1

PID Composition in Testbeam-2 for the LAMF (06/22/2015)

The Goal is to find for both Data-Run-1&2 for Test Beam-2 Data. Some basic ideas about the ToF sytem...page-2

First Step: Analysis of the contamination beside the pion-peak (using ToF variables for the Cut and the Detector variables for the eye-scanning).....page-7

Some initial problems, hypotheses and details to look at. Distinguishing between Pions and Muons.....page-16

Other interesting TB-analysis and a ToDo List.....page-19 A Procedure to Follow in order to achieve the Goal.....page-25

Antonio Zegarra Borrero

Universidad Nacional de Ingeniera

2

First I attached the actual arrangements of the beamline elements for Minerva TB-2

3

Two different Detector configurations for Data-Run-1 & 2

4

Considering each particle has a defined momentum, and the distance between ToF-1 & ToF-2 we can find the time needed for a particle to pass through both stations.

To have a point in the ToF histograms we require that both stations are hit by a particle and the trigger fires.

All events considered are those which fulfill the following conditions: {-There is beam(In Spill),-The trigger fired, -Activity in the detector(The trigger fired in a specific slice), -All 6 ToF-PMTs sent a signal, -I have not vetoed events for the first sample analyzed}

5

From the previous equation one can find the zero (considering the 120 GeV protons) and the difference between the pion and proton peaks (vertical bars). Here some results presented in the ToF Tech Note (docDB 10750v-1 )

6

Relevant things: Contamination beside the pion-peak, no peak of Kaons, peaks from particles from subsequent buckets. Evidently, either the energy was lower or the beamline is longer. The % of p increases with E_{beam}.

7

I began by looking at ToF variables In the script I just took 2 cuts: ToF_quality == 1 and Time-cut (to select

the contamination time-interval) beside the other mandatory cuts in the code (In_spill>0.5: The Beam was on, event.n_slices>0: there is activity in the Detector and triggered: the trigger was fired for that event).

I considered All Events, so did not apply the Veto-Cut (I will show why I think that cut will wipe out many of my events, although Leo's hypothesis is that since ToF-2 and Veto don't overlap significantly the effect should be negligible....This to be tested yet).

I performed the Time-cut just using a non-formal criteria (my eye) to perform the Cut in time (to isolate the contamination) and then looked how those events appeared in Arachne and made a spreadsheet to have an idea of their composition.

I considered that almost all kaons decay so muons appear as products from the decay.

8

Results for Data-Run-2 4GeV_Pos_Pions (Contamination-Interval ~ [2000,5000 ] (ps))

I scanned those Events (114 of them) in Arachne and found that the contamination is mainly due to muons (> 50%):

9

4GeV_Neg_Pions (Contamination-Interval ~ [2500,4000 ] (ps))

*I scanned those Events (83 of them) in Arachne and found that the contamination is again mainly due to muons (>~ 50%):

10

During the scanning procedure I found many events that I call party of particles, I would not like to wipe them out (muons are almost always the soul of the party) so I did not considered the Veto-cut:

TB_00001522_0073_cosmc_v09_1504250719_mergedDST.root gate=523 (Gate 523: Triggered Slice: 4)

TB_00001522_0073_cosmc_v09_1504250719_mergedDST.root gate=404 (Gate 404: Triggered Slice: 5)

11

Here I present the histograms for the other folders of Data-Run-2 and Data-Run-1 (only Pion-Folders) and the Contamination interval for each case (only for E < 8GeV). There was a problem with an argument of the DrGranCoolTool function that was solved and now I am be able to scan the events properly.

6GeV_Pos_Pions :Contamination-Interval ~ [1500,2500 ] (ps)

6GeV_Neg_Pions :Contamination-Interval ~ [1300,2000 ] (ps)

12

For E>=8GeV there is an overlap between the pion and proton peaks so ToF not useful to distinguish between these 2 species. There are some other peaks beside them that would be nice to know what they actually are (In ultrarelativistic limit the difference between the peaks is the resolution of the ToF-system ~ 200ps ).

In the pion-peak we have also muons there so to find their amount I will need to rely in other Cuts that look at Detector variables (ToF could still be useful for this?):

*E-cut (for high-E) && *LP-cut (for low-E)

13

Histograms and contamination intervals for DataRun-1:1.77GeV_Pos_Pions, Cut-Interval ~ [9000,14000] ? 2GeV_Pos_Pions, Cut-Interval ~ [7500,9200]?

2GeV_Neg_Pions, Cut-Interval ~ [9000,10000]? 3GeV_Pos_Pions, Cut-Interval ~ [7500,15000]?

14

4GeV_Pos_Pions, Cut-Interval ~ [7500,11500]? 4GeV_Neg_Pions, Cut-Interval ~ [8200,11000]?

6GeV_Pos_Pions, Cut-Interval ~ [7500,8500]? 6GeV_Neg_Pions, Cut-Interval ~ [7000,8200]?

15

7GeV_Pos_Pions, Cut-Interval ~ [7000,7500]? 8GeV_Pos_Pions

8GeV_Neg_Pions

16

Some initial problems, hypotheses and details to look at.

Explain why there are some cosmic-muons (they appear alone) in the contamination for 4GeV (Data-Run-2). Some hypotheses:

A beam-particle fired the trigger and ToF-1, it was scattered and a cosmic muon hit the ToF-2.

A beam particle fired the trigger, close both ToF stations but then a cosmic passed through the detector (?).

Maybe a pion decayed (between the start & stop stations) and the result is a muon at a certain angle w.r.t. the direction of the pion

Why there is a signal that ends at 15 ns for low-E samples? What is its composition?

What if the pion peak has a fat-tail at the right? Then the amount of actual-contamination will be reduced?

17

To find the % of pions and muons (for events in the pion-peak) I will use the E-cut and LP-cut. Add counters inside the script so it could tell me what to expect regarding the amount of these species (beside the mandatory eye-scanning). Would this increase the accuracy of particle-ID? What about the possibility of using a dE/dx analysis?

I wonder if the MC-simulation of particles inside the TB-detector could help me in increasing the efficiency of particle-ID and how to use that information.

If the MC-simulation of the 2ndary beam is finished in a couple of months, will it be useful to make comparisons?

I want to look at the 120GeV proton Data and calculate the spread of the (only) proton-peak, see also if it is symmetric (to find measurement issues or particles with unusual trajectory). Why is it relevant to compare this spread with the Pion-peak spread?

18

I am also working (beside Rob and Aaron) in an analysis of the Veto-Counters to find the spatial-distribution of the beam (to find how much it is centered). Some things we actually know:

The number of times each counter fired for each Energy The correlation-matrix for each Energy (Spatial

correlation only) The previous spatial correlation permits multiple

possibilities because there are many cases of overlaps of more than 2 veto paddles in actual physical space, then we require time information (At what time each counter fired for each event) in order to actually locate one unique point in space

Next some results of these histograms

19

First early Results for 4GeV_Pos_Pions

20

In the correlation-matrix I attached at the beginning I point each time 2 (space-correlated) counters fired, but this provides more events than the number of real physical events due to the lack of time-information (not available yet in the DSTs). Once time-info available we will be able to locate the real-points in space.

For this reason Rob suggested me: For each counter that fired pick a random one among those correlated to it...this obviously gives more points that the actual ones and for me is like a simulation of the spatial beam distribution based on spatial-correlations. Next results from the 1.77GeV_Pos_Electron folder (bad-sample, low statistics).

simulation of the beam based on spatial-data

21

22

Why do we require Time-Information beside the Spatial one?

Suppose that counters 2,4,6 and 8 fired then we have the 2 possibilities for points to allocate to the correlation-matrix (considering only spatial correlations):

{(6,4),(2,8)} && {(6,2),(8,4)}

That is why I need, when counter i fires, attach counter j (!= i) only if there is spatial && time correlation between them. Taking only spatial correlation leads to 4 points where only 2 were actually hit.

23

ToDo List Find out why there are cosmic- in my contamination-sample Analyze contamination for All other Pion-Folders of Data-Run-1

and 2 Find out what is the signal that ends at 15ns for low-E samples Analyze what is the effect of the Veto, how many events are

actually lost after that cut, then.....Should I consider only Vetoed events?

Find out if the pion-peaks are really symmetric Find out the composition of other peaks beside the proton and

pion ones (what do they represent physically, particles from the 2nd bucket?)

Compare the spread of the 120GeV proton-peak with the spread of pion-peaks (Why is it relevant?)

24

Study Cora's Thesis, although she performed PID using only ToF and I am looking at the Detector beside ToF (to test the accuracy of the cuts and to be able to separate muons and pions)

Study the composition of the accidentals (peaks from particles from the second bucket?)

I have more information than Cora had: They have rejected electrons using Lead shield & Cerenkov to ID them. Now we have also information from the Detector.

An analysis of dE/dx may be useful...to compare with E & LP cuts or to go beside them?

Continue working with the spatial-distribution of the beam (time info of veto-counters still needed to continue)

25

A Procedure to Follow in order to achieve the Goal

1-I can make a fit to the proton-peaks, then the mean will be the number of protons and the uncertainty of that estimate will be the RMS.

2-I cannot do this for the pion-peak because there are many muons present there.

3-I need to develop tools (cuts) to separate the muons inside that peak: {E, LP, dE/dx}. Then I will be able to say how many muons and pions are actually there.

4-From the analysis of the contamination I can say the amount of pions and muons present there, so I can add these amounts to the numbers calculated in the previous step.

26

After this process (for a given Energy) I would be able to present a result for the numbers:

The uncertainty in the number of protons (statistical error) is quite easy to calculate by making a fit, the problem is how to calculate the uncertainty in the number of pions and muons (because they are mixed in the pion-peak I cannot simply make a fit there). Since I am only using raw-data there should be only statistical errors? What about the my error in making cuts? And what about the error associated with each specific cut?

In Cora's analysis, she actually made a fit in the pion-peak and did not take into account the presence of muons (she only used ToF and one cannot separate the muons with this mechanism, she also had an electron-contamination, which appeared as a shoulder at the left of the pion-peak).

Should I take into account the peaks from other buckets as part of the sample (as unidentified particles)?

Slide 1Slide 2Slide 3Slide 4Slide 5Slide 6Slide 7Slide 8Slide 9Slide 10Slide 11Slide 12Slide 13Slide 14Slide 15Slide 16Slide 17Slide 18Slide 19Slide 20Slide 21Slide 22Slide 23Slide 24Slide 25Slide 26