STAR HardProbes PWG

[Jan Vaněk <Jan.Vanek AT fjfi.cvut.cz>]

Hello Everyone,

we have made some small adjustments to the paper draft, addressing more of Barbara's and Yi's comments. You can find the new version of the draft (v06) on the paper proposal web page:

https://www.star.bnl.gov/protected/heavy/vanekjan/Dpm_web/Home.html

Please, let us know if you have any comments or suggestions. Thank you.

On behalf of the PAs,

Jan.

---

Od: Jan Vaněk
Odesláno: 4. července 2022 16:44:21
Komu: STAR HardProbes PWG
Předmět: Paper draft - Production of D+- mesons in Au+Au collisions

 

Hello Barbara, Yi, and Everyone,

thank you very much for your comments. I have implemented vast majority of them to a new version of the paper draft (v05) and the Run16 analysis note (v04). You can find both on the paper proposal web page:

https://www.star.bnl.gov/protected/heavy/vanekjan/Dpm_web/Home.html

In addition, you can find my answers to your questions and a few comments below:

Paper draft (v04):

General:

- I need to update figures with the new systematic errors. I am waiting for new systematic errors from Run14 which contain the new version of the statistical error correction for the TMVA cuts and pT cut variation.

- We will check what are the exact requirements for the format of the references and adjust those later.

Barbara:

L155: according to the AN you don't use DCA_PV and Delta_max in the TMVA optimization -  that should be reflected in the paper draft, and used values given.

- The lower value DCA_PV is optimized using TMVA. I am using also additional upper cut, which removes tracks which are very far from PV. This upper value is selected manually. I have tried to re-phrase the description, so that this is clear.

L165-166: An iterative process is ... -> this sentence is not clear, bit more details are needed - what do you change in the subsequent iterations, when do you stop, etc.

- I was not sure what this refers to myself. I have removed the sentence.

L203: spatial resolution of the HFT -> efficiency of topological selection criteria with HFT ?

- I wnated to highlight that one of the inputs to the simulation is the HFT spatial resolution. I have added "finite spatial resolution.."

L259-263:  which is determined by comparing the raw yield extracted from the Gaussian part of fit to the correct-sign Minv spectrum to the value obtained from a bin counting method in the correct-sign Minv spectrum after the combinatorial background subtraction. -> which is determined by extracting the raw yield by counting the D+ candidates in the invariant mass distribution within 3\sigma from the mean of the Gaussian fit, after the combinatorial background subtraction. 

- The suggested sentence probably is not quite correct in the scope of the systematic uncertainties. The systematic uncertainty of raw yield extraction is determined by comparing value calculated from the Gaussian part of the fit (used for spectra calculation) and is compared to the value from bin-counting.

L458-461: This can be probably reformulated after adding the part that I mention below.

There are assumptions, extrapolations, done to extract the shown cross section in Au+Au for different hadrons. I think it is important to discuss it and explain different steps in the text of the paper - what assumptions and extrapolations are used, how the unc. are obtained etc. If you need more space, I think some technical parts of the paper can be shortened - I also suggested a few things that can be removed.

- Xinyue can add this information.

L462-470: regarding this, I was thinking that it might be good to add sys. unc. (or at least discuss it ) on the total charm cross section that we report, related to Sigma and Xi enhanced contributions.

- I am not sure if how exactly we could quantify the uncertainty, but we certainly can add a discussion about this. My personal opinion is, that it is quite difficult to make any conclusions regarding open-charm baryon production at RHIC from the LHC data. The reason is that high multiplicity p+p events are much more probable at the LHC, than at RHIC. The conclusions from both L_c and Xi_c enhancement observed by ALICE is that it is probably a color reconnection effect in high multiplicity events, which might not be observed at RHIC at all. Unfortunately we cannot say, as we don't have any experimental measurement of L_c in p+p at RHIC energies.

Yi:

L114: (9.38 +- 0.16)% --> 9.38 +- 0.16 %

- I would prefer to keep the brackets. It looks better in the text, in my opinion.

Figure 1: Some lines are not in high resolution. 

- We will re-draw the figure. I will ask Guannan, how the figure for D0 was drawn, so that we can have similar style of topology figure.

L171 - L176: I think you can just simply say that the shape is obtained from the wrong-sign combination and there is a free parameter for the overall scaling.

- Maybe it can stay as it is?

L178 - 185: Question: If you are using a simple "linear fit", I naively think that the direct fit on the data (or on the sideband region) should give you a similar result, did you try?

- Will explain in reply to Yi. 

L178 - 185: Question: If you are using a simple "linear fit", I naively think that the direct fit on the data (or on the sideband region) should give you a similar result, did you try?

- This is a good and relevant question. The method you propose would definitely work, but I have not tried it, as I think that my method has a few sligh advantages. The first motivation to fit the wrong-sign spectrum is that it has much better statistical precision (there are 3-times more wrong-sign charge combinations than the correct ones). We can use this to reduce an error associated with the bacground, by fitting the scaled wrong-sign first to capture the shape of the background. Second advantage of this method is that I can fit in the D+- mass peak region and so better describe the background. As you noted in the comments to the analysis note, the linear fit sometimes deviates from the points far from the peak. In this case it should not be a major problem, as the background description under the D+- invariant mass peak is good.

L190: can you remind me, why 10 - 40% can go to lower pT (0.5 GeV)? Will this point give us any extra physics information?  If not, why do we use the same pT range for all centralities?

- The 10-40% centrality bin is statistically rich, while having reasonably low combinatorial background levels. For that reason it is possible to have larger pT coverage. In 0-10% the combinatorial background prevents to go below 1 GeV/c, in 40-80 it is due to combination of lower statistics and DCA resolution.

L200: efficiency --> efficiencies

- Not sure about this.

Figure 2: It would be good if we can also show 2014 mass plots here.

- Probably can add Run14 mass peaks.

L248: It would be good if more details can be shown for the 2014 efficiencies (rather than just say "The 2014 reconstruction efficiency has similar shapes and values."

- Probably can add RUn14 efficiency plot.

L285: add a reference of the TPC uncertainty

- Could you please suggest a reference? Could use the D0 paper?

L424: (0.41 +- 0.03) --> 0.41 +- 0.03

- Prefer to keep the brackets.

Analysis note (v03):

Barbara:

- Page 31: In case of TPC PID efficiency, it was decided to simplify things by using the constant function despite the shape of the pT dependency. - Was it checked how much of the effect it is if the efficiency at high pT is not assumed flat ? Could potentially be included in the sys. unc.

- Most daughters below 2Gev, plus the original fits of the nSigma distributions are probably not perfect. I tried fitting on the right half of the peak and got better results. The asymmetry on the left side likely due to contamination.

Yi:

 - p15: I don't understand the statement of "The lower cut on pT for all tracks is used in order to reduce the combinatorial background in the low pT region." I would navilly thought that having a lower pT cut will increase the combinatorial background. Could you please explain a bit more about it? 

- What I mean by "lower cut" is that we accept only tracks with pT from the cut up, and cut off everything below the cut. Not sure if that is correct in English. Maybe can re-phrase to avoid confusion (mainly in the paper itself).

 - p18: Similar to the previous question, you stated "loose value of DCA_pv < 2 mm, which helps reduce background coming from badly reconstructed secondary vertices", I thought the looser cut will include more background? 

- I am using two cuts on DCA_PV. Lower limit is optimized using TMVA which removes tracks from close to PV. The cut I refer to here, the upper limit, cuts off tracks that are really far from PV, eg. from conversion od decay of strange particles. Both cuts reduce some type of background.

 - p19: You mentioned that adding Delta_max will let the TMVA be unstable due to the large correlations to other topological variables. But, if they are largely correlated, with/without Delta_max in TMVA shouldn't change the result much, right? On the other hand, having a cut on this might change the TMVA selection. How did you determine the cut values for Delta_max and DCA_pv? 

- The decision not to include the Delta_max to training was based on comparison of the training with and without the cut. The one without has proven to be more stable and quicker. The addition of Delta_max manually has proven to improve the significance. The value was chosen based on the previous manually tuned set, but with slightly more open value in order not to use too tight cut. The DCA_PV (the upper cut) is very open (2 mm) and is chosen by hand so that it should not affect the real decays of D+-, but only reduces background from e.g. particles that decay really far from PV.

 - p25: You mentioned that the momentum resolution is much better in the newer version of embedding. I am a bit surprised that the reconstruction efficiency was not affected by the momentum resolution. Could you please add more information about it? Also, probably you need to add a systematics from this. 

- The effect is very small (less than 5%), so I don't think we need to add it. In addition these 5% fluctuations were typically within the statistical precision of the simulation.

 - p26: Did you fit the TPC efficiency (Fig. 3.5 and 3.6) and use the smooth function for your efficiency or did you get the efficiency directly from the histogram? 

I am using the histograms directly.

 - p31: Regarding the PID efficiency, for example nSigma_pi, you obtained the "pure" signal distribution and fitted it with a function then took the ratio of two different ranges (within the cut and full). I am wondering why don't you use the "tag-and-probe" method to obtain the pT-dependent efficiency since you have very nice statistics of Ks0? The method you are using depends on the understanding of the shape a lot, for example Fig. 3.9, I don't think Gaussian is a perfect one to describe the Delta(1/beta) distribution. 

- The method I used was used in many other STAR analyses, so I used it to be consistent. My efficiency is pT dependent as I am extracting the pions from K0s decays in various pT bins.

 - p60: Do you understand the strange "tail" on the left-hand side of the peak in 2.0 - 2.5 GeV/c? 

- I am not sure about the origin. As it is probably present only in this specific centrality and pT bin, I would say it is just a statistical fluctuation.

 - p63: the fit on 8.0 - 10.0 GeV/c looks very bad, did you try to constraint the width from other pT bins? (You can plot the width vs pT and extrapolate the width for this pT bin.) 

- For Run16, this pT bin is not actually significant. I have added the figure to the note by accident.

 - p64 the Corrected-sign fit (black line) looks bad, do you know why? 

- I believe that in some of the pT bins the background is not perfectly linear. The deviation is typically visible only far from the D+- peak. More directly to your question - yes I always constrain the fit parameters from the previus pT bin. I thing that the background description in the region of the D+- mass peak will be good enough for all the fits. I can constrain the fit of the background only to the peak region (I am fitting the wrong-sign first), or to try higher order polynomial (2nd should probably be enough), but I don't expect any significant difference.

Please, let me know if you have any further suggestions or comments. Thank you.

Best regards,

Jan.

---

Od: Yi Yang <yiyang AT ncku.edu.tw>
Odesláno: 26. května 2022 18:57:58
Komu: Jan Vaněk; star-hf-l AT lists.bnl.gov
Předmět: Re: [Star-hf-l] Paper draft - Production of D mesons in Au

 

Hi Jan,

I have some questions and comments on your analysis note (Run 16). 

 - p13: missing Appendix number 

 - p15: I don't understand the statement of "The lower cut on pT for all tracks is used in order to reduce the combinatorial background in the low pT region." I would navilly thought that having a lower pT cut will increase the combinatorial background. Could you please explain a bit more about it? 

 - p18: Similar to the previous question, you stated "loose value of DCA_pv < 2 mm, which helps reduce background coming from badly reconstructed secondary vertices", I thought the looser cut will include more background? 

 - p19: You mentioned that adding Delta_max will let the TMVA be unstable due to the large correlations to other topological variables. But, if they are largely correlated, with/without Delta_max in TMVA shouldn't change the result much, right? On the other hand, having a cut on this might change the TMVA selection. How did you determine the cut values for Delta_max and DCA_pv? 

 - p24: It would be good to add the fit parameters (a, b, c, and d in Eq.3.2) on Figs. 3.1, 3,2, 3.3, and 3.4.    

 - p25: You mentioned that the momentum resolution is much better in the newer version of embedding. I am a bit surprised that the reconstruction efficiency was not affected by the momentum resolution. Could you please add more information about it? Also, probably you need to add a systematics from this. 

 - p26: Did you fit the TPC efficiency (Fig. 3.5 and 3.6) and use the smooth function for your efficiency or did you get the efficiency directly from the histogram? 

 - p31: Regarding the PID efficiency, for example nsigma_pi, you obtained the "pure" signal distribution and fitted it with a function then took the ratio of two different ranges (within the cut and full). I am wondering why don't you use the "tag-and-probe" method to obtain the pT-dependent efficiency since you have very nice statistics of Ks0? The method you are using depends on the understanding of the shape a lot, for example Fig. 3.9, I don't think Gaussian is a perfect one to describe the Delta(1/beta) distribution. 

 - p40, Fig. 4.1 (and other similar plots): you can move the legend to the "right-bottom" corner (all three plots share the same legend), so you can zoom into the y-axis a bit since lots of points are squeezed in the bottom. 

 - p43: Table 4.1: Please add the centrality range in the left column.  

 - p46, 47, Fig. 4.7, 4.8: Please zoom-in to the y-axis and the title of x-axis should be N_part.

 - p60: Do you understand the strange "tail" on the left-hand side of the peak in 2.0 - 2.5 GeV/c? 

 - p63: the fit on 8.0 - 10.0 GeV/c looks very bad, did you try to constraint the width from other pT bins? (You can plot the width vs pT and extrapolate the width for this pT bin.) 

 - p64 the Corrected-sign fit (black line) looks bad, do you know why? 

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
+++++++++++++++++++++++++++++++++++++++++++++++++++

On Tue, May 17, 2022 at 11:54 AM Yi Yang <yiyang AT ncku.edu.tw> wrote:

Hi Jan,

Sorry for the late reply and thanks a lot for the nice paper draft. 

I have some comments and suggestions for your consideration (also some questions). 

I will send the comments on the Run 16 analysis note later in a separate email. 

  - Title: Measurement --> Measurements 

             D+- meson --> the D+- meson 

             midrapidity --> mid-rapidity

  - Abstract: were obtained --> are obtained   (same tense in other places) 

                   PYTHIA calculations --> the PYTHIA calculations 

                    further extracted --> further extract 

  - L36: Ds+ and Lambda_c+ --> Ds+, and  Lambda_c+

  - L48 - 51: I would move this sentence to section 3, when you talk about reconstructing D meson, say around L112. 

  - L55: p+p collisions, and <Ncoll> --> p+p collisions, respectively, and <Ncoll>

  - L58: midrapidity --> mid-rapidity (many places, should be consistent) 

  - L59: (0-10%, 10-40%, 40-80%) --> 0-10%, 10-40%, and 40-80%   (remove the bracket) 

  - L62: PYTHIA model --> the PYTHIA model

  - L64: ground state charm hadrons --> ground states of charm hadrons 

  - L70: data-set --> dataset  (many places) 

  - L77: [19]in --> [19] in   (missing a space)

  - L102: Can you remind me why you divide these 3 centralilities? Probably it is good to give a reason here.  (For example, why not 0 - 20, 20 -50, 50-80?) 

  - L109: refer to the Ref. [15] --> refer to Ref.[15]

  - Table 1: Mean number --> Mean numbers 

  - L114: (9.38 +- 0.16)% --> 9.38 +- 0.16 %

  - Table 2 and Table 3: I would suggest merging them to be a larger table including the track selections and PIDs. 

  - Figure 1: Some lines are not in high resolution. 

                   three body decay topology with topological variable --> three body decay topology and the topological variables 

  - Footnote: Please provide some explanation on why only this cut is only for Run16

  - L162: next Sec. --> next section

  - L171 - L176: I think you can just simply say that the shape is obtained from the wrong-sign combination and there is a free parameter for the overall scaling. 

  - L178 - 185: Question: If you are using a simple "linear fit", I naively think that the direct fit on the data (or on the sideband region) should give you a similar result, did you try? 

  - L190: can you remind me, why 10 - 40% can go to lower pT (0.5 GeV)? Will this point give us any extra physics information?  If not, why do we use the same pT range for all centralities? 

  - L199: D+- mesons --> the D+- mesons 

  - L200: efficiency --> efficiencies 

  - Figure 2: It would be good if we can also show 2014 mass plots here. 

  - L213: generate D+- mesons -- generate the D+- mesons 

  - L218: and TPC performance simulation --> and the TPC performance simulation

  - L240: remove "in the 2016 data-set" in the end of the sentence

  - L248: It would be good if more details can be showed for the 2014 efficiencies (rather than just say "The 2014 reconstruction efficiency has similar shapes and values." 

  - L280 - L281: 300 MeV/c and 500 MeV/c --> 0.3 GeV/c and 0.5 GeV/c  (to be consistent) 

  - L285: add a reference of the TPC uncertainty 

  - L310: error and the PID systematic error. --> error, and the PID systematic error. 

  - L317: remove one of the "as was used" 

  - L332: ofthe --> of the

  - L352: the individual contributions the procedure --> the procedure? 

  - ** Systematic ** It would be good to have a plot to summarize all the systematics 

  - L377: a measurement --> the measurement 

  - L404: compare to two theoretical calculations -->  compare to theoretical calculations

  - L424: (0.41 +- 0.03) --> 0.41 +- 0.03

  - L427: remove ", respectively 

  - L429: (H1). --> (H1), respectively. 

  - Around L448: You should put the number of cross sections in the text 

  - L480: With he extracted --> With the extracted 

  - **Summary** --> I think the summary is not "strong" enough, you have very very nice result, it can be much more stronger, and you should mention again it is the "first time" that the total charm cross section is measured in Au+Au collisions 

  - References: "et al." needs to be in italic

                         Volume number should be in bold  

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
+++++++++++++++++++++++++++++++++++++++++++++++++++

On Fri, Jan 21, 2022 at 4:13 AM Jan Vaněk via Star-hf-l <star-hf-l AT lists.bnl.gov> wrote:

Hello Everyone,

I would like to let you know that we have prepared a paper draft for the production of D+- in Au+Au at 200 GeV. You can find the latest version on the paper proposal web page:

https://www.star.bnl.gov/protected/heavy/vanekjan/Dpm_web/Home.html

The text of the draft is ready for review. Please, let us know if you have any questions or suggestions.

I would also like to let you know that we have decided to make a few minor updates to the results, namely:

- Update the BR value from  8.89 +/- 0.28 (2018 PDG) which was used up to this point to 9.38 +/- 0.16 (latest PDG)

- Change the TMVA cut variation for Run16 from +-30% to +-50% to be consistent with Run14 and other TMVA open-charm analyses.

After those changes are finished, I will update the figures and values in the paper draft. As those are relatively minor changes, we believe that they can be done in parallel to the review in the HF PWG.

The Run16 analysis note is also available for review and can be accessed from the paper proposal web page. The latest version is up to date, except for the two aforementioned changes. I will update the analysis note as soon as I update the systematic errors and results.

The Run14 analysis note should be close to finishing. It is not available at the page yet. I will ask Xinyue to send it to me as soon as possible so that I can upload it to the page for review.

Best regards,

Jan.

On behalf of the PAs.

_______________________________________________
Star-hf-l mailing list
Star-hf-l AT lists.bnl.gov
https://lists.bnl.gov/mailman/listinfo/star-hf-l