This page combines my previous studies into the practicalities of doing a jet analysis in the endcap. Previous write ups can be found here and here. The studies presented here are focused on the questions about how to deal with the loss of tracking which occurs as one moves into the endcap region. What are the systematic changes in jet properties that occur due to the fall off in tracking? Can these effects be mitigated by using the emc only branch of the jet finder which contains no tracks? How do we recover the full Pt of the jet in regions where tracking is poor?

This study differs from the ones linked above in that I have added JP1 triggered events in addition to Dijet and Random L2JetHigh events. This study was also done after the 100% subtraction scheme bug was fixed by Pibero. A discussion of that bug can be found here.

The plots shown are from the first 20 runs in Pibero's 'golden list' which were reproduced to fix the 100% subtraction bug.

Kinematic Quantities:

The first set of plots compare basic kinematic variables. The pdf file contains nine pages, the first four show jet pt, jet eta, jet phi, and jet phi vs eta for 12 point branch jets. The next four show the same quantites for emc only branch jets. The last page shows a comparison of the jet yields between the 12 point and emc only branches as a function of jet eta. This plot is shown in figure 1. Each page contains four pannels. The upper left contains all jets for events in which at least one jet is matched with a fired jet patch. The upper right and lower left pannels contain jets which fired an L2JetHigh or JP1 jet patch respectively. The lower right pannel contains jets which did not fire a jet patch, although the event must contain at least one jet which did fire a jet patch.

The pdf containing the plots of kinematic quantities is here.

Figure 1: Plots comparing the jet yield from the 12 point branch of the jet finder (black) to the emc only branch of the jet finder (red). Ploted vs jet pseudorapidity.

As can be seen, we take a significant hit in statistics when we move from the 12 point branch to the emc only branch. Overall, the emc only branch contains 59% of the jets that the 12 point branch contains. For jets which match a jet patch that fired the L2JetHigh trigger, the emc only branch contains 98% of the jets that the 12 point branch contains. For jets which match a jet patch that fired the JP1 trigger, the emc only branch contains 72% of the jets that the 12 point branch contains. Finally, for jets which do not match a fired jet patch, yet have at least one jet in the event which did, the emc only branch contains a mere 8% of the jets that the 12 point branch contains. The low yield for unmatched emc only jets would not be a problem in an inclusive measurment where you only look at jets which match a fired jet patch, but in a dijet analysis it is fine to look at events where one of the jets matches a trigger patch and the other does not. If a dijet analysis is done using the emc only branch, it will be severely biased against dijets in which one of the jets is unmatched as compared to the 12 point branch unless a way is found to increase the emc only jet yield.

'Jet Stability' Study:

The so called Jet Stability study is focused on quantifying how jet properties change as they move into different regions of the detector. Jet quantities such as Pt, Phi, number of tracks, number of towers, the track and tower contribution to the total jet Pt, among others are ploted vs the detector eta of the jet. Also included are plots which show how the psuedorapidities of the track and tower components of a jet contribute to the overall eta of the jet as a function of the jet's detector eta.

I divided the study into four sections. There are plots looking at the 12 point branch and the emc only branch. Each of these are further divided into jets which match a fired jet patch and jets which don't match a fired jet patch.

The pdf containing plots for jets from the 12 point branch which matched a jet patch is here.

The pdf containing plots for jets from the 12 point branch which did not match a jet patch is here.

The pdf containing plots for jets from the emc only branch which matched a jet patch is here.

The pdf containing plots for jets from the emc only branch which did not match a jet patch is here.

Figure 2: This plot shows jets from the 12 point branch which match a jet patch. The top left pannel shows the average number of tracks in a jet as a function of the jet's detector eta. The top right pannel shows the average number of towers in a jet and the bottom two pannels show the number of barrel and endcap towers. Each pannel has four curves, these represent a binning in jet Pt. The black curve contains jets with Pt between 5 and 8 GeV. The red curve contains jets with Pt between 8 and 15 GeV. The blue curve contains jets with Pt greater or equal to 15 GeV. The green curve contains all jets with no Pt cut applied.

Figure 3: This is the same information as found in figure 2 but now for jets from the emc only branch which match a fired jet patch.

Figure 2 shows that the number of tracks in a jet starts falling off around a detector eta of 0.5. This is not too suprising considering that the jet cone size is 0.7, so even jets with a detector eta of 0.5 will overlap the endcap somewhat. The figure also shows that the number of towers in a jet is roughly constant over the detector.

Figure 3 shows that the number of tracks in emc only jets is zero as expected. If we compare the average number of towers in a 12 point jet to the number of towers in an emc only jet, we see that the emc only jet contains ~2-3 more towers on average. I believe that this is primarily due to the 100% subtraction scheme used to avoid double counting energy. If a tower has a track pointing to it, 100% of the track's energy is subtracted from the energy of the tower. If this subtraction removes all the energy from a tower, the tower is not included in the jet. So in practice, jets in the 12 point branch should have fewer towers on average because some will have tracks pointing to them which take all of their energy. The effect is not present in the emc only branch because there are no tracks.

Figure 4: This plots shows jets from the 12 point branch which match a jet patch. The top left pannel shows the average Pt of a jet as a function of the jet's detector eta. The top right pannel shows the average jet phi. The bottom left pannel shows the track contribution to the jet Pt. The bottom right pannel shows the tower contribution to the jet Pt. Again the black curve is for jets with Pt between 5 and 8 GeV, the red curve is for jets with Pt between 8 and 15 GeV, the blue curve is for jets with Pt greater or equal to 15 GeV, and the green curve is for all jets regardless of Pt.

Figure 5: This is the same information as found in figure 4 but now for jets from the emc only branch which match a fired jet patch.

In the top left pannel of figure 4 we can see the drop in total jet Pt as the jets move toward the endcap. This decrease in Pt corresponds to the decrease in total track Pt seen in the bottom left pannel of the same figure as well as the decrease in the number of tracks as seen in the top left pannel of figure 2. Looking at the bottom right pannel of figure 4, which shows the total Pt of the towers in the jet, we see a rise starting around a detector eta of 1. I believe that this is mainly due again to the track energy subtraction scheme. As one moves forward in the detector, the number of tracks starts to decrease meaning there will be fewer tracks hitting towers and so there will be fewer towers having energy subtracted from them, increasing the average. In figure 5 we see the same plots for jets from the emc only branch. As we would expect, the track component is zero and the total jet Pt is given completely by the tower component. We also see that the total tower Pt is flat across the detector when there are no tracks anywhere in the detector.

Figure 6: This figure shows where the tracks and towers are located with respect to the total jet and with respect for eachother for jets in different regions of the detector. The top left pannel shows the difference between the total jet pseudorapidity and the pseudorapidity of all tracks in the jet vs the detector eta of the jet. The top right pannel shows the difference between the total jet pseudorapidity and the pseudorapidity of all towers in the jet vs the detector eta of the jet. The bottom left pannel shows the difference between the pseudorapidities of the tracks and towers in the jet as a function of the detector eta of the jet.

The plots in figure 6 are my attempt to see how much the fall off in tracking affects the measured jet pseudorapidity. One could imagine that the deficit of tracks at higher eta could cause jets to have lower pseudorapidities as compared to what they would have had if tracking efficiency was uniform across the detector. In the upper left pannel, we can see a rise in Jet - Track pseudorapidity as jets move into the endcap caused by the drop off in the number of tracks as you move forward in the detector. The total jet pseudorapidity is pushed higher by the tower contribution but the track component does not extend as far. In the upper right pannel, we see no bulk shift, but we do see a decrease in the range of the positive side of the distribution as the jets move forward. This just represents the fact that when tracking dies out, the track component cannot push the full jet to higher psuedorapidities than the tower component.

Branch Comparison Study:

The branch comparison study is focused on investigating how well jet properties for a jet reconstructed in the emc only branch reproduce the properties of a jet reconstructed in the 12 point branch for the same physical 'jet object'.

The pdf containing the plots for the branch comparison study can be found here.

Figure 7: These plots compare quantities for correlated 12 point and emc only jets

Figure 8: These plots compare quantities of the tower component of 12 point branch jets to emc only branch jets for correlated jets.