Test 2: compare the efficiency on vertex separation cut for $2\nu\beta\beta$

[Shoram444]

GOAL

To see whether some option vertex separation cut should be used for 2 electron topology data.

PROCEDURE

We should use the findings on the reconstructed vertex uncertainties in #2 to find the efficiency as:

• function of relative separation $d = \sqrt{(y1-y2)^2 + (z1-z2)^2}$ for signal and background processes
• function of separation in y $d_y = |y1 - y2|$ for signal and background processes
• function of separation in z $d_z = |z1 - z2|$ for signal and background processes
• function of separation given by an ellipse with the semi major axes given by $\Delta y, \Delta z$ uncertainties from Test 1: Comparison of vertex reconstruction uncertainty (CAT vs TrackIT vs true vertex) #2

For each scenario listed above:

• create a histogram of passed evenets as a function of distance
• create a plot "fraction passed vs distance" together with signal/background ratio (see here )

SIMULATION SETUP

• Magnetic field off
• 1M+ events of:
  • generated vertices on source foil bulk:
    • $2\nu\beta\beta$, $0\nu\beta\beta$, Bi214, Tl208, Pa234m, K40
  • genertated vertices on wire surface:
    • Bi214 (Bi-Po from Radon)
  • genertated vertices on foil surface:
    • Bi214 (Bi-Po from Radon)
  • genertated vertices on PMT glass bulk:
    • Bi214, Tl208, K40
  • generated vertices outside:
    • Bi214, Tl208, K40
    • neutrons?
    • high energy gammas? (@xaguerre)
• data-cuts: 2 particles tracks, 2 OM hits, 2 foil vertices, internal/external probability
• for CAT make sure that the reco.conf file only reconstructs "line":
  • in reco.conf file change: fitting_models : string[1] = "line"

RELEVANT INFO

• Main: TrackReconstructionTests/Test2
• Scripts: TrackReconstructionTests/Test2/scripts
• Data (on CC-LYON): /sps/nemo/scratch/mpetro/Projects/PhD/TrackReconstructionTests/Test2/data

RESULTS

• figures, calculated numbers (i.e. $\varepsilon$, FWHM...)
• a short summary report or a jupyter notebook

[miroslav-macko]

I would extend the distance cut from circle metrics $d = \sqrt{(y1-y2)^2 + (z1-z2)^2}$ to elliptic metrics: $d = \sqrt{\frac{(y1-y2)^2}{b^2} + \frac{(z1-z2)^2}{c^2}}$. This adds two parameters into game but we also know we won´t have circles.

We could use two approaches:

1. We could study the separation in two dimensions (y and z) separately and then once we would have an idea what is the ratio of the "ideal cuts" in each dimension we fix this ration and we would have only one parameter left.
2. We could study this as 2D problem from the beginning and make 2D cut maps. It is more rigorous and also it would produce very nice pictures.