CASE STUDY: LHC COLLIDER DETECTORS, ATLAS AND CMS

Thousands of scientists from many countries all over the world are collaborating to find the answer to the fundamental question in elementary particle physics: What gives particles their mass? Physicists have noticed that some particles are light, whereas others are heavy. The effort to understand why this is so will move significantly ahead at two enormous detectors called the A Toroidal LHC Apparatus (ATLAS) and the Compact Muon Spectrometer (CMS) located at the Large Hadron Collider (LHC). These detectors are being built at the European Laboratory for Particle Physics (CERN) in Geneva, Switzerland. The detectors operate in large underground caverns hundreds of feet below the surface, since the LHC is located underground. These detectors have evolved from a series of earlier collider detectors, which made many earlier discoveries at lower energies. However, each of these detectors is so large in scale that they may be thought of more appropriately as "laboratories" since there are thousands of physicists there conducting a large number of diverse precision measurements and searches for new phenomena.

What Questions Do the LHC Detectors Address?

The LHC detectors are intended to explore the source of electroweak symmetry breaking (EWSB): why the massless photon is in the same family as the massive W and Z bosons. Scientists know that some new physics relating to EWSB must appear at the TeV mass scale (10¹² electron volts [eV]). If there were no such new physics, including no Higgs boson, then the Standard Model would become inconsistent. In particular, the cross section for longitudinally polarized W 's would exceed the so-called unitarity bound. The simplest possibility is the single Standard Model Higgs boson. However, it is known that the Standard Model is incomplete because it does not include gravity. In this simplest possibility, one would not know why the Higgs mass is ∼100 GeV (10¹¹ eV vs. 10²⁸ eV associated with gravity). There are three classes of models that have been widely discussed to resolve this question: (1) Extend the Standard Model to a larger symmetry called supersymmetry; (2) replace the Higgs boson with a dynamical condensate (technicolor is the prototype); or (3) extend the four-dimensional space-time at short distances to include extra dimensions.

In order to determine whether one or another of these possibilities exists, experiments will be performed to search for new particles produced in proton-proton collisions. The two large detectors must be sufficiently versatile to detect and identify what is produced when the protons collide and interact.

Although it is not known exactly what the new physics will be, there is confidence that measuring the products of the collisions will lead to the discovery and measurement of the properties of the new physics. In particular, each detector must be capable of measuring the momenta and directions of constituents of the proton (quarks and gluons), electrons, muons, taus, and photons and be sensitive to energy carried off by weakly interacting particles such as neutrinos that cannot be directly detected.

The new particles being sought are likely to be heavy themselves—in the 1 tera electron volt (TeV), or 10¹² eV, range—so the LHC must have sufficient energy to produce these particles. The LHC is a proton-proton colliding beam accelerator. Two counterrotating beams of discrete bunches of protons of 7 TeV traveling in almost circular orbits collide at four locations around the 16-mile ring. The energy of the proton-proton collisions is 14 TeV. When the protons collide, the hard scattering of the constituents of the protons, the quarks and gluons, becomes of interest. Those constituents have only a small fraction of the momentum of each 7-TeV proton; however, masses of several TeV can be easily produced. Collisions of the proton bunches will occur every 25 ns (nanosecond 10^-9 s) with an average of twenty-five collisions in every bunch crossing when the accelerator reaches its design goals. This means that in a year of data accumulation, each of these detectors would witness 10¹⁶ collisions, while only a tiny fraction of those collisions would provide evidence of the search for the Higgs, for example. In order to reach the physics goals in this challenging environment of unprecedented energy and collision rates, these detectors are larger and more complex than previous detectors. Both experiments are under construction and expect the first collisions to occur in 2007.

How Does an LHC Detector Work?

Figure 1 shows schematically how different aspects of a collider detector are used to measure the

FIGURE 1

products of the collision. In a collision of the two protons of interest, hundreds of electrically charged and neutral particles are produced at the collision point and travel outward through the detector. The detectors is seen schematically as having many separate layers, one outside the next, each with different capabilities. All charged tracks are measured in the tracking chamber consisting of several layers. A critical part of the tracking detectors is the innermost pixel layers, which can identify b quarks (the next to heaviest of the six flavors of quarks). The presence of b quarks is crucial in certain scenarios. The pixels are made from tiny rectangles of silicon, which identify the position of a charged track very close to the collision point. The entire tracking volume is immersed in an axial magnetic field so that the charged particles bend as they emerge from the collision point and traverse the tracking detectors. The tracking detectors accurately measure the position of the charged particles, which pass through them so that the curvature and direction of the particles can be measured. The curvature is proportional to the momentum of the particle. The electromagnetic (EM) calorimeter identifies and measures the energy and direction of the electrons and photons. Both electrons and photons produce localized energy deposits in the EM calorimeter. However, photons have no charged track aiming at the energy deposit, while electrons have a charged track, with the momentum matching the energy seen in the EM calorimeter. The hadron calorimeter measures the direction and energy of the hadrons. The muons penetrate these inner layers and are identified as a charged track outside of the calorimeters. The momentum of the muon is measured by determining the curvature of the muons in the outer muon detectors, which are located in a magnetic field. The collisions also produce neutral particles such as neutrinos, which are only inferred by taking measurements of all the other particles in the collision and calculating what is missing.

ATLAS (see Figure 2) and CMS (see Figure 3) use some similar and some different detector technologies to accomplish these functions. ATLAS will be as tall as a seven-story building and weigh about 5,500 tons (note the relative size of the people in the figure!). Both have silicon pixel and strip detectors as the first element of the tracking region. The pixels

FIGURE 2

in the case of ATLAS are about 10⁸, 50 × 400 μm², arranged in cylinders and disks all within a radius of about 8 inches from the beam axis. These small dimensions are used to search for charged tracks from heavy quarks (b or c quarks) that result from secondary vertices very close to the primary vertex or collision point. In CMS the entire tracking volume outside the pixel layers is composed of silicon-strip detectors, whereas ATLAS uses silicon-strip detectors to a radius of 22 inches. ATLAS then employs a Transition Radiation Tracker (TRT) beginning at the outer radius of the silicon-strip cylinders and disks to a radius of about 42 inches. The TRT not only measures the position and curvature of charged tracks (adding to the information obtained from the silicon layers) but also can identify electrons by the transition radiation (X-ray photons) they produce when traversing layers of differing indices of refraction.

A primary difference between the two experiments is the EM calorimetry. CMS uses a large array of lead tungstate crystals that produce light proportional to the EM energy. ATLAS uses a liquid argon calorimeter consisting of radial layers of accordion-shaped stainless-steel-coated lead plates separated by thin layers of liquid argon and electrodes. The movement of the free electrons produced by the electro-magnetic shower in the liquid argon is measured as a current on the electrodes. Each technique has advantages and disadvantages. The energy resolution is superior for the crystals, whereas the liquid argon has a more stable response. Also, the liquid argon

FIGURE 3

calorimeter is subdivided into more transverse and longitudinal layers to allow better photon and electron identification. Both detectors use scintillator tile hadronic calorimeters. These tiles produce light when particles pass through them. When the hadrons interact in the absorber material of the calorimeter, the light produced is proportional to the energy of the hadron. This light is channeled to a photomuliplier tube, which transforms the light into an electrical signal.

The names of the two detectors refer to the different methods used to identify and measure muons. CMS has a larger central field and measures the momentum of the muon in the inner tracking volume while using the outer muon layers for identification

FIGURE 4

and triggering. ATLAS employs a large air-core toroidal magnet system to measure muons independently outside of the inner tracking volume. Both detectors have sophisticated trigger systems. The purpose of the trigger is to make a fast selection (in a few microseconds) as to whether the particular collision is likely to contain an event that may be a signature of one of the physics quests. For example, events with two EM energy depositions above 30 GeV with no charged track pointing to them would be a candidate for a Higgs particle. Another event with three or four muons with momentum greater than 10 GeV would also be a Higgs candidate decaying into two photons. An example of how one particular event appears in the ATLAS detector can be seen in Figure 4. Much of the capability of these detectors follows from advances in technology. For example, both detectors use superconducting magnets, which allow a higher magnetic field than conventional magnets. The miniaturization of electronics allows the subdivision of the detector into very small parts to allow measurement of the complex collisions.

The main reason to have two detectors with different technical approaches is to ensure that when a potential discovery is observed, it is not the result of an instrumental effect. An independent experimental method and team can verify the discovery or reject the data.

How These Collaborations Work

The ATLAS collaboration includes about 1,850 physicists and engineers from 175 institutes in 34 countries. CMS has a similar list of participants often from the same countries, but it does not completely overlap. Most institutes participating in the LHC have joined only one of these two collaborations. Each institute has specific responsibilities as formalized in a "Memorandum of Understanding." Financial support comes from the respective governmental funding agencies.

A heavily documented process exists for first establishing the objectives of each detector and then developing detailed technical specifications. Design is next completed with a full prototyping of each component. Testing, installation, commissioning, and operations follow fabrication. Many committees are asked to review this project at each step of the way. Approvals are required before progressing to the next step.

Groups that have specific expertise in some aspect of detector technology form each of these collaborations. For example, one group may have expertise in growing crystals, another in testing the uniformity of light output of the crystals, another in the readout of the light from the crystals, etc. Each of the collaborations has an overall management and is then organized into subsystems. A subsystem may include the overall tracking or only the pixel detectors. Each subsystem has a series of subgroups. An example is the Liquid Argon group for ATLAS, which has subgroups for barrel mechanics, endcap mechanics, cryogenics and feedthroughs, front-end electronics, control of voltages and temperatures, etc. The central leadership for each collaboration includes a technical coordination group that addresses the overall configuration and integration (Do all the detectors fit together?), the routing of services (electric power, cooling, and signals), and installation and access of the detectors for repair.

These collaborations have organized meetings to resolve specific design issues and to divide the work. These meetings may occur all over the world, often over the telephone or through video conferencing, but they are mostly held at CERN. Decisions and technical specifications are documented in technical design reports, drawings, and other documents available on the World Wide Web (which was, in fact, invented at CERN by particle physicists).

These state-of-the-art detectors rely on advanced technology that can withstand a harsh radiation environment, which is highest at a small radius and in the forward and backward directions. Every component of the detector must be able to survive in the radiation levels predicted for its location. The scientists designing these experiments have worked closely with industry to use existing electronic technologies as well as develop new ones that will operate in the inner regions of the detector.

Leading industrial companies from all over the world fabricate the components of the detector. Many of these components are assembled at various collaborating institutes. Final installation and commissioning of each component occur at CERN with the participation of the collaborating teams.

Computing and Data Analysis

Once a collision occurs, the data from each detector must be held in a buffer until the trigger logic can decide whether to keep the event for future study. A complete event in ATLAS may contain 1 megabyte of information. If there were no trigger, one would need to analyze 10¹⁵ collisions times 10⁶ bytes = 10²¹ bytes per year! This amount of data would be impossible to deal with and so a selection— the trigger—must be made. The purpose of the trigger is to reject the "normal" events while retaining with good efficiency those events that may contain rare new physics signatures such as the Higgs. The multistep trigger reduces the data sample to about 100 events per second, which must be recorded for

FIGURE 5

further analysis. One hundred events recorded per second times 1 megabyte per event equals 100 megabytes per second. This means that 1 petabyte (10¹⁵ bytes) will be produced by each experiment for each year of 10⁷ seconds of data taking.

The computing power needed to analyze this huge amount of data is larger than what is available. These experiments are actively participating in the development of a data grid to facilitate such analysis. The word "grid" as used here is analogous to the power grid: users who need computing resources, data, or computational power will be able to request this from their home institutions, which will in turn be connected to all the collaborating institutions within a tiered structure (see Figure 5).

International groups will analyze the data. Members of these will be in constant electronic contact. When an analysis reveals something new and exciting, presentations will be made to the group and eventually to the entire collaboration. When the collaboration is sure that the new piece of physics is convincingly demonstrated to exist, a scientific paper will be written and published. No one knows for sure where this new understanding of the universe will lead, but one has only to look back 100 years to the discovery of the electron to realize how much of everyday life changed with the electron's discovery.

See also:CERN (European Laboratory for Particle Physics); Detectors and Subsystems

Bibliography

Adams, S. Particle Physics (Heinemann Educational, Oxford, UK, 1998).

ATLAS. <http://pdg.lbl.gov/atlas.html>.

Barnett, R. M.; Muehry, H.; and Quinn, H. The Charm of Strange Quarks: Mysteries and Revolutions of Particle Physics (Springer-Verlag, New York, 2000).

Close, F.; Marten, M.; and Sutton, C. The Particle Explosion (Oxford University Press, New York, 1987).

CMS. <http://cmsdoc.cern.ch/cms/outreach.html>.

Fraser, G. The Quark Machines: How Europe Fought the Particle Physics War (Institute of Physics Publishing, Bristol, UK,1997).

Kane, G. The Particle Garden: Our Universe As Understood by Particle Physicists (Addison-Wesley, New York, 1996).

Weinberg, S. "A Unified Physics by 2050?" Scientific American281 , 68–75 (1999).

Howard A. Gordon