CASE STUDY: LONG BASELINE NEUTRINO DETECTORS, K2K, MINOS, AND OPERA

Neutrinos are elementary particles, which are still very poorly understood, mainly because their interactions with matter are very weak and thus difficult to study. Neutrinos come in at least three different species, electron, muon, and tau. If neutrinos have mass, they can oscillate between different species as they propagate through vacuum or matter, that is, a beam of initially pure muon neutrinos can develop an electron or tau neutrino component as it travels through space. The distance over which this happens, that is, characteristic wavelength of oscillations, is inversely proportional to difference in mass squared between different kinds of neutrinos and directly proportional to the neutrino energy.

There is evidence for the possible existence of oscillations from studies of neutrinos produced in the Sun and of neutrinos produced in the Earth's atmosphere. Efforts are under way to study both of these phenomena in more detail and under more controlled conditions. The phenomena observed in solar neutrinos appear to have mass squared difference such that they are best studied with reactors, which produce low-energy neutrinos. The atmospheric neutrino phenomena appear to have a mass squared difference such that they can be best studied with accelerator neutrinos and source to detector distance of the order of several hundred kilometers. They are generally referred to as long baseline neutrino oscillation experiments.

Currently there are three accelerator experiments, running or in preparation, designed to study atmospheric neutrino phenomena. They are K2K in Japan, OPERA in Europe, and MINOS in United States. Even though their general goal—study of oscillations in the atmospheric neutrino region—is the same, their specific goals are different, and thus they require different beams and detectors.

Neutrino Production

Neutrino production in an accelerator begins with acceleration of protons. All three experiments under discussion use a circular accelerator: KEK Proton Synchrotron (PS) in Tsukuba, Japan, with a peak energy of 12 GeV; Super Proton Synchrotron (SPS) in the European Laboratory for Particle Physics (CERN), spanning the French-Swiss border region near Geneva, with a peak energy of 400 GeV; and the Main Injector (MI) at Fermilab in Batavia, Illinois, with a peak energy of 120 GeV.

Once the peak energy is reached, the proton beam is extracted from the accelerator and allowed to impinge on a target. The resulting proton interactions with the nuclei in the target produce secondary particles, the most numerous of which are pi mesons, and they are the parents of most of the neutrinos used in the experiments. Most of the secondaries produced travel in a rather wide cone in a forward direction, and thus if nothing were done to contain them, they would disperse over a large area. To avoid this, a focusing system is constructed for the pions with energy of interest, which compresses those with energy of interest into a much narrower cone, much like the parabolic mirror in a flashlight focuses the beam of light from its lightbulb. The focusing system consists of magnetic horns, which are specially shaped conductors through which a current is pulsed when the beam arrives. The resulting magnetic field bends most of the desired pions into a relatively parallel beam.

The pion beam is then allowed to enter a decay volume, an evacuated pipe with length of the order of 100 to 1,000 meters. As the pions travel through this pipe, some of them decay into a muon and a muon neutrino. The neutrino is emitted into a very narrow cone along the pion direction. The length of the decay pipe is generally proportional to the chosen pion energy, which in turn is related to the wanted neutrino energy: the very forward decay neutrinos take 42 percent of the pion energy. The decay volume is generally followed by earth shielding to stop the undecayed pions and the decay muons. Neutrinos, being very weakly interacting, pass easily through the earth shield and continue on to the detector. Besides the far detector, located several hundred kilometers away, long baseline experiments also frequently have a near detector, located just downstream of the absorber shield. Its purpose is to measure the properties of the neutrino beam before the neutrinos had a chance to oscillate.

Detectors

Neutrino detectors tend to be very massive so as to obtain a sufficiently large sample of neutrino interactions in spite of their very weak interactions. Large size is even more important in long baseline experiments since their detectors are far from the source, and thus the neutrino beam has diverged significantly before it arrives there. The detailed design of the detector depends on the specific goal of the experiment and sometimes on special circumstances.

All three experiments under discussion here locate the detectors deep underground, that is, on the order of a kilometer or more below the surface. The reason for this is to suppress the cosmic ray background, which on the surface is about 100 particles/s m². Because there will be relatively few neutrino interactions, one must suppress all possible sources of background as much as possible. Additional background suppression is obtained from the directionality of the events observed and the time at which they occur, since the accelerator beam is pulsed and on only a fraction of the time.

Detectors are built by the collaborations doing the experiment. Once its general structure is agreed on and the design complete, different institutions assume responsibilities for different subsystems of the detector. Frequently those subsystems are built and tested at home institutions far away from the detector site. The lifetime of a typical neutrino experiment is long, several years or more, and this required longevity is one of the factors influencing the design.

Collaborations

The formation and operation of collaborations performing long baseline neutrino oscillation experiments is qualitatively very similar to what happens in other large-scale particle physics experiments. These experiments are complex and of long duration; it takes a large group of scientists and engineers to perform them. A collaboration typically involves 100 to 200 people coming from fifteen to forty different institutions, primarily national laboratories and universities. The collaborations tend to be international in scope with the majority of institutions from the general region where the experiment is being performed.

The experiments tend to be initiated by a smaller number of scientists brought together by a common scientific interest. Frequently these people have worked together before or are working on a common experiment at the time of the new proposal. This group will generally be too small to construct and carry out the experiment but will be large enough to do the initial preliminary design to establish viability of the experiment. It may then take several years to obtain scientific approval and the required financial resources from the host laboratory and/or the appropriate funding agencies. These steps are usually somewhat different in different regions (United States, Western Europe, or Japan) and are frequently influenced by the potential existence of some relevant infrastructure, for example, beam line or detector.

Once this initial phase is completed, the collaboration will grow to the required size. Other groups may be invited to join since they may possess required expertise for some subsystem of the detector; alternatively, groups can express interest in participation on their own. In parallel, a detailed design of the required apparatus and software is made and responsibilities for different subsystems are assigned to specific individuals and/or groups. During the next phase (research and development [R&D], prototyping, testing, and construction), the work is done at home institutions. The different groups then deliver their hardware to the beam or detector site where the whole system is put together with the participation of members of various institutions. The checkout and subsequent data taking follows afterward and is a collaborative effort with subsystem experts playing a prominent role.

Early in its life the collaboration adopts a constitution defining governance and decision making. This will vary in different collaborations, but generally there will be one or more spokespersons who act as representatives of the collaboration vis-à-vis the outside world and as CEOs of the collaboration, a policy-making board composed of senior members of the collaboration, and an institutional board where all institutions have a voice. The role of these groups, the method of their selection, and their term of office are spelled out in the constitution. In addition, different individuals are given formal responsibilities for different technical components, and the overall technical direction may reside in a group of individuals forming a technical board.

Frequently there is a parallel structure with fiduciary and management responsibilities, which is appointed by and reports to the host laboratory that provides the funding. This organization is headed by a project manager, frequently reporting directly to the laboratory director. The project manager is responsible for the appointment of individuals to head the work on different subsystems. The project thus relieves the collaboration as an organization from most of the management responsibilities in the technical area even though the individuals assigned these responsibilities will be members of the collaboration and come both from the host laboratory and other participating institutions. The project organization is generally formed during the latter part of the design phase and exists for the duration of the construction of the experiment.

The communication between members of the collaboration occurs in a variety of ways: collaboration meetings that generally occur several times per year as well as meetings, conferences, and workshops organized by smaller subgroups of the Collaboration. They can be face-to-face meetings or via telephone or videoconferences. The subgroups tend to be organized around specific technical subsystems during the construction phase and around a physics or software topic during the data taking and analysis phase. The Internet plays an important role in collaboration communication.

K2K Experiment

The K2K experiment uses the neutrino beam created by the KEK (National Laboratory for Particle Physics in Japan) proton synchrotron and the Super-Kamiokande detector about 250 kilometers away. The latter is located in a working zinc mine with about 1,000 meters overburden of rock and earth above it. Super-Kamiokande consists of a tank filled with 50 kilotons of purified water, covered on its inside surface with about 11,146 twenty-inch photo-multipliers (PMTs). Neutrino interactions produce charged particles, most of which emit Cerenkov light in a cone around the trajectory of the particle. This light is detected by the photomultipliers, and the nature of the event is subsequently deduced from the pattern of the PMT hits.

The origin of this experiment is somewhat different from those of MINOS and OPERA insofar that the detector already existed, having been built previously for other reasons: the study of solar and atmospheric neutrinos, the detection of neutrinos from future supernovas, and the search for nucleon decay. The additional elements required were a neutrino beam line and a near detector, requiring significantly less resources than the Super-Kamiokande demanded originally. Thus the experiment is quite cost effective. The K2K Collaboration, composed mainly of Japanese and U.S. groups, has a very strong overlap with the Super-Kamiokande collaboration. Thus, many of the steps generally required for initiating an experiment were avoided in this case since they had been taken earlier for the Super-Kamiokande experiment.

The main goal of the experiment is verification of the existence of oscillations in the atmospheric neutrinos with a well-controlled accelerator experiment. The atmospheric neutrino studies indicate that muon neutrinos, produced at the top of the atmosphere, are depleted as they travel through the Earth and/or atmosphere. The dependence of the effect on the zenith angle of the neutrinos, equivalent to the total pathlength traveled, strongly favors the oscillation interpretation. The K2K beam is essentially a pure muon neutrino beam, whose characteristics, such as intensity and energy, can be calculated and also verified in a near detector. The researchers try to see whether the observed neutrino interaction rate is different from the no-oscillation prediction and if it is, to determine the energy dependence of the effect.

The projected event rate is modest, about two hundred observed events in a four-year run without oscillations, less if oscillations exist. The modest rate is due to the relatively low intensity of the KEK accelerator and low energy of the neutrino beam, peaking around 1 GeV. The experiment commenced taking data in 1999 and was scheduled to run through 2003. Toward the end of 2001, an implosion occurred in the Super-Kamiokande tank that destroyed over 50 percent of the photomultiplier tubes used in the detector. Restoration of the experiment, with a less dense photomultiplier coverage, is underway. Data taking is scheduled to resume at the beginning of 2003.

OPERA Experiment

The OPERA experiment involves a neutrino beam produced by the CERN SPS and a detector to be located in the Gran Sasso Laboratory in Italy, 732 kilometers away. Both the beam and detector are being constructed specially for this experiment.

The goal of the OPERA experiment is the observation of interactions of tau neutrinos, which are expected to be the main end product of muon neutrino oscillations. In other words, a relatively pure muon neutrino beam produced at CERN slowly develops a tau neutrino component as it travels through the Earth. The main challenge for the experiment lies in the fact that tau neutrino interactions are difficult to identify. The unambiguous signature of a tau neutrino interaction is the production and decay of a tau lepton. Because the tau lepton is very short lived—in the OPERA experiment its typical length is of the order of 1 millimeter—the detector has to have very good spatial resolution. At the same time it has to be massive to observe a significant number of events. Accomplishment of these two goals simultaneously is difficult.

The basic elements of the OPERA detector are modules composed of sandwiches of sheets of iron and photographic emulsion coated plastic. The iron provides target material; the emulsion provides a detecting medium with one micron resolution and hence the capability of observing tau events. The experiment presents a number of challenges of which the most formidable are identification of the interaction within a small volume (of the order of a few cm³) and the need to process—scan and measure— very large volumes of emulsion. The former is handled by interleaving electronic detectors in between the layers of iron/emulsion modules, information which allows one to locate the vertex. The ability to handle the second challenge is the result of many years of development of automated emulsion scanning and measuring techniques, principally by a group in Nagoya, Japan.

A neutrino interaction is identified by the electronic detectors and in most cases can be localized to a given iron/emulsion module. Periodically, roughly once a day, the modules with neutrino interactions are pulled out of the detector and developed underground so as not to contaminate the emulsion with cosmic rays. The emulsion in the vicinity of the identified vertex is then scanned quickly to search for possible evidence of tau production and decay. The potential tau candidates are subsequently subjected to additional and more sophisticated analysis.

Simulations show that tau events can be identified with negligible background from other, non-tau, neutrino interactions. This is essential to the success of the experiment since the tau neutrino production rate and detection efficiency is such that for the oscillation parameters suggested by the Super-Kamiokande results, one can expect only about twenty observed and identified events in five years of running. It is hoped that the experiment will start data taking in 2005, but the financial situation at CERN may necessitate delay. The collaboration consists of thirty-three groups as of 2002, mainly from Europe and Japan.

MINOS Experiment

The MINOS experiment uses the neutrino beam from the Main Injector accelerator at Fermilab and a detector in the former iron mine in Soudan, Minnesota, 700 meters underground and 735 kilometers away. The mine is currently run as a state park, and the experiment relies on the infrastructure provided by the park; an additional cavern was excavated to house the MINOS detector.

The main goal of the experiment is to measure the oscillation parameters by studying the disappearance of muon neutrinos. A nearby detector on the Fermilab site is used to measure the properties of the neutrino beam. Those measurements allow prediction of what the beam will be at the Soudan location. Deviations from the predicted intensity and energy spectrum would be evidence for oscillations, and their quantitative study would determine oscillation parameters (for example, difference in mass squared of different neutrino states) with a precision of about 10 percent.

The neutrino beam is designed so that it can provide beams with different energies by changing the location of the target and the focusing elements. The energy range that can be covered extends from 1 to 20 GeV. The initial running will be with a low-energy beam configuration that has a spectrum centered at about three GeV, since that configuration appears to be best for the measurement of oscillation parameters suggested by the Super-Kamiokande experiment. One expects to see about 1,000 neutrino interactions per a year.

The 5.4-kiloton far detector will be composed of 486 alternating layers of 1-inch steel plates and scintillator planes in the shape of 8-meter wide octagons. The scintillator planes consist of 4.1-centimeter wide strips and enable both positional information and energy information to be obtained. The iron will be magnetized to enable the muons produced in the neutrino interactions to have their energies measured by curvature.

The detector assembly commenced in July 2001 and should be completed in 2003. The conventional construction at Fermilab should be finished sometime in 2003 at which point the installation of the beam technical components and the near detector will start. The first beam is expected toward the beginning of 2005.

The MINOS Collaboration is composed of about 175 scientists and engineers from thirty institutions in five countries. The majority of the institutions are in the United States and in the United Kingdom.

Data Analysis

In most particle physics experiments the data arrive as electronic signals. After some processing and filtering online, the data are stored on some kind of mass storage device for offline analysis. Generally the host laboratory acts as a repository for the data, but the data are readily available to the collaborating institutions. This pattern will apply to both K2K and MINOS experiments. Because the number of events involved will be much less than for a typical particle physics experiment, the data handling issues here are relatively simple.

In the OPERA experiment the situation is more complex because the essential information consists not only of the digital data from electronics but also of the pattern of developed grains on the photographic emulsion. Thus the data analysis will involve a considerable amount of scanning and measuring of emulsions before all the data can be reduced into digital format. The scanning and measuring phase will take place at the home location of several of the collaborating institutions.

The collaborations as a whole organize the data analysis with specific responsibilities assigned to different institutions and/or individuals. This division is generally organized around physics topics, but in the initial stages different groups focus on work needed to understand different subsystems and features of the detector, such as efficiency, resolution, systematic uncertainties, etc. Graduate students generally assume responsibility for some physics topic, which then becomes the subject of their Ph.D. thesis.

See also:Detectors and Subsystems; Neutrino

Bibliography

Ahmad, Q. R., et al. "Measurement of the Rate of ν_e d → p + p + e ^- Interactions Produced by ⁸B Solar Neutrinos at the Sudbury Neutrino Observatory." Physical Review Letters87 , 071301-1–071301-6 (2001).

Apollonio, M. et al. "Limits on Neutrino Oscillations from the CHOOZ Experiment." Physics Letters , B466 , 415–430 (1999).

Bahcall, J. N.; Calaprice, F.; McDonald, A. B.; and Totsuka, Y. "Solar Neutrino Experiments; The Next Generation." Physics Today49 (4), 30–36 (1996).

Boehm, F., and Vogel, P. Physics of Massive Neutrinos (Cambridge University Press, Cambridge, UK, 1992).

Fisher, P.; Kayser, B.; and McFarland, K. S. "Neutrino Mass and Oscillation." Annual Review of Nuclear and Particle Science49 , 481–527 (1999).

Fukuda, S., et al. "Solar ⁸B and hep Neutrino Measurements from 1258 Days of Super-Kamiokande Data." Physical Review Letters86 , 5651–5655 (2001).

Fukuda, Y., et al. "Evidence for Oscillation of Atmospheric Neutrinos." Physical Review Letters81 , 1562–1567 (1998).

Kayser, B.; Gibrat-Debu, F.; and Perrier, F. The Physics of Massive Neutrinos (World Scientific, Singapore, 1989).

Kim, C. W., and Pevsner, A. Neutrinos in Physics and Astrophysics (Harwood Academic Publishers, Langhorne, PA, 1993).

Klapdor-Kleingrothaus, H. V., and Staudt, A. "Neutrino Oscillations" in Non-accelerator Particle Physics (Institute of Physics Publishing, Philadelphia, 1995).

Nakamura, K. "Status of K2K." Nuclear Physics B (Proc. Suppl.)91 , 203–204 (2001).

Rubbia, A. "ICANOE and OPERA Experiments at the LNGS/CNGS." Nuclear Physics B (Proc. Suppl.)91 , 223–229 (2001).

Wojcicki, S. "Status of the MINOS Experiment." Nuclear PhysicsB (Proc. Suppl.)91 , 216–222 (2001).

Stanley G. Wojcicki