1. PS Module Gluing Tests USCMS Outer Tracker Workshop Ulrich Heintz, Meenakshi Narain, Bill Patterson, Sinan Sagir, Adam Lanman, Eric Spencer, Juan Trenado, Bianca Hinojosa Brown University, Providence, RI, USA Trying to produce a thin, thermally conductive bond between the CF base plate and the pixel sensor

2. Gluing Process Apply glue to stencil Squeegee acrossMix and degas epoxy 2

3. Gluing Process Continued Set glass slide on topPlace glass and weights on topRemove Stencil 3

4. Epoxy Degassing Vacuum Recommended 28-29 in Hg, but our pump only capable of 25.5 in Hg. Removes large bubbles, but produces many smaller bubbles. Heat (peltiers on copper plate and heat gun) Works well, but heat decreases viscosity and reduces working time of epoxy. Ultrasonic Bath Works well and is our preferred method. Other Issues Because epoxy is too viscous to pour, removing from mixing cup introduces additional bubbles. Epoxy bubbles after mixing After 20 minutes in sonicator 4

5. Stencils Kapton Stencil (75 um thick) Issue with epoxy getting squeegeed under stencil. → Resolved by wetting glass with alcohol to keep stencil flat Grid Pattern ripped during squeegeeing. Stainless Steel Stencil (50 um thick) Very marginal amount of epoxy gets squeegeed under stencil. Stencil still in good shape although care has to be taken to not crinkle the stainless steel. Tear in Stencil 5

6. Squeegeeing Issues Many factors affect quality of epoxy film. When squeegeeing across, care must be taken to apply no downward force as that causes the squeegee to flex resulting in less epoxy in the middle than the edges. If rate squeegee is dragged across stencil is too slow, especially when attempting to apply no downward force, the squeegee can almost get stuck and you get ridges. Viscosity of epoxy, which changes with time, is also a factor. Higher viscosities produce better epoxy films. Ridges from slow squeegeeing 6

7. Epoxy Uniformity When gluing glass slides on top we were creating large air pockets This was due to unevenness of the epoxy layer. 7 Significant air pockets in the middle of both the grid square and the rectangle

8. Epoxy Uniformity (Improved squeegeeing) 8 Epoxy Film Glass Slide on Top

9. Epoxy Thickness Measurements Method A Used stencils to apply layer of Araldite to glass with both stencils. Measure thickness with profileometer (touch probe) Good sensitivity, but can only measure 2 cm into the sample. 9 Method B Glue glass slides to glass base. Cut glass in middle of slides Examine under microscope Difficult to get cut in slides and cut in base to be aligned, and when uneven microscope can not focus on everything at once. Profileometer

10. Epoxy Thickness (Method A, ~40 um) Rectangle Grid Kapton Stainless Steel (Improved Squeegeeing) 10

11. Epoxy Thickness (Method B, ~60-70 um) Using known thickness of the glass (3/32”) calculate thickness of epoxy layer (Araldite). Likely due to uneven cuts and issues with the focus and neither the piece nor the microscope being strictly vertical there was a variation between 40-80 um for the measurements. Median thickness for kapton, 60 um, and for stainless steel, 70 um Kapton, Rectangle PatternKapton, Grid Pattern 11

12. Stamping Create and fill epoxy reservoir with tape (150 um) Stamp pattern onto glass base Set glass slides on top and place weights same as was done with the stencils Can’t really clean between grooves so the stamps slowly gum up. Epoxy reservoir for stamps Stamp Patterns 12

13. Stamping Except for the epotek, epoxy is too viscous to merge together the patterned stamps to create a uniform layer. Solid stamp does not produce a solid layer of epoxy. Awaiting the result of the glue thickness measurements for the stamps. Stencils preferable to get a solid uniform layer. 13 Solid Stamp Epoxy doesn’t merge between gaps

14. Gluing Aluminum “Sensors” Also tried 250 um thick Al pieces cut to the size of the sensors. Thin aluminum much less rigid than 3/32” glass (although also less rigid than silicon) Even placing foam under the weights to better distribute the pressure still results in air pockets. 14

15. Epoxies EpoxyWorking Time (hours) Cure TimeViscosity (cps) Thermal Conductivity (W/m-K) CTE (ppm/C) Glass Transition (C) Cost Araldite 2011212 hrs @RT45,000.22-.258563C$23/50ml Epotek 30182 days @ RT225-425N/A6180 C$36/30ml 3M TC-2810 (Boron Nitride) 124 hrs @RT40,000 - 80,000 1.0-1.46260-100C*$68/37ml EP48TC1.5-24-7 days @RT Paste2.88-3.613-15149C?$550/30ml $800/240ml Ablefilm 550KN/A30 m @125C 2 hrs @150C film0.850102C$170/ft 2 * Dependent on curing temperature (60C at RT curing) 15

16. Conclusions Higher viscosity epoxies (Araldite and 3M) produce better films with stencils or stamps. So far Boron Nitride 3M epoxy has the best thermal properties and viscosity for using with the stencils. Process can still be improved to eliminate air pockets as much as possible. 16

17. Further Studies Try using film adhesive, Ablefilm 550K Try paste epoxy with excellent thermal conductivity and CTE (EP48TC) Build a jig to set down mock sensors and apply pressure to replace setting by hand and pressing down with weights on top of glass. Start gluing silicon to carbon fiber base plates and do thermal conductivity experiments. 17