How to build a particle detector
(also part 1 in an occasional series)

Why am I writing this? Well, I've learned quite a bit on the subject in the last couple of years, and I wanted to share that knowledge in a simple, easy-to-read form. If this fails to interest you, feel free to ignore it.

Anyway, let's consider the purpose of a particle detector. You have one beam of particles coming in from one side, another beam (usually of antiparticles) coming in from the other side, and they collide, hopefully producing new and interesting particles. Your job is to detect as many of these new particles as possible, measuring their properties as completely as you can.

The first consideration is the overall shape. Obviously, when an interesting collision occurs, the new particles will fly off (more or less) randomly in all directions. So, your first thought would be that you want your detector to be a sphere, centered around the point of interaction. This would be true, if collisions only occurred in one solitary point. However, for a variety of reasons, collisions typically occur in an "interaction region" along the beamline a couple of meters long. (To briefly summarize, this increases the number of collisions and thus the odds of actually getting something interesting.) So, your typical particle detector is actually mostly cylindrical centered around this interaction region (usually with caps on the end to catch any stray particles).

The four detectors I know the most about (CDF and D0, the two detectors at Fermilab, the world's current highest-energy collider, and ATLAS and CMS, the two detectors at LHC, which will be the world's highest-energy collider when it opens) all follow the same basic pattern (more or less). If you want to see some layouts, you can see them here:

CDF (or a much nicer PostScript version), D0 (only PostScript, sorry), ATLAS, CMS.

In any case, the general layout in all cases is as follows:

* Closest to the beamline is a silicon detector. The silicon detector is designed for very-high-resolution tracking. This is extremely important to the physics, since some of the particles produced in the initial interaction have very short lifetimes and will thus travel a short distance before decaying again. Identifying the location of the secondary decays ("secondary vertices" in the parlance) allows us to see the presence of these short-lived particles.

The technology used in these silicon detectors is not too different from the technology used in ordinary digital cameras. The detector is made of several layers, each containing a large number of appropriately-prepared silicon strips or pixels. When a particle hits a strip, it knocks off some electrons. By applying a voltage to the strip, these electrons will move to one end, where they can be collected and measured. The big difference between this and your digital camera, of course, is that these detectors operate in an extremely demanding environment. In order to operate at the necessary speeds, very high voltages are are used (typically a few thousand volts), which in turn generates a lot of heat (the silicon can literally melt within minutes if the cooling system fails). Furthermore, the high dose of radiation received this close to the interactions puts a high strain on the electronics used. All in all, it's not an easy job.

The silicon detector is typically not very large. The original CDF silicon detector used is maybe the size of two coffee cans on top of each other; the current one is approximately a foot in diameter and three feet in length (about the size of a large trash can), with some additional layers further out. This is partially because most of these secondary vertices are very close to the interaction point, and partially because the silicon is very expensive.

In general, the amount of silicon is usually limited by cost considerations; over the past 10 years, technology improvement has allowed most detectors to considerably expand their silicon coverage. CMS, in fact, uses silicon for all of their tracking (so they don't have a general-purpose tracker as described below).

* Outside of the silicon tracker is a larger, general-purpose tracker. This is typically a few meters in diameter and a few meters in length. The details vary from detector to detector, but the basic principle is almost always a "drift chamber": you have a large cylinder filled with gas (usually argon or some mix), and some wires running along the length of a cylinder. You apply a positive voltage to some of the wires and a negative voltage to some of the other wires. Again, particles passing through the gas will create some ions, which will drift toward the wires (hence the name) where they can be measured. In CDF, many wires are present in a "cell", while in ATLAS, each wire is isolated in its own "straw".

Drift chambers are a sturdy, (relatively) simple, and (relatively) cheap technology, and they provide good resolution, though obviously not as good as the silicon detectors.

The tracking chambers are typically enclosed in a powerful magnetic field. This magnetic field bends the path of charged particles, so that their momenta can be measured by how much the track is curved (higher-momentum particles will curve less).

* Calorimetry: In contrast to the trackers, where a goal is to disturb the particles as little as possible so that their track can be measured as accurately as possible, the calorimeters have the exact opposite goal: to absorb all of the energy of the particle so that the energy can be accurately measured. Unlike the tracker, where silicon has emerged as the dominant choice, there are a wide variety of technologies used in calorimetry.

The most straightforward way, conceptually speaking, is to use a material which has strong stopping power and emits light as the particles deposit their energy in the material. Then the amount of energy that the particle originally had can be measured by the amount of light emitted. Unfortunately, such materials don't grow on trees; CMS uses crystals of lead tungstate, but these are expensive to fabricate and maintain.

A compromise solution (used in CDF, D0, and ATLAS) is to alternate slabs of a material with strong stopping power (typically lead or steel) with slabs of a material which emits light (either a plastic scintillator or liquid argon). This is much cheaper than the first alternative, since all of the materials are easily available, but at the cost of some resolution. This is called a "sampling" calorimeter.

Calorimetry is, while an extremely valuable technique, inherently limited in its precision: as a particle interacts with the material of the calorimeter, it produces a large "shower" of secondary particles created by the interaction, and measuring the energy from all particles in a shower is inherently imperfect.

Typically, calorimeters are divided into two parts: the inner, or "electromagnetic", section absorbs particles which deposit their energy rapidly (electrons, photons, and pi-zeros), while the outer, or "hadronic" section absorbes heavier particles (hadrons) which lose energy less quickly.

* Muon chambers: Ideally, in the calorimeters everything is absorbed, with two exceptions: neutrinos, which can't (practically speaking) be detected by anything in an ordinary particle detector, and muons. A muon, which is a heavier relative of an electron, is extremely penetrating, and will make it through the lead or steel of the calorimeter without being terribly affected. So, typically, outside the calorimeter there's another tracker which detects the muons (and, occasionally, incoming cosmic rays). This tracker is usually another set of drift chambers, like the central tracker, but with much less demanding specifications (the central tracker has to deal with hundreds of particles in a very small area, while the muon chambers typically only have one or two muons to detect in a much larger area). Muons are very useful as triggers, since the presence of a muon almost always signals that something interesting has happened, so the muon chambers (which tend to be pretty slow) are usually supplemented with fast scintillators (which don't give you much position information, but which do tell you that a muon has passed by) to provide a trigger.

There are, of course, lots of other, smaller, systems involved in a detector, but these are the principal ones.