1. Simon Kwan - FermilabCMS Tracker Upgrade Workshop – June 3, 2009 1 Update on the Phase 1 FPIX Half Disk Design Simon Kwan Fermilab on behalf of the USCMS Pixel Mechanics R&D group

2. Simon Kwan - FermilabCMS Tracker Upgrade Workshop – June 3, 2009 2 Current FPix Detector Experiment needs 672 plaquettes (7 types and 5 types of sensor module) assembled on 192 panels (4 types). Two panels placed back-to- back on Al cooling channel to form a blade 192 panels are needed

3. Simon Kwan - FermilabCMS Tracker Upgrade Workshop – June 3, 2009 3 Current Phase 1 BPIX / FPIX envelope definition

4. Simon Kwan - FermilabCMS Tracker Upgrade Workshop – June 3, 2009 4 Phase 1 FPIX Disk layout requirements 1.Fits within Phase 1 FPIX envelope definition 2.Modules oriented radially (requires only 2x8 modules, and slightly improves resolution compared to the layout of the current detector) 3.Locates all outer radius sensors as far forward and out in radius as possible (to minimize the gap in 4-hit coverage between the end of the 4th-barrel layer and the forward-most disk) 4.Maximize 4-hit coverage between end of 4th layer barrel up to eta = 2.5, for particles originating at the IP +/-5cm, using a minimum number of modules 5.Keep the same 20 degree tilt as the current detector 6.Individual modules and/or module-support substrates removable and replaceable without disassembling other modules on the disks 7.Identical substrates (blades are the same) 8.Minimizes the amount of material required for cooling and module support (assuming cooling using CO 2 ) 9.(Highly desirable) Delta T < 5C across a single module 10.(Desirable) Separate inner from outer rings for easier replacement of blades on the inner ring (with earlier radiation damaged modules).

5. Simon Kwan - FermilabCMS Tracker Upgrade Workshop – June 3, 2009 5 FPix Phase 1 Upgrade Plans Baseline: 3 disks in each endcap Layout: –Current detector has 7 module types and panels placed between r=59.7mm to 144.6mm (total 84 modules per half disk or 540 ROCs) –Upgrade detector: arranged modules radially; one module type (2x8) placed between r=45mm to 161mm (total 56 to 60 modules per half disk or 896 to 960 ROCs) –Modules divided into an outer ring of 34-36 modules and inner ring of 22-24 modules –Both outer and inner assemblies contains identical blades with 2 plaquettes (1 plaquette placed on each side of a substrate). – C0 2 cooling ; use thin-walled SS tubing 316 L and the size is tentatively chosen (1.638 mm OD, 1.435 mm ID) based on getting a continuous loop providing enough cooling power for each blade assembly. – ultra light weight mechanical support and with cooling tube (aim at material reduction of about a factor of 2)

6. Simon Kwan - FermilabCMS Tracker Upgrade Workshop – June 3, 2009 6 Upgrade Design for FPIX MODULE LAYOUT  12 identical half-disks  only 2x8 modules  One 2x8 module placed on each side of a substrate for all outer and inner radius blades  Substrate uses Thermal Pyrolytic Graphite – material with excellent in-plane thermal conductivity Inner Outer

7. Simon Kwan - FermilabCMS Tracker Upgrade Workshop – June 3, 2009 7 291 396 All Identical disks (1 st and 2 nd disks in locations to maximize 4-hit eta coverage) 6 disks = (6x68) outer + (6x44) inner = 672 2x8 modules (10752 ROCs) Note: distance units in mm current FPIX 4 disks at (mean) Z: ±355 and ±485 mm η = 1.3 η = 1.6 η = 2.1 η = 2.5 2x8s Z loc. TBD suggest 491mm from IP 30 60 161 45 64.8

8. Simon Kwan - FermilabCMS Tracker Upgrade Workshop – June 3, 2009 8 Basic Design of the Pixel Blade Solid TPG (0.88 mm thick) encapsulated with carbon- fiber facing (0.06 mm thick). All blades are identical with one module on each side. (Only 2x8 module is used.) Cooling is arranged at one end of the blade in which good contact with the ring is kept. Extra tab is provided to facilitate in handling. provisions (threaded screw) allows the blade to be attached/removed from the ring so no need to remove neighbors for removal (repair).

9. Simon Kwan - FermilabCMS Tracker Upgrade Workshop – June 3, 2009 9 Basic Design of the Half Disk Half disk consists of one inner blade assembly and one outer blade assembly and they are assembled next to each other. Inner blade assembly consists of 11 or 12 blades and is supported by the outer blade assembly. Outer blade assembly consists of 17 or 18 blades. All blades are supported by 2 rings; Cooling tube is embedded at least in one of the rings.

10. Simon Kwan - FermilabCMS Tracker Upgrade Workshop – June 3, 2009 10 11 blades with Z offset =4.0 mm, Blade separation = 5.0 mm Ring width = 45 mm 0.7 mm thick carbon fiber ring for support use only 3.0 mm thick CC ring for cooling and support uses Inner Blade Assembly Note: through holes on rings for fastening blades not shown)

11. Simon Kwan - FermilabCMS Tracker Upgrade Workshop – June 3, 2009 11 0.7 mm thick carbon fiber ring for support uses only 3.0 mm thick CC ring for cooling and support uses Outer Blade Assembly Note: through holes on rings for fastening blades not shown) 17 blades with Z offset = 2.2 mm, arranged in 2 rows Blade separation = 5.46 mm Ring width for inner ring = 34 mm Ring width for outer ring = 62 mm

12. Simon Kwan - FermilabCMS Tracker Upgrade Workshop – June 3, 2009 12 Cf skin to enclose the tubing Ss tubing Machined groove to house tubing One piece ring made of CC The Edge Cooling Concept -- capture the cooling tube inside the ring. Cooling tube: simple to fabricate and less temperature difference from inlet to outlet

13. Simon Kwan - FermilabCMS Tracker Upgrade Workshop – June 3, 2009 13 The Half Disk Cf supporting spokes

14. Simon Kwan - FermilabCMS Tracker Upgrade Workshop – June 3, 2009 14 Current vs Proposed upgrade FPIX detector CurrentNew Half disk812 Coverage (r)6-15cm4-16cm Z-position (geom. mean) 35.5/48.5cm32.5/41.5/50.5cm Modules672 (7 types)672 ( all 2x8s) Module Orientation LateralRadial Blades96336 ROCs432010752 Radiation length3.48% (blade) Reduce by x2 (goal)

15. Simon Kwan - FermilabCMS Tracker Upgrade Workshop – June 3, 2009 15 Material Budget ItemCurrent (%RL)Upgrade (%RL) Sensors+ ROCs0.50 VHDI+substrate+ Components 0.930 HDI+substrate+ components 0.800.70 Cooling channel + coolant 1.250.04 Outer & inner ring0.680.48 Total per blade4.161.72 Seems feasible that we can achieve the goal of reducing the material by x2

16. Simon Kwan - FermilabCMS Tracker Upgrade Workshop – June 3, 2009 16 FEA Check on Blade with Two 2x8 Modules Blade thickness: 0.06 mm cf + 0.88 mm TPG + 0.06 mm cf Multi-chip thickness (overall 0.790 mm): Adhesive:.050 mm ROC:.200 mm Bump-Bond:.020 mm Sensor:.270 mm HDI:.200 mm Simplified model: ROC were a continuous layer instead of 16 tiny ones; Bump-bonds were modeled as a continuous isotropic layer; HDI was modeled as a continuous isotropic layer; Flexible silicone glue was used for all adhesion layers Temperature was set fixed at the end(s) of blade at -30C 150% heat load, 7.3 W on blade.

17. Simon Kwan - FermilabCMS Tracker Upgrade Workshop – June 3, 2009 17 Configuration (1) Heat sink on outer edge only HDI being the outermost within module Front SideBack Side 0.06 cf + 0.88 TPG + 0.06 cf Blade 150% heat load, 7.3W per blade; sensor: 0.6 W; ROC: 6.7 W ∆T = 5.2C across model

18. Simon Kwan - FermilabCMS Tracker Upgrade Workshop – June 3, 2009 18 ∆T = 4.6C Across CF ∆T = 4.4C Across TPG Configuration (1) Heat sink on outer edge only HDI being the outermost within module 0.06 cf + 0.88 TPG + 0.06 cf Blade 150% heat load, 7.3W per blade; sensor: 0.6 W; ROC: 6.7 W

19. Simon Kwan - FermilabCMS Tracker Upgrade Workshop – June 3, 2009 19 ∆T = 3.4C Across sensor ∆T = 3.4C Across HDI ∆T = 3.5C Across ROC Configuration (1) Heat sink on outer edge only HDI being the outermost within module 0.06 cf + 0.88 TPG + 0.06 cf Blade 150% heat load, 7.3W per blade; sensor: 0.6 W; ROC: 6.7 W

20. Simon Kwan - FermilabCMS Tracker Upgrade Workshop – June 3, 2009 20 ∆T = 1.0C across the mid-cut section Configuration (1) Heat sink on outer edge only HDI being the outermost within module 0.06 cf + 0.88 TPG + 0.06 cf Blade 150% heat load, 7.3W per blade; sensor: 0.6 W; ROC: 6.7 W

21. Simon Kwan - FermilabCMS Tracker Upgrade Workshop – June 3, 2009 21 OTHER OPTION BEING CONSIDERED Identical 1 st and 2 nd disks, different 3 rd disk to reduce number of modules and material 6 disks = (4x72) outer + (4x48) inner + (2x56) middle = 592 2x8 modules (9472 ROCs) 291 396 η = 1.3 η = 1.6 η = 2.1 η = 2.5 461 2x8s

22. Simon Kwan - FermilabCMS Tracker Upgrade Workshop – June 3, 2009 22 Inverse Cone Geometry An alternative design is to have the pixel modules arranged in a cone We had studied this geometry (see Jan ’09 presentation) and found out that the resolution wasn’t that good Re-visited this geometry during the last couple of weeks With inner and outer assembly, it’s possible to optimize the layout of each separately to obtain excellent resolution in both the azimuthal and radial direction Latest (evolving) idea is to explore concept of inverting the cone (apex towards the I.P) and combining it with the 20 deg tilt for the inner assembly to get better radial resolution for large eta Cone-1 Cone-2 Cone-3 IP

23. Simon Kwan - FermilabCMS Tracker Upgrade Workshop – June 3, 2009 23 Summary We are making good progress to have a conceptual design in the summer Prototype work will follow Design needs to be verified by simulation –Resolution vs eta –Material distribution –Acceptance In good position to have a TDR by next March