The text below is the second part of what could have become "Chapter 13" of the book "Anomaly! Collider Physics and the Quest for New Phenomena at Fermilab", which I published in 2016. For part 1 see here.

Collaboration membership from CDF to CDF II

Although from the outside one might have perceived a substantial continuity between the Run 1 and Run 2 collaborations operating CDF, change was the rule rather than the exception also from the point of view of manpower. In truth, very few collaborators actually left the experiment right after the end of Run 1. The collected data would take many more years to be fully analyzed; that was a fantastic bounty of high-energy proton-antiproton collisions still waiting to be milked for new precise measurements. Abandoning a healthy and highly productive scientific collaboration at such a juncture was not a very sound move. The peak in production of scientific publications signed by the CDF authors was in fact reached in 2000, and the last paper based on Run 1 data from CDF was submitted for publication as late as 2005: despite five years without new data, the scientific output of the experiment flew without interruptions. While hanging on, many members of the collaboration started however to gradually ramp down their involvement in the experiment. In combination they ramped up other commitments –in particular by joining the ATLAS and CMS experiments at CERN, which were entering a crucial phase of construction, or the BaBar experiment at SLAC, which had started data taking in October 1999. There was no rule managing the potential conflict of interest of scientists contributing to two different experimental groups: this was left to personal judgement, trusting that physicists would abide to some unwritten code of conduct. Only after 2010, when the search for the Higgs boson became a race between the Tevatron and the LHC, the point would become a subject of discussion, with moral suasion attempts of the leaders of experimental groups targeting those researchers who had chosen to keep a foot in CDF II while spending most of their time in the analysis of LHC data. On a personal note, I joined the CMS collaboration at CERN in 2001 and only left CDF in 2012, with a heavy heart. Many of my CMS and ATLAS colleagues however are still official members of CDF to this day.

Going back to the transition between CDF and CDF II, a large influx of new collaborators set on by 1999. I remember attending physics meetings at the usual fortnightly rate, finding out each time that there were at least a couple of new faces I could not attach a name to. The proposals of new institutions to join the CDF II experiment were discussed routinely at each meeting of the executive board of CDF. In order to be accepted, the proposals had to be presented twice at successive executive board meetings –usually by the head of the prospective new group. The ritual included a presentation of the physicists, their past experience, and most of all a list of the endowment offers that the proponents were making: new wealth for the experiment in the form of manpower, tasks that would be covered, new hardware equipment, commissioning work, maintenance tasks, responsibility for the development and upkeep of software tools. The second presentation would be followed by a direct vote by the board members, which would decree the acceptation of the new institution within the collaboration, or its refusal (the latter was a quite rare occurrence). Besides institutional commitments and “Memoranda of Understanding” –the documents where institutions detailed their foreseen contributions to the experiment – all members of the new collaboration, regardless of their provenance, had to individually provide a direct contribution to the experiment by fulfilling shift duties, irrespective of whether they were staff persons from a new institution or students which had been in the experiment since long before, formerly working for other research groups. Nobody was exempted.

Shift duties are a rule during the running of an experiment, when the detector has to be attended 24 hours a day for seven days a week. In CDF, shift crews constituted by three or four members of the collaboration were routinely organized to man the CDF control room. There, scientists had full control of the whole apparatus and, in collaboration with sub-detector experts who carried pagers and could thus be reached at any time, were tasked with ensuring a smooth data taking. However, with a surprising move by the management of the experiment, hard-pressed to meet deadlines, in 1999 working shifts were requested also during the commissioning phase of CDF II, to “save money” and speed up the completion of the upgrade program. That was a debatable concept, although the urgency it addressed was hard to question. Perhaps I should explain this better.

Above, the author during a Scientific Coordinator shift in the CDF Control room

While data-taking operations cannot be delegated to technicians, who usually lack the necessary understanding of particle physics to be trusted to take the right decisions and to operate the data acquisition system properly, the technical activities which upgrade shift crews were asked to engage in were a different matter. To exemplify the situation, let us take full professor John Smith, belonging to an institution located ten thousand miles away from Fermilab. According to the new plan, John would be required to fly over to participate in the construction of the detector or of its infrastructures by pulling cables, checking electronic connections, or soldering circuit boards. This job would keep him busy nine to five, for nine straight days. The financial aspects of the operation are easy to assess: those days of paid foreign mission for Prof. Smith were going to cost his institute something of the order of four to six thousand dollars; seventy-two hours of a trained technician (who would incidentally do a much better job at the soldering table, or in tying up bundles of high-voltage cables) would cost the laboratory about five times less. Undoubtedly, Fermilab did save money by assigning those tasks to scientific personnel from the participating institutions: but it did so by having those institutions paying a lot more in its stead!

Despite the questionable logic of the whole concept, shifts were useful because they forced new collaborators to make themselves available for construction jobs, and by doing so they would “touch with their hands” the detector they would learn to love and cherish for the coming decade. In more prosaic terms, serving shift duties allowed newcomers to get acquainted with technicians and other personnel, and to get to know the low-level organization of the lab and the infrastructures surrounding the detector.

Run 2 starts

Finally, with over one year of delay with respect to initial schedules, the upgraded Tevatron restarted operations. It was a slow start: the new accelerator complex demanded an intensive learning phase. The finding of good beam parameters and settings is a trial and error game, as witnessed by the fact that the instantaneous luminosity produced by the machine in the CDF and DZERO collision points remained very low in the first few months. It was a slightly frustrating period for experimentalists, who were eager to get new data in their hands. Graduate students in particular were those suffering the most from the delays: they needed fresh new data to produce a physics analysis for their theses, and the delays caused them to remain in a limbo, forced to continue working on simulation studies or on the final touches to the detectors they had helped upgrade.

Finally, on March 1st 2001 the first official Run 2 started, and stores with significant numbers of circulating protons and antiprotons begun to be produced. It was a very exciting moment for physicists around the world, as expectations ran quite high on the possible new discoveries that the CDF and DZERO experiments would be producing. As an example, in an interview to the Fermilab press office Joseph Lykken gave the following statement:

Two recent results from other experiments add to the excitement of Run II. The results from Brookhaven's g-minus-two experiments with muons have a straightforward interpretation as signs of supersymmetry. The increasingly interesting results from BABAR at the Stanford Linear Accelerator Center add to the importance of B physics in Run II, and also suggest new physics. I will be shocked and disappointed if we don't have at least one major discovery.

Lykken was expressing a very common view among experts in the field: the foreseen large integrated luminosity that the new accelerator complex promised to yield, combined with a few experimental hints from other measurements, convinced all but a few that CDF and DZERO would soon deliver some groundbreaking new results. Curiously, the same attitude would be common in the very same players eight years later, when after no big results at the Tevatron all the attention and hopes for a discovery of Supersymmetry or other new phenomena turned to the start-up of the Large Hadron Collider at CERN.

Besides the easy excitement of theorists, the new collisions brought strain to the nerves of experimental physicists responsible for the inner tracking systems, which were the ones closest to the beams and thus the most exposed to beam instabilities and losses: despite the use of radiation-hard materials, the electronics of silicon detector elements could indeed be damaged by large instantaneous doses of radiation. In CDF the procedure at the start of a run envisioned a very careful assessment of the properties of the beams and its stability by the shift crew in the CDF control room. Only after beams had been properly collimated and “scraped” of particles traveling off the perfect orbit, and beam losses sunk below pre-defined safety levels, silicon detectors could be powered on and the real data taking started. However, even those precautionary measures could not protect the detectors from the combination of independent, adverse factors.

On March 30th 2001 CDF withstood its worst beam-related accident of the whole Run 2. The malfunctioning of a radio-frequency cavity caused a partial “debunching” of the proton beam. This means that particles are not contained any longer within short packets along the beam; when that happens, the electromagnetic controls used to keep them in the proper orbit cannot properly operate. Right then, a kicker magnet failed, spraying the CDF detector with a huge dose of radiation. Perhaps it is useful to explain how these devices work – and how they failed on that occasion. Besides being populated by bunches of particles alternated with 118 meter gaps (equivalent to 392 nanoseconds of particle travel time) between them, the beam timing structure contains a longer empty section –called the “abort gap” – whose purpose is to give time to kicker magnets to ramp up during a time interval when no particles are traveling through them. When the need arises, kickers are powered up by large capacitors to steer the beam into proper dump areas and thus abort the run: they have a short amount of time to ramp up to full field between the crossing of the last bunch and the leading one following the abort gap. Those large capacitors have some small probability of firing up by accident, but this by itself would not be enough to cause any damage to the detector directly downstream of them –which was CDF. What happened on that day was that due to the partial debunching that the beam had withstood, the accidental firing of the kicker directed a large amount of protons toward the detector. On that occasion alone, the damage took by the SVX2 was significant: 7% of the readout chips died because of the irradiation. The loss had a non-negligible impact on the b-tagging potential of the silicon detector system in CDF. DZERO, on the other hand, withstood no damage: being downstream of CDF with respect to the proton beam, it was effectively “shielded” by its competitor.

(to be continued)