sPHENIX HCal discussion

[Xiaochun He <xhe AT gsu.edu>]

Dear Shuhang,

Thanks for sharing the nice comparison plots!

Cheers,

Xiaochun

--

Xiaochun He

Regents Professor

Department of Physics & Astronomy

Georgia State University

Atlanta, GA 30303

Email: xhe AT gsu.edu

---

From: Shuhang Li <sl4859 AT columbia.edu>
Sent: Thursday, September 22, 2022 11:55 AM
To: Xiaochun He <xhe AT gsu.edu>
Cc: pinkenburg <pinkenburg AT bnl.gov>; W.A. Zajc <zajc AT nevis.columbia.edu>; sPHENIX-HCal-l <sphenix-hcal-l AT lists.bnl.gov>
Subject: Re: [Sphenix-hcal-l] Birks constant for G4_POLYSTYRENE

 

Dear Xiaochun,

I think there are comparisons between our newest build of HCal(new geometry with G4_POLYSTYRENE used for the scintillators) to the old HCal simulation automatically generated by Jenkins, for example: https://nbviewer.org/github/sPHENIX-Collaboration/QA-gallery/blob/jenkins-sPHENIX-test-calo-single-qa-3646-pi+_pT30_Sum8/QA-calorimeter.ipynb , the references are done with the old HCal.

The sampling fraction agrees pretty well.

Best regards,

Shuhang

On Thu, Sep 22, 2022 at 9:20 AM Xiaochun He <xhe AT gsu.edu> wrote:

Dear Chris and Shuhang,

 

Thanks for the info which is indeed very helpful. I was just thinking about to ask Shuhang to run a comparison since he did inform me the difference in energy loss between the two materials while I was at Edinburgh. My apology for not being able to attend the calibration meeting yesterday because of the ePIC general meeting.

 

Shuhang, are you interested in creating the HCal scintillating material by copying G4_POLYSTYRENE since you are one of the consumers? This will save time to get this done once for all?

 

Best,

Xiaochun

--

 

 

From: sPHENIX-HCal-l <sphenix-hcal-l-bounces AT lists.bnl.gov> on behalf of pinkenburg via sPHENIX-HCal-l <sphenix-hcal-l AT lists.bnl.gov>
Date: Thursday, September 22, 2022 at 8:57 AM
To: sPHENIX-HCal-l <sphenix-hcal-l AT lists.bnl.gov>
Subject: Re: [Sphenix-hcal-l] Birks constant for G4_POLYSTYRENE

Dear Xiaochun,

Shuhang gave a presentation in the last calorimeter calibration meeting on Tuesday:
https://indico.bnl.gov/event/17284/contributions/68508/attachments/43393/73111/Hcal_SF_VSF.pdf
it is admittedly comparing the corrected EJ200 for the new hcals with G4_POLYSTYRENE for the old hcals but the sampling fraction is very similar. The change of the birks constant seems to yield some 5%-10% less light

Chris

On 9/22/2022 8:39 AM, Xiaochun He wrote:

Dear Chris,

 

My request is straightforward:

 

1. I was told that we used G4_POLYSTYRENE in our old simulation for HCal. We now use the same material for the gdml-based simulation. Can we take a quick run and see how consistent they are before making the change? For example, the sampling fraction.
2. Which specific detectors in sPHENIX are using this material?

 

As you said in other emails, the Birks constant is a fudge factor. What value of this factor is in the older HCal simulation? I have not played with Fun4All simulation for ages. If it is necessary, I will ask Murad to copy G4_POLYSTYRENE to a “new” material for HCal use.

 

Thanks,

Xiaochun

--

 

From: sPHENIX-HCal-l <sphenix-hcal-l-bounces AT lists.bnl.gov> on behalf of pinkenburg via sPHENIX-HCal-l <sphenix-hcal-l AT lists.bnl.gov>
Date: Thursday, September 22, 2022 at 7:56 AM
To: sPHENIX-HCal-l <sphenix-hcal-l AT lists.bnl.gov>
Subject: Re: [Sphenix-hcal-l] Birks constant for G4_POLYSTYRENE

Dear Xiaochun et al.,

I am not sure what is being asked here. So far the only place where we want (or consider) to change the Birks constant is for the new hcals which haven't been run yet (and the switch to G4_POLYSTYRENE is 2 days old). On the other hand everything we have done in the past (including other uses in sPHENIX, ECCE and the EIC) used the G4 value. Changing this would have effects all over the place. Dynamically changing it is likely not possible, if it was - it would have to be done in the stepping actions which would result in some real overhead (those are called for every single G4Step) and we would be confusing ourselves ever since.

The problem we want to solve is that we have a material which is closely resembled by polystyrene but want to change one physical characteristic. G4_POLYSTYRENE is a material provided by the system which you just don't change. The cleanest solution is to make a new material which has exactly the characteristics we think it should have. We have a place in our framework to do this (might need an addition to add the modified Birks constant - the current implementation leads to a duplication which I am not sure is healthy) and this will not affect anything else we are doing (including test users may run using G4_POLYSTYRENE as material). If I am not mistaken - there is even a copy ctor which copies an existing material to a new one with a new name.

Chris

On 9/21/2022 5:19 PM, Xiaochun He wrote:

Dear Chris and all,

 

I personally prefer to stick to G4_POLYSTYRENE if possible, especially for consistency with older simulation unless someone had intentionally tweaked the Birks constant for HCal. 

 

Who else is using this material in sPHENIX? 

 

It seems less necessary to copy G4_POLYSTYRENE to a new name and change one constant in the whole material class. I assume that we could always set the constant back to its default for other to use. This seems a cleaner solution.

 

My two cents,

Xiaochun

--

 

Xiaochun He

Regents Professor

Department of Physics & Astronomy

Georgia State University

Atlanta, GA 30303

Email: xhe AT gsu.edu

---

From: sPHENIX-HCal-l <sphenix-hcal-l-bounces AT lists.bnl.gov> on behalf of pinkenburg via sPHENIX-HCal-l <sphenix-hcal-l AT lists.bnl.gov>
Sent: Wednesday, September 21, 2022 1:23 PM
To: sPHENIX-HCal-l <sphenix-hcal-l AT lists.bnl.gov>
Subject: Re: [Sphenix-hcal-l] Birks constant for G4_POLYSTYRENE

 

Dear Xiaochun et al.

point taken - in this case the material should be implemented in PHG4Reco::DefineMaterials(). Then it's available when the hcals are build. But don't use a G4_XX name - I would like to keep those reserved for actual G4 defined materials. Uniplast (or Uniplast_scintillator) should be just fine.

Who will do this?

Thanks
Chris

On 9/21/2022 11:55 AM, Xiaochun He wrote:

Dear Chris and all,

 

My preference is the following, if we decide to make changes to the default parameter:

1. Define new material, maybe called it as G4_POLYSTYRENE_Uniplast and update the constants.
2. This new material can then be referenced in gdml files.

Changing optical parameters in a gdml file is a tricking business to me which may require tedious tests and ultimately further delay our work. Do we know which subsystems use this material type? 

 

Best,

Xiaochun

 

Xiaochun He

Regents Professor

Department of Physics & Astronomy

Georgia State University

Atlanta, GA 30303

Email: xhe AT gsu.edu

 

---

From: sPHENIX-HCal-l <sphenix-hcal-l-bounces AT lists.bnl.gov> on behalf of pinkenburg via sPHENIX-HCal-l <sphenix-hcal-l AT lists.bnl.gov>
Sent: Wednesday, September 21, 2022 10:05 AM
To: sPHENIX-HCal-l <sphenix-hcal-l AT lists.bnl.gov>
Subject: [Sphenix-hcal-l] Birks constant for G4_POLYSTYRENE

 

Hi folks,

let's put this discussion here to sort out how to move forward.

Here is the composition of the uniplast scintillators we use for the hcal:

1. Polystyrene     98.46%
2. Paratephenil    1.5%
3. POPOP            0.04%

based on this Polysytrene as material is good enough. G4 has polystyrene among its predefined materials, G4_POLYSTYRENE should be the way to go (and the gdml files for the hcals have been changed accordingly). It is one of the very few materials where the Birks constant is defined with a value of 0.07943mm/MeV. based on M.Hirschberg et al., IEEE Trans. Nuc. Sci. 39 (1992) 511

In our hcal test beam analysis we found that a Birks constant of 0.2mm/MeV best describes the data. There is nothing wrong with this - Makoto Asai, one of the G4 gurus told me that the Birks constant is a bit of a free parameter one can use to adjust simulations to data.

As you can see there are also dependencies on the physics list (and with that on the G4 version) which is different now than what was used here (FTFP_BERT instead of QGSP_BERT_HP) but we have known this for years and I don't see anyone spending time to get into this now (and we are about 7 months behind schedule when it comes to the sims).

We can change the Birks constant for G4_POLYSTYRENE easily but the problem is that this is a global change which will affect everything using G4_POLYSTYRENE. Just grepping I see that G4_POLYSTYRENE is used in the zdc and the emc test beam sim (not sure about the cemc, that's kind of opaque to me). Since the Birks constant is used to adjust the sims to test beam data our value is likely very specific to the outer hcal - I wouldn't even be sure it applies to the inner hcal which is now Al, not steel as it was in the test beam. But certainly for other detectors it'll be off.

My proposal would be to introduce some hcal polystyrene as a new material and set the Birks constant for this material and leave the G4_POLYSTYRENE as is. This looks to me like the safest option so we do not confuse ourselves when using G4_POLYSTYRENE somewhere else. This would mean changing the material in the gdml files and define it there (or add it to PHG4Reco::DefineMaterials()) and then add the Birks constant where it is currently set as part of the hcal detector construction (I am not sure if this can be added to the material definition in PHG4Reco::DefineMaterials(), there is a funny initialization order in G4 which might prevent this)

Chris

-- 

*************************************************************

Christopher H. Pinkenburg ;    pinkenburg AT bnl.gov

                          ;    http://www.phenix.bnl.gov/~pinkenbu

Brookhaven National Laboratory    ;    phone: (631) 344-5692

Physics Department Bldg 510 C     ;    fax:   (631) 344-3253

Upton, NY 11973-5000

*************************************************************

-- 

*************************************************************

Christopher H. Pinkenburg       ;    pinkenburg AT bnl.gov

                         ;    http://www.phenix.bnl.gov/~pinkenbu

Brookhaven National Laboratory  ;    phone: (631) 344-5692

Physics Department Bldg 510 C   ;    fax:   (631) 344-3253

Upton, NY 11973-5000

*************************************************************

-- 

*************************************************************

Christopher H. Pinkenburg      ;    pinkenburg AT bnl.gov

                               ;    http://www.phenix.bnl.gov/~pinkenbu

Brookhaven National Laboratory ;    phone: (631) 344-5692

Physics Department Bldg 510 C  ;    fax:   (631) 344-3253

Upton, NY 11973-5000

*************************************************************

-- 

*************************************************************

Christopher H. Pinkenburg      ;    pinkenburg AT bnl.gov

                               ;    http://www.phenix.bnl.gov/~pinkenbu

Brookhaven National Laboratory ;    phone: (631) 344-5692

Physics Department Bldg 510 C  ;    fax:   (631) 344-3253

Upton, NY 11973-5000

*************************************************************