IBL TDR Addendum

ATLAS Project Document No: / Page:1 of 12
Error! Reference source not found. / Rev. No.: Error! Reference source not found.
/ IBL TDR Addendum
ATLAS Project Document No: / Institute Document No. / Created: 17/03/2012 / Page: 1 ofxx
ATL-SYS-XX-XXX / Modified: / Rev. No.:1
IBL TDR Addendum
Abstract
This document is an addendum of the ATLAS IBL TDR of September 2010. This document focuses on the “Mixed Scenario” where staves are populated in the centre with two-chip planar sensor modules and with single-chip 3D sensors at the two extremities. We also describe the diamond beam monitor that uses the diamond sensors originally developed for the IBL.Planar sensors will be two chips in module size and 3D single chip.
Prepared by:
IBL Management Board + D. Ferrere, C. Gemme, W. Trischuk. / Checked by: / Approved by:
G. Darbo, H. Pernegger, M. Nessi, B. Di Girolamo
Distribution List
ATLAS IBL Collaboration for comments
ATLAS Project Document No: / Page: 1 of 14
ATL-IP-XX-XXX / Rev. No.: 1
History of Changes
Rev. No. / Date / Pages / Description of changes
1
1.1
1.2
1.3 / 17/03/2012
30/04/2012
4/05/2012
7/05/2012 / All
All, Sec.2.4
Sec. 2.3, 3
All / First Draft
Overall editing. Added section 2.4.
Added section 2.3 and 3.
Improvement from W. Trischuk.
ATLAS Project Document No: / Page: 1 of 14
ATL-IP-XX-XXX / Rev. No.: 1

Table of Contents

1INTRODUCTION

1.1Sensor Qualification and Production Status

2Mixed Sensor Scenario

2.1Module Layout

2.2Stave Layout

2.3Module Loading on Stave

2.4Electrical services

3Diamond Beam Monitor (DBM)

4References

1INTRODUCTION

At the time the IBL TDR [1] was submitted in September 2010, the sensor technology to go in the detector was not decided yet. Three technologies were considered for the IBL: planar sensors (n-on-n and n-on-p), 3D sensors (with active or slim edge) and diamond sensdetectors. A program of extensive testing of detector assemblies irradiated toIBL fluence of 5x1015neq/cm2 and 250 Mrad was carried on inout through the Spring of 2011. Results were presented to the a sensor technology review in July 2011.

The review panel recommendation was to investigate a “mixed scenario”, in which the 3D technology populates the forward regionhighest eta region of the IBL where the tracking could take advantage of the electrode orientation to give a better z-resolution after heavy irradiation. The implications of the IBL “mixed scenario” areis presented in section 2of this Addendum document with reference toemphasis on how this affectsthe module, stave and services design.

The Diamond Beam Monitor (DBM) [5]is presented in section 3[3], which is a spin-off from the IBL technology with prototype assemblies of FE-I4 and diamond sensors. The DBM is a detector that is constructed by the IBL collaboration and will only be installed, if the existing ATLAS Pixel detector will beis brought to the surface for replacingin 2013 to replace the Quarter Service Quarter Panels (nSQP project [6]).

1.1Sensor Qualification and Production Status

Between the end of 2010 and early 2011, the plans for construction of the IBL were substantially modified by two eventsfacts: the change in the LHC long shutdown planning (necessary to install the IBL) and the very good results of the FE-I4A front-end chip, which. The latter raised confidence that only minor changes would be needed for the production version of this chip.

The LHC shutdown to install the IBL, assumed in the IBL TDR, was for the end of 2015. The decision of LHC (Chamonix 2011) to have a long shutdown in 2013/14 created foroffered ATLAS the serious possibility to install the IBLtwo years earlierat this time. From all the sensor technologies under study for IBL it was considered the two that were more advanced and at asufficiently more mature stage for possible production were: planar n-in-n and double sided 3D sensors.It was therefore decided to restrict the qualification to these two technologies and develop FE-I4 modules from these two to be fully qualifiedy in the test beam studies and at full IBL radiation dose.To fulfil this “speed up schedule” it was decided, in January 2011, to launch a pre-production of planar sensors from CiS[i] and of double sided 3D sensors from CNM[ii] and FBK[iii]. A further The idea behind these production runs was to : already have between 30% andto 50% of the sensorsavailable by the time of the qualification phase deadline andthe subsequent sensor review.

The path chosen in shorteningto streamline the sensor prototyping and decision, the success of the version A of the FE-I4 with minor needs for modification together with the high production yield, which shortenedalso streamlined the production of the FE-I4version B by making engineering and production in an singleunique run. Together these saved over one year in the IBL schedule. Further optimizations and reduction of the originally generoushigh contingency in the schedule permits to gained the needed time to be ready for the phase 0 LHC shutdown in 2013/14. Figure 1 compares the IBL TDR schedule with the present schedule for the major detector items. The FE-I4B and sensors production weare almost completed as of t March 31st2013,,meetconfirming the new schedule for thesesuch critical items.

Figure 1: Ccomparison between the IBL TDR schedule (version v.3) and the version prepared for the IBL installation in the 2013/14 LHC shutdown. The comparison is made for the major production items going into the detector.

The recommendation from the sensor review panel, in July 2011,weare that both sensor technologies fulfilled the IBL requirements, and that there wais an opportunity to populate the forward region with 3D where the tracking could take advantage of the electrode orientation to give a better z-resolution after heavy irradiation.

The IBL collaboration, following thate recommendation from the review panel, decided to complete the production of planar and 3D sensors and endorsed the proposal to build enough modules for a mixed IBL sensor scenario where 25% of 3D modules populate the forward and backward part of eachvery stave. TheFull production of planar sensors will also be made toallow coverage of 100% of IBL in case thatis is needed. The fractions of planar and 3D sensors that can be put in the IBL are quantized by the granularity of the high voltage services, which individually bias a group of four FE-I4 equivalent area of sensors (i.e. 4 FE-I4 out of 32 in a stave). Preserving bBackward/forward symmetry of the IBL restricts only to a multiples of 25% the fraction of planar/3D that can populate each stave.

Table 1 andTable 2show a summary of the planar and 3D sensor production as ofthe 31st of March 2012. There are enough sensors from both technologies to fulfil the mixed scenario, considering the expected overall yield for the module production and stave loading.

Batch # / 1 / 2 / 3 / 4 / 5 / 6 / Total
Received wafers / 20 / 22 / 18 / 20 / 17 / 22 / 119
Good DC sensorstiles / 69 / 76 / 64 / 70 / 62 / 83 / 424
Yield / 86.3 % / 86.4 % / 89.9 % / 87.5 % / 91.2 % / 94.3 % / 89.1%

Table 1: Status of two-chip planar sensor production at the end of March 2012. The IBL in the 75% of planar sensorsscenario has 168 sensorstiles.

Status / Produced Wafers / Selected Wafers / Yield on selected / Good sensorstiles
FBK-A10 / Completed / 20 / 12 / 60 % / 58
FBK-A11 / Completed / 12 / 4 / 44 % / 14
FBK-A12 / Completed / 16 / 13 / 60 % / 63
FBK-A13 / In proc. (backup batch) / - / -
CNM-1 / Completed / 19 / 18 / 60 % / 86
CNM-2 / Completed / 17 / 15 / 71 % / 85
CNM-3 / In proc. / - / - / - / -
Total / 62 / 62 % / 306

Table 2: Status of planar single-chip 3D sensor production at the end of March 2012. The IBL, in the 25% of 3D sensors scenario, has 112 tilesensors.

In aAdditionally to the sensors qualification and production, thin modules have been developed with both sensor technologies, making single and double- chip assemblies. The prototyping wascarried out with 100 µm and 150 µm thin FE-I4A chips. For theaggressive IBL installation plan, it was decided to stay withuse the 150 µm thickness to be on the safe side to avoidunexpected yields issues.

Table 3summarizes the production of thin modules. Several such modules have been dressed with the flex hybrids and are assigned ready to be installed on for “stave 0” use.

FE-I4
Thickness / Planar (200µm)
Single Chip Double Chip / 3D (230µm)
Single Chip / Total
FE-I4
100 µm / 4 / 22 / 7 / 55
150 µm / 16 / 17 / 20 / 70
Total / 20 / 39 / 27 / 125

Table 3. Modules bump-bondedat IZM[iv]with thin FE-I4A chips to qualify the stave assembly procedure.

2Mixed Sensor Scenario

This section describes the main changes from the IBL TDR design to executefulfil the mixed sensor scenario. The impact of the mixed sensor scenario on the IBL constructionis on the following items:

• Module design.
• Stave layout.
• Module loading on stave.
• Electrical services.

2.1Module Design.

The IBL moduleoutlines for two-chip modules and one-chip modules are geometricallycompatible;the physical size ofa two single-chip 3D sensor assemblies ihas the same width as a single planar two-chip module. The differences in rbetween the twoof bothsensor technologies are compatible with the overall IBL envelopes. InTable 4 are listed the geometric parameters for the sensors used in the IBL mixed scenario.

Bare module assemblieswill beare dressed by gluing a flex hybrid circuit on the sensor side; there arewill be two circuits, one for single-chip 3D assembly and anotherone for the planar double-chip assembly. Such circuits are shown in Figure 2. They arewill be mechanically different, but electrically very similar once the test pigtail is cut. The reason for Hhaving two flex-circuits is for easingeases the assembly procedure, but does and not change thefor electrical connections, which will basically remainstayseparate for each of the two FE-I4 chips in the double-chip version. In the case of the double-chip module flex, there will beare two individual separate wire-bonding fieldsconnectors that bring the signals from the stave flex “wings” together. The connection step, once the modules are on the stave, is the same fortwo single and one double modules. In this way the stave with the stave flex becomes connections are compatible for either single or double chip modules.

Structure / Planar / 3D
Gap b/w modules / 205 µm / 205 µm
Sensor thickness / 200 µm / 230 µm
Module width (along z) / 41 315 µm / 20 450 µm
Bias tab / guard-ring extension (in r) / 630 µm / 1 230 µm

Table 4:Mmain geometrical parameters of the IBL sensor used for planar and 3D sensor modules.

A few special precautions have been used in the design of the module flex:

• The back of the flex (which is glued on the module) has useduses a 25µm thick polyimide film based coverlay[v] which is rated to stand 100V/µm to hold the 1000V needed by the planar modules once they have received their full integrated radiation dose. It was sensible considered to maintain do the same for the 3D modules, even if it is planned to have between 160 and ÷180 V as the maximum operation voltage. Increase inThe extra radiation length is very smalllow.
• The high voltage capacitorswill beare encapsulated using an isolating resin. The primeOne candidate is a Polyurethane resin (PUR)[vi] that was used for the ATLAS SCT.

Figure 2: Module Flex Hybrid for single chip (left) and double chip module (right). The flexes come with a pigtail and a test connector, which is cut away before loading on stave. The stave flex wing is glued on the module flex and then connections are provided by wire bonding. The double chip flex is electrically equivalent to two single chip ones.

The clock and data signals, which are individual lines on the stave flex, are routed separately on the double module flex to the input of eachtwo FE-I4s. Each line is terminated with a 160Ω resistance. This is acceptable asbeing the two stubs only 4 cm long and the frequency of the signals are 40/20 MHz for the clock/data. Single-chip modules use the same routing topology on two separated flexes.

• The two FE-I4 chips on a single module are differentiated by having their ID addresses are differentiated byconnected to a pull-up wire-bond connectedion to VDD. This is needed to differentiate the two FE-I4 chips in a module. For the double-chip module, thean additional wire-bond is used for only one of the two chips. When the In case of single-chip modules are mounted on a stave, the wire-bond of one of the two chips making a logical module will have the ID address wire-bond removedpulled out.
• The module flexes are produced with a surrounding frame having precision holes for positioning pins in the mounting jigs. When the pigtail is cut, thesuch frames are removed and the modules can then picked up by vacuum tools. This is done at the last moment, before loading to the stave.

All tooling and jigs for assembling the modules are made such as to be compatible with both designs.

The module testing is done using the USBPixreadoutR/O system. Two USBPpix systems are connected in master/slave configuration to a double-module using the test connector and an adapter card. Both single and double-chip modules uses the same adapter cards; double modules use an additional pin for the extra signal on the test connector.

2.2Stave Layout

The mixed sensor scenario stave layout is shown in Figure 3. The 3D sensors populate the 2 extremities. The area covered with planar and 3D sensors is, respectively, 75 % (equivalent to 24 FE-I4 chips) and 25 % (equivalent to 4+4 8 FE-I4 chips). The modules have a fixed gap of 205 µm. The planar and 3D modules differ slightly in thickness: fromrespectively, 200 µm toand 230 µm; and in the rwhere the 3D modules are being 700 µm widerlonger. The 3D design is floorplan was made compliant for the double and single sided design (active edge), having the second one the high voltage connection being madeis foreseen on the same side as of the bump-bonding. For this compliant reason the sensor needs to extends beyondd over the FE-I4. TheFigure 4 shows a cross-section of the stave, (top) at the position of a planar double- chip module and (bottom)at the a position ofwhere a 3D single- chip module is situated.

Figure 3: Stave layout for the mixed sensor scenario. 3D sensor modules populate the two stave extremities. The gaps betweenneighbouring modules is fixed at 0.of 205mm.

Figure 4:Sstave cross sections at the position of a planar module (top) and a 3D module (bottom).

2.3Module Loading on Stave

The module loading consists of integratesing the stave andplusstave-flex together with the planar and 3D detector modules while targeting for the highest quality in term of working pixel and modules and , as well, the long term reliability. The 16 procedural e steps, which are followed by the module loading and QA sites are:

1. The reception tests of the completed stave completed with the stave-flex.This is part of the QA to validate that the stave with the glued stave-flex has a conformal geometry after it is made and has been thermally cycled 10 times from -40°C to +40°C.
2. The reception tests of modules.Detector modulesqualified at the assembly sites pass visual inspection and basic electrical readout tests at loading site, before the module-flex test pigtail is cut to finally load them on the stave.
3. A “Guillotine tool” cuts the module pigtail. The next operation consists in the removing the wire bonds from the pads that will be re-used to connect the stave-flex wings. These same pads weare previously used to electrically connect the module test pigtail before loading the module to the stave (see Figure 2).
4. The modules are loaded on the first1sthalf of the stave (Figure 5).Six planar and four 3D modules are positionedusing precise mechanical references(dowel pins).Module placement accuracy is based on the sensor dicing accuracy, which is +/- 10 microns. The gapbetween modules of 205 µm (see Figure 3) is fixed by polyimide-coated shims having that have a thickness of 190 µm.
5. The modules are loaded on the second half stave with the same positioning technique.
6. The 32 flex-wings that are retracted during loading are then released and glued on the module flex with Araldite 2011 (epoxy glue) (seeFigure 5).
7. Once the wings are glued, the electrical interconnection between the module-flex and the stave-flex (wings) is done by wire-bonding. Multiple wire-bonds are used for redundancy. The connections bring out FE-I4 power and I/O LVDS signals, sensor bias and connections to NTC temperature sensors placed on the module-flexes. Test wire-bonds are then pulled up to measure pull force and control the quality of the wire-bonding process.
8. The loaded stave is electrically connected through an adapter card(“PCB saver”) to the readout system and cooled by a CO2 system. This step qualifies the stave in near to real operating conditions.Reworking is done in case of needsas needed before moving shipping to CERN for integration into the IBL.
9. The mModule position iss are surveyed with respect to stave references.
10. The complete sStaveis thermallycycleding.Ten thermal cycles from -40°C to +40°C are foreseen in the QC procedure. Assembly weaknesses and infant mortality are is detected and corrected in this wayby this means.
11. The sSurvey is repeated and the results are compared with thosee one from step 9. If displacements or distortions are seen, rework will have to be considered. Such positions are recorded and will be used for initial alignments of the detector modules in the IBL.

A cComplete functional test with cooling and readout is performedR/O. This test checks integrity and full functionality of all the modules. Bump integrity can be checked with pixel noise measurements without sensor bias: a lower noise value on pixel channels is indicator of disconnected bumps. (isn’t this a bit late to be finding bad bumps? WT)

Staying insideAdherence to the IBL envelope is verifiedrequested by the tight clearance in the IBL. This is particularly critical with respect the neighboring IBLstaves where the minimumal distance is as small as 0.8 mm. Stave-flex wings are the ones that needwill be the subject of the highest attention.

The last operations on the stave is the addare to add of an insulation spacer in the gap between module groups sharing the same sensor bias and a spacer protecting wire-bonds from mechanical damage in case offrom touching another stave during integration in IBL.