The most important effects for the operation of silicon detectors in high energy physics experiments are thermal excitations and electron-hole pair generation by charged particles. However electron-hole pair generation can also happen through via photon interactions. The signal generation by charged particles originates from ionisation of the particles with the silicon crystal lattice, creating electron-hole pair for every 3.6 eV of energy loss. However, the number of collisions and energy transferred for a given charged particle is not constant due to statistical fluctuations. The resulting energy spectrum is described by a Landau distribution, shown in Figure 6.7 for particles that have enough energy to easily pass through the detector. The peak of the distribution defines the Most Probable Value (MPV) of the energy loss. The mean energy deposited in the device is always greater than the MPV, because large energy deposits associated with δ rays (also known as δ electrons) create long tails on the distribution.
Figure 6.7: Probability distribution for charge deposition for 500 M eV pions in varying
Sensor Design Aspects
7.1 Sensor Layout Variants
The following will describe variations in silicon sensor layouts, commonly used for high energy physics particle detectors. Only silicon technologies that readout electrons of the n doped side are discussed, n-in-n and n-in-p technologies are discussed in Section 7.1.1 and Section 7.1.2 respectively. Another popular technology variant widely used in high energy physics, is thep-in-n type. However, this variant reads out the hole current signal. The work in this thesis is focused on radiation tolerant devices intended for use within ATLAS at the HL-LHC. The latter technology is not be discussed, because of its rapid deterioration in performance after irradiation. Further details on the enhanced radiation tolerance of n-side readout detectors are given in Chapter 8.
All the work within this thesis focusses on planar silicon detectors. Planar refers to the electrode implantation depth compared to the thickness of the silicon substrate. Typically, electrodes are implanted with a depth of a few microns. When electrode depths are comparable to the substrate thickness, devices are referred to as 3D.
7.1.1
n-in-n
A planarn-in-n sensor consists of a lightly n-doped (1013cm−3) substrate, with segmented
n+ implantation on the frontside and one large p+ implant on the backside. In this
context n+ and p+ implants have high doping concentrations (1015cm−3 ). A negative
bias voltage is connected to the p-doped electrode on the backplane, which is surrounded by multiple guard rings (Section 7.2).
Under reverse bias the depletion region extends from the backsidep+ implant prop- agating to the front of the device where the readout electrodes are located. The n+ electrodes are only isolated from each other at full depletion. This is shown in Fig- ure 7.1(a). n-in-n geometry devices undergo type inversion of the bulk from n- to p-type after sufficient radiation damage (see Section 8.5). After this the depletion region extends from the frontsiden+ readout electrodes to thep+ implant on reverse side of the detector.
Now the pixelsn+ electrodes are isolated from each other even if the device is not fully
depleted. This is shown in Figure 7.1(b).
(a) (b)
Figure 7.1: Development of the depletion zone (yellow) in an n-in-n sensor before (a) and after (b) type inversion. The depletion zone grows from different sides of the
sensor. Before type inversion, the depletion region extends from the backsidep+
implant, after type inversion it extends from the frontside n+ pixel side. This
allows efficient operation ofn-in-n devices even if they are only partly depleted
[81].
n-in-n detectors collect the charge from electrons at the readout electrodes. This makes this design more radiation tolerant compared to devices that collect charge using holes, because electrons have a higher carrier mobility than holes. The effect of type inversion also helps to make n-in-n geometry more very radiation tolerant, since even when the maximum applied bias voltage is insufficient to reach full depletion the area
surrounding the readout electrodes is still depleted enabling the generated charge to be collected faster.
7.1.2
n-in-p
Planar pixel sensors which feature a p-type bulk and n+ implants on the frontside and a largep+ implant on the back side, are named n-in-p sensors [139]. There are no guard
rings on the reverse side of the device, instead guard rings are implemented on the frontside of the device. The reverse side is totally covered by a homogeneous p+ implant.
A schematic view of an n-in-p detector is illustrated in Figure 7.2.
Since the bulk material is p-type, there is no type inversion after irradiation. The junction is between the frontside n+ electrodes and the p-type bulk, with the depletion region extending from the front side. This design is simpler to produce, test and operate. The main draw back with this technology are the large surface currents that form on the frontside edges because of the conductive cut edges. This effect is negligible unless the device is irradiated and can be a major issue for pixel modules, because there is often a large potential between the surfaces of the readout chip (GND) and sensor (HV) which is illustrated in Figure 7.2.
Making radiation hard sensors that are cost effective is very important. The size of future pixel detectors will depend on the cost per unit area. For large area pixel detectors these costs need to be contained and as such n-in-p devices are extensively studied in this thesis.
Figure 7.2: Schematic view of a n-in-p sensor, illustrating the different potentials between the sensor (HV) and the front-end chip (GND) [104].