GEM Framing Structure

To Do List


Resp. Person




Design the Grid frame





Estabilish the pattern priority





Enquiring for the possibility of building the frame in pieces





Asking for cost estimates




Measure the flatness of an Asian Module

Ask for the possibility to groove one side of the frame

Evaluate the possibility of a gate GRID

Not Started


Status Summary

One of the main challenges in the production of the DESY LPTPC Module, which will take it apart from the others, allowing for a different range of studies, is the design of a new framing structure for the GEMSs which would allow for a greater gain uniformity in the GEM stack. The only other GEM Module with a fully sensitive surface built for the LPTPC, the Asian GEM module, has a GRP framing only on the upper and lower side of the GEM which causes the GEM itself to sag in the middle (the sagging has not been quantified) and be unstable on the sides.

The DESY framing and support structure will be otherwise made out of ceramics (a material which is at least 10 time stiffer than GRP) and will feature a support grid in the internal area of the GEM itself. After a few iteartion in the process of designing the frame and receiving feedback from the ceramic producers we finalized a design which will have the structur as thin as 1 mm. Thinner structures have been found to be to fragile and prone to cracking during the manufacturing process. Out of many initial possibilities, and after finalizing the design of the GEM foil, we decided for a simple cross like structure to be the internal grid of the frame. The elements of this structure will be aligned with the electrode gaps of the GEM thus almost avoiding to increase the amount of unsensitive area of the detector. In fact the 1mm size of the ceramic bar convinced us to increase to 1 mm the size of the gap from the 200 um used, for example in the GEMs for the COMPASS experiment.

The cross-like pattern has a major drawback in the presence of the vertical bar which will undermine the reconstruction of tracks superimposed to the bar itself. Nonetheless this design has been chosen also to test the effect of this vertical dead area on the track reconstruction efficiency, being this an R&D prototype. Once the first prototype of the GEM Module is up and running it's also foreseen to produce a few other frame models which can allow us to test the efficiency of different patterns. We also evaluated the possibility to metallize the external surface of the frame to solder the GEM contact on it but we decided to go for the easier solution of soldering the GEM electrodes directly on soldering pads on the PCB surface.

With the grid design finally complete we are now waiting for cost estimates from the producing company. At the moment we have found two companies which are interested in the production of this device. The first firm is Anceram and also produced the small grid-like frame used for the feasibility studies for this framing strategy performed on the standard 10X10 cm CERN GEMs. This company produces its own ceramics and laser-cut the plate afterwards. The second company is Lasercutting-ceramics, specialized in laser cutting ceramics which have to acquire the ceramic plates from external sources.

We already received a complete cost estimate from Lasercutting which is 312.50 GBP for 10 frames cutted out of a 99% alumina ceramics. The main reason for the price to be that high is the necessity of ordering custom sized ceramic plates from which to cut our frame not existing in commerce big enough standard alumina plates (the biggest being 15 X 24 cm). To solve this problem we thought about building the frame in pieces to glue afterwards. If we could build the frame out of a standard sized ceramic plate the prize would decrease to 75 GBP. I will return on this point later.

Anceram is instead still delaying the cost estimate we ask them for. A first request was answered with a proposal of 890 € to perform a test cut in a 96% alumina ceramics. It's reasonable to believe that the cost for a order of a few pieces will decrease the price per piece by a large amount but, even after repeated requests I've not received any answer yet. It's also reasonable to believe that, being able to produce their own ceramic, the sizing problem will have a smaller effect on the price of the piece. Anyway the constant delay in each and every step of our dealing with them makes me question their reliability and their interest in the project.

Coming back to the problem of building the frame in such a size that will fit the biggest ceramic plate available we have two possibilities:

The first option is not desirable because we will infact reduce the impact of our studies, not only reducing the sensitive area of the structure, but also reducing the total unsupported area of the GEM which will, in turn, decrease the importance of our studies on the possibility to ensure a good flatness and a good gain uniformity with a big sized GEM. The second option has been evaluated asking for advice to more than one expert. The unanimous opinion is that the small size of the ceramic structure (1X1 mm) will make the glued section very fragile and easy to bend. To increase the glueing surface one should device step-like structres which can be very fragile and difficult to manifacture. Building two separate frame for the upper and lower side of the GEM and gluing it in the middle, not only will increase the size of the horizontal gap to 2 mm but will also make the two sides easily to bend one towards the other decreasing the effectiveness of this ceramic structure to ensure the flatness of the GEM. Moreover it will be very difficult to make sure for the to pieces of the framing structur to be at the same level, without any stepping at the glueing point. Finally it will not be easy to remove the glue extruding from the junction keeping the flatness of the area in the desired range (less than 100 um)

My proposal, to be approved, is to proceed with the production of the ceramic frame in a single piece out of the 99% alumina ceramics from the Laser Cutting company, unless the cost estimate of the Anceram company is much lower. Infact the Anceram company has proven to be quite unreliable, contrary to the Laser Cutting company, moreover the 99% alumina is a little bit stiffer (10-20%) than the 96% alumina which will increase the performance and resistance of our frame thus justifying a moderately (10-20%) higher cost. In the same time we may also order a few frames built out of multiple pieces to test whether it's possible to produce this device this way and to compare the two different solutions.


DesyModGrid (last edited 2016-07-08 09:51:45 by GeorgVonBismarck)