Half the material budget in CLD vertex disks

[armin-ilg]

Check duplicate issues.

• Checked for duplicates

Goal

I had a closer look at the material budget of the CLD vertex detector as I wanted to use it as a first version of disk detector for FCC-SEED. Looking at the reference table 4 from arXiv:1911.12230v3:

I expected a material budget of 0.7% X0 per double layer in CLD. I however observed only half that (3x0.35=1.05% instead of 3x0.7=2.1%):

This stems from the fact that, while using the correct numbers for the thicknesses of the support and sensor:

<constant name="VertexEndcap_Sensitive_Thickness"   value="5.000000000e-02*mm"/>
<constant name="VertexEndcap_Support_Thickness"     value="28.000000000e-02*mm"/> <!-- +50% more material budget as in CLIC VTX -->

only half that thickness is used:

<module name="SiVertexEndcapModule1" vis="SiVertexModuleVis">
    <trd x1="VertexEndcap_rmin1 * tan(pi/(VertexEndcapModules))+VertexEndcapOverlap*sin(pi/VertexEndcapModules)" x2="VertexEndcap_rmax * tan(pi/(VertexEndcapModules))+VertexEndcapOverlap*sin(pi/VertexEndcapModules)" z="(VertexEndcap_rmax - VertexEndcap_rmin1) / 2" />
    <module_component thickness="VertexEndcap_Support_Thickness/2.0"  material="Silicon" vis="SiVertexPassiveVis"/>
    <module_component thickness="VertexEndcap_Sensitive_Thickness/2.0"  material="Silicon" sensitive="true" vis="SiVertexSensitiveVis"/>
    <module_component thickness="VertexEndcap_DoubleLayer_Gap/2.0"  material="Air" vis="SiVertexAir" />
</module>

Given that this has not been discovered for such a long time in all CLD versions, changing the numbers now probably doesn't make sense as discussed with @Zehvogel and @andresailer. It would, however, be good to add a comment in the CLD vertex xmls and also in the CLD readme about this. The vertexing performance in the forward direction would be significantly worse if the material was doubled.

Note: In the CLICDet, the mat. budget is correct.

Thank you,
Armin

Operating System and Version

Alma Linus 9

compiler

Standard

The version of the key4hep stack

any

Package Version

any

Reproducer

See any vertex detector xml file of CLD

Additional context

No response

[armin-ilg]

As discussed in the tracker+PID meeting yesterday, here's a comparison between the case of double the current mat budget (so in line with the CLD paper):
t_dist_double_disk_mat_budget.pdf

and the one currently in k4geo (so half of what is in the paper):
t_dist_CLD_o2_v08.pdf

As can be seen in these plots, the difference in vertexing performance is actually rather mild. I think the material of the beam pipe in front is much more relevant than the one of the vertex disks themselves.

[s6anloes]

What are @andresailer and @Zehvogel thoughts about updating these values under a new version: CLD_o2_v09? This way it's easy to make clear all previous studies were made with v08.
But if you say the change is not big enough to warrant a new version, we can also include it later, maybe if another update requires a new version.
Silently updating v08 seems not quite right, I'd say

[andresailer]

We have new detector versions for moving lumical by 100 micron. We can make a new version of this. Of course changing existing versions is a no go. More sensibly would be an actual update of the whole setup with the developments of silicon technology in the last eight years.

[Zehvogel]

and the one currently in k4geo (so half of what is in the paper):
t_dist_CLD_o2_v08.pdf

Is this the correct one? The plots say FCC-SEED.

Silently updating v08 seems not quite right, I'd say

I don't think anyone is suggesting this. ;)
We surely should make a fixed version v09. I would suggest that @armin-ilg just opens a PR with his version. :)

More sensibly would be an actual update of the whole setup with the developments of silicon technology in the last eight years.

Yes, that would be great! But I think for now we can also be happy with just this fix. :)

[armin-ilg]

and the one currently in k4geo (so half of what is in the paper):
t_dist_CLD_o2_v08.pdf

Is this the correct one? The plots say FCC-SEED.

Silently updating v08 seems not quite right, I'd say

I don't think anyone is suggesting this. ;) We surely should make a fixed version v09. I would suggest that @armin-ilg just opens a PR with his version. :)

More sensibly would be an actual update of the whole setup with the developments of silicon technology in the last eight years.

Yes, that would be great! But I think for now we can also be happy with just this fix. :)

Sorry, messed up the plots. Here are the correct ones:

t_dist_CLD_double_mat_budget_disks.pdf
t_dist_CLD.pdf

So looking at the disks, the differences indeed are minor.

Okay, then I'll make CLD_o2_v09 with the fixed mat. budget in the disks and you can make the lumical changes on top. I'll include it in the other large merge request going on (#612).