Strip Detector

The Silicon Strip Tracker (SST) of the CMS experiment covers an area of $206\,\rm m^2$, which makes it the largest silicon detector under construction. The sensors are arranged in a total number of about 20000 modules, which consist of one or two strip detectors in series together with the associated readout electronics. Depending on the position within CMS, the geometry of the sensors and the number of readout strips varies: In the barrel region, the sensors are rectangular, while the endcap sensors are of trapezoidal shape to fit together in discs (fig. ).

Figure: Layout of the silicon modules in the barrel (left) and endcap (right) regions.

Figure: Left: Prototype support structure for the barrel with mounted dummy detector modules. Right: Same for an endcap disk.

The barrel modules will be placed on the surface of cylindrical support structures as shown on the left side of fig. . To allow better area coverage, the modules will overlap like roof tiles, which causes a tilt angle of $9$ to $12^{\circ}$ out of the tangential plane, approximately resulting in an equal Lorentz shift of electrons and holes (see section , p. ). In the disks (right part of fig. ) however, there is no Lorentz shift and thus no tilt, since electric and magnetic fields have the same direction.

A carbon fiber frame holds one or two silicon sensors which are connected to the readout hybrid via a pitch adapter. Each module has 512 or 768 strip channels which are read out by four or six chips, respectively. On both ends of the frame, cooling pipes are sinking the heat produced by sensors and electronics.

Figure: Functional groups of the CMS silicon tracker: inner barrel (IB), outer barrel (OB), inner disks (ID) and endcaps (EC).

Fig.  shows the functional groups of the CMS silicon tracker. The ten strip layers in the barrel are divided into the inner barrel (IB) and the outer barrel (OB), which are numbered in ascending order with the radius. The seven rings of the disk modules are divided into the inner disks (ID) and the endcaps (EC), again numbered with ascending radius.

The properties of the barrel and disk silicon detectors are given in tab.  and , respectively.

Table: Mechanical dimensions and numbers of the CMS silicon barrel detectors. The stated radii from the beam action are average values, since the modules are tilted. Double-sided modules count only once.

Layer	Radius $\rm [mm]$	Type	Modules	Pitch $\rm [\mu m]$	Strips
IB 1	250	double-sided	336	80	768
IB 2	340	double-sided	456	80	768
IB 3	430	single-sided	552	120	512
IB 4	520	single-sided	648	120	512
OB 5	610	double-sided	504	122/183	768/512
OB 6	696	double-sided	576	122/183	768/512
OB 7	782	single-sided	648	183	512
OB 8	868	single-sided	720	183	512
OB 9	965	single-sided	792	122	768
OB 10	1080	single-sided	888	122	768

Table: Mechanical dimensions and numbers of the CMS inner disk and endcap detectors. The stated radii are measured in the center of the active area of each layer. Double-sided modules count only once. The pitch varies due to the wedge-shaped sensors.

Layer	Radius $\rm [mm]$	Type	Modules	Pitch $\rm [\mu m]$	Strips
ID 1	277	double-sided	144	$81\ldots 112$	768
ID 2	367	double-sided	144	$113\ldots 143$	768
ID 3	447	single-sided	240	$124\ldots 158$	512
EC 1	277	double-sided	144	$81\ldots 112$	768
EC 2	367	double-sided	288	$113\ldots 143$	768
EC 3	447	single-sided	640	$124\ldots 158$	512
EC 4	562	single-sided	1008	$113\ldots 139$	512
EC 5	677	double-sided	720	$126\ldots 156$	768
EC 6	891	single-sided	1008	$163\ldots 205$	512
EC 7	991	single-sided	1440	$140\ldots 172$	512

The acceptance criteria for the silicon detectors of the CMS tracker have been worked out in great detail [42], where the most important specifications common to all sensors are:

• $<100>$ crystal orientation
• $\rm p^+$ strip implants on n-type bulk silicon
• Breakdown voltage $>500\,\rm V$
• Less than $2\%$ noisy strips
• Polysilicon resistor $R_P=1.5\pm 0.3\,\rm M\Omega$
• Ratio of implant width to pitch $w/p=0.25$

Standard silicon material has been chosen since it can withstand the radiation levels, while oxygen enriched sensors (see section , p. ) are considered not yet known well enough and thus implicate a certain risk.

Table: Specifications of the silicon barrel and disk sensors. For the trapezoidal sensors, base and top edge lengths are stated together with the height. The three outermost endcap layers consist of two sensors with different geometry (otherwise they could not be chained together).

Type	Wafer size	Sensors	Sensor Area	Thickness	Resistivity
		per module	$[\rm mm^2]$	$[\mu m]$	$\rm [k\Omega\,cm]$
IB	4" (6")	2 (1)	$63.4\times 119.2$	$320\pm 20$	$1.5\ldots 3.0$
OB	6"	2	$96.4\times 189.0$	$500\pm 20$	$3.5\ldots 6.0$
ID1/EC1	6"	1	$64.1 \ldots 88.1 \times 89.5$	$320\pm 20$	$1.5\ldots 3.0$
ID2/EC2	6"	1	$88.2 \ldots 112.4 \times 90.3$	$320\pm 20$	$1.5\ldots 3.0$
ID3/EC3	6"	1	$65.0 \ldots 83.2 \times 112.8$	$320\pm 20$	$1.5\ldots 3.0$
EC4	6"	1	$59.9 \ldots 73.4 \times 117.4$	$320\pm 20$	$1.5\ldots 3.0$
EC5	6"	2	$99.0 \ldots 112.4 \times 84.0$	$500\pm 20$	$3.5\ldots 6.0$
			$112.4 \ldots 123.0 \times 66.1$
EC6	6"	2	$86.1 \ldots 97.5 \times 99.0$	$500\pm 20$	$3.5\ldots 6.0$
			$97.5 \ldots 107.6 \times 87.8$
EC7	6"	2	$74.1 \ldots 82.9 \times 109.8$	$500\pm 20$	$3.5\ldots 6.0$
			$82.9 \ldots 90.9 \times 98.8$

The sensors used in the various parts differ considerably. Tab.  gives an overview of the detector specifications. The inner barrel modules will be fabricated either of two chained 4" wafers or of a single 6" wafer resulting in the same total area. Since the inner part is exposed to a higher radiation dose, its sensors are made of low-resistivity material which reach the inversion point at higher fluence (see section , p. ). The outer modules, which are a replacement of the previous MSGC design, have longer strips, resulting in an increased capacitive load. As pointed out in section , p. , this implies a higher noise figure. To restore a reasonable signal-to-noise ratio ($\rm SNR$), the sensors are thicker, so that the higher energy loss of traversing particles can compensate additional noise. High-resistivity sensors are required since the depletion voltage scales with the square of the thickness and the inverse resistivity. The radiation level in the outer part is sufficiently low so that the material will not get far beyond the inversion point.

In the innermost layer of the CMS strip tracker, the occupancy is approximately $5\%$, decreasing to $0.2\%$ in the outermost layer. Fig.  shows that every energetic particle arising from a collision traverses between eight and fourteen detectors in the silicon tracker [43], depending on the pseudorapidity $\eta$. Double-sided layers are counted only once. The reason why the number of radial hits can exceed the number of detector layers is that there is some overlap between adjacent sensors to ensure full coverage, which occasionally results in two hits in the same plane.

Figure: The average number of detector planes hit by high energetic particles when travelling through the CMS strip tracker. Double-sided layers count only once.

Several institutes (including the HEPHY) will assemble the CMS silicon strip detector modules using automatic and semi-automatic machinery. The modules will be fabricated with an internal precision of about $5\,\rm\mu m$ in the sensor plane and approximately $30\,\rm\mu m$ in the coordinate perpendicular to that plane [44]. After a mechanical survey procedure, the absolute position of each strip in space will be known with an accuracy better than $10\,\rm\mu m$. Each assembled module will be tested in regard of certain acceptance criteria such as a limited number of bad strips. Moreover, it will be subjected to a thermal cycle (cooling down and heating up) to verify the robustness of the bond wires.