STAR Correlations and Fluctuations PWG

[Eugenia Khyzhniak <eugenia.sh.el AT gmail.com>]

Dear ShinIch,

> if this is really the case, please show
all of the correlation functions for all the various multiplicity bins and kT bins for both pAu and
dAu systems without any fitting functions or extracting the fitting parameters

I built these pictures and today I will try to show them.

> One question, have you try to make the correlation function for min. bias without separating
the multiplicity bins for these small systems? For this case, two different event mixing would
be possible, with or without making multiplicity class for your mixing. The real pair distribution
should be unchanged, however the shape of mixed distribution could change, so that the
resulting correlation functions could be somewhat different between these two cases, have
you ever compared them?

It is interesting to investigate. I will try it.

Thank you!

Eugenia

чт, 8 окт. 2020 г. в 06:57, ShinIchi Esumi <esumi.shinichi.gn AT u.tsukuba.ac.jp>:

Dear Eugenia
Concerning on your last question, from your answer on the 1st point, I’m now guessing your
measured correlation functions are well under control, if this is really the case, please show
all of the correlation functions for all the various multiplicity bins and kT bins for both pAu and
dAu systems without any fitting functions or extracting the fitting parameters. They all can even
be made as preliminary if we all agree, as some of the theorists might like to see and use the
correlation functions themselves to explain the measurements. They can be nicely presented
by overlaying the multiplicity or kT dependences with different colors etc...

One question, have you try to make the correlation function for min. bias without separating
the multiplicity bins for these small systems? For this case, two different event mixing would
be possible, with or without making multiplicity class for your mixing. The real pair distribution
should be unchanged, however the shape of mixed distribution could change, so that the
resulting correlation functions could be somewhat different between these two cases, have
you ever compared them? Since I do remember seeing some differences, although it was on
the different variables, but it could be there especially for the small system cases. Could you
please try these?
Best regards, ShinIchi

PS : It’s a little disappointing to hear you are not much interested to go for the multi-dimensional
studies even with your beautiful and large data sample, but yes, it’s your choice. 

> On Oct 7, 2020, at 17:46, Eugenia Khyzhniak <eugenia.sh.el AT gmail.com> wrote:
>
> Dear ShinIchi,
>
> Thank you for your comments!
> >Have you also tried to look at
> other experimental effects, like two-track resolution, which might affect both
> loosing some close pairs or adding some ghost pairs, including momentum
> resolutions, by varying the pair cuts?

> We are planning to investigate momentum resolution but we are expecting that this
> effect would be small for small systems (~1%). We already looked on the pair cuts
> like splitting or merging and these effects for CFs are really small (<2%).

> >1D correlation functions sometimes are
> not well described by the gaussian assumption, so that you could try other
> formula, but if you go to 3D function, it could be well described by gaussian
> with different sizes depending on the directions Rts, Rto and Rl.

> We are expecting that 3d would be better described by the gaussian than 1d,
> but in 3d we are expecting also non-femtoscopic effects that would have big
> influence on the CFs. It would be great and complicate work. We need a good
> model to describe non-femtoscopic effects in 3d but now we see that Pythia, Hijing,
> Urqmd can't do that. We didn’t plan to use 3d for now.

> >I’m just
> more interested in the robustness of the measured raw correlation function
> as experimental measurements to start with, in stead of the description of
> the correlation shape with what kind of functional shape etc.

> Could you please clarify it? What do you mean by robustness (Correlation functions
> depends on kT, multiplicity etc.)?

> Best regards,
> Eugenia
>
>
>
> пн, 5 окт. 2020 г. в 17:26, ShinIchi Esumi <esumi.shinichi.gn AT u.tsukuba.ac.jp>:
> Dear Eugenia
> Nice studies. You have tested the coulomb effects in detail, it’s good to
> see it’s insensitive to the assumed radius. Have you also tried to look at
> other experimental effects, like two-track resolution, which might affect both
> loosing some close pairs or adding some ghost pairs, including momentum
> resolutions, by varying the pair cuts? 1D correlation functions sometimes are
> not well described by the gaussian assumption, so that you could try other
> formula, but if you go to 3D function, it could be well described by gaussian
> with different sizes depending on the directions Rts, Rto and Rl. As you also
> mentioned, you would also need to care about the base line shape especially
> for small system, which might also affect the fitting formula as well. I’m just
> more interested in the robustness of the measured raw correlation function
> as experimental measurements to start with, in stead of the description of
> the correlation shape with what kind of functional shape etc.
> Best regards, ShinIchi
>
>> On Oct 5, 2020, at 16:06, Eugenia Khyzhniak via Star-cf-l <star-cf-l AT lists.bnl.gov> wrote:
>>
>> Dear All,
>> here in attachments are examples of correlation functions and their fit with difference Coulomb radius taken into account.
>> Best regards,
>> Eugenia
>>
>> пн, 5 окт. 2020 г. в 09:53, Grigory Nigmatkulov <nigmatkulov AT gmail.com>:
>> Dear Hanna,
>>
>> >Concerning your fits, using a Gaussian source here leads to discrepancies as you see between your data and fits.
>>
>> We explicitly showed how different source assumptions affect the fits. Please follow the links from the last presentation.
>> We discussed these differences during the year. As we discussed, difference source and non-femtoscopic form assumptions will be in a paper.
>> I also want to notice that depending on the kT bin and multiplicity the Gaussian vs. non-Gaussian assumption will give different results.
>> For small systems the influence of (mini)jets and resonance decays will play a larger role than for large systems.
>> We presented those estimations using the UrQMD, Hijing and PYTHIA event generators. Only two of them contain
>> hard and semihard processes. NONE of the three describe the data to the level we would like it to be.
>>
>> >I agree that one should use the same parametrization for all kT intervals but your results clearly show that for higher kT fit does not work well.
>>
>> What do you propose if MC generators fail to describe non-femtoscopic effects and underlying event?
>> We explicitly say which assumptions have been used in the analysis and what results we have.
>> For the given assumptions one will obtain certain results. That is very similar to the ONLY small system femtoscopy paper
>> from 2009-10 by Zibi and Mike.
>>
>> >For future, (if you have not done it yet) I would also suggest to see how simple Gaussian looks (without Coulomb correction) like to study the lambda parameter's behavior.
>> >It looks for higher kT interval to get with such parametrization a better agreement between data and fit.
>>
>> For small systems the Coulomb effect is small and IS TAKEN into account in systematic uncertainty estimation! Eugenia, please send the figure of merit.
>> As for the curve at small q itself, the Coulomb effect is there but non-Gaussian structure dominates. Anyhow, the first few bins have almost no influence on the fit
>> and fit quality.
>>
>> >Meantime please remove from your presentation all plots that did not get preliminary labels and discuss your results during the upcoming PWG CF meeting
>>
>> Please specify which figures do you want to be updated and what exactly do you want to be changed? Eugenia can present the results again this Thursday.
>> The talk is scheduled for Friday morning.
>>
>> Cheers,
>> Grigory
>> <CF_default_cent0.png>_______________________________________________
>> Star-cf-l mailing list
>> Star-cf-l AT lists.bnl.gov
>> https://lists.bnl.gov/mailman/listinfo/star-cf-l
>