Extended thermal cycling of ATLAS ITk strip modules with and without stress mitigating interposers

Jesse Liu · 2025

Read Extended thermal cycling of ATLAS ITk strip modules with and without stress mitigating interposers by Jesse Liu (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-12586.txt.

Introduction

A team from the ATLAS group stress tested detector strip modules by heat cycles. Their paper, from twenty twenty-five, deals with parts for the new Inner Tracker at CERN. Five pre production modules went through long heat cycles once the test rigs were leveled and locks set for lone runs. Shape checks tracked how each module bowed over time. All five bowed more as the top heat rose.

Four modules cracked their sensors when cycled to deep cold. A soft gel and film layer was then added to three more modules as a stress shield. Those showed near no bow shift and passed full checks. Readers who build particle detectors get a fix for a fail mode that kills scarce parts.

Catalog record

CREATORJesse Liu
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE51 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2507-12586

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Extended thermal cycling of ATLAS ITk strip modules with and without stress mitigating interposers Jesse Liu Abstract This paper investigates critical mechanical failures during stand-alone thermocycling of ATLAS Inner Tracker strip pre-production modules. Five modules undergo extended thermocycling after adequately levelling thermal chucks and introducing interlocks for unattended operation. Module bow evolution is tracked via regular sensor metrology. All five modules exhibit bow increases with a mean of 146 ± 27 ​ μ 146\pm 27~\mu m when raising maximum cycling temperatures from 20 ∘ 20^{\circ} C to 40 ∘ 40^{\circ} C. Four such modules exhibit sensor fractures when cycled to − 44 ∘ -44^{\circ} C. A stress-mitigating layer of silicone gel and Kapton film interposer is introduced to three further modules, with detailed quality control data establishing electromechanical viability. No significant bow change of 1 ± 10 ​ μ 1\pm 10~\mu m is observed after ten cycles between [ + 40 , − 44 ] ∘ [+40,-44]^{\circ} C relative to [ + 20 , − 44 ] ∘ [+20,-44]^{\circ} C. Two interposer modules undergo 200 thermocycles up to [ + 56 , − 44 ] ∘ [+56,-44]^{\circ} C without fracturing. This decreased sensor deformation and fracturing is interpreted as evidence for reduced thermal stress. 1 Introduction The successful upgrade of the ATLAS Experiment [ 1 , 2 ] for the High-Luminosity Large Hadron Collider (HL-LHC) [ 3 ] underpins future discoveries in particle physics [ 4 , 5 , 6 , 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 ] . Its cornerstone is a new all-silicon Inner Tracker (ITk) detector [ 16 , 17 , 18 , 19 ] . The ITk comprises pixel [ 17 ] and strip [ 18 ] sub-detectors, each designed as a barrel bookended by endcaps. The strips barrel requires 10\ 976 detector units called modules , each comprising 97\times 97 mm p-type silicon strip sensors fabricated by Hamamatsu Photonics K. K. (HPK) [ 20 ] . Industrial adhesives [ 21 , 22 , 23 ] mechanically attach sensors to custom electronics supplying front-end power, readout, monitoring and control [ 24 , 25 , 26 , 27 , 28 , 29 ] . Module assembly and testing within operating expectations, called quality control (QC) [ 30 , 31 ] , are distributed worldwide. An important production QC procedure is module thermal cycling (MTC). Modules must pass electrical tests during repeated cycles between two temperatures. The nominal operating window is between room ( +20^{\circ} C) and end-of-life temperatures ( -35^{\circ} C) to mitigate elevated leakage currents from radiation damage [ 32 , 33 ] . Every module undergoes ten thermocycles to verify electrical functionality and mechanical integrity between [+20,-35]^{\circ} C. Cooling or power failures [ 34 ] can expose modules to temperatures outside the operating window. This motivates studies beyond nominal QC temperatures. Testing outside operating expectations is called quality assurance (QA). (a) Fracture line (b) Zoom into right edge with arrows showing fracture Figure 1 : Photo of module 20USBML1234911 thermocycled between [+40,-35]^{\circ} C prior to this paper, showing a sensor fracture. (a) The white line indicates the fracture position on the sensor. (b) Zoom into the right edge has arrows indicating the fracture under the wire bonds. However, we observe critical problems when thermocycling ten pre-production modules between +40^{\circ} C and -35^{\circ} C: visual inspection identifies a sensor fracture for one module (figure 1 ) and three modules develop excess leakage currents during -35^{\circ} C tests. Independent efforts report similar failure modes but only after loading modules onto local support structures, which is called the sensor fracturing problem [ 35 , 36 ] . Simulation studies [ 37 ] suggest adverse stress from mismatched coefficients of thermal expansion (CTE) [ 38 , 39 , 40 , 41 ] between the silicon sensor (fixed to local supports with adhesives) and copper in the flex printed circuit board (PCB); the intermediate epoxy glue provides a high-modulus mechanical coupling that transfers stress. These critical failures preclude production. Thermal stress is hypothesised to be linked to prior MTC fracturing (figure 1 ), where suction cups induce a large lever arm on the deformed sensor of cooled modules (figure 2 ). Site-specific apparatus differences may also indicate why we observe sensor fracturing during MTC, while other sites only report systematic fracturing after loading modules onto local supports. It is important to test these hypotheses and resolve site-specific MTC failures as a separate problem from post-MTC fracturing in the wider community. Figure 2 : Schematic showing the stackup of a pre-production module installed in the module thermocycling apparatus at +20^{\circ} C (left) and -35^{\circ} C (right) chuck temperatures. Mismatched thermal expansion coefficients cause the flex PCB, epoxy adhesive, and sensor to contract unevenly. This induces mechanical stress at -35^{\circ} C localised at the epoxy-sensor boundaries. This adapts the diagrams in Ref. [ 37 ] to module thermocycling, where a handling frame (not shown) has suction cups fixing the sensor to the test frame rather than adhesives onto local supports. This paper presents in-house investigations and mitigations for these critical problems. Section 2 briefly reviews module assembly and the setup. Section 3 introduces custom hardware interlocks, enabling unattended operation and extended thermal cycling studies. While investigating the prior sensor fracture (figure 1 ), we find non-flat chucks impart extraneous mechanical stress on modules during MTC and propose custom shims to level the chucks. Section 4 uses this improved setup for a test-to-destruction QA study, similar to Ref. [ 42 ] , to systematically reproduce the conditions for mechanical failure in figure 1 . Permanent sensor deformations are observed after MTC when raising the chuck above 30^{\circ} C [ 43 ] . This hysteresis is hypothesised to arise from a glass transition in the epoxy [ 44 ] that “bakes in” deformations induced by stress from the CTE mismatch. We exploit this effect by using the sensor bow of free modules as the main observable in this study. Section 5 introduces stress-mitigating materials onto modules to test the thermal stress hypothesis causing sensor deformations during MTC. Simulation studies [ 37 ] predict reduced thermal stress by inserting two materials between the flex and epoxy called interposers : (i) 100 \mu m thick layer of SE-4445 silicone gel [ 45 ] whose low modulus decouples stress [ 46 , 47 , 48 ] induced by CTE mismatch between the flex PCB and sensor, and (ii) 50 \mu m thick layer of Kapton polyimide film [ 49 ] to ensure adhesion [ 50 , 51 ] with the epoxy. We propose an ad hoc procedure to attach interposers onto modules. To test its engineering feasibility, we subject these trial interposer modules to the same QC and extended QA thermocycling as pre-production modules. This probes potential adverse side effects of these materials and ad hoc assembly, including the long-standing cold noise problem [ 52 ] , while testing for reduced sensor deformation hypothesised from simulated stress mitigation. 2 Module assembly and setup We briefly review the pre-production module assembly procedure and thermocycling setup. Figure 3(a) illustrates the original pre-production barrel module design. Pre-production assembly uses Loctite Eccobond F112 [ 23 ] epoxy resin 1 1

Read the complete text