Here I detail my exploration of the high pt tail seen in the previous study.

 

Figure 1: This figure shows the jet pt spectra from my previous dijet analysis. A deviation from the expected exponential fall can be seen in the high pt jet spectrum in the upper right hand plot.

 

The plot in the third row in the right hand column shows the away side jet pt vs the same side jet pt. I define away side and same side as follows: For each jet I look at the total TOWER pt, if the jet with the higher tower pt fired a trigger it is called the same side jet and the jet with the smaller pt is the away side jet. If the jet with the higher tower pt did not fire a trigger, the jet with the lower pt is the same side jet and the jet with the higher pt is the away side jet. Note that the condition that at least one jet fired a trigger has already been applied, so at least on jet fired the trigger.

 

In order to study the cause of the high pt tail, I have tried to isolate a sample of dijet events which look 'tail dominated'. The criteria I use are: high pt jet pt >= 50GeV && low pt jet pt <= 30GeV. You can see these regions in the third row of plots in figure 1.

 

All the plots shown below and several more can be found in the following pdfs. This pdf gives plots for dijets in the tail region. This pdf gives plots for dijets in the non tail region. And this pdf gives plots for several coincidence measures.

 

Figure 2: This plot shows the runs which contain high pt tail events as I define above. The x axis is run index, a mapping between run number and run index can be found here.

 

Figure 2 shows that the events are not localized to a few runs which may be expected if the problem were a hot tower, but there is definitely an increase in the frequency around a run index of ~450 (roughly day 146). Another thing to look at would be if all the tail events were localized in eta or phi.

 

Figure 3: This figure shows the eta and phi distributions for tail and non tail events. The left column shows events which are in my definition of the tail, the right column shows events which are not in my definition of the tail. In each plot, the red curve is for the high pt jet and the blue curve is for the low pt jet.

 

Figure 4: This figure shows the eta-phi distribution of tracks within the jets. The top two plots show jets which are in the tail region, the first plot is for the high pt jet and the second plot is for the low pt jet. The bottom two plos show jets which are not in the tail region, the same pattern for the high and low pt jets holds.

 

Figure 5: This figure has shows the eta-phi positions of towers within the jets. The left column is for the high pt jets and the right column is for the low pt jets. The first row shows the barrel towers for the jets in the tail region, the second row shows the endcap towers for the jets in the tail region. Rows three and four follow the same pattern for the jets not in the tail region.

 

Figure 3 clearly shows that the tail events are strongly concentrated in phi and to a lesser extent in eta. Figures 4 and 5 show that the track and tower eta-phi distributions match this pattern as expected.

 

Our next task should be to get an idea of what is causing this behaviour. Usually, concentrated events like this can be attributed to a hot tower. It could also be a result of a miss calibration in a TPC sector which causes some tracks to be reconstructed with the wrong curveture and thus giving very high track pt.

 

Figure 6: This figure shows the pt of each track inside a jet. The top plot shows jets in the tail region and the bottom plot shows jets in the non tail region. In both cases the red curve is for the high pt jet and the blue curve is for the low pt jet.

 

There is obviously an excess of very high energy tracks in the high pt jets in the tail section. The fact that I don't see similar behaviour in the tower energy and pt plots and the fact that the number of tracks and number of towers plots don't show any huge discrepencies (see pdf files) leads me to think that this may be a problem with the TPC calibration.

 

Figure 7: This figure shows several track, tower, and jet ratios for the tail sample and non tail sample. The top left plot shows the ratio of the total track pt in a jet to the total pt of the jet. The top right plot shows the ratio of the total tower pt in a jet to the total pt of the jet. The bottom left plot shows the ratio of the total track pt in a jet to the total tower pt in a jet. The top three plots are for the tail sample and the bottom three plots are for the non tail sample.

 

Figure 8: This plot shows the pt balance between the high jet and the low jet in an event. The red curve is this ratio for the tail dijets and the blue curve is this ratio for the non tail dijets.

 

Figure 9: This plot shows the dijet invariant mass for dijets in the tail region and in the non tail region. The red curve is this ratio for the tail dijets and the blue curve is this ratio for the non tail dijets.

 

 

Now I want to go back and look at the time structure of the tail events. Figure 2 shows the frequency of high pt tail events as a function of run. We see that although the tail events occur throughout the running period, there is a sharp jump in the number which occur per run starting around run index 460. Run index 459 corresponds to run 10146091 and run index 460 corresponds to run 10147124, between these two runs the polarity of the magnet was switched. The feature in figure 2 seems too sharp to be a coincedence, so I decided to remake the plots above for the RFF runs (1-459) and the FF runs (460-952) seperately.

 

This pdf contains the RFF tail plots. This pdf contains the RFF non tail plots. This pdf contains the RFF coincedence plots. This pdf contains the FF tail plots. This pdf contains the FF non tail plots. And this pdf contains the FF coincedence plots.

 

Figure 10: This figure shows phi and pt distributions. The top four plots are for the RFF runs and the bottom four plots are for the FF runs. For the RFF and FF plots, the top left plot shows the phi distribution of tail events and the top right plot shows the distribution of non tail events. The red curve is for the high pt jet and the blue curve is for the low pt jet. The bottom left plot shows the low pt jet pt vs the high pt jet pt and the bottom right plot shows the away side jet pt vs the same side jet pt.

 

Obviously we see very different behaviour between the RFF and FF runs. In the RFF runs the tail events are spread equally in phi whereas in the FF runs, the tail events are concentrated in one phi region. We also see a change in the non tail phi distribution with a suppression in one of the phi regions in the FF runs.

 

Figure 11: This figure shows the high and low jet pt spectrua for the RFF sample (top) and FF sample (bottom).

 

 

Figure 12: This figure shows the track pt in the tail jets (top) and non tail jets (bottom) for the RFF runs.

 

Figure 13: This figure shows the track pt in the tail jets (top) and non tail jets (bottom) for the FF runs.

 

 

Summary: There is an obvious change in the behaviour the tail events at the time of the magnetic field change. The tail events in the FF runs are concentrated between 0 and -1 in phi. This corresponds to sector 20 in the TPC which is known to have had calibration trouble. In addition to the tail events between 0 and -1, there is a suppression of non tail events between -1 and -2 in phi which begins with the field change.