverized, the grains were too small to provide convincing evidence one way or the other.

  The criticism was particularly difficult for me to accept because I was painfully aware that it was undeserved. The truth was that Luca had taken a full series of electron microscope images. The problem was that they had been lost. After Luca took the appropriate series of images, his Florence laboratory was hit with a crippling one-two punch—the electron microscope broke and the hard drive crashed beyond repair. Luca’s lab proceeded to replace the microscope and the hard drive. The remains of the broken equipment were unceremoniously shoved into a corner and abandoned. Luca’s photos were lost on the mangled hard drive.

  Luca had immediately told me about the disastrous events and was understandably upset. His greatest fear was that the problems would give Lincoln and Glenn the impression that his laboratory was substandard and amateurish. Luca thought explaining the truth, that he was the victim of a random mechanical breakdown, would sound pathetic, along the lines of “the dog ate my homework.” So he swore me to silence, having decided it would be better for him to suffer a barrage of criticism from the red team rather than to offer a lame-sounding excuse.

  I respected Luca’s decision, but suggested we make a quiet effort to see if a data recovery specialist could reclaim any of the lost images. Unfortunately, the expert we consulted did not offer much hope. He thought there was little chance of success because significant portions of the hard drive had been irreparably damaged by the crash. Disheartened once again, the two of us turned our attention back to the main investigation and forgot all about the salvage effort.

  Months later, just as the search for Tim the Romanian mineral smuggler was reaching its depressing conclusion, Luca received an unexpected message from the computer wizards. They had somehow managed to recover a handful of images from the crippled hard drive.

  The good news about the recovered images was a convenient ice-breaker for me to use to try to reestablish rapport with Lincoln and Glenn. They were surprised to hear the full truth about Luca’s secret equipment failure. And they were fascinated by the set of recovered images, one of which is shown below.

  Lincoln and Glenn were probably expecting clear evidence that the aluminum-copper alloys were artificial, just as they always suspected. But instead, as Glenn noted in an email, the images were a bit of a surprise:

  There is vastly more complexity here than seen in any of previous photos. The term ‘dog’s breakfast’ applies!

  Glenn has a penchant for picturesque language. And from that point on, “dog’s breakfast” was adopted as part of our team’s internal jargon. The phrase itself is British slang for a meal so badly ruined that it would only appeal to the four-legged m
ember of the family. That, of course, implies quite a mess indeed, given that traditional British fare includes such things as blood pudding and jellied eels.

  In this instance, Glenn was trying to convey his impression that the images were a jumble that was hard to interpret. They were unlike anything he had ever studied before. But he did not point out anything in the images that caused him to reconsider his opinion that the Florence sample was slag.

  Luca and I, on the other hand, detected several significant features that Glenn appeared to have missed. First of all, slag typically contains certain telltale features, such as bubbles or bits and pieces of other common industrial materials. Yet none of this was visible in the recovered images.

  Secondly, the interfaces between the metal seen in the light-colored materials, and the silicate seen in the darker materials, included some straight edges. The surrounding silicate, containing mixtures of silicon, oxygen, and other components, was also crystalline. This configuration between two minerals could only occur if both materials were first entirely melted into a liquid mixture and then slowly cooled.

  We knew from standard tables used by engineers and geoscientists that the cooling silicates would have been the first to crystallize, at about 1500° C. The copper-aluminum alloys would have crystallized later, at about 1000° C. That provided us with quantitative information about the high temperatures the dog’s breakfast had been subjected to.