CMS Tracker System

The CMS Tracker is completely made of silicon detectors, which are the best choice for tracking purposes in the LHC environment. In present and past experiments, large-volume gas detectors were a (cheaper) alternative to silicon, but they have a slower response time, so that the LHC timing requirements do not allow their usage.

The tracker consists of a central (barrel) part with three pixel and ten strip layers and the disk and endcap sections with two pixel and twelve strip layers [38]. A cross-section of one quadrant is shown in fig. . The pixel layers in barrel and endcap parts are shown in purple, while the strip layers are drawn in red (single-sided detector modules) and blue (double-sided detector module). The double-sided modules are made of two single-sided detectors mounted back to back with a strip inclination of $5.7^{\circ}$ against each other. Thus, these ``stereo'' modules deliver two-dimensional hit positions.

Figure: One quarter of the CMS Tracker layout. Pixel detector layers are shown in purple, while strip detectors are in red (single-sided) and blue (double-sided). The origin denotes the collision point and the numbers on top and right give the angle in units of pseudorapidity $\eta$, which is a function of radius $r$ and the distance $z$ along the beam axis (eq. , p. ).

The number of detector layers is a tradeoff between tracking efficiency, material budget and cost. On one hand, the number of hits increase with the number of layers penetrated, which makes the track reconstruction easier. On the other hand, the amount of material within the tracker should be kept as low as possible, because multiple scattering, which spoils the tracks, is proportional to the amount of material traversed by the particles. An even tougher constraint is the cost of the tracker, which reduces the number of layers to an affordable design. Simulations on various tracker configurations finally led to the geometry shown in fig. .

Figure: Simulated CMS event in $r\phi$ (top) and $rz$ (bottom) projections. A Higgs with a mass of $150\,\rm GeV$ decays into four muons. It is difficult to spot the muons in the tracker, but they are clearly identified in the muon detector, which is the outermost subsystem (see section , p. ).

In an average event, about 750 charged particles arise from each bunch crossing, which produce a few thousand hits in the tracker. Fig.  shows a simulated CMS event where a Higgs boson with a mass of $150\,\rm GeV$ decays into four muons in two projections. Physicists claim that they can extract and identify single particles out of the detector data. In fact this seems possible when keeping in mind that the tracker granularity is very small and thus the occupancy is still reasonably low, while fig.  only shows two-dimensional projections. Most of the particles are of low momentum (below $1\,\rm GeV/c$) and thus of no interest with respect to the physics goals. Due to the high magnetic field of $4\,\rm T$ in the tracker, their tracks are bent with a small radius (according to eq. ) such that many of them will not be able to exit the tracker at all. The helix traces of these particles are displayed as circles or sine curves in the shown projections.

The simplest approach to track reconstruction from a set of hit points is to start with a pixel hit in the innermost layer and project a cone onto the next layer in radial direction. If no hit can be found there, the starting point was either noise or a particle of very low energy which get stuck or was deflected by multiple scattering, so the original hit can be discarded. Otherwise, the procedure can be repeated until finally the full track through all planes is found. Of course, the procedure is much more complicated in reality: Dead or inefficient regions have to be taken into account (e.g., by skipping a layer) and the magnetic field bends the tracks depending on the particle momentum. Since there is a lot of low-momentum background in the innermost part of the tracker, a more advanced concept starts its track search from the outside. With this approach, a preselection of interesting tracks is provided by the first-level trigger, which is obtained from calorimeter and muon detector data.

The operating temperature of the CMS tracker will be $-10^{\circ}\,\rm C$. This is required by the silicon sensors, which suffer from radiation damage. Defects are ``frozen'' so they can not gradually decrease the detector quality, as discussed in section , p. .