B1pi tests Sep. 2011

In the following the GlueX Monte Carlo and and analysis code revision 8233 as of end of August 2011 is used. 50000 b1pi events are generated for two geometries. One is the current standard geometry with a model of the BCAL readout including a copper cooling plate. A second geometry is identical with the exception that the BCAL readout part has been removed. This was done to see the impact of the BCAL readout on the reconstruction of these events. The event vertex is randomly distributed within the 30cm long target along the z-axis.
As one can see there is no obvious difference in the reconstructed mass distributions for pi0, omega, rho, b1 and X regarding the presence of the BCAL readout in the geometry of the simulation.

The pi0 mass (left=standard geometry, right=no Bcal readout):

Here and in the following the black histograms are the thrown Monte Carlo values however the photon energies and charged particle momenta are smeared before the masses are reconstructed.

The omega mass (left=standard geometry, right=no Bcal readout):

The b1 mass (left=standard geometry, right=no Bcal readout):

The X mass (left=standard geometry, right=no Bcal readout):


To investigate the effect of the BCAL readout we look next at the distributions of photons and charged tracks as a function of angles as well as 2-dim. distributions of energy vs. angle for the different particle tracks.

Photon angular distribution (left=standard geometry, right=no Bcal readout):

In the above plot the black are the thrown MC photons and in red are the reconstructed photons. Since 50k events are thrown we have 100k photons thrown. About 30% more photons are reconstructed. While there are always more photons reconstructed at all angles a significant overshot is seen at angles above 10.5 degree. This is the effect of shallow impact angles into the BCAL. More on this later. Also there is a small bump above 40 degrees. This is caused by protons that are miss identified as photons. This will become apparent when examining more plots below. The difference between having a BCAL readout in the geometry (left) or not (right) is very small and concentrated around the angle of 10.5 degree. So the next plots are a zoom into this region.
Photon angular distribution zoom (left=standard geometry, right=no Bcal readout):

The location of the sharp rise in reconstructed photons at around 10.5 degrees is the same in both plots. However the structure of the peak is slightly different. Without a BCAL readout in the geometry the bump is rather flat over about 4 degrees and then falls off rapidly. With the BCAL readout in the geometry this bump has some structure. The bump is about 3% higher and has a dip around 12 degrees but still is about 4 degrees wide before falling off rapidly. However even at angles of 20 degrees and beyond there is a significant overestimation of the number of reconstructed photons. This might be related to protons that failed to be reconstructed resulting in a calorimeter cluster without an associated track. Below 10.5 degrees the reconstructed number of photons is roughly 10 to 15 % higher than the generated number of photons except at very low angles where the acceptance of the FCAL cuts off.

At this point a look at the proton reconstruction is usefull. Below the angular distribution of generated and reconstructed protons is shown where on the left the geometry include the BCAL readout and on the right this is removed.

As one can see there is a substantial lack of reconstructed protons. In fact taking the bare numbers one can state taht not event 20% of the generated protons are reconstructed. In addition there is this peak at about 5 degrees whose origin is unknown at this point but are most likely miss identified pions. To get more insight into this issue we look at the 2-dim distribution of these events, the energy vs. angle distribution of protons. Note also that in not even 10% of the events a proton is reconstructed!

It is apparent from these plots that the reconstruction of protons fails in most of the cases at angles below 40 degrees. The majority of reconstructed protons at these smaller/forward angles (red dots) are most likely miss identified pions while the true protons (black), if the track is reconstructed, are probably miss identified as pions.

This leads us to the positive pions.(left=standard geometry, right=no Bcal readout):

While there is no obvious effect from the geometry seen there is a clear bump above 40 degrees seen. This bump is readily identified with the generated protons seen in the previous plots. There is a significant amount of proton that are miss identified as pions. Note that at about 30 degrees there are the same amount of protons and positive pions generated, namely about 200!
The next plots are a zoomed in version of the previous pi+ distributions

The low energy cut off around 1 degree is the acceptance of the FCAL and the tracking detectors. The distribution levels off between 2.5 degrees and 5 degrees exhibit kind of a shoulder before increasing again to a peak around 9 degrees and then falling off at a slower pace than the generated distribution.
The following plots show the negative pions:

The distributions for negative pions look very similar to positive pions except at large angles where the positive pion distribution is contaminated with protons. The origin of the shoulder around 3 degrees is not known at this point.