Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
6 . Instr uments and Detectors for Elementary-Particle Physics INTRODUCTION Elementary particles cannot be seen directly; their path and energy as they come out of an accelerator or out of a collision must be determined indirectly. It is also important to identify the type of particle: electron, muon. proton, or photon, for example. Thus the detectors, which determine the path, the energy, and the particle type. are often complex. The development and construction of these detec- tors, and the analysis of the data produced, are the province of the particle experimentalist. The strong interest in rare processes (such as W and Z production in the recent CERN experiments) and the need to characterize events completely have led to the development of detec- tors sensitive to almost the total solid angle. At all accelerators but particularly at particle colliders it is essential to provide detectors capable of making the fullest use of the particle beams. A wide range of detectors exists including the small but often sophisticated instru- ments designed for fixed-target work, the large detectors used tor recording rare events such as the interactions of neutrinos or the decays of nucleons, and the large collider detectors that provide almost complete angular coverage and characterization of the interactions occurring in these machines. This latter class of collider detectors 1` among the most costly and demanding. Their technological problems and solutions are, in large part, shared with the other classes fit 132
INSTRUMENTS AND DETECTORS 133 detector. The following discussion. for simplicity. will largely concen- trate on the development, construction, and needs of this collider detector class. However, later in this chapter in the section on Detectors in Fixed-Target Experiments we present some highlights of the detectors used in these experiments. And in the final section of this chapter, we discuss experiments that do not use accelerators. As an introduction to large detectors. we present the Mark 1 detector first used at SPEAR in 1972. It was the first electronic detector for a particle collider with close to full angular coverage and with a magnetic field to provide momentum and energy measurements for charged particles. A view of the Mark 1 is shown in Figure 6. 1, and a reconstruction of a ~ (psi) event is shown in Figure 6.2. This detector was sophisticated for its era, allowing the discovery of the ¢, the tau lepton. and charmed particles. A bare decade later comparably important results have begun to flow from the detectors at the CERN proton-antiproton collider with the discovery of the Z and W particles. Again these discoveries and the realization of the potential of the accelerator are only made possible by the sophistication of the detectors. The enormous advances in detector technology over the last decade are best illustrated by contrasting the view of a ~ event in the Mark 1 detector with the enormously more detailed picture from the UAI detector at CERN of an event with a Z°, shown in Figure 6.3. The actual configuration of the immense 5000-ton UA1 detector is shown in Figure 6.4. Fixed-target detectors usually are more selective than the large collider detectors. The neutrino detectors are vast instrumented tar- gets, usually incorporating magnetic analysis of produced muons and calorimetric measurement of produced hadrons. The target of a Fermilab neutrino detector, shown in Figure 6.5, has an instrumented mass of 690 tons, consisting of 20-cm iron slabs interleaved with drift chambers, followed by 420 tons of momentum-analyzing toroidal magnets. Such large masses are required to achieve a reasonable interaction rate from the weakly interacting neutrinos. In fixed-target experiments, the kinematics of high-energy relativistic collisions re- sults in most of the final-state particles from an interaction being thrown forward into a relatively narrow cone. Consequently, fixed- target experiments generally appear as a linear sequence of detector elements downstream from the target, as Figure 6.5 clearly illustrates. We return to fixed-target detectors in a later section. The larger detectors constitute major facilities, with a lifetime of usage of typically more than 10 years. They cost in the range of $10 million to $50 million and compete for resources with even the large
134 ELEMENTARY-PARTICLE PHYSICS MLION SPARK CHAMBERS FLUX RETURN SHaNER COUNTERS ENn UP , . 'it -coal ~~ At',, \ \ COMPENSATING SOLENOID \ ~VACWM CHAMBER ~LLtMl'JOSITY MONITOR TRIGGER COUNTERS SPARK CHAMBER PIPE COUNTER ~ 1 MUON WIRE CHAMBERS I I RON ( 20 cm ) ////~ ~ / / /; ~ // '~ I meter (a) _SHOWER COUNTERS (24) D/~f~ TRIGGER COUNTERS (48) ad////< ^/f,~ ~ CYLINDRICAL \~ ~ ii<~ J WIRE CHAMBERS ~ ~ \~ N~\ ~ W' 1l 11 1l '~TRIGGER BEAM PIPE ~JJJJ¢~' COUNTERS ( 4 ~ //oil lJ //},_ \\k / \ ///D (b) FIGURE 6.1 (a) The first general-purpose particle detector built for use at a particle collider was the Mark I detector shown here. (b) Cross-sectional view of the Mark 1 detector.
INSTRUMENTS AND DETECTORS 135 . 77 Ape+ - ·/ . _~ _ I-' It N_ / FIGURE 6.2 A computer reconstruction of an event found by the Mark I detector in which a +' particle decays to a ~ particle plus two pions; the ~ then decays to two electrons. The ¢' particle was in the small circle at the center. and the four lines coming out indicate the paths of the four particles produced in the decay. The lower-energy pion tracks are curved more strongly by the magnetic field. This particular picture became well known because the four paths also happened to form the Greek letter ¢. parent accelerators. Unlike the early pioneering experiments of parti- cle physics, a modern experiment may well require the simultaneous collaboration of several hundred physicists from 20 or more institu- tions. These major facilities require resources comparable with those used in the construction of the parent accelerator. DETECTOR REQUIREMENTS AND PHYSICAL PRINCIPLES OF DETECTION A detector system should be able to measure, as completely as possible, all the characteristics of the produced particles within an event. This implies that the detector should function over the largest possible angular range, measure with the best attainable precision, be provided with instrumentation to identify particle characteristics, and simultaneously provide a wide range of cross checks to protect against measurement artifacts. Additionally for use with hadron colliders, detectors must be able to extract interesting classes of physics events from backgrounds of events perhaps a hundred million times more frequent. These challenges must be met while simultaneously keeping
136 ELEMENTARY-PARTICLE PHYSICS EVENT 7939. 1 00 1. ~ a) ~ TV I \ ,~// l EVENT 7933. 1001. b) \ ~ . ~ /e. FIGURE 6.3 A computer reconstruction of an event produced at the CERN' proton- antiproton collider; this event includes the production of a Z° particle. in a) the tracks of all the particles produced in tile event are shown. In b) the two tracks of the electron and positron produced in the event are shown by themselves; this electron-positron pair comes from the decay of the Z°.
INSTRUMENTS AND DETECTORS 137 /// ., , \ ~_~ / '- -, , ~ 1 ~ 'a lr ,1] IF . 1 la 1. 1~= - -1- 1 FIGURE 6.4 The UAI defector et CERN which produced the event shown in Figure 6.3. Note the enormous size of the detector compared with that of the person standing at its lower right side. the combined costs of construction, operation, and data-reduction within reasonable bounds. The basic physics underlying detector operation can be summarized as follows: · Charged particles lose energy by ionization processes and leave a track or trail of ionized atoms and electrons as they pass through gasses, liquids, or solids. A wide range of techniques serves to measure the position and magnitude of these ionization trails. The magnitude of this energy loss per unit length is a measure of particle velocities. · Velocities of particles may be determined from the time interval required to pass between two points. In general this technique differ- entiates velocities only for relatively small particle energies. · In the presence of a magnetic field, charged particles are deflected into curved orbits. Measurement of this curvature permits the mo- menta of these tracks to be determined. The particle's energy can be calculated from its momentum if the mass is known. · Characteristic, but weak, radiation is coherently emitted by par- ticles passing through material Cerenkov radiation; or by particles as
138 · to - o S ~ - of £ o o ~ S EN 8 ~ - . ~ ° ._ ~ ._ o ~ ~ or ·11 TO or l IN Ct C-] O ~ C} Z ¢8 \ PP tD O O tD \ - C' - ._ a, C - 3 c C) ._ U' CC - E - U-. ~ C ·D - ._ - E c ._ o a. ~ C ~ _ ._ C it; C, Z o ~ _ = 4.} Z _ Ct 4) ,,0 Ed . - ~ ~, - Ct
INSTRUMENTS AND DETECTORS 139 they cross an interface between different materials transition radia- tion; or as they pass through magnetic fields synchrotron radiation. The intensity and characteristics of these radiations can serve as the basis of a velocity measurement. · When electrons or photons pass through matter they produce characteristic electromagnetic cascades of secondary radiation, which in turn leave an intense core of ionized atoms. The energy of the original electron or photon is ultimately completely converted into ionization by these processes. Measurement of this converted energy in a sufficiently thick block of material determines the total incident electromagnetic energy and constitutes a calorimetric shower-energy measurement. A detector constructed to make use of this property is an electromagnetic calorimeter. · Hadrons, such as protons and mesons. interact strongly as they pass through matter, producing secondary hadrons. This again results in the production of intense cores of ionized atoms. The total energy of the incident particles can be measured from the total energy deposited in the form of ionization. Hadronic cascades can be differentiated from the electromagnetic cascades that develop in much thinner layers of material. The technique of measurement is known as hadron calorim- etry. Typically a hadron calorimeter might require a thickness of 3 to 4 feet of instrumented steel with a total weight of hundreds of tons. · Energetic muons are uniquely characterized by the property that, although charged, they penetrate large thicknesses of material and emerge with a relatively small change of energy at the outside of a detector. DETECTORS FOR COLLIDER EXPERIMENTS Modern detectors typically make use of all or many of the above properties to characterize detected events. The characteristics of these detectors have many features in common and share similar design architectures. The detectors for collider experiments are based on a series of concentric shells or layers, one behind the other, each of which is devoted to some particular aspect or aspects of the detection process. The initial detector layers are used to characterize the charged-particle component and are designed to be nondestructive, i.e., to contain little material so that charged particles will not interact or degrade in energy and thus will maintain their identity while traversing the layers. The outer detector layers deliberately use large amounts of material in order to materialize the neutral particles and to
140 El EMENTARY-PARTICLE PHYSICS convert the energy carried by the particles into detectable ionization; such a device is known as a calorimeter. . The outer layer of the calorimeter or an additional detection layer is frequently used to detect muons. As mentioned earlier, energetic muons are usually the only particles that can reach this outermost layer. Inevitably as the layered levels of detection systems are built up, a detector will become large and correspondingly complex and expen- sive. A prime objective of detector development is, therefore' to keep detection systems as compact as possible and to combine detection roles whenever possible. Additional demands are imposed on detector systems associated with hadron colliders by the high ambient radiation levels at the detector and by the fact that events of interest may be separated only by short times from uninteresting background events. Summarized below are the elements or layers constituting a typical modern detector system and some of the ongoing research and development aimed at maximizing present or future detector capabili- ties. . . . Close-in Detection: Vertex Detectors A fraction of the particles emerging from a collision point decay in flight at distances as close as 0.001 cm. Such decays provide charac- tenst~c signatures as to the nature of the decaying particles. Therefore use can be made of charged-particle detectors with high spatial resolution that are placed as close to the interaction point as possible. The first such vertex detector for collider work was recently con- structed to operate with the Mark It detector at the PEP collider at the Stanford Linear Accelerator Center. Several such detectors are now being constructed or are operating at electron-positron and proton- antiproton colliders. The first generation of such vertex detectors was h~cer1 on ~nn~ren- tional track-detection methods for charged particles, using multiwire drift chambers or time-projection chambers (TPC). (The principles of operation of these track detectors are described below.) The chambers are typically only about 10 cm in radius, are fabricated with fine subdivisions to provide separation between adjoining tracks that might otherwise overlap, are aligned with great precision (about 0.009- to 0.005-cm tolerance), and ultimately are likely to be operated under high pressures to provide sharp internal localization of the trail of ionization left by the charged particles. The limits of precision for such detection __ ~ W~ ~ v~ vim Al ~~!
INSTRUMENTS AND DETECTORS 141 are at present about 0.01 cm; this should eventually improve by a factor of 2 or 3. The second-generation vertex detectors now under development are based on modern silicon semiconductor technology. This technology has already been used successfully on a small scale in experiments designed to measure decays of short-lived particles. which decay close to the parent event. In these experiments the target is constructed of microstrips' and a detailed history of events occurring within these active targets can be recorded. Prelirrunary tests have been made and have established the feasibility of the proposed vertex detectors. Full-scale detectors should come into operation during the next 5 years. Three approaches are being tried: ( 1 ) the use of packages of long thin silicon strips with widths of about 0.009 cm (2) the use of a mosaic of semiconductor squares as currently exist in the charge-coupled devices (CCD) that are used as a basis for image detection in astro- nomical and other applications, and (3) a more speculative idea that involves drifting the ionization over a relatively long distance within the silicon. The handling, precision alignment, and electronic readout of such miniaturized devices present fascinating but soluble problems. With reasonable confidence this second generation of detectors should pro- vide previsions an order of magnitude better than those currently obtainable. Charged-Particle Tracking Chambers In a typical collider detector, beyond the vertex detector are charged-particle tracking chambers. These chambers serve to measure the directions and curvatures of the paths of the individual particles. These paths are called tracks. The principle of operation of a multiwire drift chamber is shown in Figure 6.6. The first such devices were coarse and measured only a few tracks. Modern devices are fine "rained in subdivision and may provide over a hundred measured points to a track. The resulting image is almost of photographic quality and is reminiscent, even though produced at electronic speeds, of the superb track detection provided in bubble chambers. With some additional effort and with certain possible compromises these detectors can also be used to measure the ionization of the produced tracks. An elegant variant of this detection method is to remove the fine grid of wires and to drift the ionization with a collection electric field to the end caps, where the arrival positions of the ionization are measured and also the arrival times and the degree of ionization. The arrival
142 ELEMENTARY-PARTICLE PHYSICS + H V - H V + H V JS ~ 9 no ~ APpears . - HV + H V - HV +HV - FlGURE 6.6 Most detectors use drift chambers or devices derived from drift cham- bers. In a drift chamber, slender wires are strung parallel to each other in a volume of gas. When a charged particle passes through the gas it leaves a track of ionized gas molecules and electrons. Electrons are attracted by the electrical voltage on the wires, and when they reach the wire they send an electrical pulse down the wire. Those pulses are collected and amplified electronically and recorded on magnetic tape. The position of the track is given roughly by knowing the wire that gave the signal. But a more accurate position is obtained knowing the time the electrons took to drift to the wire, hence the name drift chamber. times provide a measure of the depth at which the particle was produced, because the ionization drifts under the influence of the collection fields with a fixed velocity. This system. known as the time-projection chamber (TPC), has been implemented and is currently used in the TPC detector at PEP. Charged-particle tracking detectors are often immersed in a magnetic field in order to make it possible, from measurements of the track curvatures, to determine the signs of the charges and the momenta of the particles. Magnetic fields in the range of several kilogauss to several tens of kilogauss are used. The larger the field, the more the tracks curve, and the easier it is to measure the track momentum. To provide the highest possible magnetic fields, it is desirable to use superconducting coils to carry the required large currents. Although a number of these coils have been constructed and successfully oper- ated, the technology of fabrication is demanding, and further research is desirable.
INSTRUMEA'TS AND DETECTORS 143 Identification of Particle Types. Next in sequence beyond the tracking chamber there may be a detector layer that is used to identify the nature of the charged particles (whether they are electrons protons, pions, kaons, or muons). This region is still required to be nondestructive and thus to contain little material. A number of identification methods have been tested on small-scale devices and are under development for the next generation of detectors. These include the following: · Particle identification using Cerenkov radiation. This technique makes use of the property that particles produce light at an angle that depends on the velocity with which they are traveling through trans- parent matter. This radiation can be focused to a ring image whose radius directly measures the particle velocity. Counters photoconvert this ultraviolet radiation into ionization, which is then detected with proportional-counting techniques. This technique has been demon- strated successfully on a moderate scale but still requires considerable development to make it viable for large-scale detection. · Electron identification via the phenomena of transition radiation or, alternatively, via the emission of synchrotron radiation in the magnetic fields traversed. Calorimetric Detection and Energy Measurement The detection systems described so far do not serve to detect neutral particles, such as photons and neutrons, nor would they permit a precise measurement of the total energy in an event. The final layers of a collider detector therefore comprise thick, highly instrumented blocks of material calorimeters in which electromagnetic and hadronic particles cascade and convert their energy into ionization. These final instrumented blocks are placed at large radii, away from the point at which the beams interact, to leave sufficient space in which to insert the nondestructive low-density systems. The large radii of these blocks, coupled with the requirement of substantial thicknesses, result in calorimeters that are massive objects. The calorimeters used in collider experiments divide naturally into a front region where most of the electromagnetic cascade energy is deposited the electromagnetic calorimeter and the back region where most of the hadronic cascade energy is depositedthe hadronic calorimeter. Maximum precision is obtained when the calorimeters are constructed of materials, such as sodium iodide, in which the total
144 ELEMENTARY-PARTICLE PHYSICS deposited energy is detectable. However, such active calorimeters for collider detectors would with present technology, be prohibitively expensive. It is still very desirable to construct at least the electromag- netic calorimeter out of active material. Some of the materials under development, or in use, for electromagnetic calorimetry are heavy glasses, bismuth germanate crystals, and barium fluoride crystals. These systems are still costly, and their use at present is only made possible by leaving less space for the low-density systems in order to minimize the material requirements. Further developments in the production of comparatively low-cost materials for use in electromag- netic calorimeters would be useful. Even with substantial advances, however, most electromagnetic calorimeters and all hadronic calorimeters are likely to continue to rely on the introduction of many active sampling layers interspersed throughout the large passive calorimeter block in order to measure the ionization produced. Another design goal that is important but hard to realize is a finely divided calorimeter that is able to provide precise locations of the deposited energy. This requirement follows from the fact that the particles emitted from events in high-energy colliders are tightly bunched into jets. important information can be extracted from the angular spread and characteristics of the energy deposited by these jets of particles. This fine division or segmentation typically may require the recording of information from many hundreds or thousands of electronic channels reading out the information from the individual cells. DETECTORS IN FIXED-TARGET EXPERIMENTS The detectors used in fixed-target experiments vary tremendously in design, in size, and in complexity. Since we cannot do justice here to the range and variety of these detectors, we will only give several examples. The examples can be brief because the components of these detectors are in general just the same elements that we have been describing. The major exception is the bubble chamber, which is discussed at the end of this section. Small or Simple Fixed-Target Experiments Fixed-target experiments can sometimes be carried out with small or simple particle-detection equipment. This is often the case when the physicist is studying a simple reaction of elementary particles or studying one particular property of a particle. Two examples are given
INSTRUMENTS AND DETECTORS 145 400 MeV Pion Beom Beam Defining Collimator Magnetized Collimator Vacuum Tank and Decay Region - Removable CH2 Target E lect ron Coun ters Veto Counter Converter Array Energy-measuring Lead Glass array a Muon Cone Im ~1 : \ TO BEAM MON ITORS FIGURE 6.7 This apparatus, used at the Los Alamos LAMPF 800 MeV proton linear accelerator. studies the decay of a charged pion to a neutral pion plus a positron (e ~ ) plus a neutrino (vie). The neutral pion decays to two photons (a, and A,). in Figures 6.7 and 6.8. The first example is an apparatus used at the Los Alamos LAMPF 800-MeV proton linear accelerator. The proton beam is used to produce a charged pion beam, and the pion decay process charged pion ~ neutral pion + positron + neutrino is studied. The apparatus is relatively small and primarily uses two electromagnetic calorimeters. yet this is a fundamental measurement. Figure 6.8 shows a neutrino detector used at the AGS proton accelerator at Brookhaven to test the stability of muon neutrinos. The
146 E ~ ~ ~ E ~ _ ~ O G J ~ ~23~ ° _ ~7~ _ 1 ~e~ ~ U , I ~ ~ ~ o <D. J ~D ~ -2 1 __ ~ ~ 0,0= ~ C ,= C~ ~ o o ~ ~ _ eC ~ C ~ ~ C CL _ 4,, _ ~= ~ 4, E, Ct ~ <i ,C C~ ~ ~ C 1 . C ~ ~ Cc~c - - C 5 t~ CC t~. ~ C ~ _ :t ~ C~ , _ _ _ _ ~ _ C) _ C _ C) tV !_; ~ C c o~ 3 c - =_C C C': C '- C ,~ ~ C, ~ ~ _ ~ _ _ _ :t 4, _ ~ . _ ~ , ~ C ~ C _ _ . _ 5 . ~ · _ C C ~ . ~ _ , . _ X =, _ '; C~ =- J ~ C ~r E ~ -.=. ~ C _ _ _ .
INSTR UMENTS A ND D, TECTORS 147 muon neutrinos, produced by the primary proton beam, are allowed to travel a long distance before striking the detector. The detector is of moderately large size but simple construction; its function is to detect neutrinos and to distinguish between muon neutrinos and electron neutrinos. Large or Complex Fixed-Target Experiments A major fixed-target facility that demonstrates many instrumentation techniques is the Fermilab Tagged Photon Spectrometer, shown in Figure 6.9. It is intended primarily for studying the production of charmed particles. A photon beam produced by the primary proton beam strikes a liquid hydrogen target. Recoiling protons are identified by the recoil detector. and the produced particles are analyzed in the forward spectrometer. The spectrometer has magnetic analysis to measure charged-particle momenta; Cerenkov counters to identify pions. kaons. and protons; electromagnetic calorimetry to measure the energy of neutral hadrons; and finally a set of scintillation counters behind an iron filter to detect penetrating muons. This sequence of analysis steps is the same as that used in most collider detectors, but the target is not surrounded by all the components of the detector as it is in a collider detector. Figure 6.10 shows a rather complex detector that combines modern detector elements with the old nuclear emulsion technique. A nuclear emulsion is a thick photographic emulsion that when developed shows the paths of charged particles that have passed through it. This detector was used at Fermilab to measure the lifetimes of charmed mesons. The emulsion gave precise pictures of how the mesons decayed close to their production point. Bubble Chamber The bubble chamber, invented in the 1950s, was for many years the workhorse of elementary-particle physics experiments. A bubble chamber uses a superheated liquid, such as liquid hydrogen, neon, or Freon. Charged particles passing through this liquid leave tracks of tiny bubbles, which are photographed. The bubble chamber has gradually been replaced in most experimental applications by electronic detec- tors. The latter are more versatile, often give more information about the products of the collision, and usually provide that information in a form that can be directly used in computers. Nevertheless, the large volume and precise track information provided by bubble chambers
148 9 e. ,~ _ ~ ~ ED ~ on ._ , _ _ ~ e O I LD )) 8 ~ ~ E C' ~ . // Kit ~ D E ~ rat Cat ~ ._ 8 o 0 0 E o - CI. cO o 0 C' C' 1 ~//~ - - 3 - C~ Ct - ._ A: - c & c 3 5, At, - - C~ C at; E Ed Cat 5 o - - /~1 C - - ~, // l 1W C~ ~ ~ _ ~ C) C) o - C) ._ D ec ._ E O L _ :,` . - _ ._ o ,5 G "o - 4) - C~ ._ ~C r T ~ O C: X _ L.L
INSTRUMENTS AND DETECTORS Veto DC I -Emmy Emulsion I ~ _ ~ 40" L - SCM l 04 Mognet _. n 11111111 J hhL] DC II ._ ~ _ ~ Time of Flight Hodoscopes 100" Muon Steel Color'me er>, ~ 11 ~ I Leod Gloss v, to m d5- D 149 : 1 ~Muon' Hodoscopes _`lII FIGURE 6.10 This fixed-target particle detector combines the old technique of the nuclear emulsion with new techniques of chambers. Used to measure the lifetime of charmed particles, the emulsions give a precise location for where the charmed particle decayed. still makes them most suitable for certain types of experiments. Chief among these are the study of neutrino interactions and the study of particles with short lifetimes. Recent improvements in bubble-chamber technology include high repetition rates, precise track measurements, and holographic photography. DATA REDUCTION AND COMPUTERS Experiments in high-energy physics characteristically have pro- duced great quantities of data, whose reduction and physical interpre- tation have made up a significant component of the experimental effort. In recent years, apart from the sheer increase in scale of these experiments, data reduction has evolved to become more integrated with and intrinsic to an experiment's operation. A particular example is Monte Carlo computer simulation, which has become an important means of experimental calibration. New detector capabilities have made this evolution both possible and necessary. Very precise time resolution (billionths of a second. in some cases) has become possible over large detection volumes, allowing selective recording of particular event classes that make up only a tiny fraction of the total rate. In turn, these fast detectors can supply information in a form that can be
. 150 ELEMENTARY-PARTIC!E PHYSICS rapidly digitized and processed to provide the criteria for real-time event selection. Advanced systems for the reduction of high-energy data have unified the traditionally separate functions of trigger decision making, data logging. experimental control, and off-line event analysis. The trigger is critical to the success of many experiments because collisions may occur at a rate exceeding 106 per second. In order to select those interactions that are of particular interest in an experiment. a trigger is used. The trigger is a fast electronic decision made to record the data from a particular event on the basis of the signals received from the detector. The triggering decision itself may be based on detector input that no single computer could process fast enough. Fortunately, the repetitive nature of these calculations can be adapted to the use of fast but relatively primitive processors working in parallel. These are the first steps in what traditionally would have been termed off-line analysis. Without the results of these and other computations' the operation of advanced detectors cannot be monitored or controlled. Even the logging of data onto tape may use many levels of the data-acquisition system. For example, a single trigger may contain 100~000 characters of data (the equivalent of a 30-single-spaced-page report), which must be assembled from widely distributed local memories into the image of one event. One experiment in one year can accumulate data amounting to 10 million such reports. Because of the enormous software devel- opment required programs for real-time and off-line event analysis increasingly must share common subroutines and other features. The distinction between real-time and off-line analysis is further blurred by the scale of processing power that must be dedicated -to a single experiment, in notable cases reaching a level comparable with that of a mayor computer. The trend toward large-solid-angle general-purpose detectors at the colliding-beam facilities has been noted. As the collider energy rises, the events become more complex, and the amount of computer time required to analyze these events becomes large. For example, a Z" decay into hadrons has an average of about 20 charged particles and " neutral particles, and the off-line analysis time for such events in the detectors proposed for the SLC or LEP is of the order of 100 seconds of central processor unit (CPU) time for moderate-size computers. Millions of such events per year may need to be processed. As another example, it has been estimated that a dedicated capacity equivalent to tens of moderate-size computers will be required to process the inter- esting events from the 2-TeV proton-antiproton collider at Fermilab. it
INSTRUMENTS AND DETECTORS 151 is expected that the computer requirements for the high-luminosity SSC machine will greatly exceed the present level, as event rates, complexity, and detector sophistication will all increase. These requests for computer power are marginally met by current commercially available computers. The supercomputers tend to be machines optimized for vector or array calculations typically encoun- tered in solving large sets of partial differential equations, such as in weather prediction. These machines are not well suited to the large input-output (10) requirements of higb-energy physics nor to the large address-space requirements of the detector analysis codes. Some manufacturers have addressed the 10 problems and have reasonable CPU power but are relatively weak in modern software tools and in the system architecture to support them. These are important issues in high-energy physics, where the data production codes are being constantly improved. and data analysis involves significant amounts of programming, almost all of which is done by the physicist. Tools that improve the efficiency of the physicist are clearly valuable. Finally, the manufacturers of superminicomputers who have advanced the soft- ware state of the art do not produce machines of sufficient CPU power to analyze the data from a modern collider detector. The present situation is sufficiently serious to motivate noncommer- cial attempts to provide adequate computing power. Most of these attempts are based on the relatively large fraction of CPU to 10 activity that characterizes detector event analysis, so that many relatively simple. laboratory-designed processors can work on events in parallel, while being controlled by one commercial host computer with exten- sive 10 facilities. Examples are the emulator developments at SLAC and CERN and the multiple-microprocessor project at Fermilab. Related projects involve even more specialized processors designed for lattice gauge theory calculations or accelerator ray tracing. It is, of course, possible that commercial developments will prove adequate in the next few generations of machines, but the present situation is murky. Computing represents a significant expense at the national laboratories and universities in terms of actual hardware, support personnel, and physicist involvement. Even with the rapid decline in the unit costs of computing (crudely a reduction by a factor of 2 every 3 years)' the overall costs of computing increase. Thus computers of both large and medium size are now necessary parts of almost all elementary-particle physics experiments. Not only are they needed to reduce and study the data, but they are also used to monitor and control the experimental apparatus. The extensive use of computers has stimulated advances in some types of computer tech-
152 E~EMENTARY-PARTICLE PHYSICS nology and systems. This is because physicists have been willing to work with computers that were still in an early stage of production, interacting closely with the computer manufacturers. . FACILITIES AND DETECTORS FOR EXPERIMENTS NOT USING ACCELERATORS Most particle-physics experiments use accelerators, but some do not. In this section we sketch some of the ways in which elementary particles are studied without using accelerators. Atomic, Optical, Electronic, and Cryogenic Experiments Elementary-particle physicists are concerned with precise measure- ments of the properties of the more stable particles, such as electrons, positrons, muons, protons, and neutrons. For example, the electric charge of the proton is the same magnitude as the electric charge of the electron according to the most precise measurements that can be made. This is one of the reasons why we believe that there is a connection between the leptons (the electron is a lepton) and the quarks (the proton is composed of quarks). Such precise measurements are carried out using the methods of atomic, optical, electronic, or cryogenic physics. These include measurements of the electron g-factor, of the positronium Lamb shift, and of parity violation in atomic systems. These methods are also used to search for new types of particles in matter. Two sorts of searches have particularly intrigued particle physicists. One is the search for free quarks as contrasted with the quarks that are bound together inside protons and neutrons. The other is the search for magnetic monopoles, that is, for an isolated magnetic pole. All known particles that have magnetic properties have two magnetic poles, one north and one south, of equal size. None of these searches has produced generally accepted evidence for free quarks or monopoles. But more definitive searches are under way. Experiments Using Radioactive Material or Reactors There are experiments using radioactive material or reactors that are important in both elementary-particle physics and nuclear physics. An outstanding example is the study of the radioactive beta decay of tntium. An electron and a neutrino are produced in this decay' and the mass of this neutrino can be measured. The mass of the neutrino is a pressing question: is it exactly zero or, as indicated in a recent tri-
INSTRUMENTS AVID DETECTORS 153 tium-decay measurement in the Soviet Union. is the mass nonzero? Another example involves different forms of beta decay that have been proposed, such as the production of two electrons but no neutrino. This would violate our present theory of the weak force. and thus it is important to ascertain if such a decay exists. Reactors produce electron neutrinos, which are being used to study the stability of these particles. The most recent experiments find no confirmed evidence for neutrino instability. lncidently, evidence for neutrino instability is also being sought with solar neutrinos, as de- scribed below. as well as with neutrinos from accelerators. This illustrates how the exploration of an area in particle physics spreads over all the experimental techniques. Reactor experiments have also set important upper limits on the neutron's electric dipole moment. Experiments Using Cosmic Rays Cosmic-ray physics is concerned with three areas: the origin of cosmic rays, the use of cosmic rays as probes of extraterrestrial phenomena, and the use of cosmic rays to study elementary particles. It is the third area that concerns us here. Earlier in this century, cosmic rays were the only source of very- high-energy particles; hence substantial discoveries in particle physics were often made with cosmic rays. Prominent examples are the discoveries of the positron, the muon, and some of the strange hadrons. However, accelerator experiments have gradually displaced cosmic-ray experiments. Cosmic-ray experiments at present can only contribute to elementary-particle physics at extremely high energies, but unfortunately the flux is then small. This is shown in Figure 6.1 1. The most ambitious facility for the study of the highest-energy cosmic rays in the United States is the University of Utah~s Fly's Eye detector. Here a matrix of about 1000 phototubes in two clusters separated by 3.3 km observes, with good spatial and time resolution, the atmospheric scintillation light from cosmic-ray air showers. These air showers come from the interactions of very-high-energy cosmic-ray protons (energies above 108 GeV) with nuclei in the upper atmosphere. Because the cosmic-ray flux is so low at the highest energies (less than one per year per km' above 10'° GeV), it seems impractical to study these events in any way except through the study of air showers. Even an ambitious space station would not be able to support a detector sufficient to address this energy regime. By utilizing the atmosphere as a target, the Fly's Eye technique can explore a very large area, of the order of 100 km or greater. If the detector were high
154 ELEMENTARY-PARTICLE PHYSICS P~ P EC.m. ~ Te\J f o-2 . ~ 10 - ~ I o~6 x 10 8 1 o-lo 1 0-12 1 o-l4 10- 1 10-° 101 lo2 ~L 1 1 1 . ~ _ _ _ \ z ~ \ ~ ~ \ LLI ~ \ ~ Q > L)\ 1, <,)1 ,)1\ 2 014 1016 1018 1020 E l'\IERGY (ev ) l2 1 olo - lo8 ~ rot E 6 x J 104 2 z 10° FIGURE 6.11 The integral cosmic-ray flux. the number of primary cosmic-ray nuclei of energy greater than E. plotted versus E expressed in electron volts (eV). The vertical scale is expressed as particles per square meter per second per steradian (left) and as particles per square kilometer per year per steradian (right). The energy scale is also given in terms of the equivalent proton-proton center of mass energy, `/.s. The shading represents the experimental uncertainty of the flux determinations. in the atmosphere, the details of the first interaction would be more accessible. Again, this approach is the only access to particle physics at energies greater than even a 40-TeV proton-proton collider can provide, which is equivalent to about 109 GeV (10~8 eV) for cosmic-ray interactions. Turning to another example, current unified theories predict that there should be massive magnetic monopoles with a rest mass of about 10'6 times the proton mass and further that these should have been produced in the early universe and still be present among cosmic rays. An experiment in early 1982 reported evidence for the passage of one such monopole through a superconducting coil of a few cm' area. A large number of other experiments, using larger superconducting coils
INSTRUMENTS AND DETECTORS 155 and also searching for slow, lightly ionizing particles, have subse- quently failed to see evidence for monopoles. Larger monopole detec- tors are now being built. Another volume in this survey describes cosmic-ray experiments in more detail. with particular emphasis on their impact on questions in astrophysics and cosmology. The Solar Neutrino Experiment A major nonaccelerator facility with ramifications in particle phys- ics, nuclear physics, and astronomy is ~ Brookhaven National Labo- ratory detector for solar neutrinos. This detector is seeking evidence for inverse beta decay produced by neutrinos from the Sun. This process. whereby a neutrino produces a transition of a chlorine nucleus to an argon nucleus, would be a clean signature for solar neutrinos. After some years of operation and a multitude of careful checks, the observed rate of argon production is only about one third to one fifth of the predicted rate. Either our nuclear physics and astrophysics under- standing of the solar furnace in which hydrogen is burned in thermo- nuclear reactions is incorrect or some of the neutrinos decay before they reach the Earth (and neutrinos are unstable) or the experiment is wrong. Although this important problem remains unsolved, no new solar neutrino detectors are now being built. However. a new kind of solar neutrino detector using gallium has been proposed and designed in detail. Such a detector would be sensitive to the lower-energy neutrinos coming from the Sun. This is an advantage, since the number of these neutrinos predicted by theory is less dependent on a complete understanding of the conditions in the interior of the Sun. Searches for the Decay of the Proton Until the last decade the proton was regarded as absolutely stable; that is, it was assumed that the proton could not decay to any other particle. However, the realization that quarks, and hence the proton, are related to leptons has been growing. Therefore we have begun to consider the possibility that the proton could decay to either an electron or a muon (both leptons) plus other particles. This possibility, made quantitative by grand unification theories, has led to a first-generation family of underground proton-decay experi- ments. As of the end of 1983. the earliest very large detector, built by the lrvine-Michigan-Brookhaven (1MB) group, had seen no evidence of proton decay in about 4000 tons of water. This implies that, depending
156 ELEMENTARY-PARTICLE PHYSICS . on the mode of decay, the proton lifetime is probably greater than 103' to 1032 years. This detector employs about 2000 photomultipliers mounted on the walls of an 8000-m; water tank. The signals are from Cerenkov radiation in the water. Several large detectors in Europe, India. Japan, and the United States are or soon will be operating with only somewhat smaller masses. Some of these detectors use ionization for tracking and calorimetry' rather than the Cerenkov technique, and will therefore be more sensitive to some decay modes than the IMP experiment. Together with further 1MB data these detectors should either identify proton decay or set lower limits to the proton decay lifetime of between 1032 and 1033 years for each of several expected decay modes. SUMMARY AND FUTURE PROSPECTS In the preceding discussion a number of trends in instrumentation and detection systems for high-energy physics have been identified. Jets signatures produced by quark interactions at high momentum transfers have led to great emphasis on finely segmented detection systems and to calorimetric detectors for the measurement of energy flow. New generations of quarks and leptons with lifetimes in the range of one trillionth of a second have stimulated new progress in the development of electronic systems for high-resolution vertex detec- tion. Emphasis on rare processes has required virtually all major new detectors to be designed for efficient operation over nearly the full angular acceptance. For successful detection under conditions of high ambient rate, new levels of sophistication and integration of data- acquisition systems have become necessary. Continued progress in the development of instrumentation and detection systems is the experimental high-energy physicist's greatest challenge: these systems provide both the raison d'etre for accelerator facilities and the means by which our theoretical understanding can be confronted by experiment. Support for the development of high-energy physics instrumentation should grow in breadth and intensity. Basic research in the development of new detectors for high-energy physics increasingly should be recognized and supported as a fundamental source of much progress in the field. Correspondingly, the experimen- tal stations at our front-line accelerators should be used as effectively as available technology will reasonably permit.
