Here I take a first (very preliminary) look at the dijet asymmetries in the RFF field configuration ...

With the 2009 FF reproduction to correct the high pt track problem ongoing, I have started my dijet ALL analysis on the RFF dataset. I have done a Run QA (detailed here) to create a list of runs appropriate for a dijet analysis. The list of RFF runs I use in this analysis can be found here.

I impose a number of cuts to select good events and jets:

• valid spin4 value
• vertex rank > 0
• |Zvertex| < 90
• -0.7 <= detEta <= 1.7
• -0.8 <= eta <= 1.8
• jet RT <= 0.94 (only for jets with det eta < 1.0)
  
• jet sum track pt >= 0.5 (only for jets with det eta < 1.0)
  

The two highest pt jets meeting the above requirements become dijet candidates. To be counted as a dijet, one of these two jets must be geometrically matched to a jet patch which fired the L2JetHigh or JP1 triggers. In addition, the jets must be back to back in phi. The exact condition is: Cos(phi1-phi2)) <= -0.5 .

In my analysis, I look at 5 different dijet topologies: Barrel - Barrel, East Barrel - Endcap, West Barrel - Endcap, Endcap - Endcap, and full acceptance. The topologies are defined by detector eta, so for example an East Barrel - Endcap event will have one jet with det eta < 0.0 and the other jet with det eta >= 1.0.

Figure 1: This plot shows the 5 different dijet topologies. Plotted is the low pt jet physical eta vs the high pt jet physical eta. Note that because the topologies are defined in terms of detector eta, there is some leakage into other regions when plotting in physical eta.

Figure 2: This figure shows some additional kinematic relationships. The same side and away side jets are defined as follows: the same side jet is that which is triggered and has the larger ammount of neutral energy. All other quantities should be evident.

A pdf containing plots of these and several other kinematic quantities can be found here.

Note: The EBarrel-Endcap and WBarrel-Endcap plots in figure 1 and the high pt eta and low pt eta plot in figure 2 show that the high pt dijet is more likely to be in the barrel. This is most likely do to the drop off in tracking efficiency in the endcap. This bias will need to be explored in simulation.

Figure 3 below shows the statistical power of the dijet ALL measurement for the 367 RFF runs I have analyzed. The data are divided into 9 invariant mass bins, each 10 GeV wide. The errors are given by the following formula:

P_A and P_B are the polarizations of the beams, N++ is the number of dijets in that bin from events with the same helicity signs (N++ = N++ + N--), N+- is the number of dijets in a bin from events with the opposite helicity signs, R is the relative luminosity between helicity states, and the /Delta Ns are the square roots of the number of dijets of the given helicity configurations in that bin. The sum is over the number of runs.

Figure 3: This figure shows the raw statistical power of the dijet ALL measurement for the RFF runs as a function of invariant mass in each topological configuration. Note that the y axes have different scales.

I have also looked at several false asymmetry combinations. A pdf which defines these asymmetries and their errors can be found here.

Figure 4: This figure shows the four false asymmetries as a function of invariant mass for each dijet topology. The first group of five plots is the yellow beam single spin asymmetry, the second group of five is the blue beam single spin asymmetry, the third group of five is the like sign double spin asymmetry, and the last group of five is the unlike sign double spin asymmetry.

Figure 5: This figure shows the four false asymmetries integrated over all mass bins for each dijet topology. In each pannel, the first bin is the yellow beam asymmetry, the second bin is the blue beam asymmetry, the third bin is the like sign asymmetry, and the fourth bin is the unlike sign asymmetry.

A pdf which contains all the asymmetry plots above as well as plots which show the asymmetries and their errors as a function of run number can be found here.

The final thing I looked at were the values of the beam polarizations and relative luminosity values (R1-R6) on a run by run basis. These can be seen in the figure below.

Figure 6: This figure shows the relative luminosity values and beam polarizations as a function of run number. Note that the plot titles are wrong for R3-R6, but the Y axis lables are correct.

This is a very preliminary look at the kinematics and asymmetries, no corrections have been applied to the plots shown above. There are a number of issues which will need to be explored to move this analysis forward:

• I will need to look at simulation for a number of issues. In addition to the usual efficiency, trigger bias, etc corrections which rely on simulation I will need to use simulation to look at how jet properties change as they move into the endcap region where TPC tracking efficiency is dropping off. For example, see note after figure 2.
• Some of these jet effects can and should be investigated in a data driven way as well. Some of this work has already been done. These studies should be expanded and refined.
• I should think about how to divide the data into samples by jet trigger. Right now I only tag jets as high or low pt or away or same side.
• Probably many other things I'm not thinking about right now.