sPHENIX tracking discussion

["Frawley, Anthony" <afrawley AT fsu.edu>]

Hi All,

Here are the minutes from today's sPHENIX tracking meeting.

Cheers

Tony

Sookhyun Lee - initial Z-vertexing for MAPS+TPC
------------------------------------------------------------------
Sookhyun Lee reported on her work on initial vertexing for the MAPS+TPC configuration. This replaces using (perfect) BBC information to get the initial vertex.

Found that using all layers in the TPC did not give a sensible Z vertex.
Experimented and found that 4 layers (3 MAPS + 1 TPC) works best, gives 5 micron Z vertex resolution.
Found that 5 layers worsens Z vertex resolution, adding more TPC layers makes it worse still.
Need several hundred tracks to get good Z vertex with 4 layers.

Discussion:
---------------
Sookhyun is using 100 embedded pions in central Hijing event, but not making a distinction between embedded tracks and Hijing tracks
Suggested to use only Hijing tracks for this.

Needing hundreds of tracks to get good vertex resolution will be a problem for peripheral Au+Au and p+p and p+Au.

Not obvious why adding more TPC layers should worsen the vertex resolution.

Is the large number of tracks required because of fake tracks with 4 layers?

Suggested to try making a threshold cut on the track pT (maybe 500 MeV?)

It would be good to make a slide showing the flow of the tracking steps with and without BBC Z-vertexing

Itaru Nakagawa - longer INTT design
-------------------------------------------------
Itaru discussed the issues with lengthening the intermediate tracker (INTT) to provide eta < 1.1 coverage for events at Z = 10 cm.
There was a long discussion, I will just summarize the issues and conclusions:

Issues:
---------
Itaru showed a design that increases the length of the detector by stretching the silicon strips, so that we stay within a feasible channel count.
But increasing the length of the INTT causes a conflict with the fan-out at one end of the MAPS barrel that carries signals off the MAPS.
There are two conflicts:
   The INTT sensors for layers 1 and 2 (out of 0-3) invade the volume occupied by the MAPS readout fan-out.
   Even if that was not true, there is no room for the INTT signals to be carried off the detector on the MAPS fan-out end.
Possible solutions are:
   Make the MAPS fan-out much more compact
   Read out the INTT at one end only, opposite to the MAPS readout end
We want to consider both options.

Itaru pointed out some problems with reading out the INTT at one end only:
   The HDI bus would need to be 500 mm instead of 250 mm long. Are there technical problems? Ask the manufacturer.
   One method, doubling the width of the HDI from 40 mm to 80 mm so that it can read out the north and south sides of the ladder, doubles the mass.

Discussion:
-----------
A suggestion for reading out the INTT at one end without increasing the size of the HDI is to reduce the channel count by doubling the silicon strip pitch.
We have no evidence that 160 micron strip pitch is a problem for the job that the INTT is intended to do (pattern recognition).
Would increase strip occupancy, but only by x2. No evidence yet that this is a problem.
The HDI still needs to be 500 mm long.
Need simulations to see what effect this would have on performance of the INTT as a pattern recognition detector.

We need to explore what modifications are possible for the readout cone on the MAPS.
   Could move the apex of the cone out? Rachid said that at the end of the ladder there is an extender between ladder and ROC. Can it be lengthened?
   Explore reducing the angle of the MAPS readout cone.
Ming: We are looking at the design of the readout of the MAPS.
It is limited by the flexible PC board. Will find out if we can make a custom HDI readout

Action items:
------------------
Explore modifications to the MAPS fanout to reduce conflicts with INTT - LANL
Simulate whether 160 micron pitch for the INTT works in pattern recognition - TBD
Make simulations to address how many layers of INTT we need, and where - TBD. Need to consider live area - use existing FVTX performance (Itaru will look into this)
Explore with the manufacturer whether increasing the HDI length to 500 mm is technically feasible - Itaru

Tom and Prakhar:
------------------------
Showed results of:
a) Ansys calculations of electric fields due to mechanical distortions
b) Feeding the electric field to Garfield to propagate electrons through the electric + magnetic field

Found that our 1.5T magnetic field drops the effect of a 500 micron membrane shift to 100 microns in position resolution.
Have this tool now, will use it to establish a complete set of mechanical constraints that produce acceptable positon resolution
The goal is have 100 microns position resolution due to everything combined.

The 100 micron number is from the test beam result for the ILC TPC. That is what Alan originally put in his code