CALICE MAPS Meeting, RAL, 28/02/06 ================================== Present: Jamie Crooks, Paul Dauncey, Anne-Marie Magnan, Yoshi Mikami, Renato Turchetta, Giulio Villani, Nigel Watson Minutes: Paul Minutes of last meeting: No comments. Oversight Committee: The first meeting of the PPARC Oversight Committee (OsC) for CALICE will be on Fri 24 Mar. These will occur every six months. We have to submit a report by Tue 14 Mar which will contain information on progress in the five CALICE-UK workpackages, of which the MAPS work is one. Each workpackage will need around 2-3 pages of text and figures. Paul has a rough draft of the text but would like a figure from each of the three main areas of work; namely sensor design, sensor simulation and physics simulation. Jamie, Giulio and Nigel, respectively, will send these to Paul by the end of the week for a first release on Mon 6 Mar. The report will also contain the financial spend (which is simple as it is effectively zero so far) and schedule together with milestones. As this is the first meeting, we are free to set the milestones now but it will be harder to change them in future. The schedule was defined in the first meeting on 01/09/05 and was set to: o Feasibility study - continuing until end Dec 05 o First design - Start Jan 06 to end Dec 06 o First fabrication - Start Jan 07 to end Apr 07 o First basic tests - Start May 07 to end Jun 07 o First detailed tests - Start Jun 07 to end Mar 08 o Second design - Start Jul 07 to end Dec 07 o Second fabrication - Start Jan 08 to end Apr 08 o Second basic tests - Start May 08 to end Jun 08 o Second detailed tests - Start Jun 08 to end Mar 09 o Beam test PCB design and fabrication - Start Oct 07 to end Sep 08 o Beam test period - Start Oct 08 to end Mar 09 This would mean we should have completed the feasibility study by now and moved onto the design. However, Renato described the ISO9000 system at RAL, which formalises the changes between difference phases of a project. The move from feasibility study to design requires a Preliminary Design Review which has not taken place. This is a relatively detailed review and should specify the sensors in some detail, including pixel size, memory requirements, etc. It was thought that this could be done towards the end of Apr. There would in reality be some design work in parallel with this study so the end date for the first design of the end of the year was still thought realistic. Hence, the schedule was changed to reflect this and becomes: o Feasibility study - continuing until end Apr 06 o First design - Start May 06 to end Dec 06 o First fabrication - Start Jan 07 to end Apr 07 o First basic tests - Start May 07 to end Jun 07 o First detailed tests - Start Jun 07 to end Mar 08 o Second design - Start Jul 07 to end Dec 07 o Second fabrication - Start Jan 08 to end Apr 08 o Second basic tests - Start May 08 to end Jun 08 o Second detailed tests - Start Jun 08 to end Mar 09 o Beam test PCB design and fabrication - Start Oct 07 to end Sep 08 o Beam test period - Start Oct 08 to end Mar 09 The reviews usually become (some of) the project milestones and it was thought that this would be appropriate here. The future reviews would be roughly half-way through the two design phases and just before the two sensor fabrications. This would therefore mean there would be Interim Design Reviews at the end of Aug06, Dec06 and Sep07, and the Final Design Review would be at the end of Dec07. A Concluding Review of the whole project may be held in Mar09. The milestones were set accordingly as: o Preliminary Design Review, Apr06 (month 7). o Interim Design Review, Aug06 (month 11). o Interim Design Review, Dec06 (month 15). o First fabrication round completed, Apr07 (month 19). o Interim Design Review, Sep07 (month 24). o Final Design Review, Dec07 (month 27). o Second fabrication round completed, Apr08 (month 31). o Beam test started, Oct08 (month 37). Conferences: Renato has been approached by Marc Weber (RAL/PPD) to give an invited talk on CALICE and MAPS at the Siena conference on Innovative Particle and Radiation Detectors in Oct; see link on usual web page. This conference has an audience more towards physicists than engineers, so a talk including physics simulation results would be appropriate. Two abstracts could be submitted; one on the sensor design and simulation and the other on the physics performance. The abstract submission deadline is in Jul so it is not urgent. Giulio knows Marc well and will check with him what would be appropriate. Jamie, Giulio, Anne-Marie and Yoshi would all be available to give the talk in principle. There is also a conference on front-end electronics in Perugia in May, which is probably too early for a MAPS talk; again see the link on the web page. However, Jamie may be presenting his MI3 work and could add on a few transparencies on the applications of the MI3 design to CALICE if appropriate. The IEEE NSS-MIC conference is in San Diego in Oct/Nov and we could also submit an abstract to this; see the web page for the link for this conference. The deadline for abstracts is May. Sensor design: Jamie has been tied up with the MI3 work for the last month. However, this will be sent out to the University groups tomorrow which will allow him to start up CALICE work again. He will work 50% on CALICE during Mar and then be 100% from Apr onwards. He hopes the MI3 tests will include the SEU effects of particles in the DRAM, although this is not a high priority for the University groups. Jamie's talk in the meeting on 09/12/05 included a lot of design choices. He will choose between most of these himself as the design develops and does not need much external input. However, some things depend on Giulio's results and in addition the pixel size is the main parameter which needs to be confirmed by the simulation. Renato showed a slide of the processes and costs of three possible foundries. Foundry B has a stitchable 0.18mu process already and is close to releasing a 0.13mu process. The latter would be a factor three smaller in each direction than the 0.35mu process considered until now, which would probably be enough to allow a 25mu pixel size. However, the 50mu size remains the baseline for now. Note, foundry A has both 5 and 15mu epitaxial layers, whereas B is only available in 5mu. Unfortunately, it is necessary to choose the foundry relatively soon; the actual design needs to targetted to a particular foundry's design rules. Sensor simulation: As reported in the last meeting, Giulio is unable to do simulations for 50mu pixel sizes due to the 30GByte disk limitation. Since then, 3TBytes of new disk have been ordered and should be installed in about four weeks. Meanwhile Giulio has continued the 25mu pixel simulations, aiming to do all combinations of single and multi-diodes with 15 and 20mu epitaxial layers; see his slides on the usual web page. Completing these simulations will take effectively all the time up to the installation of the new disks. With a 15mu epitaxial layer, he still sees that a threshold around ~50-60% of the MIP signal is optimal. The actual charge collected is slightly higher for 15mu than 20mu, despite the shorter track length. The MIP signal corresponds to ~450e-, meaning the threshold is around 250e-. We are aiming for a threshold around 4-5sigma, so this would require the noise to be ~50e-, which is a reasonable target. Giulio will scan the epitaxial layer thickness to find the optimal value when the new disks are installed. It may be that a 5 or 10mu thickness is better, despite the shorter length. The noise should be independent of the epitaxial layer thickness to a large degree so the choice of the optimal thickness depends only on the charge collected and the degree of diffusion. Meanwhile, we should continue to assume a baseline epitaxial thickness of 15mu. From the simulation side, the 5mu mapping which Giulio produces is quite fine and a coarser granularity would be easier if it is not too bad an approximation. Giulio thought up to 10mu would be reasonable, but beyond that, it is not negligible compared to the typical diffusion lengths. Physics simulation: Nigel went through the issues which have arisen in starting to implement a MAPS simulation in GEANT4/Mokka; see his slides on the usual web page. Paul described the issues downstream, when trying to convert the GEANT4/Mokka output into the equivalent of the raw sensor data (the "digitisation" step). This will run in Marlin using LCIO. His slides are also on the usual web page. In both cases, the number of elements to be considered and their very fine granularity cause problems. The optimal solution is not yet clear and some work to understand how the internals of GEANT4 might handle high numbers of cells needs to be done. In principle, there will be ~5mu of silicon and metal above the epitaxial layer but as this is insensitive material and will have little effect on the shower, it can safely be neglected. Also, the epitaxial layer thickness is not well-defined; there will be some charge collected from the substrate so using a thickness of 20mu in the simulation might better model a 15mu epitaxial layer plus much of the region of the substrate which contributes. One complication with Giulio's charge diffusion mapping is that it in principle depends on the depth at which the charge is deposited. He may need to make a 3D mapping table, where the epitaxial layer is divided into 5 or 10mu planes in depth. Next meeting: 1pm on Thu 30 Mar in R76 as usual. Jamie will not be able to attend.