# 2006 di-jet Sivers PRL paper

Measurement of Transverse Single-Spin Asymmetries for Di-Jet Production in Proton-Proton Collisions at sqrt(s) of 200 GeV
Target Journal: PRL
Principal Authors (PA): J.Balewski, I.Qattan, and S.Vigdor
Abstract+summary+figures (ver 1.2) : link to protected/spin area
Status:

presented to PWGC on December 01,
pwg approved to request GPC, Decmeber 14,
GPC phone conference , January 4

1. Draft of the PRL paper version 1.2 (posted November 16, 2006)
* from Carl , received on Nov 16
* from Hal ,received on Nov 27
*
3. Revised ver 1.3 (ps) , posted December 12,
1. Akio,Les: I guess the biggest question I had was east BEMC "problem" and your measurement on Nov 30
The 2% of the off-line data have be reanalyzed with the new (electron based) BTOW gains to verify the previous conclusions hold, as posted on  Issam's summary page, spin-hn, on January 11, 2007
4. Comments from GPC are listed below
• Peter J., Jan 4, 2007
5. Revised draft ver 2.0 (January 29, 2007)
• Comments from Mike, Feb 2,2006, below

# Mike M., Feb 02, 2007

<pre>
Hi all,

1) Overall, I really like the way the first page reads -- well done.

2) I think it's lacking a crisp explanation of why online jets are  used instead of offline, perhaps I missed it.

3) I think that Figure 2 should clearly be labeled as "Simulation".   It's in the caption, but the minute the figure gets clipped into  a  presentation the caption (and that info) are gone.

4) I would prefer the use of "Fast Monte Carlo" instead of "toy  model," for various reasons.  First, anyone in particle physics will  probably understand that Fast MC implies not doing full simulation/ reconstruction, but a quick smearing.  Furthermore, there's enough  info in the MC that you believe the results, whereas the phrase "toy"  makes it sound otherwise. Second, anybody outside of particle physics  won't know the difference between a fast MC and a full slow simulation.

5) Personally I find Figure 3 fairly confusing, especially when  dumped to BW.  I realize this comes down to matters of taste, but I  personally prefer to show the model calculations as binned histograms  (TH1::Draw("hist")).  The markers are in my mind too easily confused  with the data (given that this is really 6 plots).  Further, the  "hist" option will clearly show that the models were binned the same  as the data, and makes a nice distinction between data/theory.  By a  clever choice of line styles, you can probably even make this fairly  clear in BW.

6) In Figure 3 and in the text, would it be possible to use some  english to differentiate between A_N(>pi) vs A_N(<pi).  Maybe  introduce the phrases "quark-like" and "gluon-like"?  If you can find  a smooth way, it would sure make the p4 text smoother, and give some  intuition to fig 3.
I realize it'sa little tricky because Fig3b is kind of an orphan.  Do  you really need it?  They're straight lines , and you could quote 4  numbers in the text in a single sentence, even cutting down on some  space.  Then you could have a "quark-like" left panel and a "gluon- like" right panel.

But, again, I really like the draft.

-Mike

# Peter J., Jan 4, 2007

Hi GPC and PAs,

Here are some first comments on the paper draft:

(i) physics intro, 1st para: I find the physics intro to be a bit confusing. I looked briefly at refs [6,7] (Brodsky et al and Collins). Within my distinctly limited understanding of them I don't think the last sentence of the first paragraph is accurate, and the main physics point of interest in this measurement is missed.

Evidently, the SSA arises due to interference of left- and right-handed quark polarization states, and thus is sensitive to chiral symmertry breaking. If correct, this is important and should be featured prominently in the intro.

On the other hand, I am confused by the various claims about factorization in this process. Brodsky et al claim that the process cannot be factored into PDF and FF, while Collins claims that factorization holds but then derives a pdf $f_{1T}^\perp$ whose sign is opposite for DIS and DY (eq 3), i.e. a pdf whose value is process-dependent, which doesn't sound to me like factorization. The only thing I know for sure is that I am confused on this point and could use some guidance. I suspect that most non-experts will be similarly confused. The physics intro should be precise and clear about what the theory says.

(ii) p 1 left col 2nd para bottom: what is the specific relevance to this measurement of the inclusive jet cross section being described by (factorized) pQCD? I guess if it didn't work for the inclusive yields one could stop immediately. Is that the only point to be made here? Can one say more about constraints on PDFs and FFs?

(iii) p 1 right col line 9: I know nothing about Siberian snakes. What are the limits on possible non-vertical polarization states?

(iv) End of that para: give errors on polarization: 59\pm{xx}\% (57\pm{yy}\%).

(v) jet reconstruction: nowhere do you actually describe what "jet reconstruction" you do. P 2 left col line 6 talks about "jet clusters at level 2" and the caption of Fig 1 talks about "full jet reconstrcution" but the reader is left hanging about how a jet is actually defined. Is there some peak-finding with a cut-off radius, or what? I know that you use EMC energy only but the non-expert reader will not know what this implies, i.e. all of the EMC energy plus perhaps 30% of the charged hadronic energy, with some charged-track dispersion in the magnetic field that is not corrected for. You need a couple of paragraphs defining the jet finding used for the analysis and giving its comparison to full jet reco, justifying why this technique is adequate for this measurement (there is currently some of that later in the text but it should be consolidated).

(vi) Fig 1d: why only 2% of the data? I think I know the answer: that's what was reconstructed at the time you were in the thick of this analysis, but evidently more has been done in the meantime. Not usable?

(vii) p 2 left col middle: I printed the paper in B&W and don't see the 6-fold L0 peaks in fig 1a. Am I missing them?

(viii) p 2 right col 2nd para: the "favoring" of qg vs gg at forward vs midrapidity is qualitative. Can this be made quantitative, e.g using PYTHIA? What is the magnitude of the variation of the two contributions?

(ix) p 2 right col middle: "while we away the time-consuming replay of the full dataset including TPC..." is a STAR detail of little interest to others, and has a limited shelf-life. I suggest simply describing what was done, saying that this rapid analysis technique (in contrast to full jet reco) is sufficient for present purposes.

(x) Fig 1b and discussion of tails in p2 right col bottom: "might reflect moderately hard gluon emission" is weak. Can this be studied with a model calculation? But I also find it confusing because I don't know how the jet finding was done. Hard gluon emission will generate an acoplanarity only if it pushes some of the energy flow out of the jet cone, otherwise momentum is conserved. So I suspect that this tail depends on how the jet is defined. Needs more discussion.

I also wonder about tails being generated by the combination of relatively low multiplicity in low energy jets and only partial jet reco (EM plus ~30% hadronic, with some funny spread in the latter due to the field). Could unfavorable, perhaps rare, fluctuations in charged vs neutral pions generate such apparent tails which are not present for full jet reco? Perhaps a model study would help here. Anyway, the toy model in which you just fit with a Gaussian with an exponential seems inadequate - can you do a more meaningful study based on PYTHIA or HERWIG?

(xi) p 3 left col top: is there a jet energy dependence to <kT^2>? More generally, the distribution shown in Fig 1d goes out to ~50 GeV if I jack it up by eye by a factor 50. Can you make a few coarse energy bins to look at the dependence of the asymmetry on jet energy? You say somewhere that you expect the ET dependence of the Sivers effect to be small, but surely it would be good to test this.

(xii) definitions of A_N and r_\pm (eq 1 and 2): it's late in the evening and I am a bit tired, but frankly these formulas are not speaking to me at the moment. A_N is defined as the ratio of ratios, which is OK, but I am not getting the purpose of the sqrt. There are too many +- and -+ subscripts and zeta>pi vs zeta<pi which are hard to distinguish. Can you find a more transparent notation, or explain the structure of the definitions a bit better?

That's all for now. I didn't read the last third as carefully, I'll do that next time.

Hope these are helpful, talk to you tomorrow.

Peter