Experiment: Search for the Higgs Boson

views updated

EXPERIMENT: SEARCH FOR THE HIGGS BOSON

The search for the Higgs boson has been the premier high-energy physics goal of the late twentieth century and continues into the twenty-first century. The Standard Model of particle physics has been incredibly successful at describing the electromagnetic and strong and weak nuclear forces. Yet, at the same time, it lacks confirmation of some key elements, most importantly direct evidence for the Higgs boson.

Simply put, the Standard Model is a nonsensical theory without the Higgs boson, or something very much like it. It solves several theoretical problems, providing mathematical consistency and experimental predictability to the theory and giving elementary particles their mass. Even though the Standard Model encompasses all known particles (quarks and leptons, which make up matter, and bosons, which transmit force between matter particles), without a Higgs boson it gives different answers to physical problems; for example, the rate of nuclear fusion reactions in the Sun, depends on how one calculates them from the starting equations. Another problem, which would show up in experiment, is that the probability of the weak bosons W and Z interacting with each other in high-energy collisions is greater than 100 percent, a clear impossibility.

The Higgs boson is actually a scalar set of fields consisting of four components, which are present everywhere in the universe. Scalar means the four field components have numerical values representing their strength for all points in space, much like a frying pan over a stove has a certain temperature at every point inside the pan. The constant interaction of Standard Model particles with these Higgs fields slows them down, making it appear that they have a mass (massless particles must travel at the speed of light). Three components are absorbed by the weak bosons, giving them mass, while the fourth component becomes a physical particle with a mass of its own. Furthermore, the weak force becomes weaker than electromagnetism (hence the name) as well as acting over only very short distances, the size of atomic nuclei, which is an experimental fact. Thus mass, as perceived by us, is not an inherent property, but a result of interactions between matter and the Higgs field.

This has several consequences. First, calculations show that for the Higgs boson to solve the high-energy weak boson scattering probability problem, the Higgs particle cannot have a mass greater than a certain value, called the unitarity constraint, about two million times the electron mass. This value is low enough that physicists are almost certain to see it in the next generation of experiments, starting in 2007 (if not sooner), making the theory experimentally relevant. Second, the Higgs boson interaction strength with particles is proportional to their mass. Third, the Higgs boson is spinless since it is a scalar field and does not have electric charge. These and other quantum properties are all precisely defined and can be tested. To fully understand the Standard Model, scientists must find the Higgs boson, measure its properties, and compare them to theory. The motivation is thus to test current scientific knowledge and learn more about the structure of the universe at a fundamental level.

The LEP Search for the Higgs Boson

To produce a Higgs particle in an experiment, matter must be collided at energies above the mass of the Higgs boson (mass and energy are equivalent). As collider experiments have become more energetic, lack of a Higgs boson observation has pushed the allowable mass to higher values. The limit is still far below the unitarity limit, but the gap is closing fast and will be completely accessible by about 2010.

The most recent collider engaged in this quest was the Large Electron-Positron (LEP) collider at the European Laboratory for Particle Physics (CERN) near Geneva, Switzerland. LEP was a circular collider in a 27-kilometer tunnel 100 meters underground. Small bunches of electrons and positrons (antielectrons) were accelerated around the ring in opposite directions, passing through each other at four interaction regions, where a detector belonging to one of the four experimental collaborations, Aleph, Delphi, Opal, and L3, observed the collisions.

High-energy physics experimental collaborations are an interesting study in their own right. Each LEP group consisted of a few hundred physicists, from all over the world, mostly employed by universities or institutions other than CERN. There is really no way to organize from the top such a diverse group of people, but high-energy physicists are adept at forming self-organizing structures.

The four collaborations took different approaches toward building their detectors, so each had a different performance level of particle identification and measurement, and had to be carefully calibrated, a tedious, months-long process frequently relegated to graduate students. The detectors themselves were enormously large, complicated machines. L3, for example, was four stories tall and had almost as much iron in it as the Eiffel Tower. Supplied continuously by about the same amount of power as a lightning strike, it had millions of electronic components measuring each collision, which occurred 44,000 times every second. Only about three of those collisions every second were potentially interesting, and each event recorded required about one hundred kilobytes of data storage. Computers running the detector had approximately one hundred thousandth of a second to initially judge the value of each potential collision. The other detectors had a similar task and contained about the same number of electronic instruments. Figure 1 shows an illustration of the Aleph detector. Combining the results of the four experiments was a challenge even beyond that of organizing the individual collaborations. For the Higgs boson search this effort was led by CERN physicist Patrick Janot and in the end turned out to be successful. The key to this success is the scientific process itself, the ability to cross-check results with those previously established.

Particles and antiparticles annihilate when they collide, generating a slew of other particles that stream out in all directions from the collision point, or vertex. Many of these outgoing particles are unstable and decay successively in a chain until stable final states such as electrons, protons, and photons are reached. Physicists study collisions for deviations from the expected Standard Model behavior of these collisions, which signals new phenomena. This may consist of anomalous events from production of new particles or may show up as subtly different emission patterns of known particles.

The vast majority of collisions are uninteresting, involving thoroughly studied phenomena. The uninteresting stuff is called background, while new physics is called signal. The purpose of the detector is to identify and measure all of the outgoing particles in a collision, while data analysis separates signal from background; in this case, sifting through data to find the handful of expected Higgs particles amongst the millions of background events.

For the Higgs particle, the experiments at LEP looked for events where the electron and positron annihilated to form a virtual Z boson, which could then radiate a Higgs boson to become a real Z boson. (Virtual means not real and, in this case, too heavy. Real photons, for example, are perceived as light, while virtual photons comprise electric and magnetic fields, which transmit force.) This would happen very rarely, perhaps one in every thousand Z events. Then experimentalists had to consider how the Z and Higgs bosons might be expected to decay. Only some of the decays would be observable, with some decays having larger background than others, mostly from Standard Model processes faking a Higgs boson event by producing the same final state particles in a similar pattern.

Since the Higgs boson mass previously had been excluded to be larger than twice the bottom quark mass, and the Higgs couples preferentially to heavy particles, any Higgs boson LEP could create was expected to decay most of the time to a pair of bottom quarks. This would be fortuitous because bottom quarks are relatively long-lived and will decay a small distance away from the vertex, a fraction of a millimeter to a few millimeters. The experiments could see this separation with devices called micro-vertex detectors. Such an event is often easily recognizable (see Figure 2). The Z boson may decay into any pair

FIGURE 1

of matter particles, and all of these channels were considered separately. Neutrinos are nearly impossible to detect, so they leave the detector invisibly, but since the beam energy is known, a deficit of total outgoing energy can easily tag the event as an invisibly decaying Z boson. Thus, experiments looked for Z boson events accompanied by a pair of bottom quarks that together had enough energy to come from a very heavy object, the Higgs boson candidates. The total number of events in all decay channels was compared to the total number expected from all Standard Model processes using a complex statistical procedure to determine if there was a significant excess of events.

Data analysis is a long and painstaking procedure. Experiments must compare measured particle energies against machine performance at the time of the event, since the accelerator energy and collision rate change slightly over time. These changes affect the interpretation of individual events and take additional computer time to analyze. Data must also be corrected for known problems in the detector, mapped out during detector calibration. Extraneous photon radiation from the colliding beams can alter

FIGURE 2

energy measurement, throwing off comparison to Standard Model expectations. Cosmic rays can sometimes penetrate the earth even to the experimental halls 100 meters underground and register in the detectors. The complicated electronics will even occasionally "hiccup," producing the exceedingly rare event with strange characteristics that just happens to look like new physics. Data analysis software, written by the physicists, can have hidden bugs that might not ever affect most data but can affect the search for new physics. Many of these problems can be dealt with by careful calibration and comparison with the other running experiments. But in these types of experiments, one cannot take a single event that looks like a Higgs boson to be proof of its existence. This turned out to be exactly the problem with which LEP was faced.

Before LEP, the Higgs boson mass had been excluded to be so high that if LEP could produce it at all, it would be only barely. As the LEP energy was increased, lack of observation meant that the new mass limit was essentially the machine energy, minus the Z boson mass, minus a little bit more—LEP physicists were always hunting right at the edge of accelerator output. The number of Higgs bosons produced at the limit of machine energy would be very small, so the experiments had to rely on analysis of low-statistics events. This is a dangerous situation for an experiment, as the probability for background to give just a handful more events that look signal-like turns out to be not so small.

By the fall of 1999, LEP had been pushed nearly to its design limit of about 200 GeV. (1 GeV = 1 billion electron volts, the energy an electron gains when accelerated through a potential of 1 volt. The mass of an electron is 0.00051 GeV.) Still the Higgs boson had not been found. Since the Z boson has a mass of about 91 GeV, this put the Higgs boson limit at about 105 GeV. However, other indirect but very precise data suggested that for all the parameters of the Standard Model to fit together properly, the Higgs boson must be very light, probably near the energy where LEP was hunting. However, LEP was scheduled to be shut down to make way for the construction of the Large Hadron Collider (LHC), a much more powerful proton-proton collider to be built in the same tunnel and scheduled to turn on in 2007. The LHC could not only discover a Higgs boson of any mass but also measure many of its quantum properties. However, in the meantime the proton-antiproton accelerator Tevatron at the Fermi National Accelerator Laboratory (Fermilab) outside Chicago, Illinois, was being upgraded for its second run, which would start in March 2001. This was a sensitive issue to LEP, because if the Higgs boson was indeed only slightly heavier than the energy LEP was able to access at that time, as data suggested, discovery might go to another laboratory.

The LEP collaborations petitioned for a delayed shutdown and presented a scheme to squeeze additional energy out of the machine by reducing the number of electrons and positrons being accelerated. Energy was more important than number of collisions for the Higgs search. An ingenious pattern of running was devised that boosted the energy when electrons and positrons were lost as the beams circulated. However, this made data analysis much more difficult as the exact number of particles in the beam and beam energy had to be tracked very precisely as a function of time. Finally, the acceleration cavities would be pushed well beyond their design limits, possibly to failure. The rationale for this was that LEP was going to be dismantled anyway, so there was nothing to lose—"go for broke," literally. CERN management approved the plan and granted a one-year extension.

By October 2000 the mass limit had been pushed several GeV higher, and the collaborations still had not discovered the Higgs boson. But in the last month of extension they announced tantalizing hints of a possible signal. Not all of the collaborations' data agreed, but this was reasonable given the small number of candidate events. L3 had one promising candidate event, while Opal and Delphi saw none, but Aleph had about three candidates, at a mass of about 115 GeV. The Aleph group's Higgs boson search subgroup was led by Professor Sau-Lan Wu of the University of Wisconsin at Madison, who had years before been involved in the discovery of the gluon (the carrier of the strong nuclear force). Her postdoctoral assistant Stephen Armstrong and graduate student Jason Nielsen had been among the first LEP experimentalists to devise a real-time analysis program, to search for candidate events as the data came in, eliminating some but not all of the postrun data analysis that normally took so long.

The collaborations petitioned for and received another month of running but produced no additional candidates. Despite vigorous protest from some members of the collaborations, CERN Director General Luciano Maiani made the decision not to delay LHC construction any longer, and LEP was permanently shut down. This caused a great deal of acrimony among some members, some going so far as to ridicule the director publicly—this was completely in tune with the history of CERN as the experiments there attracted some very colorful personalities with oversized egos. Often these experiment-management clashes became publicly visible. However, the director had made the decision based on the lack of confidence in the result among members of the LEP committee, the scientific advisory board. This perception was mirrored by physicists outside of CERN. When Chris Tully, an assistant professor of physics at Princeton University and member of the L3 collaboration, presented LEP summary results for the Higgs boson search at Fermilab in December 2000, the general reception was skepticism that LEP had seen any signal at all. But in a final rebuke to the shutdown, Aleph made holiday greeting cards from one of their Higgs candidate event displays and sent them to colleagues around the world.

In July 2001 LEP presented an update that lessened confidence in the candidate signal. More thorough analysis had reduced the statistical significance of the data almost to nothing. L3 was decidedly less confident about its event, as the decay particles went into a region of the detector known to have measurement problems.

Tevatron and the LHC

The Fermilab Tevatron began its second run in March 2001, with an upgraded machine that promised to deliver 20 times more collision and 10 percent greater beam energy than its first run. LEP's supreme final performance, however, set a new mass limit that was considerably higher than anticipated. This will make it difficult, but not impossible, for the Tevatron to find the Higgs boson. If the Tevatron performs well before the LHC can analyze its first data, perhaps in 2007, then its two detectors CDF (see Figure 4) and D0 have good potential to observe a Higgs boson up to twice the Z boson mass. However, many aspects of the new machine and detectors' performance are not yet known well enough to determine the Tevatron's true potential.

Higgs boson search channels at a proton collider are very different, primarily in the production mode. The largest rate would come from a gluon pair fusing to form a Higgs boson. (Massless gluons do not couple directly to the Higgs boson but can produce a virtual top quark pair, which do couple. This is known as a loop-induced process, a feature of quantum mechanics.) Another method is an incoming pair of lighter quarks annihilating to form a real top quark pair, one of which may radiate a Higgs boson. While this rate is quite small, top quarks are very distinctive and have a much smaller background. This has been called the "Cinderella discovery mode" for a Higgs boson by Fermilab theorist Stephen Parke because at first glance the low rate is uninteresting, but the process appears beautiful when one considers it more carefully.

Proton-antiproton collisions are much messier than electron-positron collisions: there are more

Standard Model backgrounds, and they are more complicated to calculate. The tradeoff is the ability to reach much higher energies. The Tevatron has the energy advantage over LEP but must deliver enough collisions to avoid the situation of too few candidates that LEP experienced. Three or four candidates in the Tevatron environment may be less clean than those at LEP and must be considered more carefully.

While it might take six years to confirm the LEP candidate, the Tevatron can rule it out in about two years. The lack of candidate events is a more powerful statistical signal of no Higgs particle than the existence of a few events is of a possible Higgs boson.

The LHC is expected to turn on in 2007. Its role is not just discovery, however, but also to measure a candidate Higgs boson's properties for comparison against theory. This is possible because of the LHC's significantly higher beam energy, seven times that of Tevatron, and enormously greater data taking capability.

While the Higgs boson as discussed in the first section is the most anticipated signal, there are several variant theories for the Higgs mechanism. Some models have two sets of Higgs fields, resulting in additional particles that would be produced, with slightly different properties. Other theories incorporate supersymmetry as well. The LHC has great potential to distinguish these different types of Higgs bosons or even more complicated scenarios. It is also possible that the LHC won't find a Higgs boson but instead will discovers a different mechanism for giving mass.