Objective:

Explore the pt-dependent mean mass position in data and MC and perhaps draw some conclusions about the quality of our simulations.

Details:

Data Vs T2 platinum MC (see here for explanations)

Bins in Pt {5.5, 6., 6.75, 7.5, 8.25, 9.25, 11., 13., 16., 21.}

Cuts:

•     Events pass L2gamma software trigger and (for data) online trigger.
•     candidate pt > 5.5
•     charged track veto
•     at least one good strip in each smd plane
•     Z vertex found
•     | Z vertex | < 60 cm

Plots:

1)

The above plot shows data (black) Vs. MC (red) for Zgg for my 9 pt bins.  The MC plots are normalized to the data so that the total number of entries is equal in both distributions.

2)

Apologies for the poor labeling.  The above left plot shows the mean mass per Pt for data (black) and MC (red).  These means are calculated by fitting the mass spectra to a gaussian between .1 and .2 GeV/c^2. (see here for more)  In addition to the cuts listed at the top of page, I make a Zgg < .7 cut and an | particle eta | < .7 cut on all pion candidates.  The PDG mass of the pi0 is shown in blue.  The above right plot shows the ratio of Data/MC for the mean masses, again as a function of Pt.  This plot is fit to a flat line and the fit parameters are shown.  

Conclusions:

The most basic conclusion we can draw is that the simulation is recreating the floating mass peak quite well.  The data is never more than 3% higher or lower than the simulation and a flat-line fit is really darn close to one.  Couple this with the Zgg comparison and I think we can say that we are simulating the EMC (and SMD) response correctly, at least as it concerns reconstructing neural pions between 5 - 20 GeV/c.  Of particular interest is the Zgg comparisons at relatively higher Pt, as then the distance between the two reconstructed photons is smaller than the size of one tower, and we rely on the SMD to correctly identify the daughter photons.