Impact of Cold Noise on the tracking performance of ATLAS ITk short strip barrel modules using a charged particle beam

TL;DR

This study evaluates how Cold Noise (CN), arising from vibrating capacitors on the flex PCB, impacts the tracking performance of ITk short-strip barrel modules under HL-LHC conditions. Using threshold-scans and beam tests on one non-irradiated and one end-of-life irradiated module, it quantifies CN effects via $Q_c^{no}$, $Q_{50}$, and $Q_c^{eff}$, showing CN narrows the operational threshold window and degrades per-strip and global tracking, especially after irradiation. Non-irradiated modules can operate within spec with adjusted thresholds, but irradiated modules often fail to meet the 99% efficiency and 0.1% occupancy targets unless masking or shielding is employed. The results motivate shielding improvements (the interposer concept), which initial tests indicate can restore CN-affected modules to specification and preserve ITk performance in CN-prone regions at end-of-life.

Abstract

The inner tracking system of the ATLAS experiment will be upgraded to a full silicon detector in 2030 for HL-LHC. The new tracking system is called ITk, the Inner Tracker. It is required to be operable with efficiency higher than 99\% and noise hit occupancy smaller than 0.1\%. During the pre-production phase of the ITk project, many short-strip modules were observed to exhibit so-called "Cold Noise (CN)", wherein clusters of strips displayed very high noise when the modules were operated at temperatures below~$-35\degree$C. To investigate the CN impact and ensure the quality of module production, huge amount of effort have been put in by the collaboration. This paper focuses on the impact of CN on the tracking performance by examining two short strip modules that exhibit CN: one is non-irradiated, while the other one has been irradiated to the maximum expected end-of-lifetime fluence. For each module, the global and single strip tracking performance are evaluated.

Figures (14)

• Figure 1: Exploded view of a short strip barrel module with all relevant components. Taken from CERN-LHCC-2017-005.
• Figure 2: (a) A physical short strip module attached to its test frame. (b) Numbering scheme for the SS module, with the DC-DC converter highlighted in a yellow rectangle.
• Figure 3: Positions selected for threshold scans of both the non-irradiated and irradiated modules. The purple dots indicate the center of the beam spot.
• Figure 4: The detection efficiency of selected single strips (shown in blue) and noise occupancy (shown in red) are plotted as a function of the charge threshold for a normal strip and a strip affected by CN in both non-irradiated and irradiated modules. The efficiency points are fitted using Equation \ref{['eq:eff_fit']} up to 5.0 fC and extended to 6.0 fC for the non-irradiated module. The dashed blue line indicates the required efficiency level, while the dashed red line represents the required noise occupancy level. The pale green region illustrates the operating window if $\mathrm{Q_c^{no}<Q_c^{eff}}$.
• Figure 5: The relationship between of $\mathrm{Q_{50}}$/$\mathrm{Q_c^{no}}$ and $\mathrm{Q_c^{no}}$ using data from strips of the non-irradiated and irradiated modules. The dead strips and their neighbour strips are excluded. Good strips are strips with $\mathrm{Q_c^{eff}>Q_c^{no}}$. Failed strips are strips with $\mathrm{Q_c^{eff} \leq Q_c^{no}}$.