The Telescope in the Ice

by Mark Bowen

  John Wiley compares the building of IceCube to the building of the Great Pyramids or the Panama Canal.

  * * *

  It was Bruce who came up with the idea for the “circus train.” Wotan was modular, all “plug and play,” and all the connecting pieces, even the heating plant and hose reel, were mounted on sleds. The reason they called it a circus train, says Morse, was that “when it comes to moving stuff around, setting up and moving quickly and getting the hell outa town, nothing beats a circus.” In other words, they could move quickly from one hole to the next.

  But this was also the first point of impact between the anarchism of AMANDA and the sober management that would be required to build and operate IceCube. It was a new ball game, with hundreds of millions of dollars on the table, while Bob and Bruce were quintessential start-up guys. Remember Morse saying that he loved to start new things but tended to be elsewhere by the time “the messy details” of finishing them up came around? Well this was the time for attending to details. Bob Paulos and his colleagues at the Space Science and Engineering Center imposed rigorous procedures like design reviews, testing the drill against specification before sending it to Pole, and so on. The difference in cultures is perhaps best signified by the fact that the SSEC folks didn’t see the humor in naming the drill Wotan nor the sense in referring to it as a circus train.

  “Well, hell, I mean, ‘circus train’? What does that mean, you know, really?” asks Paulos. “That’s a nice phrase, but it doesn’t mean anything.… I never called it that. And none of the people that actually built the thing called it that.”

  But it did mean something. Speed was important. In what Morse calls their “drearily competent way,” the SSEC engineers named it the Enhanced Hot Water Drill.

  The SSEC folks were quite competent, however. Paulos brought in a superb SSEC engineer named Mark Mulligan, and this team of four proceeded to design the drill.

  Bruce was still the guru and the driving force behind the new design. Morse says his “paw prints” are on every aspect of it. “What we would do was, we would just give physical reasons for Bruce’s tremendous intuition, heh, heh. That was our job in life, to interface between Bruce and the ordinary world. Bruce was a genius.” In part because he seemed disorganized, however, and in part because he wasn’t one to argue, the SSEC folks never fully understood or appreciated his contributions. Paulos says, “He wasn’t like a lot of engineers that I know and work with, right? At all.”

  It was an interesting dynamic. Bruce and Bob were used to inventing things on the fly—they were good friends, and they’d been doing this together for more than a decade—while the others weren’t as comfortable with that but knew how to control the process and keep things from getting as helter-skelter as they’d gotten in the AMANDA days. Mulligan contributed inventions of his own as well, and there is plenty of creativity involved in making a huge, complicated machine efficient, reliable, safe, and capable of running for years. Morse may not have appreciated Mulligan’s contributions as much as he should have, while Bruce, who was remarkably lacking in ego, probably straddled the divide. Bob talks about how Bruce used to react to the thick engineering reports that the SSEC engineers would produce. “He never seemed to be that interested, but there was nothing that was going on with the drilling that he didn’t know about.… At some point you’d see him … looking at one of these reports, and he’d grab some scraps of paper—everything was always sort of pencil scrawls and pieces of paper—and every point that they’d come up with, Bruce had already made the rough, pretty accurate calculation of what they were going to do and what they would need. So it was amazing.”

  It is to Paulos’s credit that he got them rowing in the same direction. “The lesson to be learned,” he says, “at least by me, is it takes a number of good people with different approaches and attitudes to make these things really work out.… You put those guys together and then shake it up, and it really worked out quite well.”

  * * *

  Meanwhile, all through what would turn out to be the long and harrowing passage to IceCube, AMANDA was very much alive. Indeed, in a scientific sense it had just been born. While the huge management and applied physics problem of designing and building IceCube would require more effort overall and eat up most of the resources, intellectual progress in actual particle astrophysics was being made on AMANDA.

  Both efforts were taking place in an academic milieu, of course, and physicists were at the helms of both. New graduate students and post-docs were joining in constantly, and you don’t get a Ph.D. in physics by working on instrumentation alone. You need to do some physics: figure out some new way of analyzing the data or look for neutrinos from, say, active galactic nuclei—whatever you and your adviser dream up. By the time AMANDA would eventually run its course, it would produce more than fifty doctoral dissertations and forty-nine refereed journal articles. (And, since it has been observed that a Ph.D. physicist generates about $1 million in economic wealth on average, through inventions, starting new businesses, and so on, AMANDA paid for itself several times over.)

  For example, during the twenty-odd months that the Swedes had been steering clear of the competition between Zeuthen and Madison, they had been focusing on a point source that might come as a surprise: our own planet. This is an area where neutrino astronomy intersects both cosmology and particle physics.

  Supersymmetry, the extension to the standard model on which the particle physics community is pinning its greatest hopes, posits a heavy sister or brother particle for each of the particles in the standard model (all of which have now been seen, thanks to the discovery of the Higgs boson). Several of the lightest of these hypothetical siblings also happen to be the most promising candidates for the unseen cold dark matter, which constitutes some 85 percent of the mass of the universe and is one of the great unsolved mysteries in cosmology.

  As a class, these candidates have been given the name Weakly Interacting Massive Particle, or WIMP, a name that may sound humorous but isn’t as highfalutin as it might also sound. It’s basically just a description of two general qualities that any dark matter candidate must have. Like neutrinos, they must be uncharged and unaffected by the strong nuclear force; they will feel only the weak force and gravity (so they should be at least as hard to detect as neutrinos). Unlike neutrinos, they must be heavy in order to exert the strong gravitational pull, affecting the shapes of spiral galaxies, for example, that has been observed by astronomers. In supersymmetry, the lightest WIMP candidate is named the neutralino.

  There shouldn’t be very many neutralinos floating around in empty space. If they exist at all, however, celestial bodies like the Sun and Earth are expected to gather them into their hearts by force of gravity. Once they’ve been brought into close proximity with each other through this process, they are expected to collide with one another every once in a while and mutually annihilate, giving birth to other particles that subsequently decay to produce high-energy neutrinos. Thus, if supersymmetry is correct—and that’s a big if—the Sun and Earth might be high-energy neutrino sources.

  Recall that one of the earliest and most convincing neutrinos detected by AMANDA was the so-called Eva event, discovered by Eva Dahlberg, a doctoral candidate in Stockholm. It was a dead-vertical muon running parallel to one of the strings in AMANDA-B10. That meant it was coming from the precise direction of the center of the Earth.

  Since AMANDA was shaped like a cylinder, taller than it was wide, and its axis was as perfectly vertical as Bruce Koci’s drill had been capable of making it, the instrument happened to be most sensitive to neutrinos coming from this direction. As it also happened, the Swedish contingent in AMANDA included several strong WIMP and dark matter theorists, so they were in an excellent position to conduct a neutralino search.

  Did Eva’s neutrino come from a dark matter annihilation? Impossible to say, because neutrinos don’t come with labels indicating where they were born, but it was most likely an atmospheric neutrino, gener
ated by a cosmic ray hitting the atmosphere directly above the North Pole. The key to detecting Earth-based WIMPs would be to discern a hot spot significantly brighter than the atmospheric background in that direction. Indeed, this is the trick to finding a point source anywhere in the neutrino sky. The significance of the Nature letter was that it gave a first measure of the atmospheric neutrino background over the full northern hemisphere: the background that any point source in the northern sky must rise above. Owing to AMANDA’s odd location, this included the center of the Earth.

  According to a paper that the collaboration eventually published in 2002, AMANDA’s first search for neutralinos yielded “no excess over the expected atmospheric neutrino background”: they found no evidence for WIMPs. This was a disappointment, to be sure, but perhaps a not-too-surprising one, since such evidence would have been one of the most important physics discoveries in half a century. Not only would it have provided a clue to the cold dark matter, it would also have been the first evidence ever obtained in support of supersymmetry.

  In spite of coming up empty, this “search” (a code word in physics jargon meaning they had searched and not found), did break new ground. By setting a new limit on the maximum brightness of a possible neutrino source at the center of the Earth, it provided ever so loose a constraint on one prediction of supersymmetry. As limits of this sort become more stringent, they either lead to a discovery or rule a theory out.

  Over the next couple of years, the collaboration carried out other groundbreaking searches with AMANDA, incorporating data from the years following 1997. They searched for supernovae. They searched for a point source anywhere in the northern sky. They searched for a diffuse flux of cosmic neutrinos that might be reaching our planet from every direction, in the manner of cosmic rays. They searched for neutralinos coming from the Sun, which happens to be more a promising candidate than Earth. Fourteen years into it, they were finally doing bread-and-butter particle astrophysics.

  * * *

  They were also breaking ground in applied physics. The 2000–2001 Antarctic season saw what was probably the game-changing breakthrough on the digital optical module.

  One of the most troubling aspects of AMANDA was the difficulty in calibrating the instrument, especially in time. In order to nail down the track of a muon traveling through their array at nearly the speed of light, they needed to have a precise understanding not only of where every optical module was, but also the precise moment when any photon emitted by that muon reached a module. Since light travels about one foot in a billionth of a second, or nanosecond, the positions of the modules had to be known to within a few feet, and the modules had to be synchronized to within a few nanoseconds. Both of these problems were messy ones, not only because the modules were so far underground, but also because they communicated with the surface through wires or fiber-optic cables of varying lengths and transmission characteristics. It was the old “telescope in a darkroom” problem.

  Steve Barwick and company were taking a month or so every year to synchronize the modules, using lasers on the surface and fiber-optics, and the previous year he and four accomplices had not had time to complete the task before the end of the season. This was no reflection on their abilities or work ethic; it was a huge job, even with AMANDA’s skimpy number of modules—667 to be exact. If IceCube was really going to comprise its planned 4,800 modules, it was looking as though it would be impossible to calibrate even in an entire season.

  A year or two earlier, Dave Nygren had done a back-of-the-envelope calculation that told him that they should be able to program the DOMs to calibrate themselves with no human intervention at all. He admits that he was “barking up the wrong tree,” however, on a way of implementing his notion. It was Bob Stokstad who actually figured out a way to do it. The basic idea was to send an electrical pulse down the mile and a half of copper wire to the module, ask the module what time it thought it received the pulse, have it wait for a specified interval, and send a pulse of the same shape back to the surface. Since this was something like sending a flash of light to a distant mirror and measuring how long it took to make the round-trip back, it allowed them to not only measure the length of the wire, but also to synchronize the clock in the module with a master clock on the surface. And since it could all be done with computers, even five thousand DOMs could be synchronized in the space of five or ten millionths of a second!

  A team from Lawrence Berkeley Laboratory consisting of Jerry Przybylski, a software engineer named Chuck McParland, and a physics post-doc named Azriel Goldschmidt traveled to Pole in early 2001 to test out various new ideas on the digital string, number 18. After replacing the data acquisition system that hadn’t worked the previous year with one that had been designed at LBL in the meantime, they put Stokstad’s solution in place and proceeded to demonstrate that the forty-odd modules on string 18 could be synchronized to within less than five nanoseconds.

  Goldschmidt also put in an improvement on the science side that ended up demonstrating, with data taken over the subsequent winter, that they could reconstruct the path of an up-going muon with just that one digital string—an impressive accomplishment. And they were already capable of sitting at a computer in the northern hemisphere and programming an individual DOM deep in the Ice more than half a world away. The digital technology was more than meeting its requirements, it was blowing its analog counterpart away.

  Goldschmidt and his partners were working under a certain amount of pressure, since Francis (in what he characterizes as his “main contribution in leadership, if I ever had one”) was at that time trying to herd the collaboration into a decision on which optical module to use in IceCube. He scheduled a two-day meeting for late February, before an external advisory committee chaired by the widely respected Barry Barish, the experimental physicist from Caltech whom NSF had brought in to save the Laser Interferometer Gravitational-Wave Observatory after its initial management failure. To make it easy for everyone to attend, the meeting was held in a conference room at the Chicago O’Hare airport.

  The LBL team performed with Hemingway-esque grace, delivering a report on the DOM just five days beforehand. And amazingly, since this was awfully far from theoretical physics, Francis himself wrote up a preparatory technical document. (Some time later, when he saw his name on the author list for a journal article about the DOM, he told me that he had now “lost all self-respect as a theorist.”) Dave Nygren led the digital side of the discussion, and Christian Spiering and Albrecht Karle presented the case for the latest version of the analog/fiber-optic module. Bob Paulos remembers Nygren as being extremely helpful: he “more-or-less represented the digital approach but was very evenhanded and open-minded about discussing the merits of either.” And even though Albrecht had invented the fiber-optic method, he didn’t fight either; he just wanted the best decision. “And so we selected … digital,” he says, “and everybody pulled together.… There were no hard feelings on any side.… That was a good day for the collaboration, I would say.”

  * * *

  The distracting personal melodrama that Steve Barwick decided to serve up that year was to drop out of IceCube and begin lobbying for an upgrade to AMANDA that would, in effect, be direct competition. The previous September, he had held what was ostensibly an IceCube planning meeting on his home turf in Irvine, which turned out to be an advertisement for his desire to become principal investigator and run the project out of his school. When it became “clear,” according to Buford Price, “that nobody else outside of Steve would support him being the head, he resigned in a huff.” And this was his response. At the collaboration meeting in Delaware in March 2001, Steve turned up in the role of “interloper” and gave a whiny, downbeat talk in support of his renegade idea. “Well, we’re all trees in the forest, fighting for a space in the sunshine,” Bob Morse observed.

  Since the no-new-starts policy had thrown IceCube’s funding prospects back into question, Steve’s gambit caused concern among the other PIs. It wa
s an obvious vote of no confidence, and several read his move as a ploy to undermine the larger instrument and maintain what little influence he had, while hoping for IceCube to fail.

  Steve would continue to champion what he called AMANDA++ (an inside joke from the computer language C) until 2005, when the AMANDA collaboration was formally dissolved and merged into IceCube. His gambit failed miserably, since it ensured that he would have zero influence on the larger instrument, which was where the train was headed.

  This was a disappointment for Francis, who had hoped Steve could play a major role. In his opinion, the experiment outgrew Steve. When he could no longer analyze the data on his personal computer or hold the budget in his head, “he lost it,… because, you don’t have it in your head, you rely on other people, and he just couldn’t do that.” In the early days of AMANDA, his contributions were crucial, and IceCube would not exist without him, but he has never had a role in the larger instrument.

  The general opinion, not only in AMANDA but also in the community at large, was that Steve and George Smoot’s antics arose largely from their desire for the physicist’s version of fame. When they gave talks at conferences and colloquia, each had the irritating habit of representing himself as principal investigator of AMANDA (or IceCube, if that was the subject) and making it seem as though his institution was leading it. This was very much in line with George’s behavior on the COBE experiment.

  Since COBE, George had been kicked out of a different collaboration, BOOMERANG (Balloon Observations Of Millimetric Extragalactic Radiation ANd Geophysics), for the same sort of behavior. And some years later, Steve would be kicked out of ANITA (ANtarctic Impulse Transient Array) for the same thing. (BOOMERANG flew cosmic background radiation instruments in balloons over Antarctica, and ANITA flew balloons equipped with antennae tuned to detect neutrino interactions in the ice sheet below.)