[Rgc] PbPt analysis for 4 recent NH3+ runs and 4 recent NH3- runs

[Sebastian Kuhn]

Dear all,

now that Maxime has kindly shared his code, I would like to make a few suggestions to improve the precision of the analysis (as well as it can be done without the benefit of a full alignment of the FD and CVT and subsequent calibration):

1) Beam energy: I would like to propose that instead of using the nominal beam energy (10.547 GeV), we determine the beam energy from one of the NH3 runs, by plotting 

E = 1/TAN(theta_3/2)/TAN(theta_p)-1)*0.93827

for a clean sample of H(e,e'p) events. By “clean” I mean applying the delta-phi cut on the proton and requiring W < 1.2 for the electron. By plotting this quantity “E” as a histogram over all events in a run, you should see a clear peak, whose centroid would be the “effective” beam energy. (This may still not be the “true” beam energy, since it could be distorted by misalignment and incomplete calibration of the tracking devices; however, by using this number instead of 10.547, we can be sure that all variables that we want to cut on will be centered on their expected value. 

2) Cuts: Once we have Ebeam, we can calculate an E’ cut: Plot 

Delta_E’ = E’_meas - E’_exp = E’_meas - E/(1 + 2*E/0.93827*(SIN(theta_e/2))^2)

(again, for all events with “clean” protons) to find the elastic peak. Then apply a +/-2 sigma cut on this peak for further analysis. This should remove only 5% of good events but reduce the background significantly.

3) After this, plot Delta-theta and fine-adjust the cut on that (again, I recommend +/- 2 sigma to optimize signal/background). 

4) Finally, plot again Delta-phi to determine the optimum cut (+/- 2 sigma, NOT MORE), and to find the optimum scale factor for the C background.

5) For the C background, combine ALL carbon runs that have already been cooked and produce the same Delta-phi spectrum as in 4), with the same cuts 1-3. It should be relatively featureless. Adjust the scale factor “by hand” so that the spectrum coincides with the wings of the NH3 or ND3 spectrum for both helicities combined. I am not sure a Gaussian fit + this C background to Delta-theta is the best choice, but you can do that as an alternative (to check for systematics). 

6) Calculate the difference for helicity+ - helicity- for the NH3/ND3 spectrum within all cuts, and divide by the helicity SUM minus the fitted C background:

Asym = (N^+(NH3) - N^-(NH3)) / (N^+(NH3) + N^-(NH3) - N(C)), 
where N(C) is scaled to the wings of the delta-theta distribution.

7) The final correction would be for the beam charge asymmetry - determine this from the helicity- and life-time gated FC scaler for each run, and subtract the beam charge asymmetry from the measured physics asymmetry.

The final result can be divided by the theoretical asymmetry to get the product of Pb*Pt. (In the case of the deuteron, I would have to take a look at the distributions of all the variables plotted in 1-4 to determine if there is a correction factor due to the fact that we are cutting off some of the high-momentum tail. In that case, the true DEUTERON polarization would be slightly smaller than that given by the measured quasi-elastic (e,e’p) asymmetry, because of the D-state of the deuteron. But I expect this to be a small correction which is not relevant right now.

- Sebastian

> On Jul 19, 2022, at 5:24 PM, silvia at jlab.org wrote:
> 
> Dear all,
> thanks to Maxime, who left us his codes with very good documentation, I
> managed to run the PbPt analysis for two sets of recently cooked NH3 runs,
> for the two target polarization signs (4 runs per sign).
> The results are here in attachment, the error bars are still pretty big -
> assuming a Pb of ~84% we get:
> - Positive pol. Pb ~ 0.64 +- 0.11
> - Negative pol. Pb ~ 0.59 +- 0.18
> I have not done any optimization to the channel-selection cuts, just ran
> the code as it was.
> Best regards,
> Silvia
> 
> <Asy_compar_negative.png><Asy_compar_positive.png><PbPt_negative.png><PbPt_positive.png>_______________________________________________
> Rgc mailing list
> Rgc at jlab.org
> https://mailman.jlab.org/mailman/listinfo/rgc