Minute of the Tracker Technical Meeting September 13, 2000 Present: SLAC, UCSC, Pisa, Trieste, Hytec Michela reported via email some progress on thermal testing of the Agile tray configuration. The following paragraph is her report: We have started making trials of gluing tungsten with our CERN colleagues. We have used tungsten sheets 300um thick with a dimension of 12x38 cm**2. They have been glued on a sandwich 11.465mm thick (honeycomb + 500um carbon fiber). One of them has been glued with Dow Corning 3145 and one with Araldite 106. The thickness of the silicone glue is 0.379mm (which is too high; the weight we applied on the tungsten was too low to guarantee a thickness of 100um which is what we want to reach) while for the epoxy we have a thickness of 0.182mm. In the case of the silicone, we tried to have a uniform layer, with the result that there are air bubbles. In the next trial, we will deposit the glue in half circles one inside the other instead of a uniform layer. Our CERN colleagues have then measured the variation in thickness as a function of the temperature, at 21 and 60C. We have measured 40 points on the tungsten sheets (20 on the border and 20 in the center of the sheet) to understand the behaviour in the different regions. The piece glued with silicone presents a variation in thickness between 21 and 60C which does not exceed 15um and in average is around 1um. For the one with epoxy, we have a variation which exceeds 100um and in average is 27um. We are going to make the measurement also at -20C and then go on in the gluing of a full W sheet (38x38cm**2) on one of the AGILE sandwiches (12.7mm of Al honeycomb 5056 3/16 .0.0015 CRIII, and 500um of carbon fiber on each side 985 T300 obtained gluing 4 skins with Redux 312 UL). Ossie reported that a silicon wafer glued to an aluminum slab with a high modulus adhesive (2216 epoxy) were tested up to the high end of the environmental spec. The maximum stress in the silicon, inferred from the observed strain, were a factor of 4 below where failure is expected. The will run the test to low temperatures today, going down 15C at a time, down to the low end of the environmental spec. The results still are somewhat rough and need to be corrected for the CTE of strain gauges. The monitor strain every ½ second, but the system could go up to as high as 1kHz. During the test, air flow was causing vibration of the reference piece. They now have a cardboard box with holes covering it and will rerun the curves today before going to low temperature. They used a complete, uniform coverage of adhesive for bonding the silicon to aluminum. Pisa said that they would like to see the data from SLAC, which Ossie said will be forthcoming in a report. Ossie said that SLAC is now ready for the baby detector coupons. It takes about 2 days to get strain gauges mounted. Gwelen will bring them over to SLAC next week. SLAC has temperature monitors available in their system via thermocouples. SLAC will repeat the tests with dot glue patterns in place of the uniform coverage. They also want to test with a low-modulus adhesive. It was decided that Ossie should buy more tungsten samples, of 3.5%, 15%, and 25% radiation lengths (or in those neighborhoods, depending what is available in standard thicknesses). He will order 20 of each thickness for each site (SLAC and Pisa). He will buy them oversized and wire EDM at SLAC to size. Another option is to buy large sheets and cut at SLAC. It has been decided to baseline tungsten for the thick converters. Both lead and tungsten remain as options for the thin converters. There was some inconclusive discussion of whether the tungsten should be pure or have a slight (i.e. 3%) amount of copper. Ossie remarked that we need to look into uniformity of the converters. There is a spec on uniformity, but Robert remarked that the spec should be taken with a grain of salt and compared carefully with what is economically available. Robert presented a note (attached) on three concepts that are under serious considerations for detector attachment: 1. Dots of high modulus non-conducting adhesive, with localized regions of low- modulus conductive adhesive for back biasing. 2. Dots of low-modulus adhesive, with gold-copper alloy springs in localized areas for back biasing. 3. A low-modulus adhesive between the converter and bias circuit, with a conductive adhesive (possibly high modulus) between the bias circuit and silicon detector. Robert advocated testing the first concept as soon as two candidate adhesives have been identified, using the existing 300um carbon-fiber face sheets, bias circuits, lead and tungsten foils, and full-size ladders made from soon-to-be-available aluminized silicon wafers. Robert reported that UCSC is planning to test thermal degradation of the electrical contact in the second concept, by way of thermal cycling and accelerated aging at high temperature. It was recommended that we use a non-oxidizing atmosphere for the aging test. Tom said that such contacts need a degree of built in wiping, to keep oxidation scraped away. He suggested to make the test as real as possible by using materials with the expected CTE. Robert said that the 3rd option probably mainly concerns the thick-radiator problem, so Pisa should consider it in their investigation and modeling. Alessandro spoke about his plans to make prototype pitch-adapter Kapton circuits for the 90-degree interconnect. He can obtain 10 samples in short time from CERN. He believes that they can design it to bend without breaking strips, even with a (non-Kapton) cover layer. Robert recommended making the first prototype to the BTEM dimensions to allow testing with existing hardware. UCSC is working on a drawing of the tooling holes needed in the circuit. Alessandro expects about a 1 week turn-around. He is also working on getting a quotation for production. He is doing the same for the detector bias circuit. He said that they can make very thin Kapton. Robert remarked that the IPC rules will likely demand much thicker Kapton between the bias layer and the ground-potential shield layer. Ossie said that dicing of the test silicon wafers should be done this week. He has an aluminizing person lined up. He estimated 2 weeks before good mechanical samples with coating and laser marking will be in hand. The gluing machine recquisition has gone to purchasing. It may still be in purchasing. Ossie will check up on it. John reported that he has materials and a design for the edge dipping machine and will start making parts now. There was a discussion of the detector long term testing program at UCSC. Gwelen said that he would like to do for each item one test with N2 flushing, one without. Elliott expressed concern that the tests should really be done in vacuum. Tom agreed to make recommendations on the long-term testing environment. Tom reported that he has a name of someone who worked for many years in the solar panel industry. Tom and Ossie will talk to him about experience with attaching silicon to panel structures. Andrea reported that at Pisa they are collecting data on materials and are starting to study the payload issues. He needs data from Hytec on the baseline facesheet choices. He is working with the Pisa university aerospace department. They already are making models and are studying candidate face-sheet materials. Alessandro reported that at Pisa they are reading baby detectors with CMS electronics (fast integration), both from AC pads and from DC pads. They are working well. They need about 1 more week to obtain a VME-PCI interface to be able to use the GLAST TEM board from a PC. Nick and Gloria have been looking at Wilko's code for testing the system. Andrea reported that they have some problems with importing BJ's drawing files. Alessandro would like to obtain a copy of the ladder gluing tool. John said that he has the drawing, which he made himself. BJ is working on making fab drawings of the jig. When those are ready they will be sent to Pisa, so that the Pisa machine shop can make a copy. For work on the ladder placement tool Alessandro needs input on reference points on trays. There is an outstanding issue of what we need in the way of secondary fiducials. It was agreed that Sandro would design using the pinholes for alignment. Sandro asked how will converters and kapton be installed? Robert replied that responsibility for developing that procedure will probably fall on John at SLAC, since he did the work for the BTEM. However, Robert stressed that we must make sure that the procedures can be carried out efficiently by an average technician, and the procedures must be thoroughly documented (which applies to all of the assembly procedures, of course). Gwelen asked Hytec if the tray has a feature that sticks up above the silicon. He said that there is an issue with the Pisa wire bonder, which doesn't have as much clearance as the UCSC one (unless the 30-degree head is changed to a 60-degree one). Issues: 1. What tungsten alloy should be used for the converters? 2. Clearance of the wire-bonding head with the corner posts. 3. What secondary reference fiducials are needed on the trays? 4. Pisa has difficulties importing BJ's drawings. 5. What environment should be used for detector long-term testing? 6. Violation of IPC rules in the current bias-circuit design. 7. What is the uniformity needed (or easily obtained) for the converters? Decisions: 1. Use the corner holes for alignment of the ladder mounting fixture. Action Items: 1. Ossie et al.: complete thermal tests of Si on Al, to low temp, with adhesive dots, and with a compliant adhesive. 2. Gwelen: take baby-detector coupons to SLAC for strain-gauge mounting and testing. 3. Ossie: get more information on pricing and properties of tungsten alloys. 4. Gwelen and Alec: make test coupon with gold-copper alloy spring contact between bias circuit and silicon. 5. Gwelen: send drawing of tooling holes for 90-interconnect fabrication to Pisa. 6. Gwelen: look into the IPC rules for the bias circuit. 7. Gwelen: work the bonder-head clearance issue offline. 8. John: start prototyping the edge dipping tool. 9. Tom: recommend environmental conditions for long-term detector testing. 10. Tom, Ossie: talk to the solar-panel expert about mounting silicon onto panels. 11. Erik: send Hytec facesheet data to Pisa. 12. BJ: complete fabrication drawings of the ladder assembly fixture. 13. Ossie: check up on the glue-machine procurement.