Experiment: Discovery of the Tau Neutrino

views updated

EXPERIMENT: DISCOVERY OF THE TAU NEUTRINO

Most scientific theories originate and evolve from experimental observations. The success of a theory is then based on its experimental application and verification. Often many years may pass between the time that a theory is put forth and experimental measurements are made to confirm or refute the theory. Occasionally, theories become accepted as true or complete before all of the experimental evidence to confirm them has been collected. Nevertheless, it is the responsibility of scientists to follow the scientific method strictly so that no part of a theory is accepted without proof. The discovery of the tau neutrino is just this kind of commitment to the scientific method.

The Elusive Neutrino

In the everyday world observed matter is described by the arrangement of only three subatomic particles: protons, neutrons, and electrons. In the early 1930s the existence of a fourth subatomic particle was postulated by physicist Wolfgang Pauli to explain an apparent nonconservation of energy observed in the radioactive beta decay of nuclei. (Beta decay is the term used to describe the process of a neutron constituent of a nucleus transforming into a proton and emitting an electron, formerly called a beta particle.) Though Pauli's particle could be described as having energy and momentum, it had no easily observable characteristics such as charge or mass. Several years later, physicist Enrico Fermi named the ghostly particle neutrino, Italian for "little neutral one." Fermi subsequently developed the theory of weak interactions that explained the interaction, or force, between the observable electron and its unseen neutrino partner. The essential difference between the electron and its neutrino partner is that the electron has an electrical charge of -1, while neutrinos are chargeless (charge = 0).

Although not detected in normal experience, the electron-neutrino is nevertheless a very common particle. (The identification of this neutrino with the prefix "electron" is an important distinction that will soon become apparent.) It is produced from many natural and humanmade sources, including the Sun, nuclear reactors, and particle accelerators. The difficulty in observing neutrinos is that they rarely interact with matter. Each second, both day and night, 60 billion neutrinos from the Sun pass through every square centimeter on Earth. Essentially all solar neutrinos incident on the Earth pass through it entirely. Thus it should not be surprising that it was not until 1956 that the neutrino partner of the electron was observed experimentally! When an electron-neutrino hits the constituents of a nucleus (as might happen when a solar neutrino is passing through the Earth), the result is that in addition to the nuclear constituents getting slightly rearranged, the neutrino can be "absorbed," and an electron can be produced. It is the observation of the electron that identifies the neutrino interaction.

The Subatomic Zoo

In the course of the twentieth century, many particles in addition to the proton, neutron, electron, and neutrino have been observed and classified by physicists. These particles are created naturally in very energetic reactions of interstellar protons smashing into atoms of the Earth's atmosphere and also in manmade collisions of high-energy particles in particle accelerators. The most observable characteristic of these particles is that they only exist for a fraction of a second before they decay into lower-mass, less energetic particles carrying off some of the parent particle's original energy. In the 1940s scientists studying the decays of these particles discovered one that had characteristics similar to an electron except that it was 200 times as massive. It was dubbed a muon. Like an electron resulting from a beta decay, the muon also appeared to be accompanied by a neutrino. In 1962, it was conclusively demonstrated that the neutrino partner of the muon was unique. It was not the same as the neutrino partner of the electron. This distinction was able to be made by an experimental verification that when a muon-neutrino interacted with a nucleus, a muon, not an electron, was produced. It was concluded that these elusive particles needed to be distinguished from each other and uniquely paired with their partners the electron and the muon. Collectively these four particles were labeled leptons. In the decades between 1940 and 1970, so many new particles were discovered that the term "subatomic zoo" became the best way to describe them.

In 1964 the theoretical physicist Murray Gell-Mann postulated that the baryons and mesons (general classifications of the subatomic particles that were not leptons) were composed of constituents called quarks. It could be shown that all of the observed particles could be constructed out of three different kinds of quarks, which whimsically were named up, down, and strange. Not surprisingly, it was found that everyday protons and neutrons were made out of the up and down quarks while the strange quark was a building block of the exotic short-lived particles. Experiments conducted in the 1970s confirmed the distributions of quarks in the proton and neutron. The proton contained three quarks, two with charge +2/3 (up quark) and one with charge -1/3 (down quark). The neutron, which had no charge, was composed of two down and one up quark. Amid the chaos of the subatomic zoo, a simple organization of fundamental constituents was emerging. This organization became more important as particles continued to be discovered and needed to be classified. Before the end of the 1960s the discovery of new particles required that the quark model be expanded to include a fourth quark which became a partner of the strange quark, just as the up and down quark were partnered. In the mid-1970s additional new particles required yet another extension to the quark and lepton picture.

The Third Generation and the Standard Model

In 1975, a third-generation charged lepton was produced and detected at the Stanford Linear Accelerator Center (SLAC) near San Francisco. This third lepton had properties that related it closely to the electron and muon with the exception that it was 3,480 times heavier than the electron and it lived for a very short time before decaying to lighter particles. The leader of the team of experimentalists, Martin Perl, named it the tau lepton (tau is a Greek letter, τ, and a convenient symbol for this lepton). As soon as the discovery of the tau lepton was announced, particle physicists extrapolated from experience and assumed that it would have a neutrino partner, the tau neutrino. Confirming the existence of the tau neutrino would be the obvious next step, but as will be discussed, a variety of circumstances would hold this confirmation at bay for more than two decades! Along with the realization that a third-generation lepton existed, it became obvious that the observed spectrum of baryons and mesons had expanded to need another set of quarks as well. The fourth quark, which had come to be called charm, was joined by a fifth named bottom. And although no particles had been observed needing a sixth quark to be added, the beauty of symmetry lead physicists to postulate the existence of a sixth quark that they named top, since it was the partner of the bottom quark. Like the tau neutrino, confirmation of the top quark was also going to take some time!

Setting experimental difficulties aside, it was clear that the model of fundamental quarks and leptons provided a simple way to organize the subatomic zoo and was worthy of being the cornerstone of a more general theory that might be able to explain the observed diversity in matter with a set of fundamental parameters. Hence the Standard Model of particle physics evolved.

Over the past thirty years, the Standard Model of particle physics has provided a very accurate description of the properties and interactions of the elementary particles called quarks and leptons. In the model there are six quarks and six leptons. Each quark or lepton is a member of a pair, and a pair of quarks plus a pair of leptons make up a generation. In addition, there are the particles that mediate the interactions between quarks or between quarks and leptons: the photon, and the W and Z bosons. With these particles one can "construct" all of the observed constituents of matter and describe the physical transformations that can occur, such as radioactive decay and nuclear fission and fusion processes. The Standard Model provides an explanation for the origin of the mass of the elementary particles as well as a logical framework for their classification. Quarks are fractionally charged particles (2/3 or 1/3) that are never found isolated and free, but are bound-up as three or two quark configurations. The leptons are electrically charged one unit or electrically neutral.

The Standard Model (SM) requires that each lepton or quark be paired with a partner. These partners share quantum mechanical properties that are incorporated into the SM. The up and down quarks form such a SM pair. The electron, the ubiquitous lepton in our world, has as its partner the electron-neutrino. The up-down quark pair, together with the electron and its neutrino partner, make up virtually all of the matter that is experienced in the everyday world.

In 1995 the top quark was discovered at the Fermi National Accelerator Laboratory (Fermilab) in Batavia, Illinois. This was a great achievement in twentieth-century physics. The top is the second of the pair of quarks comprising the third generation. It is not known if the third generation is the last one. The SM requires quarks and leptons to come in pairs, but it says nothing about how many generations there may be. Nevertheless, following the discovery of the top quark, the tau neutrino remained the only Standard Model particle whose existence had not been experimentally confirmed. By the mid-1990s the demonstrated success of the Standard Model had lead particle physicists to assume the existence of the tau neutrino. Experimental physicists realized that demonstrating its existence would be very challenging.

An Experiment Is Proposed

In order to show that a particle is unique, one needs to measure key properties that distinguish it from all others. For example, charge and mass are two key properties that identify an electron from all other leptons. But neutrinos have charge zero, and the masses of the electron and muon neutrinos are known to be very small and cannot be directly measured at all. The tau neutrino was anticipated to follow suit, so another property would be needed to be sure that one has observed a genuine tau neutrino interaction. The telltale signature for the interaction of a tau neutrino is creation of a tau lepton among other, less interesting, particles. This is necessary because both leptons of a given generation possess a quality that is preserved in the interactions of leptons. For the third generation, this characteristic quality is a "tau-ness" of each lepton in the pair, that is, the tau lepton and the tau neutrino. Similarly the second-generation leptons possess a "muon-ness" and the first an electronlike property. Physicists call such qualities additive quantum numbers, which is either possessed by a particle or not, there being no in-between. In fact, this quantum number is the distinguishing feature of each lepton generation and is strongly conserved, or kept, as these leptons interact or decay. The quark pairs have a similar quantum number for each generation, although the quarks are less strict in the conservation of it.

Thus identification of a tau lepton in the debris of an energetic interaction of a neutrino indicated that the neutrino was a third generation, tau neutrino. The problem facing experimenters was how to unambiguously determine that such a rare lepton was part of a reaction. Tau leptons are much heavier than the second generation lepton, the muon, which is much heavier still than the electron of the first generation. This fact, the massiveness of the tau (3,480 times heavier than the electron), leads to decay at a much quicker rate than the muon (207 times heavier than the electron). The muon, at rest, has a lifetime of 2 microseconds (2 millionths of a second), which is considered to be quite a long time in particle physics. The tau lepton exists for only 300 picoseconds (0.3 trillionths of a second)! A tau lepton produced at the energies available to Fermilab has a velocity very near c, the speed of light, and its apparent lifetime is measured to be about 10 times longer than when it is at rest. This is equivalent to the tau lepton traveling about 1 millimeter before it spontaneously decays into, for example, an electron, an antielectron-neutrino, and (another) tau neutrino. Since the tau lepton usually decays to just one other charged particle, a typical tau lepton event will have one track that travels about a millimeter and then will appear to change direction upon decay to another track. A rule of thumb in particle physics is that only charged particles can produce directly observable tracks, as only they can ionize the atoms in the matter through which they pass. Neutral (charge = 0) particles, such as gamma rays and neutrinos, leave no visible tracks in detectors. In summary, the detection of a tau neutrino interaction requires that the experimenter be able to recognize a tau lepton, which appears in the experiment as a track with a bend in it about 1 millimeter from the interaction point of the neutrino.

This experimental method for observing tau neutrino interactions was originally proposed at Fermilab in the early 1980s; however, the experiment was not carried out. In the peer review process of the proposal, it was deemed that the detector technology, at that time, may not have had the precision required to make a definitive observation while at the same time the construction cost of the experiment was more than $15 million. Since an important criterion in judging experimental proposals is the cost-benefit ratio, it was decided that the experiment should not be carried out at that time.

In 1994, twenty years after the discovery of the tau lepton, a group of forty physicists from the United States, Japan, Korea, and Greece proposed an experiment designed to uniquely establish the existence of the tau neutrino. As was true for the search for the top quark, the world's most powerful proton accelerator (at Fermilab) would be used to create the neutrinos. Specially designed detectors would be employed to create neutrino interactions and record their unique signature (production of a tau lepton), which would be evidence of tau neutrino interactions. This experiment was officially recognized as Fermilab Experiment 872, as it was the 872nd proposal received by a review committee. The experimental group replaced this rather dry moniker with the acronym DONUT, the Direct Observation of Nu-Tau ("nu-tau" is a short way of denoting the tau neutrino symbolically: υ_τ).

Using the technology available in the 1990s, recognition of the rare tau signature would indeed be possible, in principle, with several types of particle detectors developed for such precise measurements. The detectors at Fermilab were also almost completely transparent to neutrinos. There were two ways they could increase the number of these rare interactions: (1) Increase the number of neutrinos produced, and (2) increase the number of quarks in the experimental detector, that is, maximize its mass. The number of neutrinos created is limited by the accelerator, while the detector mass is limited by cost. It is very important to note that the detector must be much more than just mass. It also needs to be able to record a neutrino interaction with enough detail so that it may later be reconstructed accurately and so that the kind of neutrino that was captured within the detector may be identified.

The DONUT Detector

In DONUT the heart of the experiment was a detector consisting of 260 kilograms of nuclear emulsion acting essentially like photographic emulsion or film. An image of each charged particle is formed as it traverses the emulsion, which is made by ionizing silver-halide crystals. Just as in the case of photographic film, the development process amplifies the original, or latent image, so that a particle track passing through the emulsion creates a series of tiny black grains, each less than 1 micrometer (0.00004 of an inch) in diameter. This precision is critical in constructing a picture of the neutrino interaction and resolving the tau lepton track from the rest of the tracks created in the interaction. Using emulsion as the primary particle detector was the key element in the success of the experiment. This was far from the first use of emulsion in particle physics. Emulsion was used extensively in the early years, from the 1940s into the 1960s, and many important results came from the analyses of these early experiments. In the 1970s, emulsion was largely replaced as a detector because the advance in low-cost, sensitive electronic detectors that enabled the data to be easily digitized and stored for computer assisted analysis. Emulsion, by contrast, had to be scanned by human operators, which was a slow process. Nuclear emulsion is also very expensive and unforgiving in the sense that it records everything passing through it, both unwanted background particles as well as physically interesting events.

The use of emulsion in DONUT was made practical by the application of computers and fast digital imaging technology. In fact, the emulsion images themselves are rarely seen directly by eye in DONUT. Tracks recorded in the emulsion are stored on computer disk arrays and are analyzed using the same numerical techniques developed for the electronic detectors of the 1970s and 1980s. The whole process is like a delayed electronic detector, and the emulsion serves as the initial storage medium for the track data. The nuclear emulsion serves as both the recording medium for interesting interactions and as the target material for the beam of neutrinos passing through it.

The Making of a Tau Neutrino Beam

The decision to use an emulsion target and detector was a clear one for the DONUT experimental group. Physicists from Nagoya University in Japan were the world leaders in modern emulsion technology, and the target design and construction became

FIGURE 1

their responsibility. The design of the neutrino beam rested with the DONUT members from Fermilab. The basic principle in making a neutrino beam is simple: smash as many high-energy protons as possible into a large block of metal. This block is called the "beam dump" and is cooled to prevent melting from the heat generated by the intense proton beam. Neutrinos result from the decays of particles created in the collisions of the protons in the metal block. All of the rest of the particles produced in the dump are absorbed by material following the collision point. Thus, one is left with a beam of neutrinos, since they cannot be significantly absorbed by ordinary matter. In the design of a practical beam for DONUT, one that would fit within the allocated budget of $1 million, several compromises needed to be made.

First, the distance from the beam dump to the emulsion had to be made as short as possible in order to have most of the neutrinos pass through the emulsion and so have a chance at interaction. Second, because of the first point, many particles created in the beam dump will not be absorbed as there is simply not enough shielding material between the beam dump and emulsion. Most of these unwanted, non-neutrino particles were muons, the charged leptons of the second generation. If nothing special were done to eliminate these muons, there would be so many charged tracks recorded in the emulsion that it would be ruined after only one minute of exposure to protons in the beam dump. To make an analogy to photography, the emulsion detector would be overexposed. The muons were largely swept clear of the emulsion area by a large magnet located immediately after the beam dump. The magnetic field pushed positive muons to one side of the emulsion and negative muons to the opposite side. Since a failure of this magnet would be catastrophic, an electronic fail-safe protection circuit was used for this system. In addition to the magnet, a thousand tons of steel was set between the magnet and emulsion to absorb many products of the interactions in the beam dump. The total system of magnet and steel acted as a shield to protect a small area. The distance from the beam dump to the emulsion was about 36 meters. At this distance the backgrounds of muons and other particles was just tolerable, and about half of all the tau neutrinos that could be seen would be intercepted. This was considered a good compromise. Calculations based on data from other experiments indicated that 5 percent of the neutrinos in the neutrino beam were tau neutrinos, and the remaining 95 percent were the "ordinary" muon-neutrinos and electron-neutrinos, with approximately equal numbers.

The high-energy proton beam was first brought to the beam dump in November 1996, although the first emulsion module was not installed until April1997. During the first five months of beam (without emulsion), the area around the emulsion location was instrumented to detect muons, neutrons, and gamma rays to test the waters. Several adjustments were made to the shield to correct weak places in the shield. When the level of background flux passing through the emulsion area was acceptable, the emulsion target was installed, the proton beam turned on, and data were recorded. The total number of accelerated protons used in DONUT to make the neutrino beam was 5 × 10¹⁷. About 2 × 10¹⁵ neutrinos were manufactured (10¹⁴ tau neutrinos), of which about 1,000 interacted in the emulsion target (about fifty tau neutrino interactions).The emulsion was not the only detector in DONUT. Electronic charged particle detectors were used to

FIGURE 2

record tracks after leaving the emulsion. These detectors were essential to pinpointing the position of the interactions inside the emulsion because only0.01 percent of the total volume can be digitized per year using the best available technology. Only a relatively small volume, about 0.4 cubic centimeters, was digitized and recorded for each interaction. Unfortunately, about 30 percent of the time the position predicted using the electronic detectors was in error so that the interaction point in the emulsion was missed.

There were also several other factors causing interactions to be missed, all of which reduced the overall efficiency for seeing tau neutrino events. The net efficiency for finding a tau neutrino interaction within the emulsion was estimated to be 40 percent. In January 2000 203 interactions had been found in the emulsion (out of 1,000 that existed). Therefore, from the above fraction of expected tau neutrinos times the efficiency, it was expected that about five events should be in this set. Of these five events only four were long enough for both the parent tau and the daughter track to be seen in the detector. What remained in the analysis of the set of 203 events was the decay search: the hunt for a few bent tracks among approximately one thousand tracks.

Four tau events were expected, with an uncertainty of less than one event. This uncertainty was determined from earlier experimental results that were part of the estimate for the number of taus. Four may seem to be a small number, but it is really quite significant in the sense that the probability of seeing zero events when the expectation is four is only 1.8 percent. Conversely, the probability of seeing at least one event is 98.2 percent: very good odds. A more important factor in DONUT was the amount of background that accompanied the true tau events. Great effort was taken to understand all the details and nuances in the analysis so that the number of events that only looked like tau neutrino interactions, but really were not, was small and well understood. In May 2000, the analysis was nearly completed, and four events, each with the telltale bent track, were the result. This was in comfortable agreement with the predictions, but what was the level of background within this signal? A carefully constructed software model was used to determine the background, simulating the neutrino interaction physics as well as the DONUT detector. The number turned out to be 0.36 events. This number is to be viewed as a statistic. Given this average number of background events in a sample of 203, what is the probability that the actual number of background events is four? That is, what is the likelihood all the tau events were actually just background, just a fake? This is easy to calculate, and turns out to be about 4 in 10,000. The ratio of the number of signal events divided by number of background events is an important figure of merit. This ratio is large in DONUT, and the physicists on DONUT unanimously announced the discovery of the tau neutrino in July 2000 in a special seminar at Fermilab.

For twenty-five years particle physicists had assumed that the tau neutrino existed as the neutrino partner of the third-generation lepton, the tau. The tau neutrino was the last of the leptons and quarks to be confirmed, and the Standard Model remained untarnished. There are still pieces of the Standard Model that are not yet in place, and some important questions remain. There are many properties of the ghostly neutrinos waiting to be discovered, one of the most compelling being "Do neutrinos have mass?" It is safe to assume that new results in neutrino physics will continue to be announced, each providing one more clue about an elusive, but important, part of our universe.