Progress towards a measurement of A_{LL} for charged pion production opposite a trigger jet using data from the 2006 STAR run. Goal is to present this measurement as a function of z, defined here as p_{T}(pi) / p_{T}(trigger jet).

Update 2008-08-28: added data/MC comparison plots

Cuts

• Runlist (297 long2 runs)
• BJP1 HW+SW trigger (137221, 137222)
• BBC timebin 6-9
• |η_{π}| < 1.0
• |DCA_global| < 1.0
• nHitsFit > 25
• recalibrated nσ(π) in [-1.0, 2.0]
• trigger jet p_{T} in [10.0, 30.0]
• trigger jet detector η in [-0.7, 0.9]
• trigger jet neutral energy fraction < 0.92
• trigger jet ϕ within 36 degrees of fired jet patch center
• Δϕ(trigger jet-pion) > 2.0

Plots illustrating the effect of each of these cuts will be at this link. The plots currently on that page are for inclusive charged pions … I’m updating them with the away-side versions and additional jet plots this weekend.

I could also study each of these cuts in low-z and high-z regime …

Comparison of z Calculation Algorithms

I looked into a couple of alternative formulations of this measurement, one that defined z as the fraction of the 2nd jet’s momentum carried by the pion and one that used pT ratios but counted more than one pion per event if those pions fell in different z bins. At high z the statistical precision of the three measurements is essentially identical:

The z distribution of the measurement which requires full dijet reconstruction (the red curve) is somewhat flatter than the other two, and is somewhat better described by an exponential fit. Note that p_{T} cuts are applied to the jets (trigger jet in 10..30) and the pions (2..10) in those figures, so the distributions aren’t expected to be pure exponentials.

Single Spin Asymmetries

The following are summary results (val ± err and χ2) from straight-line fits to single-spin asymmetries versus fill (37 degrees of freedom):

π-

π+

(I don’t know why Drupal’s code module always botches those fixed-width data tables). The plots themselves can be found here. I also have SSA plots versus z.

Data / Monte Carlo Comparisons

I finished the first batch of data / Monte Carlo comparison histograms. I used the integrated luminosity for my runlist (5.43 pb^-1) to scale the simulations, but I reduced this normalization factor by a factor of 0.82 when comparing the track-level quantities to account for PID inefficiency in the data (I use Pythia PID information in the simulations because dE/dx is unreliable there).

Some general conclusions:

• the normalization looks really good for track comparisons, but in event-level comparisons the data is high by ~50%. Weird.
• there’s a clear deficiency of high-z / high-pT π- … it’s more dramatic on the near-side, but it’s there on both sides
• the z-distribution for π+ agrees extremely well for z>0.25; below that the simulations fall off (falloff is also the case for π-)

Estimated Analyzing Power from PYTHIA Method of Asymmetry Weights

Same deal, need to redo this calculation once the 2006 MC MuDsts are online again.

Relative Luminosity Systematic

9.4e-4 (Murad’s detailed documentation)

Non-longitudinal Beam Components

TODO Originally calculated bias vs. pT, need to redo using histos vs. z.

PID Background

TODO, will also look into applying Yichun’s rdE/dx corrections to see if they change the yields in any significant way.

Jet Reconstruction Bias

TODO Merge results from inclusive jet analysis, since my cuts match

Jet Trigger Bias

TODO I’m hopeful that this can be neglected in my analysis. The major effect of the trigger bias is to change the subprocess mixture as a function of jet pT. We’re not plotting this measurement against jet pT, but against z. We’re also counting jets ~2 GeV above the trigger threshold.

Other Checks ?