sPHENIX discussion of physics

[Anthony Frawley <afrawley AT fsu.edu>]

Hello Cameron and Ming,

We have to decide what we would like to count as a track for centrality purposes. The CA seeder finds roads in the TPC and they are used as the initial track list. If we throw 100 pions in an event, we may end up with 105 tracks, since the seeder may split some of them, or it may find secondary particle tracks.  The splitting is being worked on by Michael Peters. The split tracks tracks may not be associated with the silicon correctly, and may be rejected in the final fit.

The bottom line is that if a track is not fitted by Acts, there was a reason.  For the moment, I think the safest thing to do is count reconstructed tracks and assume that they are momotonically related to the number of primary tracks from the collision.

As I said, ultimately I think we should use the silicon tracklets, but for the moment we are using truth information for that, so it is a (fairly near) future thing.

Tony

---

From: Ming Liu <ming AT bnl.gov>
Sent: Monday, November 23, 2020 11:58 PM
To: Dean, Cameron <cdean AT bnl.gov>; Anthony Frawley <afrawley AT fsu.edu>; sphenix-physics-l AT lists.bnl.gov <sphenix-physics-l AT lists.bnl.gov>
Cc: Ming Liu <ming AT bnl.gov>
Subject: Re: [Sphenix-physics-l] Centrality selection in MDC analysis

 

Hi Cameron,

 

That sounds good. A vertex object containing (MVTX+INTT) tracklet with DCA information to the reconstructed primary vertex would be an excellent tool for event multiplicity (or centrality) measurement for all collisions, pp, pA and AA.

 

Cheers,

Ming

 

-- 

Ming Xiong Liu

P-25, MS H846                     TEL: 505-667-7125 

Physics Division                            631-344-7821(BNL)

LANL                                               630-840-5708(FNAL)

Los Alamos, NM 87545      FAX:  505-665-7020

 

 

From: "Dean, Cameron" <cdean AT bnl.gov>
Date: Monday, November 23, 2020 at 9:16 PM
To: "Ming Liu (BNL)" <ming AT bnl.gov>, "Frawley, Anthony" <afrawley AT fsu.edu>, "sphenix-physics-l AT lists.bnl.gov" <sphenix-physics-l AT lists.bnl.gov>
Subject: Re: [Sphenix-physics-l] Centrality selection in MDC analysis

 

Hi Tony, hi Ming,

When I was iterating over the track map, I did notice that the last track ID was typically higher than the number of tracks stored in the event which I attributed to tracks that had been generated but not reconstructed. It could be good for me to also store the last track ID as well as the number of reconstructed tracks as a sort of efficiency measurement . This obviously assumes that the final track was reconstructed so this would introduce a bias. I like Hugo’s idea of using the impact parameter to access the true multiplicity so I’ll look into this.

I’ll copy the code for counting tracks into the vertex reconstruction and make that the default. The drawback from my point of view is that, in some cases for the HF TG, we won’t associate a candidate to a primary vertex. This was a requested feature in KFParticle so it’s simple enough to use the vertex measurement as default and fall back to the track counting approach if the candidate has no associated PV. As you say, the MDC will give us a good feel for how all this looks so it’ll be better to have this in place then remove it rather than rush to implement it the day the data comes out.

Cheers,
Cameron

 

From: Ming Liu <ming AT bnl.gov>
Date: Monday, November 23, 2020 at 8:57 PM
To: "Dean, Cameron" <cdean AT bnl.gov>, Anthony Frawley <afrawley AT fsu.edu>, "sphenix-physics-l AT lists.bnl.gov" <sphenix-physics-l AT lists.bnl.gov>
Cc: "Liu, Ming" <ming AT bnl.gov>
Subject: Re: [Sphenix-physics-l] Centrality selection in MDC analysis

 

Hi Tony and Cameron,

Good discussions. I thought for (low collision rate) AuAu, the MBD is the default one for centrality measurement with high efficiency (adding ZDC hits) and minimal autocorrelation with observables in the central rapidity;

The challenge is for high rates p+p and p+Au collisions since MDB has low efficiency, for p+p MB, eff ~ 27% (?), and there is also high probability of multiple collisions per bunch crossing.

A combination of (MVXT + INTT) tracklet associated with a given reco’d primary vertex and MBD hits/timing information could be used to for centrality estimation.  

 

Cheers,

Ming

 

-- 

Ming Xiong Liu

P-25, MS H846                     TEL: 505-667-7125 

Physics Division                            631-344-7821(BNL)

LANL                                               630-840-5708(FNAL)

Los Alamos, NM 87545      FAX:  505-665-7020

 

 

From: sPHENIX-physics-l <sphenix-physics-l-bounces AT lists.bnl.gov> on behalf of "Dean, Cameron" <cdean AT bnl.gov>
Date: Monday, November 23, 2020 at 3:42 PM
To: "Frawley, Anthony" <afrawley AT fsu.edu>, "sphenix-physics-l AT lists.bnl.gov" <sphenix-physics-l AT lists.bnl.gov>
Subject: Re: [Sphenix-physics-l] Centrality selection in MDC analysis

 

Hi Tony,

Your idea here sounds effectively like what we (I) plan to do in the Heavy Flavor TG.

For each event, I use the SvtxTrackMap to iterate over all the tracks in the container, turn them into KFParticle objects and count how many there are. If our selection then wants to save the track, this count is passed to the output class and written to the file (example attached). Currently all the examples I have only contain one collision vertex but the code could be moved into the vertex unpacking rather than track unpacking.

My only hesitation on this was I was unsure if the SvtxTrackMap contained all tracks or whether there was a pre-selection that biased this count.

Cheers,
Cameron

 

From: sPHENIX-physics-l <sphenix-physics-l-bounces AT lists.bnl.gov> on behalf of Anthony Frawley <afrawley AT fsu.edu>
Date: Monday, November 23, 2020 at 3:22 PM
To: "sphenix-physics-l AT lists.bnl.gov" <sphenix-physics-l AT lists.bnl.gov>
Subject: [Sphenix-physics-l] Centrality selection in MDC analysis

 

Hi All,

 

I would like to understand how the topical groups plan to make centrality selections when analyzing the MDC data.

 

What are people planning to do?

 

One could use the vertex object associated with each collision. The existing code (PHVertexFinder) will create a vertex object for each collision vertex and place it on the node tree in SvtxvertexMapActs. Each collision vertex will have a list of track keys associated with that vertex. That list can be used to access the tracks from SvtxTrackMap. The most basic centrality selection could be just the number of tracks associated with the collision vertex. More advanced selections could involve applying cuts on the associated tracks.

 

Cheers

Tony

 

_______________________________________________ sPHENIX-physics-l mailing list sPHENIX-physics-l AT lists.bnl.gov https://lists.bnl.gov/mailman/listinfo/sphenix-physics-l