The Silicon Tracking System (STS) is the core CBM detector [54]. Its main task is to reconstruct tracks of charged particles (with high efficiency>95 % for p >1 GeV/c) and to measure their momenta with high resolution (≈1.5% for p >1 GeV/c). Such limitations are required for successfully reconstruction of particles of the physics interest. To fulfil this requirement, the STS has to provide high hit reconstruction efficiency (close to100 %) and high spatial resolution (≲20µm), which leads to the fine granularity of the detector. The presence of a strong magnetic field is needed in order to measure the particle momentum; the curvature of its trajectory is inversely proportional to the momentum. The necessity of the high momentum resolution requires a low material budget.
In order to match the CBM physic program, the STS has to cope with high hit rates (up to 700 charged particles per central Au+Au event at the highest CBM interaction rate 107Hz) without a hardware trigger (see section 1.3). Sensors with sufficient granularity together with fast free-streaming electronics are being devel-oped to fit the requirement mentioned above.
The STS design allows to fulfil all the constrains mentioned above. There are 8 detector stations (sensor layers) placed between 30 cm and 100 cm downstream the target in the1 Tdipole magnetic field. They cover polar angles from2.5◦ to25◦ (see fig. 1.6). Total material budget lays within1.3 %X0 per station including both routing cables from the sensor to electronics and the support structure (see fig. 1.7).
Figure 1.6: Concept of the STS. Tracking stations are placed between30 cmand100 cmwith10 cm gap between closest neighbours. Radii of the stations are schematically indicated [54].
Figure 1.8: Sensor (prototype CBM03) corner from p-side under microscope.
Inclined white lines are the p-strips, four thin horizontal lines — the second metal layer [54].
Figure 1.9: STS-XYTER chip in PCB (Printed Circuit Board) under microscope [55].
The sensors are double-sided microstrip silicon sensors with 58µm strip pitch and about 300µm thickness made from n-type high-resistive silicon. There are 4 main sensor sizes involved: 6.2 cmwide and 2.2,4.2,6.2,12.2 cm long. Each one has 1024 strips per side. One sensor side is read-out by 8 ASICs (Application-Specific Integrated Circuit). Smaller sensors with shorter strips will be placed in the inner region of the stations close to the beam pipe, where the hit density is high. This will reduce the hit rate per sensor to ease the hit reconstruction. A few narrower sensors will cover the space around the beam pipe. All sensors overlap in order to
Figure 1.7: Material budget map for each station including all support structures and routing cables.
minimise dead space in between.
Measuring two coordinates simultaneously with the double-sided sensors allows minimising the material budget. Strips on the n-side are vertical and those on the p-side are tilted by 7.5◦ with respect to the n-side strips. The angle was chosen in order to suppress the number of fake hits (see section 3.2) yet keeping the spatial resolution sufficiently high: 10−20µm in the bending plane and 100−200µm in the perpendicular plane. The second metal layer connects the short strips (near the sensor edge) on the p-side (schematically shown in fig. 3.2). This feature allows to preserve the constant number of 1024 strips per side but complicates the hit position reconstruction (see section 3.2). The chosen topology allows to read the sensors out from one edge in the vertical direction. It allows to connect sensors with electronics from only one side and place the electronic out of the detector acceptance.
The radiation tolerance is a vital quality condition of the silicon microstrip sensors, regarding the severe radiation conditions in the STS environment. The radiation hardness was confirmed up to twice the expected lifetime fluence [40, 56, 57]. The lifetime fluence for the STS operating under the SIS300 conditions is 1014cm−2 in 1 MeV neutron equivalent for the innermost sensors. Charge collection efficiency drops only by15−20 % after irradiation to twice this level.
The self-triggering fast readout electronics is placed outside of the acceptance to minimise the material budget inside. The electronics is based on the custom designed ASIC — STS-XYTER (STS X and Y coordinate, Time and Energy Read-out chip) [58]. Its technological predecessor, n-XYTER (neutron-X-Y-Time-Energy Readout) [59], was used when STS-XYTER was at the design stage. STS-XYTER can cope with hit rates up to47 Mhit/s per ASIC, which converts to average maxi-mum rate of 0.37 Mhit/s per channel. This is much below the maximum hit rate of
Figure 1.10: Simplified block diagram of the STS-XYTER ASIC. CSA — Charge Sensitive Ampli-fier, PSC — Polarity Selection Circuit, DISCR — discriminator, ADC — Analogue-to-Digital Con-verter [55]. The chip architecture features fast and slow branch with double discriminator logic for low noise performance.
0.12 Mhit/s that corresponds to hit density 10 MHz/cm2 for the innermost sensors.
Each of 128 channels of the STS-XYTER chip has two pulse shaping amplifiers to achieve a good amplitude resolution and a low noise rate in combination with a sufficient time resolution:
• the slow shaper with 80 ns shaping time provides amplitude information. It has a 5-bit continuous-time flash Analogue to Digital Converter (ADC), which gives precision of amplitude measurements ≈3 % of the ADC dynamic range;
• the fast shaper with 30 ns shaping time provides time information with time stamp 3.125 ns.
The dynamic range of STS-XYTER is 15 fC that is roughly 4 times more than the most probable charge created in 300µm silicon by a Minimum Ionising Parti-cle (MIP).
The readout electronics is connected to the sensors with ultra-thin micro-cables [54].
They are covered with the grounded shielding layer in order to reduce the noise level. For the signal transmission, two layers of micro-cable with the aluminium strips with a pitch of116µmare used. They are separated with a meshed spacer in order to reduce the parasitic inter-layer capacitance. The supporting material for the aluminium strips is20µm polyimide.
The sensors are mounted onto the lightweight space frames with end supports;
these structures are called “ladders” [54]. The frames are made from carbon fiber,
Figure 1.11: Cross section of a micro-cable that reads one side of a sensor [54].
which provides mechanical stability keeping the low mass in the physical accep-tance [41]. The contribution of the material of supporting structures is found to be small comparing to the impact of the sensors and the micro-cables.
Figure 1.12: Mechanical prototype of 3 mock-up sensors and corresponding micro-cables mounted onto a ladder’s carbon fiber support [54].
The cooling system for the STS has two requirements:
• keep the sensors at the temperature not higher than −5◦C, to reduce shot noise level due to leakage current for the irradiated sensors: the innermost sensors around the beam pipe will dissipate in operation around 1 mW/cm2 after accumulation of the lifetime dose;
• remove the heat dissipated by the front-end electronics boards (FEB): each FEB produces about 20 W, resulting in the power dissipation of about 40− 45 kW for the total system.
The heat transfer inside the acceptance should be realised with a minimal amount of material. This induces usage of gas convection in the STS volume. At the same time, the humidity must be kept low to avoid condensation on the sensors.
For the cooling of electronics, evaporative heat transfer based onCO2 was chosen, because of its high volumetric heat transfer coefficient [41]. The cooling blocks will be tightly connected to the boxes holding the front-end electronics at the top side and bottom side of the STS [41].