Large-area tracking systems with a good space resolution require modern techniques not only in the production of sensors but also for the signal readout of a working detector module. Modern sensors with a high density of independent sensitive areas (strips, pixels etc) as well as the corresponding readout-chips can only be produced in state-of-the-art industrial production lines. But although these most delicate constituents are manufactured in Industry, it is still up to us physicists, to put things together, to make a high resolution position sensitive detector-module out of some pieces of silicium.
Obviously, the mechanical assembly of a detector module is a chapter of its own, but this page is dedicated to the simple task of making the electrical connections.
If you are new in this field you are certainly wondering why grown-up physicists make fuss about such a simple task and that they even dedicate a written document to it. The answer is in the following three statements:
- the available space to make one connection is extremely small: you have only about a hundred micometers from a connecting wire to its neighbours;
- the total number of connections to be made is very high: alone the Karlsruhe contribution to the CMS bonding amounts to several million connections;
- most important of all is that a connection should not only look like a connection, but it has to fulfil certain quality requirements which have continuously to be checked
Item (1) determines the diameter of the wire: 25 micrometers. Item (2) determines how to approach the task: buy an automatic Industry-standard machine which is fast and reliable. Item (3) means: check each bonded detector.
Our preferred wire is 25um Aluminum (1%Silicium).
As bonding tools (wedges) we use mainly the multipurpose tool UT45A-W-2020-1"-C (SPT Roth Ltd, Switzerland).
We also experimented with the fine-pitch tools FP45A-W-2020-1"-C and FP45A-W-2020-1"-CM, but there was no evidence of worse or better performance of any of the three.
The bond-strengh is monitored through pull tests. We obtain quickest feed back with a small hand-held pulltester (Correx 2-15 from Bruetsch/Ruegger AG Switzerland)
Furthermore we can perform a systematic exploration of the bond-parameter space with a high precision pull-tester belonging to the Forschungszentrum Karlsruhe (IPE). Since it is much more time consuming to go to another building carrying a pile of bond-samples, this is only done after bigger changes in the system.
 |
Bonder Hesse&Knipps |
- CMS
- DELPHI
- various R&D
The CMS silicon Tracker consists of about 15000 detector modules of various shapes and sizes. They are produced by an international collaboration of research institutes. The Karlsruhe share of the total bonding load is the bonding of all (1600) modules for Ring-5 of the Tracker-End-Caps. Ring-5 is one of the rings to provide r-phi information, so there are modules of two different geometries (800 + 800) which will be mounted back to back (after bonding and testing).
Each module consists of two silicon sensors with 768 stips each. So the bonder has to make the connections from the readout-electronics (pitchadaptor) to sensor-B and the connections from sensor-B to sensor-A; in total 1536 connections plus some HV and ground connections per module. In total the production of Ring-5 requires 2.5 million connections, within a time range of 1.5 years!
- Bonding Homepage at CERN
- Talk on first pulltest results (CERN, 28.Sept.2000)
| Name Profession / Beruf |
Phone | |
| Tobias Barvich Techn. Angestellter |
barvich@iekp.fzk.de | +49 7247 82-3819 |
| Hans-Jürgen Simonis Wissenschaftler |
simonis@iekp.fzk.de | +49 7247 82-4300 |