sPHENIX bulk physics topical group

[Milan Stojanovic <milan.stojanovic AT cern.ch>]

Hello Jamie,

Thanks for looking at our results and your feedback! Please find below our responses to each item.

1. Please find the enclosed CEMC tower distribution. The 0.005% events are above 70. We do not apply an upper cut, but we can add one as a cross-check. However, we have verified that there are no sudden bumps or discontinuities, as were observed before applying the latest calibration. We have not estimated the purity yet, but we can do so if needed. Additionally, we are checking the impact of changing or removing the PU filter to assess its effect on the correlation function.
   
   
   >>From the correlation plots, I suspect that some form of background / streak events are sneaking in for the 0.005% selection -- where you see a dramatic narrowing of the distribution.   

      Yes, this is something we were thinking about. We are working on implementing streak events filter. For now, we have checked the effect of removing all jet-triggered events, as most streak events are expected to fire the jet trigger.

      >>I am also curious if you considered using the sEPD for event selection, and EM Towers for the correlation as a comparison.

      

      Not at the moment. We are open to exploring that possibility if you have a specific selection in mind. 

2. We did some initial checks of the delta phi projections with different |delta eta| cut, in particular we did |delta eta| > 1.6, which did not show a significant effect. We chose a value of 1.2 because we aim to reach extremely high multiplicities, where statistics become crucial, but we can apply a tighter cut if needed. As for the away side, as you pointed out, handling it is much more challenging. For this reason, we did not prioritize it, at least not within the time frame of QM. Additionally, this is closely related to item 3—what we ultimately aim to measure.
   
   
3. Yes, precisely measuring vn​ using calorimeter data alone is challenging because we cannot strictly control the phase space of the selected particles. That is why our primary goal is to qualitatively assess how the correlation function evolves with multiplicity and whether any collective effects emerge at some point. To estimate detector effects, we plan to compare our results with measurements from the silicon trackers (I don't know if you saw Yuko's results [1]) and HCAL. Additionally, we aim to determine what average charged track multiplicity corresponds to a given calorimeter-based multiplicity. On a slightly longer time scale, we may also attempt to select events based on the total energy in the underlying event, which would provide a more direct comparison to theoretical models. The check with peripheral Au+Au collisions sounds like a good idea, and we will proceed with it. However, if we do not observe a clear signal in the data, we can conclude that, within the limitations of our statistics and methodology, no significant modification is evident. That said, this does not rule out the existence of a signal—only that we are unable to detect it with our current approach. I am not sure of the exact luminosity of the data STAR collected using their high-multiplicity trigger, but if it is significantly higher than our minimum-bias data, it would not be surprising if they observe a signal while we do not.

Please let us know if you have further comments. 

Best regards,

Milan

[1] https://indico.bnl.gov/event/26572/#2-update-on-correlations-in-pp

---

From: sphenix-bulk-l-request AT lists.bnl.gov <sphenix-bulk-l-request AT lists.bnl.gov> on behalf of Jamie Nagle <sphenix-bulk-l AT lists.bnl.gov>
Sent: Tuesday, February 18, 2025 3:00 PM
To: sphenix-bulk-l AT lists.bnl.gov <sphenix-bulk-l AT lists.bnl.gov>
Subject: [[Sphenix-bulk-l] ] Input on pp-correlation analysis

 

Hello Milan, Joern, Yuko, Rachid (cc sphenix-bulk list),

Dennis asked if I would take a look at the latest results from your very interesting (and challenging) pp correlations analysis.

https://indico.bnl.gov/event/26572/contributions/102833/attachments/59885/102885/ppFlow_20250211.pdf

I think you have identified two of the key issues and wanted to follow up on further quantification.

1.    Event selection and double interactions.    Your initial studies look promising.     For the later 0.5, 0.05, 0.005% selections, can you show those on the EM Tower multiplicity distribution?    Is there an upper end cut, to remove very, very high end background?     For each of these percentage selections, can you quote the estimated purity (e.g., 95%, 98%), and how that is determined.    From the correlation plots, I suspect that some form of background / streak events are sneaking in for the 0.005% selection -- where you see a dramatic narrowing of the distribution.   

I am also curious if you considered using the sEPD for event selection, and EM Towers for the correlation as a comparison.

2.    With a selection of |delta-eta| > 1.2, that is rather modest, and there are likely to be non-flow near-side jet contributions.   A near side contribution is rather problematic and one could consider an even wider cut since you have statistics.    For the away-side non-flow, a group of us implemented a number of non-flow subtraction techniques in one package:

https://journals.aps.org/prc/abstract/10.1103/PhysRevC.100.024908

Sanghoon and Blair are two sPHENIX members who have the code package, and I am sure that you can request that from them.

Then you can take the correlations on slides 9 and apply a decomposition -- where you can have a low multiplicity (LM) and high multiplicity (HM) selection.    I would be glad to help, though those two are the real experts.

3.   A third one relates to what you are really measuring.      Imagine you see no clear signal.    How does one compare that to a STAR result (note they had a special low luminosity pp running in 2024 specifically for this purpose -- presumably using their EPD)?   How does it compare to a theory prediction?     Thus, I think it is important to take some Monte Carlo with known flow and see what you would measure from the EMCal Towers -- not clear with your cuts that you are mostly measuring photons from pizero and eta decay.    You could even look at a AuAu MC sample and ask how many towers (in peripheral events) that pass your cuts are photons, and what else...   How does the v2 extracted by this method compare to some truth value -- for example pizero/eta --> photons truth v2 between 0.6-3.0 GeV...

I really appreciate your effort on this front, and am happy to clarify / help resolve any of the points above.

Sincerely,

Jamie

||------------------------------------------------------------------------------------------
|| James L. Nagle   
|| Professor of Physics, University of Colorado Boulder

|| EMAIL:   jamie.nagle AT colorado.edu
|| SKYPE:  jamie-nagle        
|| WEB:      http://spot.colorado.edu/~naglej 
||------------------------------------------------------------------------------------------

Attachment: towerMult_01.pdf
Description: towerMult_01.pdf