The Telescope in the Ice

by Mark Bowen

  As they prepared for this campaign, they fell into the annual build cycle that would drive AMANDA and IceCube for the next twenty years. You don’t carry a hot water drill or two hundred optical modules to Antarctica in your luggage. Anything that goes to Pole, for whatever purpose, even as part of a building, must be “Hercable”: it has to fit in the cargo bay of a Hercules LC-130.

  The Herc is a standard coin for working at Pole. The drill, for example, with its numerous hot water boilers and thousand-plus meters of hose, filled almost two Hercs. When you submit your proposal for your working season at Pole, you negotiate with NSF not only about how many people you need to house at the station and for how long, but also how many Herc flights it will take to get your equipment and instrumentation down there. These items are transported to McMurdo on cargo ships that leave the continental United States on non-negotiable schedules from Port Hueneme, a naval base on the California coast, not far north of Los Angeles. If you miss the boat you may very well miss your season.

  * * *

  The plan was for each of the universities in the AMANDA collaboration to become expert in most of the hardware tasks, so that whatever collection of students, faculty, and other personnel happened to find themselves on the Ice at any given time would have a large enough reservoir of knowledge to deploy the hardware and get it running. The optical modules were assembled in Madison, Irvine, and Sweden, while Berkeley made other critical parts. Since Steve Barwick’s institution, Irvine, was closest to Port Hueneme and he had access to large workspaces on campus, Irvine served as the staging area, where the hardware was integrated, packed in wooden crates—or, in the case of the hose and the cables for the strings, wound onto large reels—and shipped by truck to the port. Once it reached McMurdo, it would be transferred to Hercs and flown to Pole, on an unpredictable schedule that was highly dependent on weather. The trip from Port Hueneme to Pole usually takes between seven and eight weeks, so equipment for the start of the season needs to reach the port by the end of September. Ships continue to depart on a regular schedule until about the first week of January.

  Barwick was a powerhouse. He became the hardware guru. Among other things, he designed a data acquisition system that would run on an Apple Macintosh, to collect and computerize the raw electrical signals as they streamed to the surface from the phototubes below. (This so-called DAQ was small enough to be carried south in his hand luggage.) He was in a strategic position, of course, and this became both a blessing and a curse—especially when money got tight, which it did right away.

  * * *

  The management of an academic collaboration is very different from that of a corporation. For one thing, it’s more democratic. For another, its finances are subject to the whims of capricious elected officials—in this case, the United States Congress. Thus the annual funding cycle for U.S. science agencies resembles a rollercoaster ride. NSF, DOE, NASA, and the other science agencies prepare budgets in advance of the fiscal year; they are modified over some months, first by the White House, then by Congress (the “appropriators”); agreement is usually reached; budgets are usually approved; and money becomes available at some indeterminate date, usually in the late summer or early fall. Depending upon political winds, there can be excessive delays. The infamous budget stand-off in the mid-nineties between Bill Clinton and Newt Gingrich stretched into the following year and wreaked havoc on countless scientific programs. In the end, an agency’s budget may actually exceed its original request, but it usually falls short—as it did for NSF in the first year of the first AMANDA grant.

  Wisconsin was the lead institution, Berkeley and Irvine were listed as subcontractors, but the budget called for each institution to get about the same amount of money. Now, some of that money was earmarked for salaries: 50 percent of Bob Morse’s salary, for example (since he was a research scientist, rather than a professor, he lived on such “soft money”), and two months of Steve Barwick’s, to cover the summer. Students were partially covered as well. Without going into the arcane details of NSF money management, let’s just say that Lynch and Wilkness changed the ground rules only a month or two after the grant was funded. They were happy to pay for any and all hardware and for salaries at Madison, but decreed that Berkeley and Irvine should pay their salaries from other grants. (Buford Price had a longstanding NSF grant for nuclear physics research that was more-or-less rubberstamped every three years.) This upset Buford and Steve tremendously, and they began making frantic requests for money, both to NSF and to Bob and Francis in Madison. Steve’s were particularly emotional. In an e-mail written in July 1993, scant months before their second deployment season would begin, Francis wrote to Buford that “after talking to Steve (for hours) I have the impression that there might be a misunderstanding concerning the AMANDA funding. Steve reacts with his heart although his brains know the facts and, given the circumstances, this is hardly a criticism.”

  As the hard deadline for getting their equipment to Port Hueneme approached, Steve upped the ante with what would become another annual rite: he held the project hostage. He threatened not to deliver the final two strings until he got more money.

  * * *

  Considering the University of California’s reputation as the premier public university system in the world and its long history of running huge programs like the Los Alamos bomb project, it is surprising how much more adaptable the University of Wisconsin proved to be in dealing with these vicissitudes.

  The reasons go back to the beginning of the twentieth century, when the president of the state’s university system, a man named Charles Van Hise, first enunciated what became known as the Wisconsin Idea: that the purpose of the university should be to serve the people of the state by promoting public service and the search for truth. “I will never be content until the beneficent influence of the university reaches every family in the state,” he declared. The Idea is now enshrined on a bronze plaque on a granite boulder on a hill on the Madison campus, and the people of Wisconsin support it across the political spectrum. It may sound lofty, but it has had immense down-to-earth ramifications.

  In 1923, a Madison biochemist named Harry Steenbock discovered that ultraviolet radiation increased the vitamin D content of foods. A vitamin D deficiency can cause rickets, a crippling bone disease that affected mostly children at that time. Inspired by the Wisconsin Idea, Steenbock filed for a patent, declined a licensing offer of $1 million from the Quaker Oats Company (quite a lot of money in those days) and instead rounded up a group of likeminded friends to found the Wisconsin Alumni Research Foundation (WARF) in order to use the proceeds from the patent to support the university. It was licensed to Quaker Oats and a pharmaceutical firm. The irradiation process is especially effective in milk. It essentially eliminated childhood rickets worldwide within about twenty years, and it brought oodles of money to WARF.

  Over the decades, the foundation has licensed thousands more patents, developed an equity fund of about $2 billion, and disbursed the better part of $1 billion to the Madison campus. (Another huge moneymaker was Warfarin, which is named for the foundation. It is the most widely prescribed oral anticoagulant in North America. One of its trade names is Coumadin.) In an average year, WARF distributes $50 to $70 million. The money can be used for pretty much anything related to its broad mission: funding research projects, building buildings, purchasing land, endowing fellowships, subsidizing visiting scholars. Vernon Barger used WARF money to bring Francis Halzen to Madison in 1971, for the visit that still has not ended.

  Twenty years later, when the money crunch hit AMANDA, it so happened that the man in the strongest position to steer WARF funds was John Wiley, Bob Morse’s old drinking buddy from the days of the 602 Club. They’d been graduate students together and even part-time roommates.

  Wiley had earned his doctorate in solid state physics, gone off for a six-year stint at the legendary Bell Telephone Laboratories, returned to a faculty position at Madison, and proceeded to rise like cream. By 1993, he
was dean of the graduate school and by virtue of that position the university’s primary liaison with WARF, which is a separate, non-profit entity. The foundation’s profits “are funneled through the graduate school,” he says, “so the dean of the graduate school is kind of Santa Claus on campus.” Like Peter Wilkness, John Lynch’s boss at NSF, Wiley loves the big idea.

  According to Lynch, “Smart guy like Francis goes and talks to a smart guy like Wiley, things can happen, you know?” The dean responded to AMANDA’s money crunch by arranging a million-dollar line of credit through WARF. The hope was that NSF grants would eventually right the balance—and eventually they did—but in the meantime the credit line would buffer the project against the rollercoaster of federal funding.

  With their newfound flexibility, the Wisconsinites did everything they could to keep the project and the Californians afloat. Francis even sent an installment from his endowed professorship, the Gregory Breit Chair in the Madison physics department, to Steve Barwick, to pay for salaries. Steve also got what he wanted to finish the last two strings. (The Swedes kicked in some money for hardware as well.) By the time AMANDA headed south for the 1993–94 drilling season, Bob and Francis owed WARF a quarter of their credit limit, and WARF had also guaranteed a $100,000 loan that Buford had obtained from Berkeley.

  * * *

  In the two-year break while the physicists went through this exercise, Bruce Koci went off on other adventures. During the summer after the Bucky-1 debacle, he and Bill Barber joined Lonnie Thompson on a three-month expedition to the highest icecap in the remote Kunlun Shan range in far western China, where they drilled on a wide snow dome at 20,000 feet and recovered an ice core that held a climate record reaching back three-quarters of a million years. Bruce did some non-AMANDA hot water drilling in Antarctica the following winter, and over the summer of 1993 joined Thompson on Huascarán, the highest mountain in Peru’s Cordillera Blanca. Camping for more than fifty consecutive days in the 19,850-foot col between the mountain’s twin summits, the drilling team recovered two long cores to bedrock that altered the course of thinking in climatology. These were the first tropical ice cores to reach back to the last ice age, 20,000 years ago.

  A couple of months after descending from Huascarán, he followed what was becoming his own annual pattern by returning to Pole to take a second stab at AMANDA.

  * * *

  The other key personnel on the Ice that year were Bob Morse, Steve Barwick, Per Olof Hulth, and Serap Tilav, a Turkish woman who had earned her doctorate under Tom Gaisser at Bartol and was now working as a post-doc under Francis. Adam Bouchta, a first-year grad student in Stockholm, and Pat Mock, a post-doc of Steve’s at Irvine, also joined in. And the Swedes provided two drillers.

  They got a little more organized this year and put Pat in charge of string deployment. He, Bruce, and Bob caught one of the first Hercs south, reached Pole at the end of October, and began getting the ball rolling all around. Bob still has the hand-drawn map he made from his survey of where to drill the holes. Since compasses don’t work at the pole, he made a homemade sundial by planting a wooden pole in the snow and estimating directions based on the direction of the shadow it cast under the ever-present sun and the time of day on his wristwatch.

  Through a beneficial alignment of the political stars at PICO, Bruce was given clear leadership of the drilling team that year, so the drilling went well.

  But that doesn’t mean it was easy. The weather was miserable, unusually cold, with blizzard after blizzard. Drilling and deployment again took place in the wide open, with the one saving grace that the computers that ran the drill had to be kept warm, so the humans could retreat to a small plywood shack equipped with a space heater when it really got rough. Again there were only two drilling crews, tag-teaming twelve-hour shifts, and to save time and assure that reinforcements would always be handy, the drillers slept in a large unheated tent out by the drilling site.

  Bruce recalled that it was actually colder inside the tent than out and that they kept the door open in order to let the “liquid air” out. Their meals were delivered from the galley in the dome, half a mile away. “It was pretty barbaric conditions,” he said. “Of course, we were pretty barbaric, too. Still are.”

  Their first task was to assemble the drill by hand in wind chills Bruce described as “below a hundred below.” “There were six or eight of us huddled around any kind of shelter we could find by the garage, trying to assemble this stuff, freezing our tushes off.” (I believe this was the only time I ever heard Bruce complain about the weather.) And the hot water boilers for the drill burned diesel fuel, which doesn’t want to ignite even in good weather, so it took them three or four hours simply to light them, which they had to do whenever they started a new hole.

  The basic message that Bruce carried home that year was that humans are more resilient than machines. Drifting snow covered the cables and hoses snaking around on the snow, and even the large reels that held them. The machines broke down constantly. Sometimes the drillers had to lie on their backs on the snow to fix them, often without gloves. The ice in the hole lost about 8 percent of its volume as it turned to water (ice is less dense than water, the most obvious consequence being that ice cubes float), but they had to keep the holes topped up to the level of the firn layer in order to keep them from collapsing. So they had to shovel about 50,000 pounds of extra snow, by hand, into their hot water reservoirs for every hole that they drilled. Everyone helped shovel, including Per Olof.

  The benefits of Bruce’s new drill, and especially the larger hose, demonstrated themselves immediately. They had no problem reaching a thousand meters, and they also drilled faster and used less fuel than they had two years earlier: 4,500 gallons per hole, as opposed to 12,000 for the one successful hole on the previous attempt, which probably hadn’t reached a thousand meters. And progress was steady. They began drilling the first hole on December 20 and reached a thousand meters four days later, on Christmas Eve. Then the physicists started in on deployment—a process Bruce monitored closely, since it was basically the report card on how well he had drilled his holes. He didn’t get much sleep.

  You might be surprised to learn that it takes about forty hours for the water in a hole surrounded by minus fifty degree ice to freeze, the main reason being that water is an insulator. Bruce often said that no matter how many times he did the calculation he couldn’t convince himself that it wouldn’t freeze instantly. This meant that the physicists had about two days to drop their strings of pearls into the ice before they would freeze in place forever. They had only one shift, so they worked non-stop until the job was done.

  * * *

  For both drilling and deployment, a tall metal arch is set in place over the hole by means of a crane. It supports a large pulley or sheave, fifteen or twenty feet off the ground (see photograph 10). The huge reel that holds either the hose for the drill or the cable that provides mechanical support and electrical connection to the optical modules (the string for the pearls) is placed on an axle, like the reel to a fishing rod, sitting on the Ice some distance away from the sheave. The hose or cable runs across the Ice from the reel and up to the sheave and then drops straight down into the hole under the pull of gravity. A powerful winch is used to lower and raise the cable or hose, by rotating the reel.

  The cable has so-called breakouts at the specified interval—ten meters that year—where the pearls are attached one-by-one as the cable is lowered into the hole. The first thing that goes on the cable is a weight at the bottom end, six or eight hundred pounds, to keep it straight and pull it downward. The cable is lowered to the first breakout and stopped, the first optical module is attached mechanically and electrically, the cable is lowered ten meters to the next breakout, and so on. Owing both to the cold and the two-day time window, the entire process needs to be thought out in advance. A few of the steps are intricate and require bare hands. Bruce remarked that “the physicists suffered equally,” since they had no shelter either.
<
br />   Having supported quite a few science experiments in his day, Bruce had observed that at the beginning of a project the engineering always seems more difficult than the science, but that sooner or later a crossover occurs: the engineering reaches a point where the scientists can begin facing their own challenges, and these, he was humble enough to admit, are “orders of magnitude more difficult” than the engineering. Sometime in January 1993, that crossover occurred. AMANDA began producing its first live data—and confounding the scientists, not for the last time.

  * * *

  Along with the main electrical cable, they had run individual optical fibers to each of the detector modules on every string, so that they would be able to pipe light down to the modules from a laser on the surface. The fibers terminated in nylon balls that broadcast this light more or less evenly in all directions. By sending pulses of laser light through the different fibers they could send flashes between different modules and strings and thereby measure the optical characteristics of the ice. Not long after they’d deployed the first few strings and got them running, they realized they had a problem on their hands.

  The strings on the perimeter of the array were a hundred feet from the string in the center and about sixty-eight feet from each other. (Bob Morse had scandalized the Swedes by doing his survey in feet and inches, because that’s what he had on his tape measure.) That meant that the flashes should have taken about a tenth of a microsecond, one ten-millionth of a second, to travel from string to string. Instead, they were taking a full microsecond and more—more than ten times what the scientists expected. The flashes were changing shape as well. An initial short square pulse would arrive at the next string as a longer rounded pulse: it would spread out in time, an effect known as dispersion.

  When something goes wrong in a physics experiment it can be hard to tell whether your instrument is acting up or you’re measuring a real effect. Steve Barwick, the hardware guru, checked what he could; they decided the instrument probably was working correctly … and slowly it began to dawn on them that something was wrong with the ice! The light was clearly bouncing around down there—the technical term is scattering. Bob Morse remembers thinking things through with Per Olof and “arguing with Steve, which is always an irrational event.” Within “thirty minutes it hit with a painful clarity that I remember today. It was bubbles … lots of bubbles.… We had not gone deep enough—SHIT!!!”