STAR HardProbes PWG

[Yi Yang <yiyang AT ncku.edu.tw>]

Hi Yuanjing,

Since we are moving to the new mailing list, I copy all the discussions (your newly released note and draft + Barbara's latest comments) to here. 

**********  From Yuanjing (May 11)  ********** 

Dear convenors,

Again, thank you for your valuable comments. We have updated our 
notes/paper draft and prepared responses to your comments. Please find 
the details below. We would like to get your sign-off and move on to 
GPC.

Paper draft: 
https://drupal.star.bnl.gov/STAR/system/files/Heavy_Flavor_Electron_v2_at_27_and_54_4_Au_Au_Collisions_May10.pdf

Note: 
https://drupal.star.bnl.gov/STAR/system/files/NPE_v2_at_Au_Au_27_and_54_4_GeV_analysis_note_May10.pdf

Website: https://www.star.bnl.gov/protected/heavy/jiyj/NPEweb

Response to convenors:
Comments to Note:
response to Sooraj: 
https://drupal.star.bnl.gov/STAR/system/files/Sooraj_NPE_v2_note_May10.pdf
response to Barbara: 
https://drupal.star.bnl.gov/STAR/system/files/Barbara_NPE_v2_note_May10.pdf
response to Yi: 
https://drupal.star.bnl.gov/STAR/system/files/Yang_NPE_v2_note_May10.pdf
Comments to Paper:
response to Barbara: 
https://drupal.star.bnl.gov/STAR/system/files/Barbara_NPE_v2_paper_May10.pdf
response to Yi: 
https://drupal.star.bnl.gov/STAR/system/files/Yi_yang_NPE_v2_paper_May10.pdf

Best
Yuanjing

**********  From Barbara  (May 29) **********  

Hi Yuanjing,

thanks for the answers and the new version of the AN and the draft.

I have a few remaining comments, please see below. And I will send my comment to the paper draft soon.

Cheers,

Barbara

Replies:

(18) - Have you considered relaxation of the mean and width parameters for the sys. unc. estimation? 

- In systematic uncertainty estimation, we also directly take the normalized nSigmaE histograms, instead of gaussian functions, as the templates to carry out purity fitting. So the uncertainties from the description of the particles’ nSigmaE shape, including mean and width, will be taken into account. And considering the statistics under most of the momentum bins, uncertainty from mean and width are quite small. Also when carrying out template fitting to extract purity, we have already had 5 free parameters (particle yields), so it is not suitable to add more free parameters

Yes, but at very low pT the sys. unc. on the purity are not even visible with the current variations. Maybe then, even if the means and withs from the fits to the pure hadron and electron samples have small uncertainties, they will give a not negligible effect compared to the other variations that you use currently for the sys. unc. estimation. My point is not to relax the mean and the sigma fully, but from your fits to the pure samples you get values that have some uncertainties, you can then put limits on the mean and width in the total fit to e.g. mean +/- 3unc. Also, up to 0.33 the merged pions don't fit so well in the total fit. 

AN: 

- nSigmaE fits for 27 GeV: here I spotted that in some bins the electron fit is taken over by hadrons. Between momentum of 1 and ~1.1 GeV/c and then in the 0.53-0.55 GeV/c range the electron gaussian is replaced by the Kaon gaussian. This is probably because you don't constrain the yields for 27 GeV and shouldn't change your results (I think all of these cases are in the excluded regions), but still it would be good to update it for future. And I think it's better to have consistent methods for 54 and 27 GeV.

- Purity estimation: what is the reason to take histograms for pions instead of constraints based on the gaussian fits to these distributions? The gaussian shape looks to describe the pion distribution well. Taking this variation as the sys. unc. introduces quite large unc. at higher pT where it seems to me that it might be driven by the statistical fluctuation in the pion histograms.  For the cases where you use histograms for pions in the total fit, could you please add stat. unc. on the purity distribution ? I'm making this point also because the purity for 54 GeV at 2-2.5 GeV/c is larger than 95% and if the unc. were smaller there, this range could be usable for the v2 calculation. This drop is not visible for 27 GeV when using histogram when you use wider bins, so I wonder if this sys. for 54 GeV is not driven by the statistical fluctuations in the pion sample - you don't have many entries in the tails of your pion distribution and then when you normalise it for the total fit, these fluctuations are enhanced. 

- L366: missing figure number

- L433: missing figure number

- "The 𝑁𝑐𝑜𝑙𝑙 in Au+Au 39 and 62.4 GeV are also taken from PHENIX direct photon paper [20] and PhD thesis on this measurement [34]." - why do you take Ncoll from the PHENIX paper, not from independent Glauber calculations ?

- Fig. 20: are these plots for 62. 4 GeV ? If so, it's missliding to label them "Au+Au 54.4 GeV", I would change it to "Au+Au 62.4 GeV" and explain in the text that these spectra are used for the 54.4 GeV analysis.

- Fig.26 and 27: you've changed binning for some of the sys. unc. sources, which could be fine, but a bit confusing because different sources now have different binning and it's also not clear how then the total sys. unc. in the fine binning is obtained. I would use the same binning for all the sources, it should correspond to the binning that you have for the reconstruction efficiency.

Also, there is no unc. for the lowest pT bin for nHitsFit.

L483: uncertainty for is given in Fig. 26. Figure 29 shows - there's a missing word "for ... is", also Fig.26 and 27, and Figures 28 and 29 show.

With the reconstruction efficiency obtained from the embedding simulation -> is it the same as the combined reconstruction efficiency that you describe ? What's the difference between efficiencies in Fig.28,29 and fig. 37?

L578: flatten Fig. ??(a) -> missing figure number

Fig. 63, 64: why for 54.4 you have Npho stat (data) and Npho stat (embed) while for 27 GeV there's only one contribution: Npho stat, is it only from data ? And why the Npho stat (embed) for 54.4 has quite a big contribution, larger than sys. unc. in some bins ? 

Fig. 80: second HFe v2 point at 27 GeV is missing.

**********  From Yi (June 8) **********  

Dear Yuanjing,

Thanks a lot for the updated note, paper draft, and replies. 

I needed to go through them again since it was almost a year ago and I forgot some details.

I think the note and draft are ready to move forward. 

I have some very minor comments for your consideration: 

  - Analysis note: 

    - The replies are good to me.

    - L366: Fig. ?? --> Fig. 9

    - L433: Fig. ?? --> Fig. 14

  

  - Paper draft: 

    - Fig 2 (a): consider to use points for data (not black histogram) 

    - L114: you use "opposite-sign" and "unlike-sign" at the same time, it would be good to pick one of them. 

    - L120: similar for "same-sign" and "like-sign".

    - L131: Au+Au and p+p --> The Au+Au and p+p 

    

Cheers,

Yi

+++++++++++++++++++++++++++++++++++++++++++++++++++
Yi Yang, Associate Professor
Department of Physics
National Cheng Kung University
Tainan, 701 Taiwan
E-Mail: yiyang AT ncku.edu.tw
Tel: +886-6-2757575 ext.65237
Fax: +886-6-2747995
Group Web: http://phys.ncku.edu.tw/~yiyang
+++++++++++++++++++++++++++++++++++++++++++++++++++