3.2 Silicon Detectors Technology
3.2.1 Sensor types
As explained in the previous section, a silicon detector consists basically on an asymmetric
pn
structure. This structure would have a large doping concentration on one side of the juntion, for instance a heavily dopedn
-type material (n
+). The other side will be lightly doped, for example, ap
-type region (p
−, shortened simply top
).In this case the depth of the depleted region on the
n
+-side is small compared to the depth on the weakly dopedp
-side. The electric field always grows from then
+ implant. Increasing the reverse bias the electric field can be extended far into thep
bulk. Hence the+
implant may be made only microns wide and the depletion region in thep
bulk silicon can be a few hundred microns wide.High Energy Physics (
HEP
) experiments require high segmentation in the tracking systems for accurate position and momentum measurements. This is the reason for using silicon microstrip detectors in this kind of experiments. The segmentation of the sensor is achieved by dividing the diode into small parallel regions calledstrips
. Each strip-bulk junction acts as an individual silicon detector. A schematic cross-section of such type of detectors is shown in figure3.8.3.2 Silicon Detectors Technology 75
Figure 3.8: Schematic view of a silicon microstrip detector. The bulk isp-type silicon and the electrodes aren+implants. Holes drift towards the p+back-plane, while electrons towards then+implants. An insulator (SiO2) is used to protect the silicon of the wafer.
The strips are connected to the readout electronics through an aluminum layer. With this configuration electrons are registered by the readout.
Depending on the type of the implants and the silicon bulk, the microstrip sensors can constitute different structures:
• p
-on
-n
The silicon bulk in
p − on − n
sensors isn
-type withp
+ strip implants on the sensor surface. The back implant isn
+ so the abrupt junctions are between the strips and the bulk silicon. An oxide layer (SiO
2) is used as passivation layer to protect the silicon bulk. The connection of the implants to the readout electronics can be made following two configurations: a direct connection between the aluminium traces and the implants (DC) or distributing a secondSiO
2layer on top of the implants (AC). In the DC case, the leakage current flows directly into the readout electronics. On the other hand, with an AC configuration the implants and the aluminium strips are separated by the oxide layer. This layer acts as a capacitor, therefore, a polysilicon resistor is needed to provide a voltage reference to the strips. Figure3.9shows twop −on−n
sensor sketches with both configurations.In these devices the depletion region grows from the strips to the backplane allowing the sensor to operate partially depleted. The readout electrodes will collect holes. Under radiation exposure, due to the lower mobility of holes, the trapping effects will be more probable and for the short collection times at HL-LHC the charge collection efficiency will be negatively affected.
Figure 3.9: Drawings of twop-on-nsilicon microstrip sensors. The bulk sensor isn-type while the strip implants are p+doped silicon. The AC configuration uses aSiO2layer as a capacitor between the aluminium traces and the implants (left) while the DC uses a direct connection between them (right).
They will also suffer from type inversion with the change of the effective doping concentration due to radiation damage. With radiation exposure the
n
material becomes lessn
-type and can be turn top
-type. Therefore, the junction will dissapear from the strips and the bulk and will migrate to the sensor backplane.These effects will be explained in detail in section3.3.
• n
-on
-n
These sensors have
n
-type doped silicon bulk andn
+implants. Thep
-n
juction is created at the backplane with ap
+ implant. In this case, the depletion region grows from the backplane to the frontn
+implants so the device must be fully depleted to achieve good charge collection efficiencies. Nevertheless, the electron collection by then
+implants provides higher signal collection efficiency under trapping effects than inp
-on-n
sensors. Radiation damage will also cause type inversion on these devices (see section 3.3), however this results in the bulk silicon becoming lightlyp
-doped and turning the sensor ton
-on-p
. They will be able to operate partly depleted.These sensors need isolation structures that will be explained after. This will be needed in both sides of the sensor and the fabrication requires aligned double sided processing (for the inclusion of guard ring structures near the junction before irradiation) which increases the complexity and cost of such devices.
Figure3.10shows a drawing of an
n
-on-n
sensor where its components can be distinguised.3.2 Silicon Detectors Technology 77
Figure 3.10: Sketch of ann-on-nsilicon microstrip sensor. The bulk sensor isn-type while the strip implants aren+doped silicon. Thepnjunction is created at the backplane with ap+implant.
• n
-on
-p
In this kind of sensor the detector bulk is
p
-type and the strip implants aren
+ placed above thep
-type silicon surface. Then
+ strips readout electrodes will collect electrons that will suffer less charge trapping than holes allowing higher signal integration in the short collection times at HL-LHC. This results in a higher charge collection efficiency [64].In
p
-type sensors the depletion region grows from the implants to the backplane.This allows the sensor to operate partially depleted since the
p
–n
junction is always on the signal collecting side, making the sensor highly radiation-tolerant.Furthermore
p
-type sensors do not suffer from type inversion with irradiation since an increase in acceptors only increases the depletion voltage as explained in section3.3.3.2.A sketch of the sensor components can be seen in figure3.11.