87
Part Two · Quantum fields and particles

Detectors: tracking, calorimetry, triggering

Inside ATLAS and CMS, bunches of protons cross forty million times a second, and each crossing produces dozens of collisions. Recording all of it would mean storing about a petabyte every second. So within four millionths of a second, custom electronics decide whether each crossing is worth a second look, and within a fraction of a second more, computer farms decide whether to keep it. About one crossing in ten thousand is written to disk. Everything else, including any physics no one thought to look for, is gone forever. This chapter is about how a particle detector sees, and what it chooses to remember.

On the plate: a slice through a collider detector. Send different particles through it.

Layers, and what each sees

A general-purpose detector is an onion. Innermost, a silicon tracker records the path of every charged particle in a magnetic field. Next, an electromagnetic calorimeter absorbs electrons and photons; then a hadronic calorimeter absorbs protons, pions and neutrons; outermost, muon chambers catch the muons that pass through everything. Send , , , or . Each leaves a distinct signature. The neutrino leaves nothing; it is inferred from the momentum that is missing when everything else is added up. Established

Momentum from curvature

A charged track in field BB curves with radius RR; its momentum across the field is pT[GeV]=0.3 B[T] R[m]p_T[\text{GeV}] = 0.3\,B[\text{T}]\,R[\text{m}]. High-momentum tracks are nearly straight, so precision depends on measuring a tiny sagitta: the resolution worsens in proportion to pp and improves as BL2BL^2, with LL the track length. That is why CMS chose a 4-tesla solenoid, and ATLAS a large volume. Set the momentum to 5.0 GeV in a 4 T field.

How particles lose energy, and proton therapy

A charged particle crossing matter ionizes atoms along its path. Hans Bethe’s formula, refined by Felix Bloch, gives the loss per distance: it rises steeply as the particle slows, roughly as 1/β21/\beta^2. So a proton deposits little energy where it enters, and most of it just before it stops: the Bragg peak.

That shape is the reason for proton therapy. Choose the beam energy, 150 MeV, and the peak moves to a depth in water (and roughly in tissue) set by it: about 16 cm at 150 MeV. Tumour gets the peak; tissue beyond it gets almost nothing, unlike X-rays, which deposit dose all the way through. Established

Showers

A high-energy electron in dense material radiates photons; each photon makes an electron–positron pair; each of those radiates. The cascade doubles roughly every radiation length X0X_0, 0.56 cm in lead and 0.89 cm in the lead tungstate crystals of the CMS calorimeter, until the particles fall below about 10 MeV and stop. The total light collected measures the original energy, with a resolution that improves as 1/E1/\sqrt E. Hadrons shower too, over the longer nuclear interaction length, about 17 cm in iron, and less precisely, because some energy is lost breaking nuclei.

The trigger

Forty million crossings a second, a first hardware decision in microseconds, about 100,000 kept for software, and about one to three thousand a second written out. Set a first-level threshold on the energy of the highest-energy object, 25 GeV. Lower it and the rate exceeds what the readout can handle; raise it and you lose Higgs bosons decaying to photons of about 60 GeV. Trigger design is a choice of what physics to make possible. Neural networks running on fast programmable chips now make some of these decisions within the microsecond budget.

What you now have

The layered detector and particle signatures, momentum from curvature and its resolution, Bethe–Bloch energy loss and the Bragg peak with its medical use, electromagnetic and hadronic showers, and the trigger chain from 40 MHz to kilohertz.

· · ·

Three questions to carry with you.

How is a neutrino detected at a collider?tap
As missing transverse momentum: the imbalance after summing every visible particle.
Why does proton therapy spare tissue beyond the tumour?tap
Protons deposit most energy in the Bragg peak near the end of their range, and almost none after.
Why must triggers discard most collisions?tap
The raw data rate is far beyond what can be read out or stored.

And one to work out. A track curves with radius 3.33 m in a 4 T field. What is its transverse momentum, in GeV?

p_T =GeV

Sources: C. Grupen and B. Shwartz, Particle Detectors (2008). Particle Data Group, “Passage of particles through matter” and detector reviews. ATLAS and CMS technical design reports for the trigger systems. W. D. Newhauser and R. Zhang, Phys. Med. Biol. 60, R155 (2015) (proton therapy physics). Checked by code: Bragg–Kleeman range 0.0022 E^1.77 cm gives 15.7 cm at 150 MeV in water; p_T = 0.3BR.

Next · Chapter 88
Collider data analysis and statistics (5 sigma, look-elsewhere effect)

Why 5σ? Because physicists got burned many times at 3σ (p = 1.35×10⁻³ ✓).

Turn the page